Horticultural reviews 45 9781119281252, 1119281253

Table of contents :
Title Page......Page 13
Table of Contents......Page 2
Contributors......Page 15
Dedication: Jules Janick......Page 17
1 The Flowers of Fragaria × ananassa......Page 21
II. STRAWBERRY GROWTH, REPRODUCTION, AND COMMERCIAL MANAGEMENT......Page 22
III. INFLORESCENCE ARCHITECTURE......Page 34
IV. GENETICS OF FLOWER INDUCTION......Page 40
V. CONCLUSIONS......Page 52
LITERATURE CITED......Page 53
2 Small Unmanned Aircraft Systems (sUAS)......Page 60
I. INTRODUCTION......Page 63
II. AIRCRAFT......Page 64
III. SENSORS AND DATA PROCESSING......Page 69
IV. HORTICULTURAL APPLICATIONS......Page 80
V. CHALLENGES......Page 87
LITERATURE CITED......Page 89
3 Leaf Blackening......Page 100
I. INTRODUCTION......Page 101
II. VARIATION IN EXPRESSION OF LEAF BLACKENING......Page 102
III. PHYSIOLOGICAL CAUSES OF LEAF BLACKENING......Page 107
IV. THE BIOCHEMICAL MECHANISMS OF LEAF BLACKENING......Page 113
V. CONTROL OF LEAF BLACKENING......Page 115
VI. CONCLUSIONS......Page 122
LITERATURE CITED......Page 123
4 Sapota (Manilkara achras Forb.)......Page 130
I. INTRODUCTION......Page 133
II. NUTRITIVE VALUE......Page 137
III. PHYSIOLOGICAL AND BIOCHEMICAL CHANGES DURING FRUIT MATURATION AND RIPENING......Page 139
V. PHYSIOLOGICAL DISORDERS......Page 145
VI. POSTHARVEST DISEASES......Page 146
VII. POSTHARVEST TECHNOLOGY......Page 147
VIII. POSTHARVEST TREATMENTS......Page 149
IX. NON‐DESTRUCTIVE METHODS FOR IDENTIFYING FRUIT MATURITY AND QUALITY......Page 153
X. PROCESSING......Page 154
XI. SUMMARY AND FUTURE PROSPECTS......Page 156
LITERATURE CITED......Page 157
5 The Citron (Citrus medica L.) in China......Page 168
II. HISTORY AND CULTURE......Page 170
III. NOMENCLATURE......Page 173
IV. CURRENT CITRON CULTIVATION IN CHINA......Page 174
V. MAJOR CULTIVARS OF CHINESE CITRON AND SELECT CITRON HYBRIDS......Page 184
VI. GERMPLASM STATUS; REGIONAL AND GLOBAL PERSPECTIVE......Page 222
LITERATURE CITED......Page 224
6 Apple Rootstocks......Page 232
II. HISTORY......Page 234
III. ROOTSTOCK–SCION INTERACTIONS......Page 238
IV. STRESSES INFLUENCING ROOTSTOCK PERFORMANCE......Page 254
V. INTERSTEMS......Page 281
VI. INFLUENCE OF ROOTSTOCK ON FRUIT CHARACTERISTICS......Page 283
VII. GENETICS AND BREEDING......Page 286
VIII. ROOTSTOCK EVALUATION......Page 295
LITERATURE CITED......Page 299
7 Canopy Growth and Development Processes in Apples and Grapevines......Page 343
I. INTRODUCTION......Page 345
II. PHENOLOGY......Page 346
III. DORMANT BUDS IN APPLE TREES AND GRAPEVINES......Page 347
IV. WINTER CHILLING IN APPLE TREES AND GRAPEVINES......Page 348
V. BUDBREAK AND SHOOT DEVELOPMENT IN APPLE TREES AND GRAPEVINES......Page 349
VI. FRUIT GROWTH......Page 361
VII. BIOMASS PARTITIONING......Page 366
VIII. PHOTOSYNTHESIS AND THE CARBON ECONOMY......Page 368
IX. ABIOTIC STRESS EFFECTS ON CANOPY PHYSIOLOGY......Page 376
X. IMPACT OF CLIMATE CHANGE ON PHENOLOGY......Page 381
XI. CONCLUSIONS......Page 383
LITERATURE CITED......Page 384
8 Organic Acids in Fruits......Page 401
I. INTRODUCTION......Page 404
II. THE FUNCTION OF THE FLESH OF FRUITS AND ITS IMPLICATION FOR THEIR ORGANIC ACID CONTENTS......Page 406
III. ACIDS THAT CONTAIN A BENZENE RING: THE AROMATIC ACIDS......Page 407
IV. THE INTER‐RELATED ACIDS: ASCORBIC, OXALIC, TARTARIC, AND GALACTURONIC......Page 416
VI. MALIC, CITRIC, AND METABOLICALLY RELATED ACIDS......Page 433
LITERATURE CITED......Page 446
Subject Index......Page 471
Cumulative Subject Index......Page 481
Cumulative Contributor Index......Page 580
End User License Agreement......Page 605

Citation preview

Table of Contents Cover Title Page Contributors Dedication: Jules Janick 1 The Flowers of Fragaria × ananassa I. INTRODUCTION II. STRAWBERRY GROWTH, REPRODUCTION, AND COMMERCIAL MANAGEMENT III. INFLORESCENCE ARCHITECTURE IV. GENETICS OF FLOWER INDUCTION V. CONCLUSIONS LITERATURE CITED 2 Small Unmanned Aircraft Systems (sUAS) I. INTRODUCTION II. AIRCRAFT III. SENSORS AND DATA PROCESSING IV. HORTICULTURAL APPLICATIONS V. CHALLENGES VI. CONCLUSIONS LITERATURE CITED 3 Leaf Blackening I. INTRODUCTION II. VARIATION IN EXPRESSION OF LEAF BLACKENING III. PHYSIOLOGICAL CAUSES OF LEAF BLACKENING IV. THE BIOCHEMICAL MECHANISMS OF LEAF BLACKENING V. CONTROL OF LEAF BLACKENING VI. CONCLUSIONS LITERATURE CITED 4 Sapota (Manilkara achras Forb.) I. INTRODUCTION II. NUTRITIVE VALUE III. PHYSIOLOGICAL AND BIOCHEMICAL CHANGES DURING FRUIT

MATURATION AND RIPENING IV. PREHARVEST EFFECTS ON POSTHARVEST QUALITY V. PHYSIOLOGICAL DISORDERS VI. POSTHARVEST DISEASES VII. POSTHARVEST TECHNOLOGY VIII. POSTHARVEST TREATMENTS IX. NONDESTRUCTIVE METHODS FOR IDENTIFYING FRUIT MATURITY AND QUALITY X. PROCESSING XI. SUMMARY AND FUTURE PROSPECTS LITERATURE CITED 5 The Citron (Citrus medica L.) in China I. INTRODUCTION II. HISTORY AND CULTURE III. NOMENCLATURE IV. CURRENT CITRON CULTIVATION IN CHINA V. MAJOR CULTIVARS OF CHINESE CITRON AND SELECT CITRON HYBRIDS VI. GERMPLASM STATUS; REGIONAL AND GLOBAL PERSPECTIVE LITERATURE CITED 6 Apple Rootstocks I. INTRODUCTION II. HISTORY III. ROOTSTOCK–SCION INTERACTIONS IV. STRESSES INFLUENCING ROOTSTOCK PERFORMANCE V. INTERSTEMS VI. INFLUENCE OF ROOTSTOCK ON FRUIT CHARACTERISTICS VII. GENETICS AND BREEDING VIII. ROOTSTOCK EVALUATION LITERATURE CITED 7 Canopy Growth and Development Processes in Apples and Grapevines I. INTRODUCTION II. PHENOLOGY III. DORMANT BUDS IN APPLE TREES AND GRAPEVINES IV. WINTER CHILLING IN APPLE TREES AND GRAPEVINES

V. BUDBREAK AND SHOOT DEVELOPMENT IN APPLE TREES AND GRAPEVINES VI. FRUIT GROWTH VII. BIOMASS PARTITIONING VIII. PHOTOSYNTHESIS AND THE CARBON ECONOMY IX. ABIOTIC STRESS EFFECTS ON CANOPY PHYSIOLOGY X. IMPACT OF CLIMATE CHANGE ON PHENOLOGY XI. CONCLUSIONS LITERATURE CITED 8 Organic Acids in Fruits I. INTRODUCTION II. THE FUNCTION OF THE FLESH OF FRUITS AND ITS IMPLICATION FOR THEIR ORGANIC ACID CONTENTS III. ACIDS THAT CONTAIN A BENZENE RING: THE AROMATIC ACIDS IV. THE INTERRELATED ACIDS: ASCORBIC, OXALIC, TARTARIC, AND GALACTURONIC V. FATTY ACIDS VI. MALIC, CITRIC, AND METABOLICALLY RELATED ACIDS VII. CONCLUSIONS LITERATURE CITED Subject Index Cumulative Subject Index Cumulative Contributor Index End User License Agreement

List of Tables Chapter 01 Table 1.1. Map positions and genetic effect of QTL detected for the PF (perpetual flowering) and RU (runnering) traits for the female (f) parent ‘Capitola’ and male (m) parent ‘CF1116.’ QTL identification was based on compositeinterval mapping analysis with LOD > LOD threshold (3.1) (α = 0.05). From Gaston et al. 2013; reproduced with permission from Oxford University Press. Chapter 02 Table 2.1. A summary of the types of aircraft used for agriculture, forestry, and horticulture research based on a survey of literature from 2004 to 2016.

Table 2.2. A summary of the types of sensors used for agriculture, forestry, and horticulture research using UAS based on a survey of literature from 2004 to 2016. Chapter 04 Table 4.1. Nutritional components of sapota (Manilkara zapota) fruit per 100 g edible portion. Table 4.2. Chemical composition of sapota fruit juice with pH 5.36 and titratable acidity 0.16% (% citric acid). Chapter 07 Table 7.1. Differences in phenology of six Malus domestica cultivars growing in three Chilean locations (after Yuri et al. 2011). Budbreak represents the date of budbreak in days after 1 September, and flowering and harvest are given in days after budbreak. These data are all estimated from the original publication. Chapter 08 Table 8.1. Concentrations of quinic acid (QA; mg g–1 FW) in the flesh of some ripe fruits. Table 8.2. Concentrations of shikimic acid (SA; mg g–1 FW) in the flesh of some ripe fruits. Table 8.3. Concentrations of ascorbic acid (AA; mg g–1 FW) in the flesh of some ripe fruits and in some leaves. Table 8.4. Concentrations of dehydroascorbic acid (DHA; mg g–1 FW) in the flesh of some ripe fruits and some leaves. Table 8.5. Concentrations of tartaric acid (TA: mg∙g–1 FW) in the flesh of some ripe fruits. Table 8.6. Concentrations of total oxalic acid (mg∙g–1 FW) in the flesh of some ripe fruits (based on Nguy n and Savage 2013a). Table 8.7. Concentrations of malic acid (MA; mg g–1 FW) in the flesh of some ripe temperate fruits and some leaves and roots. For leaves and roots, nitrate and ammonium refer to the form of nitrogen on which the plants were grown. Table 8.8. Concentrations of citric acid (CA; mg g–1 FW) in the flesh of some ripe temperate fruits and some leaves. For leaves, nitrate and ammonium refer to the form of nitrogen on which the plants were grown. Table 8.9. Concentrations of malic acid (MA; mg g–1 FW) in the flesh of unripe and ripe fruits. For unripe fruit (except tomato) the maximum content during the period of development studied is given. For tomato, unripe fruit are 28 days after anthesis, and nitrate and ammonium denote the form of nitrogen with which the plant was fertilized.

Table 8.10. Concentrations of citric acid (CA; mg g–1 FW) in the flesh of unripe and ripe fruits. For unripe fruit the maximum content during the period of development studied is given. For tomato, unripe fruit are 28 days after anthesis, and nitrate and ammonium denote the form of nitrogen with which the plant was fertilized.

List of Illustrations Chapter 01 Fig. 1.1. Leaves and axillary meristems of cultivar ‘Portola.’ Fig. 1.2. (a) Diagram of fully developed flower cluster with (a) primary flower, (b) secondary flower, (c) tertiary flower, and (d) quaternary flower http://www.hort.cornell.edu/grower/nybga/pdfs/2012berryproceedings.pdf (from Poling 2012). (b) Picture of a flower cluster of the cultivar ‘Portola’, with (a) primary flower, (b) secondary flower bud, and (c) tertiary flower bud. Fig. 1.3. (a) Principal flower parts of cultivar ‘EvieII,’ including (a) stamen, (b) pistil, (c) receptacle, (d) petal, and (e) sepal. Photograph taken July 10, 2014 in Minnesota. (b) Crosssection of F. × ananassa showing (a) pistil and (b) receptacle. Fig. 1.4. (a) Profile of mature fruit of the cultivar ‘Amandine,’ with embedded achenes. (b) Crosssection of an ‘Amandine’ fruit, with (a) interior receptacle, (b) fibrovascular tube and (c) calyx. Fig. 1.5. Average daylengths of Minneapolis, MN and Santa Maria, CA, taken on the 20th of each month. Raw data acquired from Time & Date AS: http://www.timeanddate.com/worldclock/astronomy.html? n=3857&month=12&year=2014&obj=sun&afl=1&day=1. Fig. 1.6. Average high temperatures in Minneapolis, MN and Santa Maria, CA, taken on the 20th of each month. Raw data acquired from Intellicast: http://www.intellicast.com/. Fig. 1.7. Diagram of the mattedrow system common to Junebearing cultivars. Fig. 1.8. Dayneutral ‘Monterey’ runner, with developed inflorescence. Fig. 1.9. (a) Diagram of a single, typical strawberry inflorescence, and (b) model of ‘Seascape,’ where 1°, 2°, and 3° represent primary, secondary, and tertiary inflorescences, which developed from buds after initial planting. Fig. 1.10. Photograph of ‘Portola’ inflorescence. 1°, 2°, and 3° represent primary, secondary, and tertiary flowers. Labeled brackets indicate primary (I) and secondary (II) internodes. Fig. 1.11. Selected flower clusters of (a) ‘Annapolis,’ (b) ‘Albion,’ (c) ‘Evie2,’ (d) ‘Monterey,’ (e) ‘Portola,’ (f) ‘San Andreas,’ and (g, h) ‘Seascape.’

Fig. 1.12. Fragaria vesca shoot and flower development. (a) F. vesca YW5AF7 grown in a 10.2cm pot; (b) YW5AF7 dichasial cyme bearing yellow berries; (c) Inflorescence with primary flower (1) and two developing secondary flowers (2). Young tertiary buds (arrows) are present beneath the secondary flower buds; (d) Diagram of shoot architecture. Numbers indicate primary, secondary, and tertiary flower buds; (e) Diagram illustrating floral organ arrangement. The two outer whorls are concentric rings of five bracts (b) alternating with five sepals (s). The third whorl consists of five white petals (p). Interior to the petals are two whorls of stamens. Stamens are arranged in a repeating pattern of five tall (T) and five short (S) in the inner whorl and 10 medium length (M) in the outer whorl. The center circle indicates a receptacle topped with numerous, spirally arranged carpels; (f) Scanning electron micrograph (SEM) of a developing floral bud, illustrating spirally arranged carpel primordial; (g) Abaxial view of a typical F. vesca flower with five narrow bracts (b) alternating with five wider sepals (s); (h) Adaxial view of typical F. vesca flower illustrating a whorl of five white petals, two whorls of ten stamens each, and an apocarpous gynoecium with ~160 pistils. (i) Dissected flower illustrating the “S, M, T, M, S” stamen pattern. Scale bars: (a) 2 cm; (g–i) 1 mm. Fig. 1.13. Silencing of FvTFL1 leads to daylengthindependent flowering. (a) Phenotypes of FvTFL1 RNAi silencing and overexpression lines in the shortday (SD) F. vesca background. Clonally propagated plants (runner cuttings) of SD F. vesca and P35S:FvTFL1RNAi1 and P35S:FvTFL11 lines were subjected to SD induction treatment for four weeks followed by longdays (LDs; left), or grown continuously under LDs (right); (b), Flowering time of SD F. vesca and P35S:FvTFL1RNAi and P35S:FvTFL1 plants (RNAi and OX, respectively) in SDs and LDs. Flowering time is indicated as days to anthesis from the beginning of the treatments. Treatments and plant materials were as described in (a). Values indicate mean ± SD. n = 4 (OX1), n = 5 (RNAi2), n = 6 (RNAi1 and OX2), and n = 7 (SD F. vesca); (c,d) Expression of FvTFL1 (c) and FvAP1 (d) in the apices of two independent P35S:FvTFL1RNAi (RNAi) lines. Values indicate mean ± SD. n = 3 (RNAi1 and SD F. vesca) or n = 2 (RNAi2). Fig. 1.14. Frequency distribution of number of weeks flowering at Maryland (MD) (a) and California (CA) (b). The phenotypic values of the parents, ‘Tribute’ and ‘Honeoye’ are indicated by arrows (Castro et al. 2015). Fig. 1.15. Quantitative trait loci (QTL) for number of weeks of flowering detected in the octoploid strawberry (F. × ananassa Duchesne ex. Rozier) population ‘Tribute’ × ‘Honeoye’ evaluated in Maryland (dotted line) and California (solid line) using interval mapping (IM) (Castro et al. 2015). Chapter 02 Fig. 2.1. Google Scholar search for “UAV aircraft” and the resulting number of articles (does not include patents or citations) annually from 1990 to 2015.

Chapter 03 Plate 3.1. The various positions where leaf blackening development can first manifest in Protea ‘Sylvia’ (P. eximia × P. susannae). (a) Leaftip; (b) Leafbase; (c) Mid rib; (d) Lateral leaf margins (Windell 2012) Plate 3.2. Grading scales to visually assess the impact of leaf blackening as a quality attribute in Protea (Ekman et al. 2008). Less than 10% leaf blackening is not considered greatly noticeable, whereas greater than 50% appears fully blackened Plate 3.3. The maturity stages of a Protea ‘Sylvia’ inflorescence showing (a) a more immature, but harvestready softtip stage where involucral bracts are just starting to retract from the center and the florets, and (b) a more mature and fully opened inflorescence, but not yet senesced (Windell 2012) Plate 3.4. Protea ‘Sylvia’ displaying leaf blackening and phytotoxic symptoms in the leaves (a) and after storage at the tips in the involucral bracts of the inflorescence (b) (Windell 2012) Chapter 04 Plate 4.1. Sapota tree Plate 4.2. External and internal images of round and elongated sapota fruit Chapter 05 Plate 5.1. Sculpture of fingered citron carved in carnelian in 18th century China. Fig. 5.1. (a) Map of Citron cultivation in China. Plate 5.2. Dried ‘Chuan’ fingered citrons, for use in Chinese traditional medicine, in Xiaogu, Sichuan. Plate 5.3. Drying fingered citron slices for Chinese traditional medicine near Huaning, Yunnan. Plate 5.4. Bonsai fingered citron plants in pots for sale at a street market along a road in Jinhua, Zhejiang. Plate 5.5. In Jinhua, fingered citron growers use trellises, stakes, and ties to ensure that the limbs and fruit maintain an attractive upright growing habit. Plate 5.6. Mr. PeiYong Zhang eats a slice of common citron at his home in Dehong Prefecture, Yunnan. Plate 5.7. ‘Dog Head’ citron in Mojiang, Yunnan. Plate 5.8. ‘Jianshui Round’ citron, Bai Shi Yan village, Xizhuang township, Jianshui County, Yunnan. Plate 5.9. ‘Jinggu’ citron, Jinggu Dai and Yi Autonomous County, Yunnan. Plate 5.10. ‘Jinghong Water’ citron, Jinghong, Yunnan, with large central cavity and

small locules with numerous seeds and scanty juice vesicles. Plate 5.11. ‘Ning’er Giant’ citron weighing ca. 8 kg, held by the farmer, Li Hua Zhong, near Ning’er, Yunnan. Plate 5.12. ‘Persistent Stigma’ citron, Huaning, Yunnan. Plate 5.13. ‘Pumpkin’ citron in Xiaowan, Fengqing County, Yunnan. Plate 5.14. ‘Weishan Bullet’ citron near Weishan, Yunnan. Plate 5.15. ‘Weishan Sour’ citron near Weishan, Yunnan, China. Plate 5.16. ‘Weishan Sweet’ citron, Tuanshan village, Nanzhao township, near Weishan, Yunnan. Plate 5.17. ‘Yun’ fingered citron, showing how one fruit actually has some juice vesicles. Plate 5.18. ‘Aihua’ fingered citrons. Plate 5.19. ‘Chuan’ fingered citrons with typical closed fingers, Shuangshi village, Xinfan township, Muchuan County, Sichuan. Plate 5.20. ‘Guang’ fingered citron in Guangdong. Plate 5.21. Typical specimens of ‘Octopus’ fingered citron grown near Weishan, Yunnan. Plate 5.22. ‘Qingpi’ fingered citron plants in bonsai cultivation in Jinhua, Zhejiang. Plate 5.23. ‘Yun’ fingered citrons sold at a produce stand at the airport in Kunming, Yunnan. Plate 5.24. ‘Zhaocai’ fingered citron, Jinhua, Zhejiang. Plate 5.25. ‘Half and Half’ fingered citron at a market in Jinghong, Yunnan. Plate 5.26. ‘Muli’ citron in Baidiao, east of Muli, Sichuan. Plate 5.27. Characteristic form of ‘Muli’ citron, Muli, Sichuan. Plate 5.28. Wild citron, type 1 (with thin rind, segments and juice vesicles), from a wild forest in a valley in the mountains 2 km west of Hexinchang village, Mangshi township, Luxi City, Yunnan. Elevation: 1190 m. Plate 5.29. Wild citron, type 1, growing wild in Mangshi township, Yunnan. Top: view of the valley where wild citron grows as an understory plant. Bottom: trees of wild citron, type 1. Plate 5.30. Wild citron, type 2 (with thick albedo, no segments or juice vesicles). Plate 5.31. ‘Goucheng’ citron hybrid, Binchuan, Yunnan. Plate 5.32. ‘Yunmao Oval’ citron hybrid near Mengwan village, Dehong prefecture,

Yunnan. Chapter 07 Fig. 7.1. The process of budbreak in ‘Royal Gala’ and MM.111 apple trees.In the case of ‘Royal Gala’, days of treatment were days after 1 September and for MM.111, the days of treatment were days of forcing at 20 °C. Fig. 7.2. Rates of budbreak of ‘Delicious’ apple trees as a function of temperature. Fig. 7.3. The impact of postharvest day/night temperatures on the budbreak of potted ‘Braeburn’ apple trees in the subsequent spring. Fig. 7.4. Leaf appearance in days after budbreak of ‘Semillon’ grapevines, with and without a fruit crop load, as a function of node position along the shoot and separated into three zones. Zone 1 represents the initial cluster of leaves that appear very rapidly from preformed primordia. Zone 2 are the leaves also from preformed primordia but which appear in a linear sequence. Zone 3 leaves all originate from neoformed primordia formed by activity of the shoot apical meristem. Fig. 7.5. Rates of shoot extension for a number of grapevine cultivars shown as a function of temperature. The potted vines were grown in controlled conditions for 13 weeks from budbreak. Fig. 7.6. Expansion of selected leaves of ‘Semillon’ grapevine across the growing season with a fit of the Boltzmann sigmoid function to each. Fig. 7.7. Rates of dry weight accumulation of berries of three grapevine cultivars as shown as a function of temperature for vines grown in controlled environments. Fig. 7.8. Net photosynthesis of leaves along the shoot on a vineyardgrown ‘Semillon’ vine at different times of the growing season (DAB = days after bloom) as indicated (D.H. Greer unpublished data). Fig. 7.9. Seasonal changes in (a) rates of the net carbon acquisition (NCA), (b) the rate of carbon accumulated into biomass (CAB), and (c) net daily carbon balance (NCB) for ‘Semillon’ vines growing in outdoor conditions. Fig. 7.10. The photosynthetic light response for ‘Red Gala’ trees at a range of leaf temperatures as shown and grown in Australian conditions. The response at 20 °C was similar to that at 30 °C, and so has been excluded for clarity. The lines are a fit to the hyperbolic tangent function (see Greer and Halligan 2001). Fig. 7.11. Seasonal changes in leaf photosynthesis of ‘Braeburn’ apple trees grown in New Zealand orchard conditions. Note the marked decrease in photosynthesis after harvest had occurred. Leaf temperatures were between 26 and 28.5 °C throughout these measurements. Fig. 7.12. The effect of an 8day heat stress event on photosynthesis of ‘Red Gala’ trees before, during, and after the hightemperature period, as indicated.

Chapter 08 Fig. 8.1. Simplified scheme showing the pathways utilized in the synthesis of the main groups of organic acids found in fruits. OAA = Oxaloacetate; PEP = Phosphoenolpyruvate. Fig. 8.2. Simplified scheme showing the shikimate pathway and the synthesis of quinic and shikimic acids. PEP = Phosphoenolpyruvate; DAHP = 3DeoxyDarabino heptulosonic acid 7phosphate; DHS = Dehydroshikimic acid; DHQ = Dehydroquinic acid. Fig. 8.3. Simplified scheme showing the synthesis of phenolic acids. PEP =  Phosphoenolpyruvate; DAHP = 3DeoxyDarabinoheptulosonic acid 7 phosphate; DHS = Dehydroshikimic acid; DHQ = Dehydroquinic acid. Fig. 8.4. Simplified scheme showing the synthesis of 5caffeoyl quinic acid. CoA = Coenzyme A. Fig. 8.5. Simplified scheme showing the interrelations between ascorbic, oxalic, tartaric, and galacturonic acids. GDP = Guanosine diphosphate. Fig. 8.6. Simplified scheme showing the synthesis and catabolism of ascorbic acid. GDP = Guanosine diphosphate; DHAR = Dehydroascorbate reductase; MDHAR =  Mondehydroascorbate reductase. Fig. 8.7. Simplified scheme showing the synthesis of tartaric acid. Fig. 8.8. Simplified scheme showing the synthesis and breakdown of oxalic acid. CoA  = Coenzyme A: ICL = Isocitrate lyase. Fig. 8.9. Simplified scheme showing the synthesis of galacturonic acid. UDP = Uridine diphosphate. Fig. 8.10. Simplified scheme showing the synthesis of Krebs cycle acids. MDH =  Malate dehydrogenase; OAA = Oxaloacetate, PEP = Phosphoenolpyruvate; CoA =  Coenzyme A; PEPC = phosphoenolpyruvate carboxylase.

Horticultural Reviews is sponsored by: American Society for Horticultural Science International Society for Horticultural Science

Editorial Board, Volume 45 A.R. Ferguson J. Janick S. Nicola

HORTICULTURAL REVIEWS Volume 45 Edited by Ian Warrington Massey University New Zealand

This edition first published 2018 © 2018 John Wiley & Sons, Inc. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions. The right of Ian Warrington to be identified as the editor of the editorial material in this work has been asserted in accordance with law. Registered Offices John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK Editorial Office 9600 Garsington Road, Oxford, OX4 2DQ, UK For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com. Wiley also publishes its books in a variety of electronic formats and by printondemand. Some content that appears in standard print versions of this book may not be available in other formats. Limit of Liability/Disclaimer of Warranty While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. Library of Congress CataloginginPublication data applied for Hardback ISBN: 9781119430957 Cover Design: Wiley Cover Image: Image courtesy of Jules Janick

Contributors Franco Famiani, Dipartimento di Scienze Agrarie, Alimentari e Ambientali, Università degli Studi di Perugia, Perugia, Italy Gennaro Fazio, USDA/ARS Plant Genetic Resources Unit, Geneva, NY, USA John B. Golding, NSW Department of Primary Industries, Gosford, NSW, Australia Dennis H. Greer, School of Agricultural and Wine Sciences, National Wine and Grape Industry Centre, Charles Sturt University, Wagga Wagga, NSW, Australia Eleanor W. Hoffman, Department of Horticultural Science, Stellenbosch University, South Africa Emily Hoover, Department of Horticultural Science, University of Minnesota, MN, USA Xulan Hu, Yunnan Provincial Department of Agriculture, Kunming, Yunnan, China Gerard Jacobs, Department of Horticultural Science, Stellenbosch University, South Africa David Karp, University of California, Riverside, CA, USA Babak Madani, Tropical Fruit Research Center, Horticultural Science Research Institute, Agricultural Research Education and Extension Organization (AREEO), Hormozgan, Iran Richard P. Marini, Department of Plant Science, The Pennsylvania State University, University Park, PA, USA Amin Mirshekari, Department of Agronomy and Plant Breeding, Faculty of Agriculture, University of Yasouj, Yasouj, Iran Andrew Petran, Department of Horticultural Science, University of Minnesota, MN, USA James A. Robbins, University of Arkansas System Division of Agriculture, Little Rock, AR, USA Waafeka Vardien, Department of Horticultural Science, Stellenbosch University, South Africa Robert P. Walker, Dipartimento di Scienze Agrarie, Alimentari e Ambientali, Università degli Studi di Perugia, Perugia, Italy Nicole E. Windell, Fynbloem, Riviersonderend, South Africa Elhadi Yahia, Facultad de Ciencias Naturals, University of Queretaro, Queretaro, Mexico

Jules Janick

Dedication: Jules Janick Dr Jules Janick, James Troop Distinguished Professor of Horticulture in the Department of Horticulture & Landscape Architecture at Purdue University, is undoubtedly one the most well known and well regarded horticultural scientists in the world today. As the founder of Horticultural Reviews, and the related Plant Breeding Reviews, he is highly deserving of having this volume dedicated to him. His contributions to horticultural science have been extensive, embracing many aspects of the discipline. He has been truly international in his endeavors. Jules was born in New York City in 1931, and graduated with a B.S. in Agriculture from Cornell University in 1951. This was followed by an M.S. (1952) and a Ph.D. (1954) – both at Purdue University in genetics and breeding. He was 23 years old at the completion of the Ph.D. – an early indication of his commitment and productivity! Within his career in horticultural research, Jules and his students (13 masters and 17 doctoral) have made important advances in the genetics of sex determination, including the synthesis of heteromorphic sex chromosomes in spinach, fire blight resistance, cleistogamy, cucurbitacins, artemisisin production, anthocyanin pigmentation, plant density, in vitro metabolite production from somatic embryos, and the production of synthetic seed. In crop improvements he has been associated with the release of 21 scabresistant apple cultivars, three pear cultivars with tolerance to fire blight, delayedbolting arugula, crackresistant tomato (for Brazil), and the first release of a cultivar (pelargonium) from somaclonal variation. In a particular niche area of study, Professor Janick has made contributions to the historical aspects of horticulture, with emphasis on ancient Egypt and the New World, and has explored the relation of art and horticultural technology with special studies on the iconography of Rubus, Daucus, the Cucurbitaceae, the Solanaceae, and opening up a new approach to the study of plant diversity, origins, cultivar evolution, and diversity. He has contributed iconographic studies on Dioscorides, the Drake Manuscript, the Unicorn Tapestries, Caravaggio, Cotan, and the Raphael frescoes in the Villa Farnesina in Rome. He has written on the interrelationship of horticulture and scholarship, art, ethics, and the contributions of horticulture to human welfare. At present, he is immersed with Dr Arthur Tucker in the unraveling of the bizarre Voynich Codex, demonstrated to be a 16th century Mesoamerican herbal. Professor Janick has taught many courses, including genetics, horticultural plant breeding, seed production, plant propagation, tropical horticulture, and the history of horticulture. The last two courses are now offered continuously online with over 500 students per year. Jules has been a prolific author and editor in horticultural science. He was the editor of HortScience from 1970 to 1983 (14 volumes) and transformed it into one of the major journals in horticulture. He was editor of the Journal of American Society for Horticultural Science (ASHS) from 1976 to 1983 (8 volumes). He is the founder and editor of both Horticultural

Reviews (44 volumes since 1979) and Plant Breeding Reviews (40 volumes since 1983), one of his major achievements. From 2002 to 2010 he was the science editor of Chronica Horticulturae (International Society for Horticultural Science, ISHS), and transformed this publication into a significant magazine of world horticulture. He has edited and produced seven proceedings of New Crops symposia since 1990 that have had a deep impact on new crop information. The development of a new crop website (www.hort.purdue.edu/newcrop) has become a major world resource for information on crops. Professor Janick is the author of the book Horticultural Science (4 editions), an influential text that has been translated into Spanish, Portuguese, Arabic, and Hindi. Another text – Crop Science; An Introduction to World Crops – has gone through three editions. He has organized a number of monographs on fruit breeding, including Advances in Fruit Breeding (translated into Chinese, Russian, and Spanish), Methods in Fruit Breeding, and a threevolume update of Advances entitled Fruit Breeding. Jules has edited eight volumes of Acta Horticulturae. Altogether, he has authored, coauthored or edited over 140 volumes of books, journal volumes, proceedings, and monographs. Dr Janick also coedited the CABI Encyclopedia of Fruit and Nuts (2008). He has authored 61 book chapters and 427 papers, of which half are in refereed journals. Jules had the vision to initiate Horticultural Reviews in 1979. At that time, reviews on horticultural topics were limited in both length and scope, and did not do justice to the volume of horticultural research that had been conducted particularly in the latter half of the 20th century. This new publication fulfilled that niche and has grown to be a valued source of information for scholars, scientists, and horticulturists worldwide. A total of 349 review articles, comprising 18 936 pages in aggregate, have been included in the 44 volumes published to date (noting that in some years more than one volume was published). Each volume is dedicated to a horticulturist. Throughout that period Jules remained the sole editor, soliciting manuscripts, cajoling authors to deliver by the due date, and editing the copy received. He was tireless in each of these pursuits. In 1983 he founded Plant Breeding Reviews (40 volumes), with about half the articles devoted to horticultural crops. The impact that Jules Janick has had on horticulture has not been confined to the USA. He served two years at the Rural University of Minas Gerais in Brazil in 1963–1965 as part of a USAID Purdue contract, where he lectured in Portuguese. The list of countries that he has visited while participating in meetings, consulting, conducting research, advising, and teaching is extensive, and includes Argentina, Australia, Bahamas, Belgium, Canada, Canary Islands, Columbia, Costa Rica, Crete, Ecuador, Egypt, England, Finland, France, Germany, Hungary, Indonesia, Israel, Italy, Ivory Coast, Japan, Korea, Mexico, Morocco, New Zealand, Norway, People’s Republic of China, Poland, Portugal, Puerto Rico, Spain, Sweden, The Netherlands, Tanzania, and Turkey. Among his numerous recognitions are several awards of ASHS, the Wilder award of the American Pomological Society, corresponding member of the Italian Academy of Agriculture, and four honorary degrees [University of Bologna (1990), the Technical University of Lisbon (1994), the Hebrew University of Jerusalem (2007), and University of Agricultural Sciences

and Veterinary Medicien, ClujNapoka, Romania]. He served as President of the American Society for Horticultural Science in 1986–87, and was inducted to the ASHS Hall of Fame in 2009. In 2011 he received the Lifetime award of the National Association of Plant Breeders. He also served on the Board of the ISHS for two terms (2002–2010). He is a Fellow of the ASHS (1976), of the Portuguese Horticultural Association (1981), the ISHS (2006), and is an ISHS Honorary Member (2010). In the nomination of Professor Janick for the ASHS Hall of Fame Award in 2008, Dr Fred Bliss (Professor Emeritus at the University of California, Davis) stated the following: “Seldom, if ever, has anyone made the breadth and depth of contributions to the field of horticulture as Jules has done. Whether a person is a professional horticulturist or hobbyist, in academia or business, breeder or physiologist, author or reader, you likely would have heard about and benefited from Jules’ multiple interests. In addition to his numerous contributions to teaching and extended education in the academic setting, he is a teacher in the broadest sense by virtue of his tireless efforts to expound and promote horticulture by research, scientific publications, and wonderful oral/visual presentations.” In the same nomination, Dr Martin J. Bukovac (Professor Emeritus at Michigan State University) stated: “He is an individual with a missionary commitment to advance horticulture locally, nationally and internationally through his extensive lecturing, writing, advising and organiser of conferences. He has rekindled an interest in the art and history of horticulture – and is probably the world’s authority in both. …….Jules is a human catalyst in bringing people together from various disciplines, nationally and internationally, in organising projects, publications, and conferences and seeing them to fruition. One look at his CV confirms that Jules is an outstanding educator, scientist and horticulturist….”. Jules, the Renaissance man, is the poet laureate of the Horticulture Department and a talented artist. Throughout his career, Jules has been ably supported by Shirley, his wife of 64 years. They have a son, Peter, a daughter Robin, and four grandsons, Noah, Lee, Nathan, and Aaron. He regards his family as his greatest achievement. His credo has been that advances in horticulture throughout the centuries represent some of the greatest human accomplishments for the betterment of humanity and he is strongly committed to the view that horticulture provides food for body and soul. IAN WARRINGTON Emeritus Professor Massey University Palmerston North New Zealand In 2016, Jules passed the editorship of Horticultural Reviews to Dr Ian Warrington, and of Plant Breeding Reviews to Dr Irwin Goldman, who have undertaken to keep retain the high standards that have been set for these two important publications.

1 The Flowers of Fragaria × ananassa: Morphology, Response to Photoperiod, and Genetics of Induction Andrew Petran and Emily Hoover Department of Horticultural Science, University of Minnesota, MN, USA

ABSTRACT The common cultivated strawberry (Fragaria × ananassa) is a healthy and popular fruit throughout the world, but its octoploid genetic structure poses difficulties to breeders, and the plant’s flowering response to temperature and photoperiod has been challenging to predict, resulting in multiple flowering phenotypes throughout the commercial germplasm. This review assesses the morphology and physiology of these phenotypes, the cultural practices which are common to each flowering response, and focuses on recent efforts to map the genetic basis of dayneutrality within F. × ananassa and its progenitor Fragaria vesca. We summarize the recent consensus observing that the genetics of dayneutral flower induction in diploid F. vesca and octoploid F. × ananassa are not orthologous, and discuss the variance of findings regarding determination of dayneutrality in octoploid cultivars. KEYWORDS: strawberry, dayneutral, diploid, octoploid I. INTRODUCTION II. STRAWBERRY GROWTH, REPRODUCTION, AND COMMERCIAL MANAGEMENT A. Vegetative Growth B. Flower Structure C. Flower Induction, Initiation, and Development 1. Junebearing 2. Everbearing and Dayneutral 3. Thermophotoperiod and Temperature Effects III. INFLORESCENCE ARCHITECTURE IV. GENETICS OF FLOWER INDUCTION A. Fragaria vesca B. Fragaria × ananassa

V. CONCLUSIONS LITERATURE CITED

I. INTRODUCTION The strawberry (Fragaria spp.) is one of the most widely distributed fruit crops in the world. Production of the fruit is present in almost every continent and has exceeded 4 million tonnes per year since 2007 (Wu et al. 2012). There is considerable genetic diversity within strawberry germplasm; wild diploid through decaploid plants have been discovered (Stewart and Folta 2010). This diversity leads to genotypic and phenotypic variance even within the same strawberry species. Perhaps the most commercially important variance is that of flowering habit within the commercially cultivated strawberry Fragaria × ananassa. Because of its commercial value and popularity, the strawberry is a thoroughly documented fruit crop. The purpose of this review is to compile and contrast the morphologic and physiologic traits of F. × ananassa flowering types and review the most recent efforts to identify the underlying genetics behind flowering habit.

II. STRAWBERRY GROWTH, REPRODUCTION, AND COMMERCIAL MANAGEMENT A. Vegetative Growth The strawberry plant is an herbaceous perennial with short internodes forming a modified stem rosette (Savini et al. 2005). This modified stem is commonly known as a crown, where long petiole trifoliate leaves and axillary meristems converge spirally around its axis, ending in a terminal inflorescence (White 1927). Strawberry leaves present a typical dicotyledonous structure with long petioles and foliaceous basal stipules (Savini et al. 2005). Leaf lifespan can exceed three months in favorable conditions (Poling 2012). Axillary meristems can differentiate into branch crowns, which stay near and are structurally identical to the original crown, or stolons (also called runners), which give rise to separate daughter plants (Fig. 1.1) (Demchak 2010). Crowns typically produce one to two branch crowns in a season, but have been known to produce more than five; from a production standpoint, three to four total crowns per plant is desirable, as more can result in decreased fruit size (Poling 2012).

Fig. 1.1. Leaves and axillary meristems of cultivar ‘Portola.’ Photograph taken August 9, 2016, in Minnesota.

B. Flower Structure Inflorescences have two internodes, and develop terminally on the crown or branch crown of the plant in a structure known as a dichasial cyme (Savini et al. 2005). Dichasial cymes have a terminal, primary flower branch with opposite secondary branches beneath the terminal bud, leading to secondary flowers. In strawberry, the inflorescence is commonly known as a flower cluster, and the primary flower, known as the “king flower,” typically bears the largest fruit. Secondary branches begin at the juncture of the first and second internodes; some inflorescences also have tertiary and quaternary branches and flowers (Fig. 1.2).

Fig. 1.2. (a) Diagram of fully developed flower cluster with (a) primary flower, (b) secondary flower, (c) tertiary flower, and (d) quaternary flower http://www.hort.cornell.edu/grower/nybga/pdfs/2012berryproceedings.pdf (from Poling 2012). (b) Picture of a flower cluster of the cultivar ‘Portola’, with (a) primary flower, (b) secondary flower bud, and (c) tertiary flower bud. Photograph taken July 10, 2014, in Minnesota.

The principal parts of the flower itself are shown in Fig. 1.3. Strawberry flowers have five

sepals; fleshy green structures beneath the petals which enclose the flower at bud stage and eventually become the “calyx,” or cap of the berry. Stamens discharge pollen and fertilize the pistils, which are secured on a conical stem known as the receptacle. This receptacle becomes the full, fleshy “berry” at fruit maturity. Despite this plant’s common name, the fruit itself is not botanically classified as a berry. The seedlike organs embedded on the epidermis of the receptacle are actually modified dry fruits known as achenes. The achenes are each connected to the interior of the receptacle by fibrovascular strands, and hold the true seed within their pericarp (Fait et al. 2008) (Fig. 1.4). In F. vesca, auxin and gibberellin biosynthesis occurs in the endosperm and seed coat of the developing achenes, which in turn triggers maturity of the surrounding receptacle (Kang et al. 2013). Because the strawberry fruit contains multiple achenes, and is comprised of a receptacle in addition to its ovaries, it can be classified both as an aggregate and as an accessory fruit.

Fig. 1.3. (a) Principal flower parts of cultivar ‘EvieII,’ including (a) stamen, (b) pistil, (c) receptacle, (d) petal, and (e) sepal. Photograph taken July 10, 2014 in Minnesota. (b) Cross section of F. × ananassa showing (a) pistil and (b) receptacle. Photograph obtained, with permission, from G.D. Carr, December 9, 2015; http://www.botany.hawaii.edu/faculty/carr/images/fra_sp.jpg.

Fig. 1.4. (a) Profile of mature fruit of the cultivar ‘Amandine,’ with embedded achenes. (b) Crosssection of an ‘Amandine’ fruit, with (a) interior receptacle, (b) fibrovascular tube and (c) calyx. Photograph taken July 10, 2014, in Minnesota.

C. Flower Induction, Initiation, and Development Flower induction, initiation and development are highly variable by cultivar, and dependent on genotypic responses to temperature and photoperiod (Savini et al. 2005; Stewart and Folta 2010). These responses are commonly grouped into three flowering categories: Junebearing; everbearing; and dayneutral. Strawberry cultivars are typically classified under one of these three categories based on their photoperiodic flowering habits, and it was originally assumed these habits remained constant over a wide range of temperatures (Darrow and Waldo 1933). However, further research led to the discovery that the photoperiod response of many cultivars would be altered if temperatures were either sub or supraoptimal (Guttridge 1985; Nishiyama and Kanahama 2000; Sonsteby and Heide 2007). This interaction of temperature with photoperiod, known as thermophotoperiod , adds a quantitative factor to the original categorical classifications. Indeed, some believe it incorrect to assign broad flower habit categories to strawberry at all, as photoperiod responses appear to be cultivarspecific (Durner 2015). However, as the vast majority of strawberrybased publications use these classifications, this review will utilize them as well, with the implicit understanding of variance and interaction even within each flowering type. In this section, photoperiod response and common cultural practices of the three groups assuming optimal temperature conditions will first be discussed. The way in which the responses have been observed to change under different temperature ranges will then be explored.

1. Junebearing.  Natural flowering patterns of cultivated octoploid strawberry, F. × ananassa, are of the June bearing type (Darrow 1966). Junebearing cultivars are predominantly grown for commercial purposes in the Upper Midwestern United States, where other flowering types have historically performed poorly (Durner et al. 1984; Luby et al. 1987; Luby 1989). June bearing cultivars induce flowers under shortening daylengths, optimally from 9.5 to 13h days, depending on cultivar (Darrow and Waldo 1933). The change in daylength over time in the United States Upper Midwest (specifically using Minneapolis, MN 44.9833° N as a representative point) compared to a more southern latitude, where strawberries are also grown (specifically using Santa Maria, CA 34.5914° N as a representative point), is shown in Fig. 1.5. The figure implies that flower induction would typically occur in midSeptember for Junebearing cultivars in the Minneapolis area, until temperatures induce plants into dormancy. Savini et al. (2005) noted that Junebearing cultivars will also have flower initials before they enter dormancy. For many Junebearing cultivars the dormancyinducing temperature is a high of 10 °C (Kronenberg et al. 1976). On average, this threshold temperature will be reached in early November in the United States Upper Midwest (Fig. 1.6).

Fig. 1.5. Average daylengths of Minneapolis, MN and Santa Maria, CA, taken on the 20th of each month. Raw data acquired from Time & Date AS: http://www.timeanddate.com/worldclock/astronomy.html? n=3857&month=12&year=2014&obj=sun&afl=1&day=1 .

Fig. 1.6. Average high temperatures in Minneapolis, MN and Santa Maria, CA, taken on the 20th of each month. Raw data acquired from Intellicast: http://www.intellicast.com/. As daylength and temperatures increase the following spring, Junebearing plants stop flower induction and divert resources into flower development (Salisbury and Ross 1992; Nishizawa and Shishido 1998). This inductiontodevelopment shift leads to Junebearing plants bearing high fruit yields until the induced flower buds are depleted, typically in late June or early July. Thus, Junebearing strawberry plants can be considered to have shortday induction requirements and longday development requirements. Under high temperatures (>30 °C), Junebearing plants will experience severely reduced flower development, even in optimal photoperiods (Serce and Hancock 2005). Savini et al. (2005) also noted that the morphology and differentiation time of inflorescences is based on the thermophotoperiod that the plant is exposed to; Junebearing plants growing in warmer, shortday conditions tend to have faster and more prolific flower differentiation and shorter petiole lengths than plants exposed to longday, cooler conditions. Common cultural practices treat Junebearing strawberries as a perennial crop, typically

using a “matted row” system. Rooted plugs of the Junebearing crop are planted in the spring of the first year (the “establishment” year). Flower clusters are typically removed during this entire first season, allowing the plant to divert more reserves into crown/branch crown development, root development, and runner production (EamesSheavly et al. 2003). June bearing cultivars rarely establish runners during early season flower development. However, both flowering and runnering take place as daylength increases, and finally runners alone are developed during the hottest, longest photoperiods of the summer (Stewart and Folta 2010). Growers often arrange runners spatially from the crown to eventually root themselves, creating a thick, matted row of plants (Fig. 1.7) (Archbold and MacKown 1995). The plants then overwinter, and flower clusters induced during the short daylengths of fall are left on the plant the following spring for the first harvest. In this system, the number of leaves on each plant at the beginning of overwintering can be correlated with fruit production the following year (Poling 2012).

Fig. 1.7. Diagram of the mattedrow system common to Junebearing cultivars. 2. Everbearing and Dayneutral.  The second and third flowering types, everbearing and dayneutral, are often considered synonymous, likely due to crossover in pedigrees. Everbearing cultivars include the diploid alpine strawberry F. vesca, along with various more common octoploids (Duchesne 1766; Fletcher 1917). Cultivars categorized as everbearing both induce and develop flowers under longer photoperiods, typically 12 h or more. Sironval and El TannirLomba (1960) found that flower induction and development of F. vesca var. semperflorens was inhibited when plants were exposed to shortday treatments. Octoploid everbearing cultivars initiate most of their

flowers on unrooted or recently rooted runners during the long days of summer, leading to fall harvests (Stewart and Folta 2010). The origin of the everbearing trait appears to have occurred separately in North America and Europe, as little crossbreeding occurred between European everbearing F. vesca and North American everbearing F. virginiana cultivars (Stewart and Folta 2010). The North American everbearing phenotype is due to a single, unstable locus within the typical Junebearing genome (Stewart and Folta 2010), while the origin of the European everbearing trait is older and more difficult to identify (Darrow 1966). The first recorded instance of a dayneutral phenotype was F. virginiana sub. glauca, and this was used as a parent in commercial everbearing breeding programs in the 1930s and 1940s (Darrow 1966). F. vesca may also display dayneutrality (Iwata et al. 2012). Many everbearing cultivars such as ‘Arapahoe’ and ‘Ogallala’ have dayneutral parents present in their pedigrees, which may contribute to why everbearing and dayneutral cultivars are sometimes thought to be the same (Hildreth and Powers 1941). However, true dayneutral cultivars often exhibit flowering habits that are phenotypically distinct from their everbearing relatives. The crowns of all dayneutral genotypes have a strong tendency to fruit proliferously in their first year, as opposed to most everbearing genotypes (Ahmadi and Bringhurst 1991). Dayneutral runners can also develop inflorescences before rooting occurs (Fig. 1.8). Just as important, dayneutral cultivars are historically documented as insensitive to changing photoperiods, fruiting at the same rate throughout a growing season of dynamic daylength (Durner et al. 1984). This distinguishes dayneutral cultivars from everbearing cultivars, which display longday photoperiodism for flower induction and development. These traits, in addition to increased heat tolerance (Stewart and Folta 2010), have contributed to abundant strawberry production in California, where dayneutral cultivars perform well. Other areas of the United States, such as the Upper Midwest, did not observe the same success, as dayneutral cultivars yielded poorly in Midwestern climates and were difficult to propagate (Durner et al. 1984; Luby et al. 1987; Luby 1989). This dayneutral market advantage allows California to account for 44% of the total national strawberry acreage and almost 90% of total yields, leading to a total revenue of $US 2.12 billion in 2012 (California Agric. Statistics Review 2014; National Agric. Statistics Service 2014).

Fig. 1.8. Dayneutral ‘Monterey’ runner, with developed inflorescence. Photograph taken July 10, 2014, in Minnesota.

In environments where they are commercially viable, dayneutral phenotypes are typically

managed as annual plants in raisedbed systems with driptape irrigation and plastic mulch. An abundance of research has been conducted on cultivar/plastic combinations, with the consensus being that yeartoyear and environmental variances across sites complicate the development of a single, optimal cultural practice for dayneutral production (Himelrick et al. 1992; Hughes et al. 2013). Recently, high tunnel structures that increase air and soil temperatures offer season extension potential, and have been shown to increase total and marketable yields in dayneutral strawberry cultivars without pollination being inhibited by the closed structure (Kadir et al. 2006). However, there has been a documented increase in fungal disease incidence in high tunnel systems due to reduced air circulation (Kennedy et al. 2013). It is often considered good horticultural practice to remove flower clusters from Junebearing plants for the first four to six weeks after initial planting (EamesSheavly et al. 2003); this forces the plants to partition more metabolites into vegetative growth and runner production, making the perennial crop more productive in subsequent years. Flower cluster removal is also practiced in dayneutral production, even though dayneutral cultivars are often only grown as annuals. Interestingly, Lantz et al. (2009), when conducting a study in Garrett County, Maryland (39.2833° N), demonstrated no significant difference in total yield when day neutral ‘Seascape’ plants did not have flower clusters removed compared to treatments where flower clusters were removed two and four weeks after planting. 3. Thermophotoperiod and Temperature Effects.  There is still some uncertainty regarding the photoperiodic nature of Junebearing, everbearing and dayneutral flowering habits. While the common consensus is that June bearing cultivars display shortday flower induction, everbearing cultivars display longday flower induction and dayneutral cultivars are truly photoperiod insensitive, additional research has led many to believe that the photoperiodic tendencies of strawberry cultivars can be altered with temperature (Durner et al. 1984; Sonsteby and Heide 2007). In many cases, cultivars classified under photoperiodic categories only display their classified flowering response in moderate temperature conditions; once a certain threshold temperature is exceeded, their photoperiodic nature changes. For example, Guttridge (1985) found that flower induction of certain Junebearing cultivars can occur under any photoperiod if temperatures are 14 h) were not present. This implies that some dayneutral cultivars may display longday flowering habits under hightemperature conditions. Indeed, Sonsteby and Heide (2007) found similar results when testing the cultivar ‘Elan’, leading them to conclude that “…everbearing strawberry cultivars, in general, whether of the older Europeantype or the modern Californiantype originating from crosses with selections of F. virginiana ssp. glauca, are qualitative (obligatory) LD plants at high temperature (27 °C), and quantitative LD plants at intermediate temperatures. Only at temperatures below 10 °C are these cultivars dayneutral.” Such general statements should be avoided, however, since there is considerable variability in strawberry flowering and fruiting response to temperature, even within the Junebearing,

everbearing, and dayneutral categories (Wagstaffe 2009). For example, Bradford et al. (2010) discovered that plants of the dayneutral cultivar ‘Tribute’ required long photoperiods for flowering after a threshold temperature of 26 °C was exceeded, while plants of the day neutral cultivar ‘RH30’ required short photoperiods for flowering once the temperature exceeded 23 °C. This variance of thermophotoperiod within a flowering category suggests that study is merited on all cultivars of commercial significance, even if research has already been conducted on similar cultivars within their traditional photoperiod classification. Temperatures can also affect fruit production in ways that are not related to photoperiod. Kumakura and Shishido (1995) observed that strawberry flower buds of everbearing cultivars aborted during periods of high temperature (30 °C), while Karapatzak et al. (2012) found that everbearing cultivars exposed to supraoptimal temperatures (30 °C/20 °C) experienced severely reduced pollen viability leading to significantly reduced yields. Similar supraoptimal temperature effects were observed with Junebearing cultivars (Ito and Saito 1962; Durner et al. 1984). Yield reductions likely manifest as a result of unviable pollen contributing to poor fertilization and misshapen fruit (Ariza et al. 2011). These reductions in pollen viability appear to be dependent on high night temperatures, as supraoptimal day temperatures with cool night temperatures did not result in reduced viability (Wagstaffe 2009). The effect of supraoptimal temperatures on flowering and yield in dayneutral cultivars is less thoroughly researched, though dayneutral cultivars have previously been regarded as being more heat tolerant (Stewart and Folta 2010). Suboptimal temperatures can also affect fruit development. Ariza et al. (2015) conducted a thorough analysis of cold temperature on differentiating inflorescences, and observed that chilling events (24 h at 2 °C) can reduce pollen grain production and viability as early as 20 days before anthesis, and increase ovule abortion three to six days before anthesis. These events would be especially deleterious for Junebearing plants, as all Junebearing flower buds develop in the spring when chilling events are more likely to occur. A chilling event on dayneutral plants may also inhibit fruit production on developing inflorescences, but since dayneutral plants tend to produce inflorescences throughout the growing season it likely would not have as large as an effect on cumulative yields. In the diploid F. vesca, Davik et al. (2013) observed the accumulation of alcohol dehydrogenase, dehydrins, and galactinol as biomarkers associated with cold tolerance.

III. INFLORESCENCE ARCHITECTURE Strawberry flower cluster anatomy has been thoroughly researched, as possible differences in inflorescence architecture have been hypothesized to correlate with differences in yield and berry weight among cultivars (Webb et al. 1978). Savini et al. (2005) documented the most common flower cluster and inflorescence anatomy in an architectural model, with primary, secondary, and tertiary flowers (Fig. 1.9a). Inflorescences that follow this architectural pattern appear to display two primary internodes leading to the primary flower, secondary branch internodes that form opposite the primary node and lead to secondary flowers, and tertiary internodes that form at the node of secondary branches, leading to tertiary flowers (Fig. 1.10).

Unlike the Savini diagram (Fig. 1.9a), tertiary internodes can grow much longer than secondary internodes, making tertiary flowers appear “ahead” of secondary flowers (Fig. 1.10). Thus, the best way to distinguish secondary flowers from tertiary flowers is to compare differences in flower development; secondary flowers should be further advanced along the development path to mature fruit than tertiary flowers (Fig. 1.10). Interestingly, the formation of inflorescences from new branch crowns after planting follows the same architecture as flowers forming on individual inflorescences, with secondary branch crowns branching from the primary crown, and tertiary branch crowns branching from secondary branch crowns (Fig. 1.9b).

Fig. 1.9. (a) Diagram of a single, typical strawberry inflorescence, and (b) model of ‘Seascape,’ where 1°, 2°, and 3° represent primary, secondary, and tertiary inflorescences, which developed from buds after initial planting. From Savini et al. 2005; reproduced with permission from Taylor & Francis Ltd.

Fig. 1.10. Photograph of ‘Portola’ inflorescence. 1°, 2°, and 3° represent primary, secondary, and tertiary flowers. Labeled brackets indicate primary (I) and secondary (II) internodes. Photograph taken July 15, 2014.

There is, however, observable variability from this typical inflorescence pattern, and of the dayneutral cultivars only ‘Seascape’ inflorescences have been formally documented (Hancock 1999; Savini et al. 2005). Figure 1.11a shows the Junebearing cultivar ‘Annapolis’ displaying the most typical inflorescence architecture, with dayneutral cultivars ‘Albion’ and ‘Seascape’ displaying similar habits (Figs 1.11b, g and h). ‘Monterey’ and ‘San Andreas’ inflorescences will sometimes only form a single secondary branch (Figs 1.11d and f). Occasionally, more developed inflorescences displaying this habit will create additional secondary branches, but these branches display an alternate growth habit, as opposed to the opposite secondary branching pattern of the more documented habit typical in ‘Seascape.’ The inflorescences of ‘Evie2’ and ‘Portola’ sometimes appear to form two separate primary branches, forking off the first node (Figs 1.11c and e). Interestingly, ‘Monterey’ and ‘San Andreas,’ whose inflorescences typically only form a single secondary branch, were also the two lowestyielding cultivars in 2013 University of Minnesota trials of dayneutral cultivars, while ‘Evie2’ and ‘Portola,’ which seem to produce two primary branches, were the highestyielding (Petran et al. 2016). While causation cannot be applied, these findings do raise the question of inflorescence architecture/yield relationships for further research.

Fig. 1.11. Selected flower clusters of (a) ‘Annapolis,’ (b) ‘Albion,’ (c) ‘Evie2,’ (d) ‘Monterey,’ (e) ‘Portola,’ (f) ‘San Andreas,’ and (g, h) ‘Seascape.’ Photographs taken July 15, 2014.

IV. GENETICS OF FLOWER INDUCTION The underlying genetics that promote or inhibit flowering is complex and debated, and in order to appreciate that complexity in strawberry, the history of genetic flowering research in general can provide some background. The idea of florigen, a plant hormone (or family of hormones) responsible for flower initiation and development in all flowering plant species, was proposed by Chailakhyan (1936) after a series of grafting experiments. A quest to isolate and identify the florigen hormone took place thereafter, spanning the rest of the 20th century (Zeevaart 2006).

The existence of florigen as a universal floral initiator was doubted after genetic research discovered multiple distinct flowering pathways in different species, but this dissonance was resolved after it was seen that the each separate pathway converged to a shared set of flower promoting genes, the most wellknown being FLOWERING LOCUS T (FT) (Koornneef et al. 1991; Samach et al. 2000; Simpson and Dean 2002; Putterill et al. 2004). Thus, florigens are indeed understood to be universal flowering inducers, but the production of florigen hormones is regulated by single genes in certain species and is polygenic in others. The protein produced by FT, now considered to be a florigen hormone, travels through the phloem to the shoot apical meristem and interacts with other proteins already present in the meristem to induce flower differentiation (Abe et al. 2005; Notaguchi et al. 2008). Lifschitz et al. (2014) emphasized that, in addition to florigen, there are agents that act antagonistically to this pathway, known as antiflorigens . The protein of TERMINATION FLOWER 1 (TFL1) in the model plant Arabidopsis thaliana is an antiflorigen that suppresses termination of the inflorescence and maintains vegetative growth in shoot meristems (Bradley et al. 1997). Lifschitz et al. (2014) proposed that not only florigens but also the florigen: antiflorigen ratio determines flowering times in shortday, longday, and dayneutral flowering plants. Gene pathways that regulate florigen and antiflorigen production in strawberry must be discovered and understood before breeding efforts can be refined to select specifically for dayneutral habits. While previous reviews have stated that no genes involved in strawberry flowering have been identified (Darnell et al. 2003), subsequent research has proposed specific floral promoters and suppressors within the strawberry genome. These discoveries are first reviewed within the diploid F. vesca genome, followed by the octoploid genome of commercial F. × ananassa.

A. Fragaria vesca Fragaria vesca, commonly known as the woodland strawberry, shares morphologic and genotypic similarities with the more commercially significant, F. × ananassa. Despite its smaller size, the petal and sepal number, inflorescence architecture and early fruit development of F. vesca are notably similar to those of F. × ananassa, and the two species share a high degree of colinearity between their genomes (RousseauGueutin et al. 2008; Hollender et al. 2012). Hollender et al. (2012) performed a thorough analysis of F. vesca floral development (Fig. 1.12), linking the genetic similarities to F. × ananassa with documentation of morphologic similarities as well.

Fig. 1.12. Fragaria vesca shoot and flower development. (a) F. vesca YW5AF7 grown in a 10.2cm pot; (b) YW5AF7 dichasial cyme bearing yellow berries; (c) Inflorescence with primary flower (1) and two developing secondary flowers (2). Young tertiary buds (arrows) are present beneath the secondary flower buds; (d) Diagram of shoot architecture. Numbers indicate primary, secondary, and tertiary flower buds; (e) Diagram illustrating floral organ arrangement. The two outer whorls are concentric rings of five bracts (b) alternating with five sepals (s). The third whorl consists of five white petals (p). Interior to the petals are two whorls of stamens. Stamens are arranged in a repeating pattern of five tall (T) and five short (S) in the inner whorl and 10 medium length (M) in the outer whorl. The center circle indicates a receptacle topped with numerous, spirally arranged carpels; (f) Scanning electron micrograph (SEM) of a developing floral bud, illustrating spirally arranged carpel primordial; (g) Abaxial view of a typical F. vesca flower with five narrow bracts (b) alternating with five wider sepals (s); (h) Adaxial view of typical F. vesca flower illustrating a whorl of five white petals, two whorls of ten stamens each, and an apocarpous gynoecium with ~160 pistils. (i) Dissected flower illustrating the “S, M, T, M, S” stamen pattern. Scale bars: (a) 2 cm; (g–i) 1  mm. From Hollender et al. 2012.

To date, most strawberry genetic research has focused on F. vesca because of its relatively simple diploid genome compared to the octoploid F. × ananassa (Slovin and Michael 2011). F. vesca has an appealingly small genome size (~240 MB; x = 7), with a recently published reference genome (Shulaev et al. 2011) and several cultivars that have had success with Agrobacteriummediated transformation (Folta and Dhingra 2006). Research within F. vesca has been undertaken with the belief that any genetic discoveries within F. vesca could be used as models for subsequent research within F. × ananassa (Weebadde et al. 2007; Hollender et al. 2012). Among the most commercially desired discoveries is that of flowering habit. Since F. vesca also contains multiple lines and cultivars with dayneutral flowering habits, it has been long hoped that understanding the genetic precursor to dayneutrality in F. vesca would serve as a springboard for similar discoveries in F. × ananassa; such a discovery would be highly beneficial to breeders looking to ensure a stable dayneutral habit in potential cultivar releases. A genetic model for flowering habit in F. vesca was proposed by Brown and Wareing (1965) when F1 and backcrossed progeny of seasonal × perpetual flowering parents segregated into 9:3:3:1 and 1:1:1:1 ratios, respectively. Such results indicated that a single gene controlled F. vesca flowering habit, with seasonal flowering as the dominant phenotype. Cekic et al. (2001) went further, using a type of microsatellite analysis (ISSRPCR primer pair combinations) to identify two markers specific to F. vesca located near a single seasonal flowering locus. Finally, Iwata et al. (2012) discovered a specific antiflorigen TFL1 homolog, KOUSHIN (KSN) within the genome of F. vesca, naming it FvKSN. A 2bp deletion in the first exon of the FvKSN allele causes a frame shift, leading to a nonfunctional mutant ksn. All F. vesca strawberries displaying a continuous flowering habit were homozygous ksn/ksn, while once seasonal flowering habits were either KSN/ksn or KSN/KSN. These results coincided with the findings of Koskela et al. (2012) that a TFL1 homolog, which they named FvTFL1, served as

an antiflorigen, expressing itself in longday photoperiods. Transgenically silenced FvTFL1 lines and lines overexpressing a mutated FvTFL1 with a 2bp deletion both displayed day neutrality, supporting the findings of Iwata et al. (2012) (Fig. 1.13). It was later found that FvSOC1 activates the FvTFL1 antiflorigen gene in the shoot apex during longday conditions; thus, FvSOC1 is currently alleged as the photoperiod control center of floral differentiation in F. vesca (Mouhu et al. 2013).

Fig. 1.13. Silencing of FvTFL1 leads to daylengthindependent flowering. (a) Phenotypes of FvTFL1 RNAi silencing and overexpression lines in the shortday (SD) F. vesca background. Clonally propagated plants (runner cuttings) of SD F. vesca and P35S:FvTFL1RNAi1 and P35S:FvTFL11 lines were subjected to SD induction treatment for four weeks followed by longdays (LDs; left), or grown continuously under LDs (right); (b), Flowering time of SD F. vesca and P35S:FvTFL1RNAi and P35S:FvTFL1 plants (RNAi and OX, respectively) in SDs and LDs. Flowering time is indicated as days to anthesis from the beginning of the treatments. Treatments and plant materials were as described in (a). Values indicate mean ±  SD. n = 4 (OX1), n = 5 (RNAi2), n = 6 (RNAi1 and OX2), and n = 7 (SD F. vesca); (c,d) Expression of FvTFL1 (c) and FvAP1 (d) in the apices of two independent P35S:FvTFL1RNAi (RNAi) lines. Values indicate mean ± SD. n = 3 (RNAi1 and SD F. vesca) or n = 2 (RNAi2). From Koskela et al. 2012; © American Society of Plant Biologists, www.plantphysiol.org.

B. Fragaria × ananassa F. × ananassa contains a complex octoploid genome, yet progress has been made in identifying specific florigens and antiflorigens within the species. When examining the Junebearing cultivar ‘Nyoho,’ Nakano et al. (2015) observed mRNA accumulation of the gene FaFT3 in the shoot tip under shortday (8 h), cooltemperature (13 °C) conditions just prior to the induction of a floral meristem identity gene, FaAP1. The protein of FaFT2 was found in the flower bud shortly thereafter, and both of these appeared to act antagonistically to FaFT1, which accumulated in plant leaves under longday (16 h) conditions. FaFT1 was also associated with the upregulation of FaTFL1 in shoot tips in warmer temperatures (27 °C). From these observations, Nakano et al. (2015) proposed that FaFT1, FaFT2 and FaFT3, all of which contain amino acid residues similar to the florigen FT in A. thaliana, work to regulate flowering in Junebearing F. × ananassa, with FaFT2 and FaFT3 functioning as florigens in shortday, cooler temperatures, and FaFT1 being a precursor to FaTFL1 antiflorigen production in longday, warmer temperatures. Further work should be conducted to analyze how these genes and proteins interact in dayneutral F. × ananassa. While Nakano et al. (2015) proposed homologs of the FaFT proteins in F. vesca, their studies did little to elucidate the cause of flower habit differentiation in F. × ananassa and its relation to flower habit control in F. vesca, because it only focused on the Junebearing cultivar ‘Nyoho.’ In fact, despite its morphologic and genetic similarities, a preponderance of research has indicated that the singlegene models of flowering habit in F. vesca cannot be directly applied to dayneutrality in the more complex octoploid F. × ananassa genome. Inconsistent inheritance ratios among different field environments imply that dayneutrality may involve auxiliary genes interacting with a single, dominant locus in octoploid strawberries (Serçe and Hancock 2005; Castro et al. 2015). While this complexity certainly makes breeding for specific flowering habits more difficult, it is not a new phenomenon, as flowering is a polygenic trait in many other plant species (Hayama and Coupland 2004; Esumi et al. 2005). In an effort to map the polygenetic sources of dayneutrality in F. × ananassa, Weebadde et al. (2007) conducted an amplified fragment length polymorphism (AFLP) marker analysis of the progeny of dayneutral ‘Tribute’ and Junebearing ‘Honeoye’ in five states throughout the United States. They found several (more than five) quantitative trait loci (QTLs) that were correlated with phenotypic variation in flower habit. Interestingly, no single QTL explained more than 36% of this variation, and the ability of a QTL to explain variation changed based on the location in which the cultivars were grown. For example, a QTL on LG 28 was a significant predictor of flower variation in plants grown in the hotter climates of every central and eastern state (Minnesota, Michigan, and Maryland), but the same QTL could not predict variation in California, where average temperatures are milder. Such a finding implied that not only may flower habit be a polygenic trait in F. × ananassa, but the genetic requirements for dayneutrality change based on other environmental factors. Weebadde et al. (2007) postulated that a minimum threshold of genes favoring dayneutrality must be present in order to achieve the habit, and the strength of a gene to contribute to reaching that threshold – that is, the necessity of a gene that confers heat tolerance in hotter climates – is dependent on its

location. This genetic × environmental (GxE) theory helps understand previous research that found the photoperiodic nature of F. × ananassa to change based on temperature and light conditions (Guttridge 1985; Sonsteby and Heide 2007). Research investigating whether markers associated with flowering habit in F. vesca colocalize with flowering habit markers in F. × ananassa has yet to be conducted. There is still some debate regarding the polygenic control of flowering habit in F. × ananassa. In a recent study, Gaston et al. (2013) used marker analysis to find a QTL named FaPFRU that they identified as the major controlling locus of dayneutrality and runner production in F. × ananassa. The study found dayneutrality to be dominant over seasonal flowering, suggesting that the genetic basis for flowering in F. × ananassa was not inherited from F. vesca, where seasonal flowering is dominant over dayneutrality. FaPFRU also appeared to control vegetative development, a discovery different from the locus controlling flower development in F. vesca, which does not control vegetative growth (Koskela et al. 2012). While acknowledging the polygenic evidence provided by previous research (Hancock et al. 2002; Weebadde et al. 2007), Gaston et al. (2013) still proposed dayneutrality in F. × ananassa to be under singlelocus control by FaPFRU. Indeed, FaPFRU, located on the IVbf linkage group explained up to 59.3% of phenotypic variation in flower habit in that study (Table 1.1). Honjo et al. (2016) went on to affirm the findings of Gaston et al. (2013) while proposing their FxaACA02I08C marker, also located on LG IV, to be an accurate marker of day neutrality. Table 1.1. Map positions and genetic effect of QTL detected for the PF (perpetual flowering) and RU (runnering) traits for the female (f) parent ‘Capitola’ and male (m) parent ‘CF1116.’ QTL identification was based on compositeinterval mapping analysis with LOD > LOD threshold (3.1) (α = 0.05). From Gaston et al. 2013; reproduced with permission from Oxford University Press. Trait Linkage QTL group name a PF Idm PF LGId m Idm PF LGId m Idm PF LGId m IVbf PF LGIVb f IVdm PF

Yearb

Positione

2002

No. Markerd QTLc 1 ccta185

2002

1

gata148

21.11

2007

1

ccaa267

43.71

2002/2003/2003b/2004/2005/2007 6

gatt284

2.0/4.0

2002/2003

tgac408/tgta115 60.06/55.57

2

6.01

Vaf

RU

LGIVd m PF LGVaf

2007

1

u009180

20.01

vIcf

PF LGVIc f

2007

1

v013200

84.47

IIcm

RU LGIIC m RU LGIIIb m RU LGIIId m RU LGIVb f

2005

1

gctg207

54.59

2002

1

gtag280

54.86

2002

1

tgac108

15.51

2002/2003/2005

3

gatt284

0.0/2.0

IIIbm

IIIdm

Ivbf

a Name of the unique QTL. b Year of observation of the QTL. c Number of significant QTL. d The left marker associated with the QTL is indicated. e Position indicates the distance in cM of the QTL from the top of the chromosome. f LOD is the loglikelihood at that position. g r2 is the percentage of phenotypic variation explained by the QTL. h Mean effect on a trait mean value of the presence of one allele at a marker by comparison with the presence of the second allele. + and − indicate the direction of the additive effect. A positive effect means a higher value for the Capitola allele on the female map or a higher value for CF1116 on the male map.

The dissonance regarding genetic control of F. × ananassa flower habit in the current literature may be due to the GxE interaction effects described previously. The studies by Gaston et al. (2013) and Honjo et al. (2016) were each conducted in only one location (Villenave d’Ornon, France, 44.7806° N, 0.5658° W and Morioka City, Japan, 39.46° N, 141.8° E, respectively) and, while FaPFRU explained a majority of the phenotypic variation in the study of Gaston et al. (2013), the effect of this gene may be lessened in regions with different environmental conditions, as was observed with the LG 28 QTL by Weebadde et al. (2007). Before FaPFRU can be universally accepted as the singlecontrol locus of flowering habit in F. × ananassa, its predictive strength would also have to be tested in areas such as

Maryland and coastal California, where environmental pressures against dayneutrality appear to be high and low, respectively (Hancock et al. 2002; Weebadde et al. 2007). If FaPFRU remained a strong predictor of phenotypic variation in flower habit in each environment, the claim of Gaston et al. (2013) that FaPFRU is the major controlling locus of flower habit in F. × ananassa would be strengthened. Indeed, such a project was conducted by Castro et al. (2015). Crosses of dayneutral ‘Tribute’ and Junebearing ‘Honeoye’ were qualitatively and quantitatively scored for day neutrality in five different states in the United States with contrasting temperature conditions: Beltsville, Maryland (MD), East Lancing, Michigan (MI) and Saint Paul, Minnesota (MN), where average maximum summer temperatures are at least 28 °C; and Watsonville, California (CA) and Corvallis, Oregon (OR), where average maximum temperatures are at 22 °C and 26 °C, respectively (National Centers for Environmental Information, 2016). The authors also used 267 molecular markers in an attempt to construct a map of linkage groups associated with flower habit in each environment. As expected, even with the same parents, a higher proportion of progeny were scored as dayneutral in cooler CA than in hotter MD, reinforcing the role of GxE interaction in flower habit (Fig. 1.14). Similar to the findings of Gaston et al. (2013), the dayneutral:Junebearing inheritance ratio was 1:1 in the hotter MD, MI, and MN environments, suggesting singlelocus control. However ratios were 3:1 in CA and 5:3 in OR, implying the presence of multiple alleles at one or more loci that only express themselves in cooler temperatures.

Fig. 1.14. Frequency distribution of number of weeks flowering at Maryland (MD) (a) and California (CA) (b). The phenotypic values of the parents, ‘Tribute’ and ‘Honeoye’ are indicated by arrows (Castro et al. 2015). Through their marker analysis, Castro et al. (2015) found that dayneutrality could be mapped to a specific genetic region on linkage group (LG) IVT1 of the ‘Tribute’ map – a region with multiple QTLs – regardless of environment (Fig. 1.15). In MD, a QTL on LG IVT1 explained between 63.9% and 63.6% of the phenotypic variation, while in CA a separate QTL on the same LG explained between 51.2% and 50.1% of the phenotypic variation. These results imply that while dayneutrality may be under singlelocus control in certain environments, multiple loci within the same genetic region may be expressed in cooler environments, and “…the distortion toward dayneutral progeny found in OR and CA presents a slight challenge to the singlelocus theory and should be remembered in future research” (Castro et al. 2015). The singlelocus control apparent in warmer regions seems to mirror the genetic findings of Gaston et al. (2013) and Honjo et al. (2016), but the QTL found on LG IV T1 could not be confirmed as FaPFRU because the studies did not use common markers to identify QTLs.

Fig. 1.15. Quantitative trait loci (QTL) for number of weeks of flowering detected in the octoploid strawberry (F. × ananassa Duchesne ex. Rozier) population ‘Tribute’ × ‘Honeoye’ evaluated in Maryland (dotted line) and California (solid line) using interval mapping (IM) (Castro et al. 2015). Taken together, research on the genetics of strawberry flower induction has shown F. × ananassa to contain more complexity than F. vesca. Control of dayneutrality in F. vesca is

almost certainly achieved by a single locus (Cekic et al. 2001; Iwata et al. 2012; Koskela et al. 2012), and while the inheritance of flower habit in F. × ananassa is more uncertain, it appears that GxE interactions lead to singlelocus control in hotter regions and potential multilocus control (albeit on the same linkage group) in cooler regions (Weebadde et al. 2007; Gaston et al. 2013; Castro et al. 2015; Sooriyapathirana et al. 2015). While the understanding of induction in F. vesca was proposed as a springboard for induction research and consequent breeding efforts in F. × ananassa (Hollender et al. 2012), the support for this proposal remains unrealized; in fact, recent evidence suggests that the genetics of flower induction in F. vesca and F. × ananassa are not orthologous (Gaston et al. 2013). Regardless of this relation, continued research into the inheritance of dayneutrality in commercial F. × ananassa will assist breeders in releasing cultivars with extended yields in diverse environments throughout the world.

V. CONCLUSIONS The strawberry is an extensively documented horticultural crop. There appears to be adequate information regarding the growth and reproduction habits of commercial F. × ananassa and its progenitors F. vesca, F. virginiana, and F. chiloensis. Within the literature there is a focus on F. × ananassa flowering habit, as this response represents the aspect of growth most commercially significant to strawberry growers. The three categories of flowering habit – Junebearing, everbearing and dayneutral – each respond differently to photoperiod and have separate cultural production practices endemic to the environments in which they are most easily grown. The dayneutral phenotype is the most desired because it offers the potential for extended seasons and increased yields, but until recently commercial dayneutral production has been limited to only a few regions such as Mexico, and California and Florida within the United States. This is apparently due to older cultivars having a narrow set of environmental tolerances, but newer cultivars are showing promise to increase the commercial range of dayneutral production (Petran et al. 2016). Despite its value and popularity, little progress has been made in mapping the genetic basis of dayneutrality in F. × ananassa until recently. This is due to its complex octoploid genome and a history of inconsistent inheritance ratios even within traditional breeding efforts. Because of these difficulties, most strawberry genetic research has focused on F. vesca, which features a simpler diploid genome, inheritance ratios that implied singlegene control of flower habit, and an assumption that any discoveries made within this species could be used as a springboard for genetic research in F. × ananassa. Indeed, it was discovered that a 2bp deletion in a homolog of the antiflorigen gene TFL1 resulted in dayneutrality in F. vesca (Iwata et al. 2012; Koskela et al. 2012). Recently progress has been made in the mapping of dayneutrality in F. × ananassa, though it has little to do with any discoveries made within F. vesca; in fact, it has been proposed that genes controlling flower habit in the two species are not orthologous (Gaston et al. 2013). It appears that dayneutrality in F. × ananassa is influenced by GxE interactions, with single locus control taking effect in regions where average maximum temperatures are above 28 °C,

and potential multilocus control in cooler regions (Castro et al. 2015). This may explain observed higher proportions of progeny displaying dayneutrality in milder versus hotter climates, even when parents are the same (Weebadde et al. 2007). While no research has discovered common markers between the flowering loci of F. vesca and F. × ananassa, the results of the induction research covered in this review imply that such work may not be as useful as previously believed. Instead, efforts may be better focused on investing a higher density of markers in linkage group IVT1 of the F. × ananassa genome, where QTLs determining dayneutrality appear to be focused. Designing markers for targeted genome regions will be made easier now that a reference genome for octoploid F. × ananassa has been released (Hirakawa et al. 2014), and will allow researchers to determine if markers highly correlated with dayneutrality in ‘Tribute’ will explain the habit in other parental lines as well. Despite its complexities, recent advances in understanding F. × ananassa flower habit have been promising, and increase the potential for the production of this valuable crop to be commercially viable throughout the world.

LITERATURE CITED Abe, M., Y. Kobayashi, S. Yamamoto, Y. Daimon, A. Yamaguchi, Y. Ikeda, H. Ichinoki, M. Notaguchi, K. Goto, and T. Araki. 2005. FD, a bZIP protein mediating signals from the floral pathway integrator FT at the shoot apex. Science 309(5737):1052–1056. Ahmadi, H. and R.S. Bringhurst. 1991. Genetics of sex expression in Fragaria species. Am. J. Bot.:504–514. Archbold, D.D. and C.T. MacKown. 1995. Seasonal and cropping effects on total and fertilizer nitrogen use in Junebearing and dayneutral strawberries. J. Am. Soc. Hortic. Sci. 120(3):403–408. Ariza, M.T., C. Soria, and E. MartinezFerri. 2015. Developmental stages of cultivated strawberry flowers in relation to chilling sensitivity. AoB Plants 7:10.1093/aobpla/plv012. Ariza, M., C. Soria, J. Medina, and E. MartínezFerri. 2011. Fruit misshapen in strawberry cultivars (Fragaria × ananassa) is related to achenes functionality. Ann. Appl. Biol. 158(1):130–138. Bradford, E., J.F. Hancock, and R.M. Warner. 2010. Interactions of temperature and photoperiod determine expression of repeat flowering in strawberry. J. Am. Soc. Hortic. Sci. 135(2):102–107. Bradley, D., O. Ratcliffe, C. Vincent, R. Carpenter, and E. Coen. 1997. Inflorescence commitment and architecture in Arabidopsis. Science 275(5296):80–83. Brown, T. and P.F. Wareing. 1965. The genetical control of the everbearing habit and three other characters in varieties of Fragaria vesca. Euphytica 14(1):97.

California Agricultural Statistics Review. 2014. Fruit and Nut Crops. California Department of Food and Agriculture, Sacramento, California. Castro, P., J. Bushakra, P. Stewart, C. Weebadde, D. Wang, J. Hancock, C. Finn, J. Luby, and K. Lewers. 2015. Genetic mapping of dayneutrality in cultivated strawberry. Mol. Breed. 35(2):1–16. Cekic, C., N.H. Battey, and M.J. Wilkinson. 2001. The potential of ISSRPCR primerpair combinations for genetic linkage analysis using the seasonal flowering locus in Fragaria as a model. Theor. Appl. Genet. 103(4):540. Chailakhyan, M.K. 1936. About the mechanism of the photoperiodic response. Dokl. Akad. Nauk SSSR (1):85–89. Darnell, R.L., D.J. Cantliffe, D.S. Kirschbaum, and C.K. Chandler. 2003. The physiology of flowering in strawberry. Hortic. Rev. 28:325–349. Darrow, G.M. 1966. The strawberry. History, breeding and physiology. Holt, Rinehart and Winston, New York, New York, USA. Darrow, G.M. and G.F. Waldo. 1933. Photoperiodism as a cause of the rest period in strawberries. Science 77(1997):353–354. Davik, J., G. Koehler, B. From, T. Torp, J. Rohloff, P. Eidem, R.C. Wilson, A. Sønsteby, S.K. Randall, and M. Alsheikh. 2013. Dehydrin, alcohol dehydrogenase, and central metabolite levels are associated with cold tolerance in diploid strawberry (Fragaria spp.). Planta 237(1):265–277. Demchak, K. 2010. The midAtlantic berry guide 2010–2011. Online at http://pubs.cas.psu.edu/freepubs/pdfs/AGRS097.pdf Duchesne, A.N. 1766. Histoire naturelle des Fraisiers. Didot le jeune. Durner, E.F. 2015. Photoperiod affects floral ontogeny in strawberry (Fragaria × ananassa Duch.) plug plants. Scientia Hortic. 194:154–159. Durner, E.F., J. Barden, D. Himelrick, and E. Poling. 1984. Photoperiod and temperature effects on flower and runner development in dayneutral, Junebearing, and everbearing strawberries. J. Am. Soc. Hortic. Sci. 109:396–400. EamesSheavly, M., M. Pritts, C. Cramer, L. Bushway, I. Merwin, R. Reisinger, and S. McKay. 2003. Cornell guide to growing fruit at home. Cornell Cooperative Extension, Ithaca, New York. Esumi, T., R. Tao, and K. Yonemori. 2005. Isolation of LEAFY and TERMINAL FLOWER 1 homologues from six fruit tree species in the subfamily Maloideae of the Rosaceae. Sexual Plant Reprod. 17(6):277–287.

Fait, A., K. Hanhineva, R. Beleggia, N. Dai, I. Rogachev, V.J. Nikiforova, A.R. Fernie, and A. Aharoni. 2008. Reconfiguration of the achene and receptacle metabolic networks during strawberry fruit development. Plant Physiol. 148(2):730–750. Fletcher, S.W. 1917. The strawberry in North America: history, origin, botany, and breeding. The Macmillan Company. Folta, K.M. and A. Dhingra. 2006. Invited review: Transformation of strawberry: the basis for translational genomics in Rosaceae. In Vitro Cell. Dev. Biol.Plant 42(6):482–490. Gaston, A., J. Perrotte, E. LerceteauKohler, M. RousseauGueutin, A. Petit, M. Hernould, C. Rothan, and B. Denoyes. 2013. PFRU, a single dominant locus regulates the balance between sexual and asexual plant reproduction in cultivated strawberry. J. Exp. Bot. 64(7):1837–1848. Guttridge, C. 1985. Fragaria × ananassa. CRC Handbook of Flowering: CRC Press, Inc., Boca Raton, Florida:16–33. Hancock, J.F. 1999. Strawberries. CABI Publishing. Wallingford, UK. Hancock, J., J. Luby, A. Dale, P. Callow, S. Serce, and A. ElShiek. 2002. Utilizing wild Fragaria virginiana in strawberry cultivar development: inheritance of photoperiod sensitivity, fruit size, gender, female fertility and disease resistance. Euphytica 126(2):177– 184. Hayama, R. and G. Coupland. 2004. The molecular basis of diversity in the photoperiodic flowering responses of Arabidopsis and rice. Plant Physiol. 135(2):677–684. Hildreth, A. and L. Powers. 1941. The Rocky Mountain strawberry as a source of hardiness. Proc. Am. Soc. Hortic. Sci. 38:410–412. Himelrick, D.G., W. DozierJr, and J. Akridge. 1992. Effect of mulch type in annual hill strawberry plasticulture systems. II Intl. Strawberry Symp. 348. Hirakawa, H., K. Shirasawa, S. Kosugi, K. Tashiro, S. Nakayama, M. Yamada, M. Kohara, A. Watanabe, Y. Kishida, T. Fujishiro, H. Tsuruoka, C. Minami, S. Sasamoto, M. Kato, K. Nanri, A. Komaki, T. Yanagi, Q. Guoxin, F. Maeda, M. Ishikawa, S. Kuhara, S. Sato, S. Tabata, and S.N. Isobe. 2014. Dissection of the octoploid strawberry genome by deep sequencing of the genomes of Fragaria species. DNA Res. 21(2):169–181. Hollender, C.A., A.C. Geretz, J.P. Slovin, and Z. Liu. 2012. Flower and early fruit development in a diploid strawberry, Fragaria vesca. Planta 235(6):1123–1139. Honjo, M., T. Nunome, S. Kataoka, T. Yano, M. Hamano, H. Yamazaki, T. Yamamoto, M. Morishita, and S. Yui. 2016. Simple sequence repeat markers linked to the everbearing flowering gene in longday and dayneutral cultivars of the octoploid cultivated strawberry Fragaria × ananassa. Euphytica 209(2):291.

Hughes, B.R., J. Zandstra, and A. Dale. 2013. Effects of mulch types on dayneutral strawberry production in three distinct environments in Ontario. Intl. J. Fruit Sci. 13(12):255–263. Ito, H. and T. Saito. 1962. Studies on the flower formation in the strawberry plants I. Effects of temperature and photoperiod on the flower formation. Tohoku J. Agric. Res. 13:191–203. Iwata, H., A. Gaston, A. Remay, T. Thouroude, J. Jeauffre, K. Kawamura, L.H. Oyant, T. Araki, B. Denoyes, and F. Foucher. 2012. The TFL1 homologue KSN is a regulator of continuous flowering in rose and strawberry. The Plant J. 69(1):116–125. Kadir, S., E. Carey, and S. Ennahli. 2006. Influence of high tunnel and field conditions on strawberry growth and development. HortScience 41(2):329–335. Kang, C., O. Darwish, A. Geretz, R. Shahan, N. Alkharouf, and Z. Liu. 2013. Genomescale transcriptomic insights into earlystage fruit development in woodland strawberry Fragaria vesca. Plant Cell 25(6):1960–1978. Karapatzak, E., A. Wagstaffe, P. Hadley, and N. Battey. 2012. Hightemperatureinduced reductions in cropping in everbearing strawberries (Fragaria × ananassa) are associated with reduced pollen performance. Ann. Appl. Biol. 161(3):255–265. Kennedy, C., T.N. Hasing, N.A. Peres, and V.M. Whitaker. 2013. Evaluation of strawberry species and cultivars for powdery mildew resistance in openfield and high tunnel production systems. HortScience 48(9):1125–1129. Koornneef, M., C. Hanhart, and J. Van der Veen. 1991. A genetic and physiological analysis of late flowering mutants in Arabidopsis thaliana. Mol. Gen. Genet. 229(1):57–66. Koskela, E.A., K. Mouhu, M.C. Albani, T. Kurokura, M. Rantanen, D.J. Sargent, N.H. Battey, G. Coupland, P. Elomaa, and T. Hytonen. 2012. Mutation in TERMINAL FLOWER1 reverses the photoperiodic requirement for flowering in the wild strawberry Fragaria vesca. Plant Physiol. 159(3):1043–1054. Kronenberg, H.G., L. Wassenaar, and C. Van De Lindeloof. 1976. Effect of temperature on dormancy in strawberry. Scientia Hortic. 4(4):361–366. Kumakura, H. and Y. Shishido. 1995. Effect of temperature and photoperiod on flower bud initiation in everbearing type strawberry [Fragaria ananassa] cultivars. J. Jpn. Soc. Hortic. Sci. 64:85–94. Lantz, W., S. Frick, and H. Swartz. 2009. Optimizing day neutral strawberry growth. J. NACAA 2(1). . Lifschitz, E., B.G. Ayre, and Y. Eshed. 2014. Florigen and antiflorigen – a systemic mechanism for coordinating growth and termination in flowering plants. Front. Plant Sci. 5:465. Luby, J. 1989. Dayneutral strawberries. Northland Berry News 3(1):10.

Luby, J., E. Hoover, S. Munson, D. Bedford, C. Rosen, D. Wildung, and W. Gray. 1987. Day neutral strawberry production cultivars and production systems. Proc. Mich. State Hortic. Soc:144. Mouhu, K., T. Kurokura, E.A. Koskela, V.A. Albert, P. Elomaa, and T. Hytonen. 2013. The Fragaria vesca homolog of suppressor of overexpression of constans1 represses flowering and promotes vegetative growth. Plant Cell 25(9):3296–3310. Nakano, Y., Y. Higuchi, Y. Yoshida, and T. Hisamatsu. 2015. Environmental responses of the FT/TFL1 gene family and their involvement in flower induction in Fragaria × ananassa. J. Plant Physiol. 177:60–66. National Agriculture Statistics Service. 2014. Census of Agriculture. United States Department of Agriculture. National Centers for Environmental Information. 2016. Climate Data Online Database. National Oceanic and Atmospheric Administration, U.S. Department of Commerce, Washington, D.C. December 15, 2016. . Nishiyama, M. and K. Kanahama. 2000. Effects of temperature and photoperiod on flower bud initiation of dayneutral and everbearing strawberries. IV Intl. Strawberry Symp. 567:253– 255. Nishizawa, T. and Y. Shishido. 1998. Changes in sugar and starch concentrations of forced Junebearing strawberry plants as influenced by fruiting. J. Am. Soc. Hortic. Sci. 123(1):52– 55. Notaguchi, M., M. Abe, T. Kimura, Y. Daimon, T. Kobayashi, A. Yamaguchi, Y. Tomita, K. Dohi, M. Mori, and T. Araki. 2008. Longdistance, grafttransmissible action of Arabidopsis FLOWERING LOCUS T protein to promote flowering. Plant Cell Physiol. 49(11):1645–1658. Petran, A., E. Hoover, L. Hayes, and S. Poppe. 2016. Yield and quality characteristics of day neutral strawberry in the United States Upper Midwest using organic practices. Biol. Agric. Hortic.:1–16. http://dx.doi.org/10.1080/01448765.2016.1188152. Poling, E. 2012. Strawberry plant structure and growth habit. Proceedings from the EXPO Berry Sessions. Cornell University, New York. http://www.hort.cornell.edu/grower/nybga/pdfs/2012berryproceedings.pdf Putterill, J., R. Laurie, and R. Macknight. 2004. It’s time to flower: the genetic control of flowering time. BioEssays 26(4):363–373. RousseauGueutin, M., E. LerceteauKohler, L. Barrot, D.J. Sargent, A. Monfort, D. Simpson, P. Arus, G. Guerin, and B. DenoyesRothan. 2008. Comparative genetic mapping between octoploid and diploid Fragaria species reveals a high level of colinearity between their genomes and the essentially disomic behavior of the cultivated octoploid strawberry.

Genetics 179(4):2045–2060. Salisbury, F.B. and C.W. Ross. 1992. Plant physiology. 4th. edn. Belmont, CA. Wadsworth. Samach, A., H. Onouchi, S.E. Gold, G.S. Ditta, Z. SchwarzSommer, M.F. Yanofsky, and G. Coupland. 2000. Distinct roles of CONSTANS target genes in reproductive development of Arabidopsis. Science 288(5471):1613–1616. Savini, G., D. Neri, F. Zucconi, and N. Sugiyama. 2005. Strawberry growth and flowering: an architectural model. Intl. J. Fruit Sci. 5(1):29–50. Serçe, S. and J.F. Hancock. 2005. The temperature and photoperiod regulation of flowering and runnering in the strawberries, Fragaria chiloensis, F. virginiana, and F. × ananassa. Scientia Hortic. 103(2):167–177. Shulaev, V., D.J. Sargent, R.N. Crowhurst, T.C. Mockler, O. Folkerts, A.L. Delcher, P. Jaiswal, K. Mockaitis, A. Liston, and S.P. Mane. 2011. The genome of woodland strawberry (Fragaria vesca). Nat. Genet. 43(2):109–116. Simpson, G.G. and C. Dean. 2002. Arabidopsis, the Rosetta stone of flowering time? Science 296(5566):285–289. Sironval, C. and J. El TannirLomba. 1960. Vitamin E and flowering of Fragaria vesca L. var. semperflorens Duch. Nature 185:855–856. Slovin, J. and T. Michael. 2011. Strawberry Part 3 – structural and functional genomics. Genetics, genomics and breeding of berries. Science Publishers, Enfield: 240–308. Sonsteby, A. and O. Heide. 2007. Longday control of flowering in everbearing strawberries. J. Hortic. Sci. Biotech. 82(6):875–884. Sooriyapathirana, S.S., S. Mookerjee, C.K. Weebadde, C.E. Finn, K. Lewers, J.M. Bushakra, J.J. Luby, P. Stewart, S. Neils, and J.F. Hancock. 2015. Identification of QTL associated with flower and runner production in octoploid strawberry (Fragaria × ananassa). J. Berry Res. 5(3):107–116. Stewart, P.J. and K.M. Folta. 2010. A review of photoperiodic flowering research in strawberry (Fragaria spp.). Crit. Rev. Plant Sci. 29(1):1–13. Wagstaffe, A. 2009. The development of methods to alleviate thermodormancy in everbearing strawberry and secure season extension in the UK. Final Report. The University of Reading, Reading. Webb, R., J. Terblanche, J.V. Purves, and M. Beech. 1978. Size factors in strawberry fruit. Scientia Hortic. 9(4):347–356. Weebadde, C.K., D. Wang, C.E. Finn, K.S. Lewers, J.J. Luby, J. Bushakra, T.M. Sjulin, and J.F. Hancock. 2007. Using a linkage mapping approach to identify QTL for dayneutrality in the

octoploid strawberry. Plant Breeding 127(1):94. White, P. 1927. Studies of the physiological anatomy of the strawberry. J. Agric. Res. 35:481– 492. Wu, F., Z. Guan, and A. Whidden. 2012. Strawberry industry overview and outlook. University of Florida. Zeevaart, J.A. 2006. Florigen coming of age after 70 years. Plant Cell 18(8):1783–1789.

2 Small Unmanned Aircraft Systems (sUAS): An Emerging Technology for Horticulture James A. Robbins Professor/Extension Specialist, Horticulture, University of Arkansas System Division of Agriculture, Little Rock, AR, USA

ABSTRACT Small unmanned aircraft systems (sUAS), an emerging technology, are envisioned to be used for a variety of applications in agriculture, including horticulture. Some specific horticulture applications include crop monitoring, pesticide and nutrient applications, asset tracking and monitoring, crop inventory, and sales and marketing. Growing interest in sUAS is reflected in the increasing number and decreasing costs of aircraft and sensors. Until recently, the greatest challenge to widescale adoption of this technology has been regulatory issues. However, significant issues related to aircraft, sensors, and data processing and interpretation still require researchbased solutions. KEYWORDS: drones, remote sensing, aerial imagery, agriculture, sensors, vegetation indices

ABBREVIATIONS AGL Above ground level APS Activepixel sensor BGI1 Bluegreen index CCD Chargecoupled device CMOS Complementary metalsemiconductor COTS Commercial offtheshelf CWSI Crop water stress index ExG Excess green

FAA Federal Aviation Administration GNDVI Green normalized difference vegetation index GPS Global positioning system gRAID Geospatial realtime aerial image display IMU Inertial measurement unit JPEG Joint photographic experts group LAI Leaf area index LARS Lowaltitude remote sensing LiDAR Light detection and ranging NDVI Normalized difference vegetation index NIR Nearinfrared OBIA Objectbased image analysis PRI Photochemical reflectance index QN Total nitrogen uptake per square meter RGB Red, green, blue SAVI Soiladjusted vegetation index TIFF Tagged image file format sUAS Small unmanned aircraft systems UAV Unmanned aircraft vehicle

VIS Visible WS Waterstressed WW Wellwatered I. INTRODUCTION II. AIRCRAFT A. Rotary B. Fixedwing III. SENSORS AND DATA PROCESSING A. Sensors 1. RGB Digital Camera 2. Modified RGB Digital Camera 3. Multispectral 4. Hyperspectral 5. Laser 6. Thermal 7. Spectrometer B. Data Processing IV. HORTICULTURAL APPLICATIONS A. Crop Monitoring 1. Nutrient Status 2. Water Stress 3. Disease Incidence 4. Weed Infestation B. Chemical and Nutrient Applications C. Asset Tracking and Management D. Crop Inventory Management, Species Classification, and Crop Yield E. Marketing and Sales V. CHALLENGES

A. Regulations B. Aircraft C. Data Processing and Interpretation D. Privacy and Security VII. CONCLUSIONS LITERATURE CITED

I. INTRODUCTION Small unmanned aircraft systems (sUAS) are quickly evolving into useful remote sensing platforms that can deliver high spatial and temporal resolution imagery in a timely and noninvasive manner (Nolte 2015; von Bueren et al. 2015). Interest in unmanned aircraft vehicles (UAVs) has increased dramatically over the past 25 years (Fig. 2.1). Improvements in precise position referencing and miniaturization of electronic components are driving these developments (Salami et al. 2014).

Fig. 2.1. Google Scholar search for “UAV aircraft” and the resulting number of articles (does not include patents or citations) annually from 1990 to 2015. sUAS are forecast to play a critical role in the future of row crop and animal agriculture production (Jenkins and Vasigh 2013; Blake 2015; Simelli and Tsagaris 2015). LowAltitude

Remote Sensing (LARS) has been suggested for specific agricultural applications in developing countries (Swain et al. 2007). Zhang and Kovacs (2012) have provided a complete overview of sUAS and how they will be applied to precision agriculture. Horticulture (e.g., fruits, vegetables, tree nuts, turfgrasses, ornamentals, Christmas trees) offers some unique challenges as a result of the likely smaller production areas and diversity of crops compared to row crop agriculture. From a regulatory standpoint, horticulture/agriculture applications of sUAS should be viewed as a lowrisk (minimally intrusive) flight environment. Users have access to a wide variety of aircraft and sensors, each having their strengths and weaknesses, which need to be considered to achieve optimal data acquisition requirements (von Bueren et al. 2015). Costs of aircraft and sensors have declined as a result of increased demand, improvements in manufacturing processes, and the development of new technologies that are adapted to lightweight aircraft (Myers et al. 2015). It is clear that the use of sUAS in horticulture is in its infancy, with expanded use linked to enhanced aircraft development, enhanced and userfriendly image processing, relaxed flight regulations, costeffective sensors, and practical data processing and interpretation for horticulture users.

II. AIRCRAFT Satellites and lowaltitude aircraft, although commonly used in agricultural applications (Nebiker et al. 2008; Dobberstein 2013), have been used minimally in horticulture mainly due to the smaller crop areas and greater diversity of crops. Based on a comparison of normalized difference vegetative index (NDVI) surveys of an Italian vineyard conducted by a UAV, aircraft, and satellite, Matese et al. (2015) concluded that a UAV is advantageous for small acreage crops, with the breakeven area based on a cost–benefit analysis being 5 ha (12.4 acres). Horticulture may be an ideal market/application for sUAS, since spatial scale is smaller than that of traditional row crops that are typical in agriculture. This review will focus on “small unmanned aircraft systems” (sUAS), which are defined by the U.S. Federal Aviation Administration (FAA) as having a takeoff weight of less than 25 kg (55 lb). It might be worth noting that very often the terms UAS and UAV are used interchangeably; however, UAS more often refers to the “aircraft plus sensor(s)” whereas a UAV is simply the aircraft (platform). Several recent comprehensive reviews on aircraft are available (Gupta et al. 2013; Cai et al. 2014; Colomina and Molina 2014). Slightly older reviews (Nonami 2007; Everaerts 2008; Eisenbeiss 2009) provide a good framework to compare how quickly this emerging technology has advanced. To illustrate the explosion in the number of aircraft options, on 21st December 2015 a website (http://drones.specout.com/) that compares drones listed 271 models. On 29th June 2016, the number had increased to 548, and to 1253 by 7th February 2017. The types of aircraft used in agriculture/horticulture research, based on a survey of literature from 2004 to 2016, are summarized in Table 2.1. Table 2.1. A summary of the types of aircraft used for agriculture, forestry, and horticulture research based on a survey of literature from 2004 to 2016.

Year Authors

Platform category

2014 Bellvert et al. 2013 Calderon et al. 2013 Calderon et al. 2014 Calderon et al. 2008 Chao et al.

Fixed wing Fixed wing Fixed wing Fixed wing Fixed wing Fixed wing Fixed wing Fixed wing Fixed wing Fixed wing

2008 Chao et al. 2015 DiazVarela et al. 2015 Gaulton et al. 2009 Gay et al. 2013 Gonzalez Dugo et al.

2004 Herwitz et al. 2010 Hunt et al.

Fixed wing Fixed wing 2015 Kutnjak et Fixed al. wing 2011 Laliberte et Fixed al. wing 2008 Lelong et al. Fixed wing 2013 Link et al. Fixed wing

No. of Aircraft rotors model (rotary)

–

(custom built) mXSIGHT

–

Viewer

–

mXSIGHT

–

Procerus

–

Crossbow

–

mXSIGHT

–

QuestUAV AGRI Pro CropCam

– –

– – – – – –

Manufacturer Topic studied Quantalab

Crop

Water stress Vineyard

UAV Services Disease & Systems ELIMCO Disease

Olive

UAV Services Disease & Systems Procerus Water management CrossBow Water management UAV Services Breeding & Systems QuestUAV Biomass

Poppy

Olive

Vegetation Vegetation Olive Willow

Pentagon Crop Oat Performance monitoring mXSIGHT UAV Services Water stress Almond, & Systems apricot, peach, lemon, orange Pathfinder NASA Crop Coffee Plus monitoring VectorP IntelliTech Crop Winter Microsystems monitoring wheat eBee Sensefly Plant Alfalfa composition grass mix Bat3 MLB Co. Crop Rangeland monitoring (power glider) L’Avion Jaune Nitrogen Wheat ETrainer 182

Graupner

Nitrogen

Corn

2009 Rango et al. Fixed wing

–

Bat3

MLB Co.

Crop monitoring

2012 Zarco Tejada et al. 2013 Zarco Tejada et al. 2014 Zarco Tejada et al. 2008 Lelong et al.

–

(unknown)

Quantalab

Water stress Citrus

–

mXSIGHT

–

mXSIGHT

–

Pixy

UAV Services Water stress Vineyard & Systems UAV Services Tree height Olive & Systems ABS Nitrogen Wheat Aerolight

2009 Antic et al.

Paragliding – wing 2014 Aasen et al. Rotary 8

Pixy

IRD Mikrokopter

2014 Aden et al.

Microdrones Ascending Technologies Mikrokopter Crop monitoring Mikrokopter Plant Count Mikrokopter Disease Vario Water stress Helicopter SAL Eng. Crop monitoring DJI Nitrogen

Fixed wing Fixed wing Fixed wing Motorized parachute

Rotary

4

MKOkto XL Quad

2011 Aguera et al. Rotary 2014 Anthony et Rotary al. 2015 Bareth et al. Rotary

4 6

MD4200 Firefly

8

2014 Bazi et al. Rotary 2012 Bendig et al. Rotary 2009 Berni et al. Rotary

6 8 1

2015 Candiago et Rotary al. 2016 Caturegli et Rotary al. 2015 Das et al. Rotary

6

MK Oktocopter Hexacopter MKOkto Benzin Acrobatic ESAFLY A2500 WH S1000 Octocopter Innovations S800

2010 De Biasio et Rotary al. 2013 GarciaRuiz Rotary et al. 2015 Giles and Rotary

1

8 8

3D Robotics

DJI

Crop monitoring Crop monitoring Surface temperature Nitrogen Crop height

Rangeland

Noncrop Sunflower Corn Barley Palm Sugar beet Peach Vineyard, tomatoes Turf

6

CAMCOPTER Schiebel S100 Hexacopter Mikrokopter

Apple, orange, grape Classify land Mixed usage vegetation Disease Orange

1

RMAX

Spraying

Yamaha

Crop monitoring

Vineyard

Billing 2012 Gini et al.

Rotary

4

MD4200

Microdrones

2014 Gomez Rotary Candon et al. 2013 Honkavaara Rotary et al. 2010 Jaakkola et Rotary al.

8

MKOkto

Mikrokopter

Tree Forest classification Water stress Apple

1

Logo 600

Mikado

Biomass

1

TRex 600E

Align

Tree height

2011 Kazmi et al. Rotary

1

XLC

2011 Kazmi et al. Rotary

1

Vario Helicopter Maxi

2015 Maja et al. 2013 Matthews and Jensen 2015 Misopolinos et al. 2009 Nagai et al.

Rotary Rotary

8 4

Rotary

6

Rotary

1

2008 Nebiker et al. 2008 Nebiker et al. 2014 Pena et al.

Rotary

1

Rotary

4

Rotary

4

2013 Pena et al.

Rotary

4

md41000

Microdrones

2015 Pena et al.

Rotary

4

md41000

Microdrones

2012 Pena Barragan et al. 2013 Polonen et al.

Rotary

4

md41000

Microdrones

Rotary

1

(unknown)

(unknown)

Weed monitoring Joker3 Weed monitoring Mk Octocopter Mikrokopter Plant count HawkeyeII ElectricFlights Crop monitoring Hexa Vulcan Crop monitoring RPH2 Fuji Heavy 3D Ind. mapping AG WeControl Crop monitoring md4200 Microdrones Crop monitoring md41000 Microdrones Weed monitoring

Weed monitoring Weed monitoring Weed monitoring Crop monitoring

Wheat, barley Trees Sugar beet Sugar beet Nursery Vineyard Kiwifruit Land Vineyard Vineyard Sunflower, corn, olive, poplar Corn Sunflower Corn

Wheat, barley

2012 Primicerio et Rotary al. 2012 Stagakis et Rotary al.

6

HexaII

Mikrokopter

Crop Vineyard monitoring Water stress Orange

1

Benzin Acrobatic

Vario Helicopter

2013 Torres Sanchez et al.

Rotary

4

md41000

Microdrones

Weed monitoring

Sunflower

2015 Tsouvaltsidis Rotary et al. 2015 Tugi et al. Rotary

6

3DRY6

3D Robotics

Various

4

XR Q350 Pro Walkera

2011 Turner et al. Rotary

8

Oktocopter

Mikrokopter

2014 von Bueren et al. 2014 von Bueren et al. 2015 von Bueren et al. 2015 von Bueren et al. 2012 Wallace et al. 2011 Xiang and Tian 2015 Yun et al.

Rotary

4

QuadKopter

Mikrokopter

Rotary

8

Falcon8

Rotary

4

QuadKopter

Asctec Krailing Mikrokopter

Rotary

8

Falcon8

Rotary

8

Oktocopter

Asctec Krailing Mikrokopter

soil monitoring Crop monitoring Crop monitoring Crop monitoring Crop monitoring Crop monitoring Crop monitoring Tree count

Rotary

1

(unknown)

6

F550

Weed monitoring Nitrogen

Turf

Rotary

Rotomotion Inc. DJI

2009 Zarco Rotary Tejada et al.

1

Benzin Acrobatic

Vario Helicopter

Oil palm Vineyard Ryegrass pasture Ryegrass pasture Ryegrass pasture Ryegrass pasture Trees

Hairy vetch Water stress Olive, peach, orange

Some advantages of sUAS compared to satellite or manned aircraft include: higher image resolution; autonomous flight path which can be repeated; “asneeded” use; lower flight altitude so image collection is less likely to be influenced by clouds; safe solution for dull/dangerous/dirty applications; reduction in time needed to prepare and initiate flights; and lower cost. Some current limitations of a sUAS include: battery life; aircraft reliability and endurance; effort to process (i.e., mosaicking, geocoding) images due to increased resolution; and sense and avoid capability.

The different UAV aircraft (i.e., rotary, fixedwing, hybrid) each have their strengths and limitations that need to be considered in order to optimize data acquisition requirements (von Bueren et al. 2014). Aircraft typically fall into two broad categories – rotary and fixedwing. However, hybrid aircraft (i.e., those that lift off vertically like a rotary and then reorient propellers to fly as a fixedwing) are available (e.g., FireFly6 Pro, BirdsEyeView Aerobotics, Sutton, NH).

A. Rotary The number of blades on a rotary aircraft may vary from one to eight, with four (quadcopter), six (hexacopter), and eightbladed (octocopter) being the most common for the collection of imagery. In general, increasing the number of blades from one to eight increases the stability of the aircraft, which may prove useful when acquiring highresolution images. The majority of sUAS rotary aircraft are powered by lithium ion batteries, which is one current limitation to longduration flights (von Bueren et al. 2014; Myers et al. 2015). One solution to overcoming short battery life may be the utilization of mobile charging platforms (Rezelj and Skocaj 2015). Another limitation to a rotary aircraft is that the total area that can be covered is still relatively small (von Bueren et al. 2014). Advantages of rotary aircraft over fixedwing are: the ability to hover in place/stationary measurements; vertical takeoff (no need for airstrip, launchers); lower flight altitudes; multidirectional flight; and more complex flight patterns. Research using sUAS rotary aircraft can be traced back to the early 1990s (Cai et al. 2014). According to their survey of research institutions using sUAS, 68% were using a rotarytype aircraft and 28% used a fixedwing aircraft. This agrees with the survey of literature from 2004 to 2016 (Table 2.1), whereby twothirds (45/67) of aircraft used in research were rotary and 30% (20/67) were fixedwing.

B. Fixedwing Currently, the main advantages of fixedwing aircraft over small rotary aircraft are that they can cover larger areas (Aasen et al. 2014) and achieve longer flying times (Myers et al. 2015; Woldt et al. 2016). The main disadvantage of fixedwing aircraft is requirements for takeoff and landing (Woldt et al. 2016). Takeoff for a fixedwing aircraft typically requires some kind of launch assistance (e.g., launcher) or runway.

III. SENSORS AND DATA PROCESSING Several comprehensive reviews on sensors and image processing are available (Eisenbeiss 2009; Colomina and Molina 2014; Salami et al. 2014; Gago et al. 2015). A major thrust that enabled the coupling of sensors with sUAS was the development of electronics that allow for the registration and tracking of aircraft position and orientation in a global coordinate system (Eisenbeiss 2009). Although most sUAS applications involve the use of a single sensor, there is a trend toward “sensor suites” to achieve multiple data products from a single flight (Nagai

et al. 2009; Jaakkola et al. 2010; Das et al. 2015; Ehsani et al. 2016). A summary of the types of sensor used for agriculture/horticulture research, based on a survey of literature from 2004 to 2016, is provided in Table 2.2. Table 2.2. A summary of the types of sensors used for agriculture, forestry, and horticulture research using UAS based on a survey of literature from 2004 to 2016. Year Author

Sensor type

Sensor model Manufacturer Topic studied 2014 Aasen et al. Hyperspectral UHD 185 Cubert Crop Firefly monitoring

Crop

2015 Bareth et al. Hyperspectral UHD 185 Firefly 2015 Bareth et al. Hyperspectral Hyperspectral Camera 2013 Calderon et Hyperspectral Micro al. Hyperspec VNIR 2013 GarciaRuiz Hyperspectral AISA Eagle et al. VNIR 2013 Polonen et Hyperspectral al. 2012 Zarco Hyperspectral Micro Tejada et al. Hyperspec VNIR 2009 Nagai et al. Laser scanner LMS291

Cubert

Crop monitoring Crop monitoring Disease

Barley

Specim Ltd.

Disease

Orange

(custom)

Crop Wheat, monitoring Barley Water stress Citrus

2015 Das et al.

Laser scanner UST20LX

Hokuyo

2010 Jaakkola et al. 2010 Jaakkola et al. 2012 Wallace et al. 2014 Anthony et al. 2008 Chao et al.

Laser scanner LMS151

Rikola Headwall Photonics

Headwall Photonics SICK

Barley Olive

Land

SICK

3D mapping Crop monitoring Tree height

Laser scanner LUX

ibeo

Tree height

Trees

Laser scanner LUX

Ibeo

Tree count

Trees

Laser scanner URG 04LXUG01 Modified GFDC RGB OPTIO E30

Hokuyo

Crop height

Corn

Pentax

Water Vegetation management

Apple, orange, grape Trees

2015 DiazVarela et al. n.d. Gaulton et al.

Modified RGB Modified RGB

Lumix DMCGF1 Lumix LX5

Panasonic

Breeding

Olive

Panasonic

Biomass

Willow

2012 Gini et al.

Modified RGB

DP1

Sigma

Tree Forest classification

DP1x

Sigma

Water stress Apple

2014 Gomez Modified Candon et al. RGB 2010 Hunt et al. Modified RGB 2015 Kutnjak et Modified al. RGB 2008 Lelong et al. Modified RGB 2008 Lelong et al. Modified RGB 2015 Maja et al. Modified RGB 2008 Nebiker et Modified al. RGB 2015 Tugi et al. Modified RGB 2014 von Bueren Modified et al. RGB 2015 von Bueren Modified et al. RGB 2015 von Bueren Modified et al. RGB 2015 Yun et al. Modified RGB 2014 Zarco Modified Tejada et al. RGB 2015 Zhao et al. Modified RGB 2015 Das et al. Monochrome

FinePix S3 Fuji Pro UVIR IXUS 125 HS Canon NIR EOS 350D Canon

Crop monitoring Plant composition Nitrogen

Winter wheat

DSCF828

Sony

Nitrogen

Wheat

PowerShot SX260 HS SmartCam

Canon

Plant count

Nursery

Sony

Vineyard

PowerShot XS260 PowerShot SD780 PowerShot SD780 Nex5n

Canon

SX 260

Canon

Crop monitoring Crop monitoring Crop monitoring Crop monitoring Crop monitoring Nitrogen Tree height

Olive

Canon Canon Sony

Lumix Panasonic DMCGF1 ELPH110HS Canon BlueFOX

2011 Aguera et al. Multispectral ADC Lite

Alfalfa grass mix Wheat

Oil palm Ryegrass pasture Ryegrass pasture Ryegrass pasture Hairy vetch

Water stress Almond

Matrix Vision Crop monitoring Tetracam Nitrogen

Apple, orange, grape Sunflower

2009 Antic et al.

Multispectral MCA

Tetracam

2009 Antic et al.

Multispectral ADC

Tetracam

2012 2012 2009 2013

Multispectral Multispectral Multispectral Multispectral

MCA6 MCA mini MCA6 MCA6

Tetracam Tetracam Tetracam Tetracam

Water stress Disease Water stress Disease

Vineyard Sugar beet Peach Olive

Multispectral ADC Lite

Tetracam

Disease

Poppy

Multispectral ADC Micro

Tetracam

Multispectral ADC Micro

Tetracam

Crop monitoring Nitrogen

Vineyard, tomatoes Turf

Multispectral Condor 1000 MS5 Multispectral MCA6 mini

Quest Innovations Tetracam

Classify land Mixed usage vegetation Disease Orange

Multispectral MS3100

DuncanTech

Multispectral (custom)

VTT Techn. Rsch. Cntr. Tetracam

Crop monitoring Biomass

Baluja et al. Bendig et al. Berni et al. Calderon et al.

2014 Calderon et al. 2015 Candiago et al. 2016 Caturegli et al. 2010 De Biasio et al. 2013 GarciaRuiz et al. 2004 Herwitz et al. 2013 Honkavaara et al. 2011 Kazmi et al.

Multispectral MCA mini

2011 Laliberte et Multispectral MCA6 mini al. 2015 Misopolinos Multispectral ADC Lite et al. 2009 Nagai et al. Multispectral ADC3

Tetracam

2014 Pena et al.

Multispectral MCA6 mini

Tetracam

2013 Pena et al.

Multispectral MCA6 mini

Tetracam

2015 Pena et al.

Multispectral MCA6 mini

Tetracam

Tetracam Tetracam

Crop monitoring Crop monitoring

Weed monitoring Crop monitoring Crop monitoring 3D mapping Weed monitoring Weed monitoring Weed

coffee Wheat, Barley Sugar beet Rangeland Kiwifruit Land Sunflower, corn, olive, poplar Corn Sunflower

2012 Pena Barragan et al.

monitoring Weed monitoring

Multispectral MCA mini

Tetracam

2012 Primicerio et Multispectral ADC Lite al.

Tetracam

Crop monitoring

2012 Stagakis et Multispectral MCA6 al. 2013 Torres Multispectral MCA6 mini Sanchez et al. 2011 Turner et al. Multispectral MCA mini

Tetracam

Water stress Orange

Tetracam

Weed monitoring

Sunflower

Tetracam

Vineyard

2014 von Bueren et al. 2015 von Bueren et al. 2011 Xiang and Tian 2009 Zarco Tejada et al. 2013 Zarco Tejada et al. 2014 Bazi et al. 2008 Chao et al.

Multispectral MCA6 mini

Tetracam

Multispectral MCA6 mini

Tetracam

Multispectral ADC

Tetracam

Multispectral MCA6

Tetracam

Crop monitoring Crop monitoring Crop monitoring Weed monitoring Water stress

Multispectral SixCam RGB RGB

EOS 550D OPTIO E30

QuantaLab IASCSIC Canon Pentax

2015 Das et al.

RGB

BlueFOX

n.d. Gaulton et al. 2012 Gini et al.

RGB

Lumix LX5

Counting Water management Matrix Vision Crop monitoring Panasonic Biomass

RGB

Optio A40

Pentax

2014 Gomez RGB Candon et al. 2004 Herwitz et RGB al.

DP1x

Sigma

555ELD

Hasselblad

Crop monitoring

Coffee

2010 Jaakkola et

Pike F421C

AVT

Tree height

Trees

RGB

corn

Vineyard

Ryegrass pasture Ryegrass pasture Turf

Olive, peach, orange Water stress Vineyard Palm Vegetation Apple, orange, grape Willow

Tree Forest classification Water stress Apple

al. 2015 Kutnjak et al.

RGB

IXUS 125 HS Canon VIS

Plant Alfalfa composition grass mix

2011 Laliberte et al. 2015 Maja et al.

RGB

SD 900

Canon

Rangeland

RGB

Canon

2013 Matthews and Jensen 2009 Nagai et al.

RGB

2008 Nebiker et al. 2014 Pena et al.

RGB

EOS 20D

Canon

RGB

PEN EPM1

Olympus

Crop monitoring 3D mapping Crop monitoring Weed monitoring

Vineyard

RGB

PowerShot SX260 HS PowerShot A480 EOS 10D

Crop monitoring Plant count

2013 Torres Sanchez et al. 2015 Tugi et al.

RGB

PEN EPM1

Olympus

Weed monitoring

RGB

PowerShot XS260 550D

Canon

Nex5n

Sony

Crop monitoring Crop monitoring Crop monitoring Nitrogen Water stress Tree height

2011 Turner et al. RGB 2014 von Bueren et al. 2015 Yun et al. 2015 Zhao et al. 2010 Jaakkola et al. 2013 Link et al. 2015 Tsouvaltsidis et al. 2014 von Bueren et al. 2015 von Bueren et al.

RGB

2014 Aden et al.

Thermal

Canon Canon

Canon

RGB S110 Canon RGB ELPH110HS Canon Spectrometer V10H Specim Spectrometer MMS1 Spectrometer Argus 1000NK Spectrometer STS custom Spectrometer STS custom MLX90614

Nursery

Land Vineyard Sunflower, corn, olive, poplar Sunflower

Oil palm Vineyard Ryegrass pasture Hairy vetch Almond Trees

Carl Zeiss Nitrogen Thoth Soil health Technology Ocean Optics Crop monitoring Ocean Optics Crop monitoring

Corn None

Melexis

NonCrop

Surface

Ryegrass pasture Ryegrass pasture

2012 Baluja et al. Thermal

Thermovision FLIR A40 M

2014 Bellvert et Thermal al. 2012 Bendig et al. Thermal 2009 Berni et al. Thermal

Miricle 307  K F30 IS Thermovision A40 M MIRICLE 307

2013 Calderon et al.

Thermal

2014 Calderon et al. 2015 Das et al.

Thermal Thermal

2014 Gomez Thermal Candon et al. 2013 Gonzalez Thermal Dugo et al.

2015 Misopolinos Thermal et al. 2011 Turner et al. Thermal

2012 Zarco Tejada et al. 2009 Zarco Tejada et al. 2013 Zarco Tejada et al. 2011 Kazmi et al.

temperature Water stress Vineyard

Thermoteknix Water stress Vineyard NEC FLIR

Disease Sugar beet Water stress Peach

Thermoteknix Disease Systems

MIRICLE 307 A35

Thermoteknix Disease Systems FLIR Crop monitoring Miricle 307  Thermoteknix Water stress K Miricle 307  Thermoteknix Water stress K

A65

Thermal Infrared' (TIR) Thermal MIRICLE 307 Thermal Thermovision A40 M Thermal MIRICLE 307 Time of flight Swiss Ranger SR4000

FLIR FLIR

Crop monitoring Crop monitoring

Olive Poppy Apple, orange, grape Apple Almond, apricot, peach, lemon, orange Kiwifruit Vineyard

Thermoteknix Water stress Citrus Systems FLIR Water stress Olive, peach, orange Thermoteknix Water stress Vineyard Systems Mesa Imaging Weed Sugar beet monitoring

For the most part, sensors that are used with small unmanned aircraft are no different than sensors used on satellites, manned, or groundbased systems. However, size and weight need to be reduced to make them useful on such small aircraft. Matese et al. (2015) compared NDVI

surveys of vineyards obtained using multispectral sensors on three aerial platforms (sUAS, manned aircraft, satellite). Their report outlined the pros and cons of each system, with a cost analysis. Traditional remote sensing utilizes sensors on towers over crops, where the main limitation is the fixed position from which data are collected (Gago et al. 2015). Manned aircraft and satellites are also traditional remote sensing platforms, but their temporal and spatial resolution limits their usefulness for agricultural applications. Advantages of sUAS over manned aircraft and satellites are that the area of interest can be isolated with highresolution imagery, there is easier usability, and reduced cost of data acquisition (Candiago et al. 2015). Data collected by some sensors [e.g., modified red/green/blue (RGB) camera; multispectral] can be processed to yield various “vegetation indices” (Antic et al. 2009; Bendig 2015). These indices can then be correlated to useful biophysical crop parameters such as leaf area index (LAI), nitrogen uptake (QN), and crop water status (Antic et al. 2009). Recent reviews (Salami et al. 2014; Shahbazi et al. 2014) have provided a good summary of indices and corresponding crop parameters. Misopolinos et al. (2015) presented a “proofofconcept” for a UAV system, that utilized simultaneous acquisition of highresolution imagery in the visible, nearinfrared and thermal infrared wavelengths, in order to derive NDVI for a kiwifruit orchard. The quality of reflectance measurement is dependent upon the precise calibration of sensors (von Bueren et al. 2014).

A. Sensors The following is a brief overview of the types of sensors that may be used for horticultural crops. 1. RGB Digital Camera.  A digital frame camera is likely one of the most common passive sensors used for aerial remote sensed data. In a digital camera, the standard silverhalide emulsion is replaced by a digital image sensor (Jensen 2005). A “fullframe” digital camera with a “chargecoupled device” (CCD) image sensor is used where highquality image data are required. A number of consumer and professional cameras may use an alternative image sensor called an “active pixel sensor” (APS), which uses “complementary metalsemiconductor” (CMOS) technology. Since each lightsensitive photosite (pixel) is inherently monochromatic, manufacturers use a variety of special filters to capture red, green, and blue (RGB) photons to create a visible band image (Jensen 2005). 2. Modified RGB Digital Camera.  While the use of modified commercial offtheshelf (COTS) cameras for multispectral imaging is increasing due to low cost and weight, issues with the relative spectral response need to be addressed. Berra et al. (2015) outline a detailed methodology to measure the spectral response of unmodified and modified COTS cameras. The use of modified cameras allows for image acquisition in spectral bands that are not currently used in traditional visible

wavelength photography (Lebourgeois et al. 2008). Lebourgeois et al. (2008) presented an indepth comparison of an RGB camera (Canon EOS 400D, Canon Inc., Tokyo, Japan), and the same camera modified with different filters (red edge and nearinfraredmodified) for radiometric data acquisition using an ultralight aircraft. In this single scenario, these authors demonstrated that vignetting effects were greater in the modified versus the nonmodified camera. Based solely on this singlecamera analysis, it was also noted that the operator should pay particular attention to the camera settings, and the authors strongly recommended saving images in the raw image format rather than in JPEG or TIFF formats. However, they cautioned that their conclusions may not apply beyond the camera evaluated. A more recent comparison of an RGB and a modifiedRGB camera was conducted to detect the seasonal development of understory plants in a forest (Nijland et al. 2014). For this set of circumstances, the RGB camera outperformed the modifiedRGB camera, and the underperformance was attributed to limitations in dynamic range and band separation. Simulations showed that a conversion with a redrejection dualbandpass filter would largely overcome this issue. 3. Multispectral.  In the survey of literature from 2004 to 2016 (Table 2.2), the most commonly used multispectral sensor (18 out of 32 total uses) used with sUAS was the MCA mini/micro (Tetracam, Chatsworth, CA, USA) with the ADC series (Tetracam) coming in second (9 out of 32). Results from the survey indicated that the general types of application for horticultural crops include nutrient monitoring (Aguera et al. 2011; Caturegli et al. 2016), water stress monitoring (Berni et al. 2009; ZarcoTejada et al. 2009, 2013; Baluja et al. 2012; Stagakis et al. 2012), disease monitoring (Bendig et al. 2012; Calderon et al. 2013, 2014), weed monitoring (Xiang and Tian 2011; TorresSanchez et al. 2013; Pena et al. 2014, 2015), and general crop monitoring (Herwitz et al. 2004; Turner et al. 2011; Primicerio et al. 2012; Candiago et al. 2015; Misopolinos et al. 2015). The ability of multispectral sensors to discriminate specific wavelengths makes them ideal for developing correlations with plant biophysical processes using vegetation indices (Colomina and Molina 2014). Huang et al. (2010) presented a thorough comparison of three different multispectral sensors (ADC, Tetracam; MS 4100, Geospatial Systems, W. Henrietta, NY, USA; and TTAMRSS, custombuilt by Texas Tech Univ., Lubbock, TX, USA) that could be used for agricultural applications with recommendations for the preferred aircraft platform. They concluded that the ADC was more preferable for lowaltitude unmanned aircraft, that the intermediate cost sensor (MS 4100) works well for lowaltitude remote sensing on fixed wing aircraft, and the most expensive custombuilt sensor (TTAMRSS) was recommended for highaltitude remote sensing on fixedwing aircraft. Arnold et al. (2013) outlined two methods to acquire multispectral imagery and to compare the performance between their low cost setup to a highercost system. They calculated NDVI to distinguish between different types of vegetation and soil, and concluded that the lowcost setup was effective in

identifying the vegetation present. 4. Hyperspectral.  Several types of hyperspectral sensors have been used on sUAS. In the survey of literature from 2004 to 2016 (Table 2.2), the most commonly used hyperspectral sensor was evenly split (two out of six total uses) between the UHD 185 Firefly (Cubert GmbH, Ulm, Germany) and the Micro Hyperspec (Headwall Photonics, Fitchburg, MA, USA) (also two uses). Hyperspectral sensors capture narrow spectral bands over a continuous spectral range, thus capturing more detailed information than that available from multispectral sensors (Colomina and Molina 2014). To meet the needs of the emerging field of sUAS, designers of miniaturized multi and hyperspectral sensors are challenged in terms of optics and calibration (Colomina and Molina 2014). Relatively lowcost, smallsized, and lightweight hyperspectral sensors are emerging (Salami et al. 2014). Bareth et al. (2015) mentioned a hyperspectral sensor (OCITMUAV, BaySpec, San Jose, CA, USA) introduced in 2014. Using a hyperspectral sensor on sUAS raises several challenges (e.g., image acquisition, signaltonoise ratio, spectral calibration), as outlined by Aasen et al. (2014). Understanding these challenges, Aasen et al. (2014) evaluated a “new” type of hyperspectral sensor that records a full hyperspectral image frame (full frame). Their results showed that this “new” sensor (UHD 185Firefly) is suitable for hyperspectral remote sensing of agricultural crops. Bareth et al. (2015) compared two sUASmounted fullframe sensors (Hyperspectral Camera, Rikola, Oulu, Finland and Cubert UHD 185 Firefly) for crop monitoring. While they concluded that both sensors successfully delivered hyperspectral data, spectral calibration was still a critical issue. 5. Laser.  Laser scanning involves the controlled steering of laser beams, followed by a distance measurement at every pointing direction. It is used to rapidly capture the shapes of objects, including plants. A laser scanner, which is an active measurement system, can generate two or threedimensional (3D) images. LiDAR (Light Detection And Ranging) uses laser light to target the object and produces 3D images. Likely, the first mention of UASbased LiDAR was when Nagai et al. (2004) outlined the workflow for the geoprocessing of LiDAR data. In forestry, LiDAR has been used to measure canopy cover, biomass, and tree height (Wallace et al. 2012; Anthony et al. 2014). The application of laser scanners to sUAS is a challenge, due to the tradeoff between performance and size and cost of sensors, and due to the effect of flight dynamics on the measurement process (Colomina and Molina 2014). In the survey of literature from 2004 to 2016 (Table 2.2), only four different types of laser scanner were found for applications such as 3D mapping (Nagai et al. 2009), tree count (Wallace et al. 2012), crop (apple) monitoring (Das et al. 2015), and height of corn (Anthony et al. 2014). 6. Thermal.  The most commonly used thermal sensor (seven out of 15 total uses) used with sUAS was the

Miricle 307K (Thermoteknix, Cambridge, UK) with the ThermoVisionTM A40M (FLIR Systems, Wilsonville, OR, USA) coming in second (three out of 15) (Table 2.2). Thermal imaging sensors for sUAS are becoming more common as both weight and size are reduced (Colomina and Molina 2014). Thermal sensors on sUAS are used mostly to monitor water stress (Berni et al. 2009; ZarcoTejada et al. 2009, 2012, 2013; Baluja et al. 2012; GonzalezDugo et al. 2013; Bellvert et al. 2014; GomezCandon et al. 2014) or for disease detection (Bendig et al. 2012; Calderon et al. 2013, 2014). 7. Spectrometer.  An optical spectrometer measures the electromagnetic spectrum, and can show intensity as a function of wavelength or frequency. Tsouvaltsidis et al. (2015) provided some preliminary information in their efforts to evaluate a small spectrometer (Argus 1000, Thoth Technology, Ontario, Canada) for remote spectral imaging to assess “soil health” using sUAS. von Bueren et al. (2014) evaluated four different sUASbased sensors, including a spectrometer, to monitor ryegrass pasture. The customdesigned spectrometer, which was based on the STS VIS spectrometer (Ocean Optics, Dunedin, FL, USA), delivered spectra with strong correlations to those from a groundbased portable spectroradiometer. Their continuing work (von Bueren et al. 2015) demonstrated that for vegetation analysis, the customdesigned spectrometer has the potential for feature identification in crops as well as the development of vegetation indices.

B. Data Processing While a great deal of effort has been devoted to the development of highperformance lightweight sensors, more effort needs to be placed on data acquisition and processing for the “Big Data” associated with sUASgenerated imagery. A number of authors (Nebiker et al. 2008; Colomina and Molina 2014) cite challenges in processing sUASgenerated data. There is already a move to cloud computing (Elston 2016). While reduction in size and cost are progressing well with sensors, that does not necessarily translate to data processing. In fact, sUAS will require more sophisticated processing (Colomina and Molina 2014; Gago et al. 2015; Matese et al. 2015) and, as sUAS provide lessstable platforms than manned aircraft, postprocessing (e.g., orthorectification, mosaicking) is more challenging (Rango and Laliberte 2010). Lowaltitude sUAS provide a particular challenge when trying to retrieve georeferenced aerial images based only on features in the image. To address this issue, Jensen et al. (2008) proposed a system called gRAID (GeoSpatial RealTime Aerial Image Display) which georeferenced images using only the position and orientation of the sUAS. One method to improve the exterior orientation of images obtained from a sUAS, that may help reduce the cost and turnaround time for production of orthorectified mosaics, was investigated by Laliberte et al. (2008). These authors continued to develop processing workflows that describe challenges and solutions associated with the efficient processing of multispectral imagery to obtain orthorectified, radiometrically calibrated mosaics (Laliberte et al. 2011).

A number of avenues are available to deal with sUASgenerated data. One option is that growers will process their own data. Another option is for “thirdparty” providers to supply output products that a grower can use (Staley 2015). For example, in the Central Valley of California, Raptor Maps (http://www.raptormaps.com/; accessed on 26 Oct 2016) advertises that they will provide growers with “…harvest data using the latest sensor technology” on a per acre fee basis. Another example is DroneDeploy (https://www.dronedeploy.com/; accessed on 4 Nov 2016) that offers a variety of cloudbased services to produce orthomosaics, terrain models, NDVI analysis, and 3D models. Agrisens (http://www.agrisens.com/; accessed on 4 Nov 2016) advertises image collection, processing, analysis, and visualization services for agricultural users.

IV. HORTICULTURAL APPLICATIONS Gago et al. (2015) compiled a comprehensive table of agricultural (mostly horticultural) uses for sUASs. While the compilation is impressive, it illustrates how few studies have actually been conducted using sUASs on horticultural crops. Only olive, peach, apple, almond, citrus, grape, turf, sunflower, coffee, trees, ornamental shrubs, and poppy have been evaluated (Table 2.1). Huang et al. (2013) presented an overview of the development, prospects, and limitations of sUAS technologies for agricultural production and management. It is worth noting that since this is an emerging technology, potential applications for sUAS in horticulture may not yet be fully identified. For example, preliminary investigations using a sUAS are being carried out to evaluate options for pollinating commercial Medjool date palms (eXtension UAS in Agriculture webinar, 21 Apr 2016, Kurt Nolte, Univ. of Arizona) as an alternative to the traditional method, which uses a leaf blower. A small sachet of pollen is delivered to the inflorescence, and the pollen is then distributed by the “propeller wash” of the aircraft. This concept of using a sUAS as a pollinator is representative of the many as yet thought of uses for this emerging technology. Main uses for sUAS in precision horticulture may fall into five categories: crop monitoring (including nutrient status, water status, disease, insect, and weed incidence); chemical and nutrient applications; asset tracking and management; crop inventory management; marketing and sales.

A. Crop Monitoring Crop monitoring is a broad category including such areas as monitoring for nutrients, water status, disease and insect incidence, and weed infestation (Berni et al. 2009; Dobberstein 2013). Salami et al. (2014) and Shahbazi et al. (2014) have provided thorough reviews of recent applications of sUAS imagery in the management of agricultural and natural resources. Utilizing components of reflected wavelengths, researchers have been able to: (1) classify and map crops; (2) predict crop yields; (3) determine crop condition; (4) detect weeds, diseases, and nutrient status; and (5) measure photosynthetic pigments (Antic et al. 2009). Several thermal and spectral indices have been correlated to biophysical plant parameters based on

sUAS imagery (Salami et al. 2014; Shahbazi et al. 2014). In the short term, many growers may use remote sensed imagery from a UAS to simply make a quick assessment on “general crop health.” As an indication to what might be applicable to horticultural crops, researchers have used sUASbased sensors to assess general pasture health through crop monitoring. A custombuilt spectrometer was found to deliver spectral data that matched the quality of ground spectral measurements of ryegrass pastures (von Bueren et al. 2014, 2015). Tugi et al. (2015) used a lowcost sUAV platform and a modified off theshelf digital camera to monitor oil palm tree health for smallacreage farmers. A six band multispectral camera was used to monitor vigor and health in a vineyard (Turner et al. 2011) and, similarly, a sUAS equipped with a multispectral camera was able to produce vigor maps for a vineyard based on NDVI values (Primicerio et al. 2012). Using a gaspowered sUAS equipped with a multispectral or a thermal sensor, Berni et al. (2009) produced a variety of remote sensing products (e.g., chlorophyll concentration, water stress detection) in three crops (peach, corn, olive). Candiago et al. (2015) outlined the potential of high resolution sUAS data and photogrammetric techniques to collect multispectral images and then correlate to three vegetation indices (NDVI, GNDVI, SAVI) for tomato and wine grape. They concluded that while it was possible to distinguish between healthy and unhealthy portions of a field, the results obtained were mostly qualitative since their approach lacked comparative ground radiometric measurements. Since a sUAS can fly closer to a crop, an exciting opportunity for sUASbased imagery is the measurement of steadystate chlorophyll fluorescence to monitor water stress (Gago et al. 2015). Herwitz et al. (2004) reported on a proofofconcept to use sUASbased imagery to manage several aspects (e.g., weed mapping, fruit maturity) in a coffee plantation. 1. Nutrient Status.  Extensive literature exists documenting the use of colorinfrared imagery to monitor the physiology and growth of agricultural crops (Nebiker et al. 2008). A sUAS is viewed as a critical tool to obtain remote sensed data for crop monitoring. As an example, Hunt et al. (2010) used a UAVbased modified RGB camera to evaluate the variability in two fertilized fields of winter wheat. They concluded that their system found a good correlation between leaf area index (LAI) and green normalized difference vegetation index (GNDVI) between the two variably fertilized fields. Lelong et al. (2008) presented a very thorough analysis of the methodology used to establish relationships between two vegetation indexes (NDVI, GNDVI) and fieldmeasured biophysical parameters (LAI, QN) in ten varieties of wheat. The quality of their derived relationships demonstrated that spectral ranges produced by standard cameras are suitable for remote sensing. Link et al. (2013) evaluated a sUAS equipped with a multispectral sensor to assist in crop nutrient monitoring in corn, and concluded that the system was not able to derive suitable spectral information about corn biomass, corn LAI and corn nitrogen status during the beginning of the vegetation period; however, slight correlations were obtained at the end of the growing season. A group in Finland (Polonen et al. 2013) developed a lightweight hyperspectral imager to estimate biomass and nitrogen content in wheat and barley. Their field studies seem to suggest that it is possible to create a totally automated

machine learning process to estimate those two parameters. Using a sUAS equipped with an offtheshelf RGB and one modified to collect NIR imagery, Yun et al. (2015) demonstrated a strong linear relationship between the biomass and nitrogen content of hairy vetch when using the NDVI, GNDVI, and ExG vegetation indices. Growers would be interested in a system that would allow them to monitor crop nutrient status, especially nitrogen, over a growing season, and lowaltitude sUAS that can obtain high resolution images may allow this to happen (Aguera et al. 2011). The latter authors showed with sunflower, that a sUAS with an ADC Lite Tetracam (520 to 920 nm) provided a better estimate of nitrogen applied than a groundbased system. While these specific applications have been for agricultural crops, the approaches used indicate the potential for these technologies with horticultural crops. Caturegli et al. (2016) compared spectral reflectance of three turfgrass species acquired by either a groundbased sensor (GreenSeeker R, Trimble, Sunnyvale, CA,USA) or a multispectral sensor (ADC Micro, Tetracam) mounted on a sUAS. Their results demonstrated that, for large areas such as a golf course or sod farm, the sUAS acquired data can adequately assess N status. However, a handheld crop sensor can be less expensive and a more practical option for smaller turf areas. 2. Water Stress.  Gago et al. (2015) presented a thorough review of methodologies that have been used to assess water stress in sustainable agriculture using a sUAS. Their report provides a compilation of studies and indices used for plant vigor and stress assessment. It is not surprising that water management and irrigation control might be a critical area of study with sUAS. Chao et al. (2008) outlined a roadmap for how realtime water management might be accomplished using a sUAS and a bandconfigurable sensing system. GomezCandon et al. (2014) demonstrated the feasibility of using a sUAS to monitor differences in canopy temperature between water stressed (WS) and wellwatered (WW) apple trees. The relationship between canopy temperature and waterstressed condition was dependent on date. While there was no significant difference in canopy temperature between WW and WS at the first time of measurement (4 July), the differences were highly significant on the second (1 August) and third (6 September) dates. Although SepulcreCanto et al. (2006) used a manned aircraft to conduct their studies on detecting water stress in an olive orchard, their overall approach may benefit others using a sUAS to evaluate the use of thermal remote sensing for water stress detection. Baluja et al. (2012) demonstrated that they could map spatial variability in the water status of a vineyard using thermal and multispectral imagery obtained using a sUAS. It has been proposed that high spatial and spectral resolution provided from sUASbased imagery could be used to detect deficit irrigation strategies that could affect fruit quality in a commercial citrus orchard (Stagakis et al. 2012). Their results demonstrated that both the photochemical reflectance index (PRI) and their proposed modified PRI (PRI515) were able to track water stress before any other canopy biochemical or structural changes were detected by four other standard vegetative indices. Not all studies that have tried to correlate a vegetation

index to water stress have been successful. Using a sUAS equipped with a standard RGB off theshelf camera, or one modified to capture NIR imagery, Zhao et al. (2015) investigated whether NDVI was a suitable measure of water stress in a commercial almond orchard. They concluded that canopy level NDVI does not indicate water stress in this situation. Canopy temperature has long been used as an indicator for monitoring crop water status (Idso et al. 1978). Jackson et al. (1981) proposed the concept of a crop water stress index (CWSI) that could be derived from measurements taken with an infrared thermometer. A number of researchers have evaluated sUAS for obtaining the thermal imagery needed to calculate CWSI. GonzalezDugo et al. (2013) used a thermal camera mounted on a fixedwing aircraft to estimate CWSI on five different tree fruits (almond, apricot, peach, lemon, and orange) that were either wellwatered or stressed. Overall, they concluded that thermal imagery acquired by a UAV over commercial orchards can accurately describe spatial variability in crop water status, and allow mapping of an orchard on a treebytree basis. They noted that future work will be needed to establish CWSI thresholds for different species and phenological stages. A prerequisite for precision irrigation in vineyards requires characterizing the spatial variability in water status across the field. By varying the pixel size from aerial thermal imagery, Bellvert et al. (2014) were able to show that a 0.3 m per pixel was needed to produce precise “crop water stress index” (CWSI) mapping in a vineyard. A sUAS equipped with a thermal infrared camera has been used to map soil moisture, which can be used to assess irrigation efficiency (Turner et al. 2011). Spanish researchers have studied early detection of water stress in a variety of horticultural crops using a sUAS equipped with multispectral and thermal cameras. The group demonstrated the feasibility of chlorophyll fluorescence extraction at the tree level that yielded determination coefficients r2 = 0.57 for olive and r2 = 0.54 for peach (ZarcoTejada et al. 2009). They concluded that imaging solarinduced chlorophyll fluorescence emission using the infilling method is feasible. They also assessed thermal and multispectral imagery for water stress detection in a citrus orchard, and demonstrated that crown temperature, bluegreen index (BGI1), and chlorophyll fluorescence estimates correlated best with water stress treatments (ZarcoTejada et al. 2012). Although the standard photochemical reflectance index (PRI) is sensitive to water stress, studies involving a sUAS equipped with very highresolution narrowband multispectral and thermal sensors indicated that PRI does not accurately track the diurnal dynamics of stomatal conductance and water potential in a vineyard (ZarcoTejada et al. 2013). As a result, the group proposed a new index called the normalized PRI (PRInorm). 3. Disease Incidence.  A number of researchers have demonstrated how a sUAS can be used for disease detection in horticultural crops. In an early study, Nebiker et al. (2008) demonstrated the potential of a prototype multispectral sensor and a sUAS to monitor plant health in a vineyard. Preliminary results demonstrated a strong correlation between plant health status obtained using the aerial system and groundtruth data. Calderon et al. (2013) demonstrated, using fixedwing aircraft and highresolution hyperspectral, multispectral and thermal sensors, the potential for early

detection of Verticillium dahliae Kleb., and the ability to discriminate among severity levels of verticillium wilt in olive. This same group (Calderon et al. 2014) is currently developing aerial methods to detect downy mildew (Peronospora arborescens Berk.) on opium poppy, using highresolution thermal and multispectral imagery. A serious global horticultural disease, Huanglongbing (HLB) or citrus greening disease, requires early detection and diagnosis for successful control. GarciaRuiz et al. (2013) compared imagery collected using a lowaltitude (100 m AGL) sUAS fitted with a multispectral sensor to hyperspectral images taken using a manned aircraft (590 m AGL). They concluded that the sUASbased data agreed better with ground reflectance measurements than manned aircraft data. Although the multispectral imagery was collected using a manned aircraft, the general protocol that was used to detect late blight in field tomatoes (Zhang et al. 2005) would be helpful to researchers using a sUAS. Bendig et al. (2012) used a small rotary UAS, equipped with either a thermal or multispectral sensor, to identify four sugar beet plots with varying degrees of disease resistance. 4. Weed Infestation.  In the not too distant future, aerial monitoring for weeds may be directly linked to herbicide applications applied by ground or aerial systems. With that in mind, Kazmi et al. (2011) evaluated the preliminary concept of using a team of air and ground vehicles for a holistic solution to weed monitoring and control. Several vegetation indices are used to monitor crop weed plants in agricultural research (Simelli and Tsagaris 2015). The use of aerial imagery acquired by a sUAS equipped with NIR/VIS sensors may be valuable to estimate botanical composition (Kutnjak et al. 2015), which may prove useful in weed management strategies. Creating georeferenced weed infestation maps would be useful in a successful cropweed management strategy. Spanish researchers have investigated several aspects of weed monitoring using a sUAS, changing from using a manned aircraft, to an unmanned aircraft, for sitespecific weed management using remote sensed imagery (PenaBarragan et al. 2012). Pena et al. (2013) were able to demonstrate the ability to map three categories of weed coverage in a corn field with 86% overall accuracy using a sUAS combined with objectbased image analysis (OBIA). These findings are promising for the implementation of sitespecific weed management (SSWM). This group continued their efforts (TorresSanchez et al. 2013) to help select the appropriate sensor and flight configuration to achieve early season sitespecific weed management in sunflower. They were able to demonstrate that a robust classification of sUAS images could be developed involving two steps: (1) discrimination of vegetation (weeds and crop) from bare soil using spectral information; and (2) discriminating weeds from crop rows using objectbased image analysis. Pena et al. (2014) demonstrated an approach to generate weed infestation maps in both herbaceous (corn and sunflower) and woody (poplar and olive orchards) crops. Pena et al. (2015) expanded on the previous results and highlighted some of the spectral and spatial challenges required to generate suitable weed maps. Using highresolution Kodak imagery, Herwitz et al. (2004) were able to map outbreaks of guinea grass in coffee fields.

Xiang and Tian (2011) outlined, in detail, a lightweight (14 kg) unmanned helicopter equipped with a multispectral camera that was interfaced with a ground station to collect field imagery. The system was used to monitor herbicidetreated turfgrass plots. While the sUAS system proved to be very promising for monitoring temporal changes in turf plots, the overall system would be useful to monitor other types of parameters.

B. Chemical and Nutrient Applications Japan has been a leader in the use of UAS for chemical applications. The primary aircraft used for this purpose in Japan has been the gaspowered Yahama RMAX which weighs 64 kg (141 lb) and so cannot be categorized as a “small” (50% by means of a standardized grading scale (Plate 3.2) (Ekman et al. 2008). The limit of acceptability of leaf blackening is subjective and may vary in retail according to the demand and availability of the product. Tolerance level is estimated at less than 20% of foliage which can be affected, whilst a severity of 10% or less could serve as a guideline of the tolerance maximum.

Plate 3.2. Grading scales to visually assess the impact of leaf blackening as a quality attribute in Protea (Ekman et al. 2008). Less than 10% leaf blackening is not considered greatly noticeable, whereas greater than 50% appears fully blackened The extent of leaf blackening varies widely among species (McConchie and Lang 1993a), cultivars within species (Paull and Dai 1990), within clones, and even between stems of the same treatment (Paull et al. 1980; Ekman et al. 2008), the harvest time during the day (Hoffman et al. 2014), the time of year or season (Paull and Dai 1990; Windell 2012), and from year to year (De Swardt et al. 1987; Windell 2012).

A strong relationship exists between leaf blackening and the climatic conditions that precedes harvesting. A study in Hawaii on four P. neriifolia selections (Paul and Dai 1990) revealed that early seasonharvested flowers – a time when hot and dry environmental conditions prevail – were found to be more prone to leaf blackening. A long period of drought that induces leaf blackening may relate to carbohydrate stress, as water stress may result in stomatal closure, and in turn can reduce the rate of carbon dioxide fixation and the carbohydrate levels in the leaves (van Doorn 2001). In South Africa, it is generally observed that the incidence of leaf blackening is accelerated when daily temperatures exceed 30 °C for two or more days preceding harvest. Harvesting is not recommended when temperatures exceed a threshold of 30 °C, especially when these temperatures are experienced in combination with high humidity (Malan 2012). Even though leaf blackening is generally viewed as an exclusively postharvest disorder, it was shown that in Protea ‘Sylvia,’ following an excessive warm period of weather, symptoms of leaf blackening may start to become noticeable preharvest in the field on inflorescencebearing shoots. Symptoms were first noted and were most severe within stems subtending inflorescences that were close to harvest maturity (Ferreira 2005). Windell (2012) reported in a threeyear study on Protea ‘Sylvia,’ a widely planted cultivar in South Africa due to its allyear round flowering habit, that two seasonal peaks of high leaf blackening incidences exist. The first peak was recorded from January to March, which coincided with midsummer and associated high temperatures. However, a second peak later in the year with consistently higher leaf blackening rates was reported during early to late spring (September–November), when much cooler conditions prevailed and when no or little water stress was likely to occur. During this period, extended cloudy weather close to harvest can significantly increase the susceptibility of the leaves to leaf blackening (Malan 2012). Phenological studies on ‘Sylvia’ have shown that carbohydrate concentrations in flowering shoots were significantly lowered in early spring, when rapid spring flush bud break and extension takes place (Hettasch 1999; Hettasch et al. 2001). This spring flush extension directly coincides with the time that flowering stems of ‘Sylvia’ are known to be most susceptible to postharvest blackening (Malan 2012; Windell 2012). Ferreira (2005) reported that suppression of the spring flush growth with a plant growth regulator, Paclobutrazol (Cultar™), significantly reduced postharvest development of leaf blackening by 39% at 14 days after treatment. Time of day for harvesting may be important in developing an operational strategy to lower leaf blackening, as suggested by Paull and Dai (1990), since leaves might be slower to blacken if harvested in the afternoon rather than in the morning. Windell (2012) confirmed this hypothesis and reported leaf blackening to be 40–60% less after 10 days of vase life in afternoonharvested stems compared to stems harvested in the morning, irrespective of the season of harvesting. Low sucrose and reducing sugar levels, together with high water content associated with harvesting in the early morning, was positively correlated to high incidences of leaf blackening (Hoffman et al. 2014). It is therefore recommended that, during winter or early spring, varieties prone to leaf blackening or destined for sea freighting be scheduled for harvest in the afternoon, provided that the ambient temperature is below 30 °C (Malan 2012).

In addition to these abovementioned factors, harvest stage is an imperative consideration when estimating the propensity of flowering stems to develop leaf blackening, as well as in the number of days to moderate bract curling within the flower. Leaves of flowers in the closed bud stage have been observed to blacken more rapidly than those where flowers of the harvested stems were just unfolding or were at a later stage (Malan 2012). The optimum stage for the flower at harvest was therefore considered to be when the involucral bracts had just started to retract from each other, but with all the florets still enclosed. Harvesting flowers at this stage, known as the ‘soft tip’ stage (prebract release) ( Plate 3.3a) ensures a longer total vase life than flowers picked at other stages (Paull and Dai 1990), and is also favored for export as the risk of insects contained within the inflorescence is minimized (Export Standards and Requirements 1997). The differences associated with flower maturity can be ascribed to continued floral growth in the young, closed bud, with higher respiration rates recorded in the more immature inflorescence (Ferreira 1986), and a surge of nectar production which follows just after opening of the involucral bracts (Plate 3.3b) (Dai and Paull 1995).

Plate 3.3. The maturity stages of a Protea ‘Sylvia’ inflorescence showing (a) a more immature, but harvestready softtip stage where involucral bracts are just starting to retract from the center and the florets, and (b) a more mature and fully opened inflorescence, but not yet senesced (Windell 2012)

III. PHYSIOLOGICAL CAUSES OF LEAF BLACKENING

A. Water Stress Product desiccation is listed by Malan (2012) as one of the prime factors known to accelerate the onset and severity of leaf blackening. In summer it is recommended that the time from harvesting to rehydration should not exceed 30 min, but should be kept to a minimum of 10 min, especially where cultivars that are prone to leaf blackening are concerned. Rehydration was even considered more important than cooling, as dehydration of the stems will shorten the shelf and vase lives of the product excessively. Water stress was previously proposed as a major cause of cell membrane damage and eventual leaf blackening development in Protea (Paull et al. 1980; Ferreira 1983; De Swardt et al. 1987; Paull and Dai 1990). Water stress can be induced by transpiration from the flower head, as water loss through the flower was estimated to vary between 25% and 50% of the total loss, agreeing with the total weight ratio between inflorescence (66%), stem (16%), and leaves (18%) (Paull et al. 2003). Water stress at the cellular level leads to an increase in membrane permeability (Ferreira 1983), which plays an important role in tissue browning. Leaf blackening can be eliminated by removal of the flower head (Paull et al. 1980), a response that was suggested to be largely due to the reduced water demand of the headless branches. However, in many flowers, water loss from the flowers or inflorescences contribute only a minor role to overall water loss (Halevy and Mayak 1979), although the Protea inflorescence is substantially larger in proportion to the stem than in most other cut flowers. Of interest is that a girdling treatment to the stem just below the flower significantly reduced leaf blackening, but had no effect on water uptake, suggesting that the effect of water loss from the head is not necessarily the major factor initiating leaf blackening. Additional evidence supporting the absence of a role for water stress in the development of early leaf blackening include experiments by Newman et al. (1990) and by Reid et al. (1989), where placing a plastic bag over the flower head to reduce loss of water by transpiration did not reduce leaf blackening. Furthermore, the presence of antimicrobial compounds such as sodium hypochlorite, trichloroisocyanuric acid, hydroxyquinoline citrate and silver nitrate in the vase solution, which would reduce water stress by eliminating bacterial blockage, had no effect on the incidence of leaf blackening (Bieleksi et al. 1992; van Doorn 2001). Finally, a short period of water stress prior to vase life where stems were held dry for about 10 h at 20  °C and 60% relative humidity did not promote blackening (Reid et al. 1989). In fact, when ‘Sylvia’ stems were dehydrated to 85% of their original fresh weight in a study by Windell (2012), leaf blackening was significantly reduced compared to fully hydrated stems. These results are in conflict with the significant relationship found between moisture and leaf dry weight loss and the rate of leaf blackening (Ferreira 1983; De Swardt et al. 1987). Alternatively, when Du Plessis (1978) continuously recut stem bases and replaced the holding solution daily throughout the vase life of P. neriifolia, leaf blackening was reduced. Leucoanthocyanins, which rapidly oxidize to form highmolecularweight polymerized tannins in water, were reported to leach from Protea stems in the vase solution, causing the blockage of xylem vessels and inevitable water stress and the onset of leaf blackening (De Swardt et al. 1987). A reduction in leaf blackening was reported when phenyl mercury acetate

(Masie 1979) and lead acetate (Du Plessis 1978; De Swardt 1979) were used to precipitate these phenolic compounds from the vase solution. However, it is not clear whether the reduced leaf blackening could be ascribed to unblocking of the stem by precipitating the tannin, or to a direct effect of the applied chemicals (Jones et al. 1995). The role of stomata and abscisic acid (ABA) in the control of leaf blackening has not been studied extensively. Evidence towards water stress contributing to leaf blackening was provided by Paull and Dai (1990), who showed that coating the leaves with antitranspirants such as “Exalt” and “Folicote” delayed leaf blackening to some extent, compared to the control. However, leaf blackening could not be delayed to the same extent than when the flowering head was removed. Still, the reduction in leaf blackening caused by ABA and other commercial antitranspirants showed promise and requires further attention. As opposed to water stress, Reid et al. (1989) observed that free water on the leaves, due to condensation, significantly increased the incidence of blackening. This finding is confirmed in the grower’s manual by Malan (2012), where harvesting of wet flowers is discouraged not only to avoid wound infections but also to minimize leaf blackening in those Protea varieties with inherent problems. Furthermore, it was recommended that, under conditions when the cold chain cannot be maintained, it is most often better not to cool the flowers below ambient shade temperatures, since condensation on the flowers will cause more damage than the few hours that the flowers were not cooled. Dew and rainfall, in addition to condensation, may thus relate to leaf blackening, although there are no data to support these claims, or to explain the mechanism of leaf blackening as a result of free water on leaves (van Doorn 2001). In light of conflicting evidence, leaf water stress has been considered a contributing factor, although not exclusively the primary cause of leaf blackening in cut Protea inflorescences (Jones et al. 1995).

B. Carbohydrate Stress The majority of studies on postharvest leaf blackening have concluded that it is caused by carbohydrate stress, mainly due to carbohydrate depletion of the leaves to sustain the continuous development and respiration rates of the flower (Brink and De Swardt 1986; Reid et al. 1989; Newman et al. 1990; McConchie et al. 1991; Bieleski et al. 1992; McConchie and Lang 1993b; McConchie et al. 1994; Dai and Paull 1995; van Doorn 2001, Stephens et al. 2001, 2005). Flowers have high rates of respiration and growth, and therefore are major sinks for carbohydrates (Reid et al. 1989). The concentrations of starch and sucrose, considered to be the metabolically active carbohydrates in Protea, have been found to decline rapidly after harvest (McConchie et al. 1991; Bieleski et al. 1992; McConchie and Lang 1993a,b; Ferreira 2005). These observations follow from experiments with girdling, that blocks the phloem transport of sugars from leaves to the flower head, from experiments using darkness or additional light (Reid et al. 1989), and during coldstorage (Ferreria 2005). In the light, leaf blackening is inhibited by the production of carbohydrates as a result of photosynthesis (McConchie et al. 1991; Bieleksi et al. 1992), or by the maintenance of an environment that prevents phenol oxidation (Jones and ClaytonGreene 1992).

Analysis of sugars and starch levels during cold storage demonstrated a significant decrease in leaf concentrations (Stephens et al. 2001, 2003b). Stephens et al. (2003b) explored the importance of glucose for the control of leaf blackening in Protea ‘Sylvia’ rather than relying on sucrose, which is commonly used elsewhere in the ornamental industry as an energy pulse. Glucose and fructose concentrations in both the leaves and flower head were observed to be higher compared to sucrose. Fructose concentrations in the flower head did not change significantly during the experiment, while leaf fructose concentration decreased significantly only after removal from storage. However, leaf glucose concentrations declined consistently and rapidly with storage at 0 °C for 3 days and during vase life at 18 °C. Furthermore, the use of 2.5% glucose as a holding solution increased vase life substantially (Stephens et al. 2003b). In a further study, where sugar concentrations were observed in Protea ‘Sylvia’ during cold storage for 3 weeks at 1 °C, glucose and starch concentrations decline more in the stems and leaves than in the inflorescence (Stephens et al. 2005). Interestingly, analysis of foliage of several Protea cultivars revealed the presence of a polymeric soluble carbohydrate present in significant concentrations, namely polygalatol (Bieleksi et al. 1992). Polygalatol (1,5anhydroDglucitol) is a sugar alcohol and a simple derivative of sorbitol (Dglucitol). After harvest, when the concentrations of the reported reducing sugars, sucrose and starch concentrations had declined in Protea foliage, levels of polygalatol remained relatively constant over time in both P. neriifolia and P. eximia (Bieleksi et al. 1992; McConchie and Lang 1993a). Ferreira (2005) found similar results for the hybrids Protea ‘Sylvia’ and ‘Lady Di’ (P. magnifica × P. compacta), where polygalatol remained fairly constant directly after harvest and throughout longterm storage. Apparently, polygalatol does not contribute to the metabolically active carbohydrate pool and is thought to play a dynamic role in osmotic buffering of this Mediterraneanadapted species.

C. Role of Respiration The respiration rate of cut flowers in general is considered high in comparison to other plant organs, such as tubers. For Protea ‘Sylvia,’ Stephens et al. (2003b) reported respiration rates of intact stems (7.43 ml kg–1 h–1) versus stems stripped of leaves (7.07 ml kg–1 h–1) to be significantly higher than that of decapitated stems (5.48 ml kg–1 h–1), thus confirming that the flower head accounts for the majority of respiration at both 0 °C and 4.5 °C. The onset and incidence rate of leaf blackening has been found to accelerate with an increase in vase life temperature (Jacobs and Minnaar 1977; Ferreira 1983, 1986). This correlation of respiration with the prevailing environmental temperature is not surprising, as there is a wellknown logarithmic increase in respiration with increased temperature; hence the importance of cooling cut flowers to decrease the rate of senescence (Kader 2002). Under conditions of coldstorage, Stephens et al. (2001) clearly showed in Protea ‘Sylvia’ that increasing storage temperatures from 0 °C to 10 °C for a 3day period resulted in an increased incidence of leaf blackening. The rate of respiration also differs greatly from one part of the Protea inflorescence to another. For example, there is a marked difference between the respiration rates of florets in different

positions on the Protea inflorescence, as well as of florets from inflorescences cut at different stages of development. Respiration rates were reported to be higher in immature florets than mature florets of P. neriifolia which, therefore, have a higher carbohydrate demand (Ferreira 1986). Furthermore, nectar production was reported to increase as flowers reached optimal picking stage (Dai and Paull 1995). The process of depletion of the substrates available for respiration begins soon after harvest of the cutflower, and eventually leads to a decrease in the rate of respiration (Ferreira and De Swardt 1980). This results in the hydrolysis of structural cell components (Halevy and Mayak 1979). In senescing tissue, this is accompanied by a decrease in membrane permeability, followed by a loss of protoplasmic compartmentalization and the mixing of cellular contents (Teixeira da Silva et al. 2003) until, eventually in the case of Protea, leaf blackening manifests. The maintenance of normal respiration, however, depends not only on the availability of substrates but also on an adequate water status in the cells to ensure normal functioning of physiological processes in the plant (Reid and Jiang 2012).

D. Nectar Production Dai and Paull (1995) concluded that leaf blackening in Protea is the result of depletion of carbohydrates used for the expansion and respiratory metabolism of the inflorescence, but most importantly due to the demand of sugars for nectar production. Since Protea species with a strong tendency to blacken, such as P. eximia, P. compacta and P. neriifolia, produce high volumes of nectar, carbohydrate depletion from their leaves by this strong carbohydrate sink has always been strongly associated with the early onset and development of leaf blackening (Mostert et al. 1980; Cowling and Mitchell 1981; Ferreira 1986; Paull and Dai 1990; Dai 1993). Protea nectar consists primarily of glucose, fructose, sucrose (in selected species), as well as xylose (Cowling and Mitchell 1981; Van Wyk and Nicolson 1995; Jackson and Nicolson 2002; Schmid et al. 2015). Dai and Paull (1995) found that after applying radioactive 14Clabeled sucrose to P. neriifolia, 50% of the radioactivity could be detected in the nectar after 24 h. Van Doorn (2001) thus recommended that the role of nectar production as a sink for leaf carbohydrates be further investigated by using inhibitors as such carbonyl cyanide mcholorphenyl (CCCP) and pchloromercuribenzenesulfonic acid (PCMBS). However, of interest is that P. cynaroides is known to produce high volumes of nectar, yet leaf blackening is seldom observed under South African conditions (G.M. Littlejohn, personal communication, 2002).

E. Light The importance of photosynthesis to prevent the onset of leaf blackening was demonstrated by Jones and ClaytonGreene (1992), when the incidence of leaf blackening considerably increased when the photosynthetic inhibitor 3(3,4dichlorophenyl)1,1dimethylurea (DCMU) was used to interfere with the photosynthetic electron transport chain of photosystem II in stems held under light. Even low levels of photosynthetically active radiation (PAR) of between 15–25 μmol m–2 s–1 were found to significantly reduce leaf blackening (Newman et al. 1990). Placing floral stems in bright light (Reid et al. 1989; Paull and Dai 1990;

McConchie et al. 1991) delayed or even prevented leaf blackening. Alternatively, leaf blackening symptoms were more rapidly expressed when the flowers were held in the dark at room temperature (Newman et al. 1990). The starch and sucrose concentrations in leaves declined in stems held in the dark compared to those held in the light (McConchie et al. 1991; Bieleski et al. 1992; Crick and McConchie 1999). Transport of potted Protea plants, particularly those transported in darkness and at high temperature, also resulted in leaf blackening. The leaves that blackened were observed to occur just below the vigorously growing top leaves. The uppermost growing leaves are likely to be a strong sink for carbohydrates and the subtending, fullgrown leaves, a main source of these compounds (van Doorn 2001). However, damage to P. cynaroides after midrange transportation were rather ascribed to high CO2 concentrations than to shortterm darkness over a period of 4 days (D’Andrea et al. 2007). Van Doorn (2001) stated, based on personal communications with growers, that leaf blackening has been known to be reduced where stems were held under incandescent lights in pack rooms, as well as in cold rooms. Phytochrome pigments, which will be saturated by continuous exposure to red light known to be emitted from incandescent lamps, will thus be converted to the biologically active farred light P fr form, which has been implicated in the control of a number of physiological leaf problems such as yellowing associated with ornamentals such as Alstroemeria (Van Lieberg et al. 1990). In a more recent study, Hoffman and Du Plessis (2013) showed that low light, as opposed to darkness, during cold storage at 2 °C for 21 days resulted in little or no chilling injury in three Leucospermum cultivars, implying that light may be important for membrane stability and chilling injury of these temperaturesensitive plants. This notion was further supported in a study by Hoffman et al. (2015), in which LED lights at an energy level of ±25 μmol m–1 s–1 significantly controlled chilling injury and leaf yellowing in Leucospermum potted plants under longterm storage at 0 °C, thus improving quality. The role of light in Protea leaf blackening needs to be further assessed.

F. Ethylene Although many cut flowers are detrimentally affected by ethylene, there is little evidence for the role of ethylene in the process of leaf blackening in Protea. Ethylene released by the leaves was not related to blackening in a number of species (McConchie and Lang 1993b). Although data were not shown, Stephens (2003) also reported no difference in vase life in terms of leaf blackening when Protea ‘Sylvia’ was exposed to a continuous air flow containing ethylene at 50 μl l–1. Also, a short pulse with 4 nmol silver thiosulfate (an inhibitor of ethylene action) prior to vase life had no effect on the time to onset of leaf blackening in P. eximia (Reid et al. 1989; Bieleski et al. 1992). When Protea ‘Venus,’ a natural hybrid between P. repens × P. aristata, was exposed to ethylene at a concentration of 2 μl l–1 for a period of 24 h, the red flowering head presented with black bract markings, although the leaf quality remain unaffected (E.W. Hoffman, personal observation). On the contrary, van Doorn (2001) suggested that the presence of fruit, such as apples, during postharvest storage and transport may have a

negative influence on leaf blackening occurrence in P. magnifica.

IV. THE BIOCHEMICAL MECHANISMS OF LEAF BLACKENING Although carbohydrate depletion being one of the primary reasons for blackening is plausible, the link between these responses is far from being unequivocal (van Doorn 2001). Cell death is excluded, as constant respiration rates have been measured on blackened whole leaves (McConchie and Lang 1993b), in comparison to pulsed, nonblackened leaves. However, as this was an isolated study, van Doorn (2001) suggested that a “falsepositive” respiration rate could have been recorded, as oxidation reactions involved in leaf blackening could also have accounted for the measured respiration rates. Following the process of carbohydrate depletion in cells, protein degradation is most often reported, leading to an accumulation of ammonium in the vacuoles (van Doorn 2001). It is, however, unknown whether this mechanism operates in Protea and is, therefore, involved in the trigger of the leafblackening process. The degree that leaf blackening can be ascribed to enzyme activity is unclear as well as the involvement of specific enzymes, if any. There is support for the hypothesis that the direct cause of the leaf blackening in Protea, induced by various possible stresses, is the oxidation of hydroxyphenols and tannins (Paull et al. 1980; Ferreira 1983). This oxidation process can occur either enzymatically or nonenzymatically (Kader 2002). Phenolic and flavonoid compounds, which are colorless in a reduced state but rapidly turn brown or black when oxidation and polymerization take place (De Swardt 1979), are known to be abundant in Proteaceae (Van Rheede van Oudtshoorn 1963). Mulder (1977) ascribed the blackening of P. neriifolia leaves primarily to the oxidation of leucoanthocyanins, while Jones and Clayton Greene (1992) associated leaf blackening with polyphenol oxidase activity. Hernández et al. (2012, 2014) also suggested that blackening may be linked to polyphenol oxidase and peroxidase activity, as stems of Protea ‘Pink Ice’ and ‘Susara’ treated with ascorbic acid presented less leaf blackening and had lower phenol and polyphenol activity compared to untreated stems. In contrast, Whitehead (1979) and Whitehead and De Swardt (1982) found air oxidation of flavonoid compounds to be the major cause of leaf blackening in P. neriifolia. Carbohydrates in leaves of Protea species susceptible to leaf blackening have been reported to be bound with phenolics in the form of an Oglycoside ester (Perold et al. 1973, 1979; Perold and Carlton 1989; Perold 1993). It has been hypothesized that a continuous demand from the large inflorescence would require the release of the sugar component, resulting in a highly reactive free phenol which is oxidized nonenzymatically (Perold 1993). This oxidation process appears to take place relatively slowly, and may explain the delay in the appearance of leaf blackening. Phenolic concentration determined on removal from storage was significantly higher than the prestorage concentration, possibly as a result of synthesis (Stephens et al. 2001). Blackening reactions in plants are often due to the activities of the peroxidase or polyphenol oxidase (PPO). When peroxidase and PPO activities were measured in vitro in ‘Pink Ice,’

either held in darkness or 12 h light (McConchie et al. 1994), no increase or significant difference was found between treatments for levels of peroxidase activity. There was, however, a 10fold increase in PPO activity for stems held in the light (and which did not blacken during the recorded period). Leaf PPO activity as measured in vitro, was recorded to be high in P. neriifolia. When PPO activity was similarly measured in Leucospermum, no activity could be detected (Dai and Paull 1997). It was thus hypothesized that Leucospermum, which does not develop leaf blackening, might contain an inhibitor of Protea PPO. In vitro assays may not be a true reflection of the in vivo processes, as in vitro analysis requires the enzyme to be released from the chloroplast or peroxisomes to react with its substrate (van Doorn 2001). Thus, membrane degradation or disruption is still required to allow contact with the phenolic substrate in the vacuole. McConchie et al. (1994) however found no association between PPO and leaf blackening, and no convincing evidence of membrane degradation preceding the appearance of leaf blackening in Protea leaves. This conclusion was made based on a zero increase detected during the onset of leaf blackening in oxidized glutathione, which is indicative of an oxidative stress, or malondialdehyde, a byproduct of lipid peroxidation, also an indication of loss in membrane integrity. An osmotic stress as the cellular cause for leaf blackening was rather suggested, as that may lead to a release of free phenol and subsequently the nonenzymatic oxidation thereof, although PPO can thus not be entirely excluded in the process of blackening and thus requires further investigation. In vitro measurements of βDglucosidase activity, which was found to significantly increase just before the onset of leaf blackening, also face similar experimental challenges, as experienced in the determination of PPO values. Jones and Cass (1996) investigated the role of glucosidase by supplying Protea cut stems with solutions containing ions of Zn2+ and Cu2+ or immersing cut flower stems in either a 20% or 50% ethanol solution, whereafter the enzyme activity was measured in vitro. βDglucosidase activity was reduced when zinc or copper was added, although a delay in leaf blackening was not reported with the addition of these compounds (Jones and Cass 1996). Leaf blackening was however significantly delayed when ethanol was used, but as ethanol is known to inhibit a range of enzymes this was not sufficient proof for the involvement of βDglucosidase activity in leaf blackening (van Doorn 2001). The lack of convincing evidence linking enzymatic oxidation of phenol and leaf blackening has resulted in the search for alternative enzymes that may be involved (Jones et al. 1995). According to the current ideas on blackening, the substrate for the reaction resides in the vacuole, while some of the enzymes that may be involved are present in the cytoplasm or in organelles such as the peroxisome (peroxidase) or the chloroplast (PPO). To establish contact between the substrate and enzyme would require a breakdown of some of the membranes, or at least that of the vacuole (van Doorn 2001). The absence of membrane degradation products and indications of a functional antioxidant system, as well as a high respiration rate, provides little evidence for the required membrane breakdown to satisfy the above hypothesis. Alternatively, the substrate or enzyme may both be localized in the same organelle, such as the vacuole, but this postulation requires a cellular trigger that would activate the leafblackening process at certain times, whereas it would be absent at other times. van Doorn (2001) also

raised the possibility of βDglucosidase activity in the cell walls and cytosol, in addition to that of the vacuole, as was the case for other wellstudied crops such as barley and oat leaves. Cellular localization of blackening, therefore, remains unresolved and requires further ultrastructural analysis by means of transmission and scanning electron microscopy (van Doorn 2001).

V. CONTROL OF LEAF BLACKENING Studies to obtain a greater understanding of the nature, cause, and mechanism of leaf blackening are all largely aimed at developing methods and techniques to inhibit, control, or reduce leaf blackening so as to provide a highquality product with commercial longevity, in order to meet the expectations of a discerning international floricultural market. A number of approaches have been followed to provide some control over leaf blackening, but to date no method can successfully eradicate this serious disorder completely.

A. Girdling In ornamental stems, girdling is a process by which a ring of bark (phloem) just below (5mm) the inflorescence, is removed. Girdling can be achieved by severing the phloem either by means of a knife or through using less laborintensive devices such as an electrified metal wire or laser beam technology (van Doorn 2001). Girdling was the first practical treatment available to producers and exporters for providing some control over leaf blackening (Reid et al. 1989; Newman et al. 1990), and is still considered a viable option to reduce leaf blackening in cultivars that are nonresponsive to pulsing with glucose. By girdling, the sink demand from the inflorescence is interrupted, but without causing any interference to the water uptake process. The result is a reduction of the rapid depletion of carbohydrate from the leaves subtending the inflorescence, and subsequently delays leaf blackening and significantly extends vase life (Brink and De Swardt 1986; Dai and Paull 1995). Caution should be applied when using girdling, however, as removing a ring of bark can weaken the stem below the inflorescence and risk the inflorescence snapping off during handling, transport, or when using in floral displays (Windell 2012). In most varieties, however, the risk is rather that the cut is not made deep enough to be effective (Malan 2012). Although girdling, as a timeconsuming and laborintensive practice, has greatly been replaced by sugar pulsing, it still is most effective when used during lowtemperature storage (Stephens et al. 2001) or in combination with sugar or ethanol pulsing (Windell 2012).

B. Sugar Pulsing Pulsing stems with sugars, shortly after harvest, has become an essential treatment in delaying leaf blackening in susceptible cultivars during longterm cold storage and transport (Jones 1991; McConchie and Lang 1993a,b; Stephens et al. 2005). Pulsing is required as soon as possible after harvest, as reserve carbohydrates in leaves on Protea flowering stems (mainly in the form of starch) rapidly decline within the first 24 h (McConchie and Lang 1993a,b; Ferreira 2005).

Pulsing, though very effective and reliable when administered under controlled laboratory conditions, has proved to be problematic to execute at the producer level. A serious disadvantage of pulsing is the unpredictable rate of uptake of the pulsing solution. Not only can recommendations vary significantly for the amount and type of carbohydrate required for loading into stems amongst cultivars and species, but it is most likely to differ between respective harvesting seasons due to microclimatic conditions, and between circumstances of different stem characteristics. Very often, pulsing duration is the only guideline provided in pulsing protocols apart from the recommended sugar concentration. However, factors such as the vapor pressure differential between the flowering stems and the environment driving transpiration, as well as the differential of the water potential of the stem and that of the osmotic potential of the solution, may seriously impact on the rate and volume of pulse uptake within a certain period of time (Windell et al. 2015). In a study by Windell et al. (2015), Protea ‘Sylvia’ flowering stems were dehydrated by 0, 5, 10, and 15% postharvest, or were pulsed with glucose concentrations ranging from 0% to 17.7% for 4 h. Pulse uptake volumes were found to increase with a decrease in stem hydration from 100% to 85%, whilst solution uptake decreased with an increase in glucose concentration. Similarly, when flowering stems were harvested at 8:00, 12:00 or 15:00 h on a relatively cool (20 °C maximum temperature) or on a relatively warm day (32 °C maximum temperature), stems harvested on both days at 8:00 h absorbed significantly less pulse solution after 4 h than stems harvested later in the day (Hoffman et al. 2014). From these results it was concluded that harvesting later in the day was more beneficial in the reduction of leaf blackening, as it facilitated a faster and more efficient uptake of the pulsing solution, due to increased xylem tension relating to decreased water potential, as was likely to be present in stems harvested later during the day. The rate of glucose uptake is however important, as a supraoptimal rate of glucose accumulation in stems may lead to phytotoxicity in the leaves. Phytotoxicity is typically observed when supraoptimal sugar concentrations are pulsed, or when stems are exposed to an optimal concentration of sugar, but over longer than recommended periods, or when accumulation takes place at faster rates than recommended (Meyer 2003; Stephens et al. 2003a; Malan 2012; Windell et al. 2015). Stephens et al. (2003a) reported phytotoxicity in all proteas pulsed with 10% glucose (at 18 °C for 24 h), except for ‘Susara,’ whereas ‘Pink Ice’ toxicity symptoms were observed at glucose pulse concentrations of 4% and above. Phytotoxicity symptoms appear similar to those of the leafblackening disorder, but with the exception that the lesions are more a lightbrown color than black, and usually occur in the leaves at the distal end where the pulsing solution is administered (Plate 3.4a) (Windell 2012). After storage or later during vase life, phytotoxicity may also manifest and be detected as bract browning in the involucral bracts (Plate 3.4b) (Windell 2012). Malan (2012) recommended pulsing as being particularly suited to seafreighted flowers, where maintenance of the cold chain can be controlled with greater success. Heavily pulsed flowers can collapse when exposed to stress situations associated with air transport. For such flowers, a pulse with a lower sugar concentration, and/or citric acid alone at a concentration of 150–250 mg l–1 is recommended, mainly to set the water pH and to prevent algae growth.

Plate 3.4. Protea ‘Sylvia’ displaying leaf blackening and phytotoxic symptoms in the leaves (a) and after storage at the tips in the involucral bracts of the inflorescence (b) (Windell 2012) Sucrose vase solutions of about 2 g l–1 have been shown to be beneficial in delaying the onset of postharvest leaf blackening in P. compacta (Haasbroek et al. 1973), P. neriifolia (Brink and De Swardt 1986; Paull and Dai 1990; McConchie et al. 1991), and P. eximia (Bieleski et al. 1992). Sucrose as an effective vaselife preservative was considered essential for maintaining membrane integrity, and thus subsequently delaying the onset of leaf blackening (Coorts 1973). Holding solutions with sucrose at higher concentrations were, however, found to aggravate leaf blackening (Newman et al. 1990; Paull and Dai 1990; Jones et al. 1991). Sucrose, when supplied as a pulse solution at 200 g l–1 also significantly delayed the onset of leaf blackening in some proteas, such as P. cynaroides (Jones et al. 1991) and P. neriifolia (McConchie and Lang 1993b). For sucrose pulsing, a 12 h pulse period was recommended (Brink 1987), as longer pulse periods of 18 h or more lead to the flower head becoming the preferred sink (Brink and De Swardt 1986). Dai and Paull (1995) confirmed that a 5% sucrose solution stimulated nectar production in Protea, in addition to reducing leafblackening symptoms. Pulsing stems of Leucadendron ‘Safari Sunset’ with either 5% sucrose or glucose prior to storage, followed by the inclusion of 2% sucrose in the holding solution, also improved the carbohydrate balance and reduced leaf blackening during sea transport (PhilosophHadas et al. 2007, 2010). These treatments, however, were used in combination with preservatives as well as controlled atmosphere (CA) technology (PhilosophHadas et al. 2007, 2010). In contrast with the above studies reporting vaselife improvements with sucrose postharvest supplementation, other authors have found that neither vase nor pulsing solutions containing sucrose extended vase life or reduced leaf blackening in certain Protea cultivars and species (Stephens et al. 2001, 2003a; Malan 2012; RodríguezPérez et al. 2012). Glucose pulsing at 1% or 2%, after a cold storage period of 10 days, significantly extended vase life in ‘Brenda,’ ‘Cardinal,’ ‘Carnival,’ ‘Pink Ice,’ ‘Susara,’ and ‘Sylvia,’ as well as in ‘White Pride’ (P. longifolia selection), ‘Lady Di’ (P. magnifica × P. compacta) and ‘Candida’

(P. magnifica × P. obtusifolia) (Stephens 2003a,b, 2005). However, glucose had little or no effect on postharvest leaf blackening of P. magnifica, P. grandiceps, P. cynaroides, ‘Ivy’ (P. lacticolor selection) and ‘Venus’ (P. repens × P. aristata) (Stephens et al. 2005). Stephens et al. (2005) associated Protea species or hybrids with one parent belonging to the section Ligulatae with a genetic responsiveness to glucose in reducing leaf blackening. Ferreira (2005) further investigated the response of various Protea cultivars and species to a range of sugars, by providing a 2% holding solution of either sucrose, fructose, glucose, or galactose. Glucose, however, consistently gave the best results in all species or cultivars evaluated. Further research on the varied responses among species and/or cultivars to glucose and sucrose supplementation, and how they metabolize these respective sugars, is needed. Trials were conducted by Ekman et al. (2008) on P. eximia ‘Duchess’ and ‘Pink Ice’ to determine whether glucose pulses (pulsed for 20 h at 20 °C) could effectively reduce leaf blackening in Australiangrown flowers. This study found all nonstored stems, pulsed with a 6% glucose solution prior to vase life, to be free of leaf blackening over 8 days of vase life. Stems that were pulsed with 3% sucrose prior to storage, for either 2 or 3 weeks at 1 °C, developed significantly more leaf blackening than any of the glucose pulse treatments. A 9% glucose pulse treatment was found to be most effective in reducing leaf blackening in comparison to the control, but caused unacceptable damage to the flowers. Pulsing with 3–6% glucose promoted ‘Pink Ice’ stem quality during coldstorage, as required for sea freight. In a study conducted by Windell (2012), as reported in Windell et al. (2015) on ‘Sylvia,’ where stems were pulsed prior to extended coldstorage with an increasing range of glucose concentrations, it was noted that 10 ml pulses of concentrations between 4.7% and 9.4% glucose not only significantly delayed leaf blackening but also maintained a positive water balance for longer during vase life. This positive water balance, associated with glucose pulsed stems, resulted from a combination of better water retention by stems as well as lower transpiration rates in the vase when compared to the control, nonpulsed stems. Stomatal conductance data suggested that pulsed glucose may play a role in protecting stomatal functionality, and therefore assist in delaying and lowering the incidence of leaf blackening. The inclusion of an antimicrobial compound is imperative to control bacterial growth in sugarcontaining solutions by preventing blockage of the stems, which would inhibit water and sugar uptake (van Doorn 2001). However, Meyer (2003) and Stephens et al. (2005) found most of the Protea cultivars to be sensitive to a standard concentration (0.5 g l–1) of the disinfectant and bleaching agent, sodium hypochlorite (NaOCl). Leaf blackening was exacerbated in nonpulsed, control stems containing hypochlorite compared to nonpulsed stems held in water only. Malan (2012) recommended wetting agents be added to the holding water up to 0.1% prior to grading and packing, in order to improve water uptake, but not aggravate leaf blackening.

C. Temperature Control Cooling Protea cut flowers rapidly directly after harvest, grading and packing to reduce the field temperature and keeping the produce cold, thereby significantly lowering the respiration

rate, can be effective in reducing the incidence of leaf blackening. Cooling can also subsequently improve the transport ability of the product dramatically (van Doorn 2001). The immediate removal of field heat from flowers was, however, only worthwhile if low temperatures were maintained en route to the wholesaler or retailer. According to Gollnow and Worrall (2010), for P. cynaroides, removing the field heat of flowers by immediate cooling, followed by either forcedair cooling or overnight holding in a cool room is important to retaining quality and maximizing vase life. Vacuum cooling followed by 3 days at 0 °C was found to extend vase life by 2 days compared to forcedair cooling (Stephens et al. 2003b). However, after long storage periods (14–21 days) the benefit obtained through vacuum cooling was not reflected in an extended vase life. The quality of most cutflowers is maintained effectively at storage temperatures of between 0 and 1 °C (Reid 1992). Several studies confirmed this also to be the case for Protea, as a significant reduction of leaf blackening could be achieved by decreasing storage temperatures. The vase life of several Protea species stored at 2 °C for 42 days was reported to be satisfactory (Haasbroek et al. 1973). Suppression of leaf blackening was also documented for P. eximia cutflowers stored at 2 °C (Paull et al. 1980). Stephens et al. (2001) reported respiration and leaf blackening of ‘Sylvia’ to be negatively correlated with temperature, while starch, reducing sugar and total carbohydrate concentrations were each decreased significantly with increasing temperature. This lower carbohydrate content concurs with reports that the carbohydrate content of P. eximia and P. neriifolia declined rapidly after harvest (Bieleski et al. 1992; McConchie and Lang 1993b, respectively). Flowerbearing stems held at 0 °C had a significantly longer vase life than those held at 4.5 °C (Stephens et al. 2003b). Although low temperatures are essential for longterm storage, leaf blackening was found to increase significantly with length of storage (Stephens et al. 2001). Ekman et al. (2008) confirmed that the storage of ‘Pink Ice’ stems for either 2 or 3 weeks at 1 °C increased leaf blackening compared to nonstored stems. Interestingly, this study reported that most stems were fully green on removal from storage, but blackening developed rapidly once stems were warmed to 20 °C and continued to progress during vase life. The vase life of P. cynaroides can be maximized if stored at 2–4 °C during storage and transport, under highhumidity conditions, while avoiding condensation which is known to exacerbate the development of leaf blackening (Gollnow and Worrall 2010).

D. Controlled Atmosphere Storage in a controlled atmosphere (CA) containing a low (1%) oxygen concentration and a high (5%) carbon dioxide concentration, delayed blackening, at least during the period of storage (Jones and ClaytonGreen 1992). However, Stephens et al. (2003b) found that CA storage at 0 °C using either 10.5 kPa O2 in combination with 10.5 kPa CO2, or 5 kPa O2 with 5 kPa CO2, or 5 kPa O2 (balance as N2) had no effect on vase life of Protea ‘Sylvia’ compared to flowers stored in air. Alternatively, CA may only be effective to suppress leaf blackening while in storage; van Doorn (2001) referred to evidence which suggested that, upon the removal of stems from CA, leaf blackening rapidly developed. For Leucadendron ‘Safari

Sunset,’ however, CA storage ensured preserved quality and reduced leaf blackening, when coupled with sucrose pulsing before transport, during prolonged sea freight (PhilosophHadas et al. 2010). Crick and McConchie (1999) evaluated the efficacy of ethanol vapor to control leaf blackening during storage (control stems that were kept in closed bags), and found decreased blackening during vase life. The reduction in leaf blackening was ascribed to the presence of elevated CO2 levels from the respiring stems that subsequently inhibited respiration and reduced the use of carbohydrate reserves. This research was based on a study by Ghahamani and Scott (1998), who showed that ethanol vapor would successfully inhibit apple scald, a postharvest physiological disorder that develops on the fruit skin of apple or pears during or after low temperature storage. In apple scald, the oxidation of αfarnesene (a naturally occurring volatile terpene in pome fruit) to its conjugated trienes and 6methyl5hepten2one (MHO) has been associated with a typical brownblack discoloration (Ingle and D’Souza 1989), similar to that of leaf blackening. Investigations into alternative control mechanisms for apple scald following the ban of diphenylamine (DPA) postharvest drenches identified 1 methylcyclopropene (an ethylene inhibitor), CA, dynamic controlled atmosphere (DCA) and initial low oxygen stress (ILOS) as technologies with the potential to inhibit superficial scald and increase the storage life of pome fruits (Robatscher et al. 2012; Zanella and Strüz 2013). Metabolomic studies showed that storing fruit under DCA and/or ILOS suppressed superficial scald development on ‘Packham’s Triumph’ pears, most likely by inhibiting the autooxidation of αfarnesene to its byproduct MHO in the fruit peel (Ramokonyane 2016). The efficacy and applicability of CA technology to control leaf blackening in Protea, similar to that of apple scald in pome fruit, thus evidently requires further research.

E. Ethanol Several studies have reported on the advantages using ethanol to reduce leaf blackening, either as a vapor (Crick and McConchie 1999), as a pulsesolution, or as part of a vasesolution formulation (Jones and Cass 1996; Cannon and McConchie 2001; Windell et al. 2010; Windell 2012). Ethanol vapor at an optimum concentration of 5.6 g ethanol kg–1 fresh weight was found to significantly suppress leaf blackening to below 20% for 2 weeks in Protea ‘Pink Ice’ held at 20 °C compared to the control. Leaves exposed to the higher end of the range at 11.2 g ethanol kg–1 resulted in toxicity by showing enhanced leaf blackening from the onset of vase life (Crick and McConchie 1999). Though the physiological process involved by which ethanol reduced leaf blackening was not explored in this trial, a possible link between reductions in PPO activity via acetaldehyde was suggested. Cannon and McConchie (2001) evaluated ethanol, both as a pulse and as a holding solution, as possible postharvest treatments, across a range of vapor concentrations. The efficacy of other alcohols, including butanol and propanol, to control leaf blackening, as well as different ethanol application methods such as slowreleasing sachets which could be commercially applicable, was included in the study. Results showed that leaf blackening reduction was best obtained in Protea ‘Pink Ice’ at a vapor concentration of 6–7 g ethanol kg–1 fresh weight.

Longerchain alcohols were found to be ineffective in delaying leaf blackening. In general, leaves appeared to be very sensitive to ethanol vapor, which proved to be toxic at high concentrations or when released at a high rate. The commercial application of gaseous ethanol would therefore have to be carefully controlled to avoid damage to stems. The use of various sachet types was thus investigated to facilitate the required release rate. Sachets constructed of polyethylene showed a consistent rate of release, but this rate was too low at 1.5% within 24 h to be effective in delaying leaf blackening. Other sachets, such as those constructed of Tyvek® or Antimold®, released ethanol rapidly and unevenly, respectively, and were thus found to be too unpredictable for commercial use. Ethanol applied through a vase solution was more consistent in reducing leaf blackening, but was considered unsuitable for commercial use as the flower quality was negatively affected. Similarly, ethanol applied as a concentrated pulse could delay leaf blackening, but led to an unacceptable accelerated loss of stem quality. In a preliminary study by Windell et al. (2010) to establish the efficacy of ethanol to reduce leaf blackening in ‘Pink Ice,’ ‘Sylvia,’ and P. magnifica under longterm cold storage, the optimum ethanol concentration for ‘Pink Ice’ stems stored for 2 weeks at 4 °C was found to be 2.5 g ethanol kg–1 fresh weight. However, longer storage durations resulted in a decreased efficacy, as leaf blackening rapidly increased over the extended storage time. In addition, ethanol was also shown to be less effective in controlling leaf blackening in other cultivars than for ‘Pink Ice.’ Followup simulated and commercial trials conducted by Windell (2012) again showed that the use of ethanol, either as a vapor or pulse treatment, varied widely as reductions in leaf blackening with ethanol were recorded in both P. magnifica and ‘Pink Ice,’ but not in ‘Sylvia.’ Both the simulated and commercial study concluded that any treatment combination which included glucose pulsing consistently yielded better results than any of the ethanol treatments alone. However, ethanol and glucose in combination were found to have a synergistic effect and resulted in a more marketable product than when treatment with only glucose was used. Thus, neither ethanol or acetaldehyde vapor or pulsing alone is currently considered a commercial alternative to pulsing with glucose in the control of leaf blackening (Windell 2012).

F. Genetic Selection and Hybridization Currently, most Protea cultivars available on the international floricultural market have been solely selected for their aesthetic qualities, with some attention being given to disease tolerance and vigor (Blomerus et al. 2010). As leaf blackening is a dominant trait associated with Protea neriifolia, P. compacta, P. coronate, and P. eximia, future breeders should focus on screening for seedlings less prone to this serious disorder (Blomerus et al. 2010). Until this goal to select for leaf blackeningresistant cultivars can be achieved, producers and exporters alike will have to rely on innovative postharvest technologies to ensure that Protea will remain a highvalue niche product, sought after on international floricultural markets. The advent of the selection of leaf blackeningresistant cultivars can be achieved with insight of the genetic profile of Protea. Recently, RNAsequencing technology has become widely used to explore patterns of gene sequence variation between biological samples (Schuster

2008). For example, RNAsequencing technology has been used to study trait variation along an environmental gradient in Protea repens (Akman et al. 2016). In the horticultural industry, the same technology has been applied to study changes associated with flesh browning in apples (Mellidou et al. 2014) and CO2 injury (Johnson and Zhu 2015) in apples under CA storage. The gene sequences yielded by RNAsequencing provide information about a biological sample which, in turn, can be related to functional metabolism, adaptation, or phenotypic expression. The technology produces expression levels of thousands of genes, often analyzed by clustering approaches to associate these with metabolic pathways. As such, RNAsequencing is useful to apply to the leafblackening phenomenon, as it will indicate which genes are repressed or upregulated during blackening. Differences between blackened and nonblackened leaf samples of the same cultivar, or between different cultivars, may be used. The profiles that result from such studies may serve in the selection of potential genetic markers that predict the incidence of this postharvest disorder. It can also be used to prevent the expression of certain genes, especially those that cause blackening, in the same way that plant breeders are developing diseaseresistant or abiotic and biotic stress tolerant crop varieties (Younis et al. 2014).

VI. CONCLUSIONS Leaf blackening in selected Protea species and cultivars is not fully understood, and remains a challenge for producers and exporters worldwide. The extent of leaf blackening is variable among species and cultivars, and is further affected by seasonal changes and time of day at which stems are harvested, apart from the most widely believed causes of the disorder which include water stress, carbohydrate stress, respiration rates, and nectar production. The majority of studies support the concept that carbohydrate depletion initiates the phenomenon, and this has led to many approaches being centered on supplying exogenous sugars to stems as sugar reserves become quickly depleted after harvesting. As such, inhibiting phloem unloading may prevent glucose and/or sucrose from being transported from the leaves to the flower head. These approaches, however, must be feasible at the producer level to be practiced commercially. Different Protea species and cultivars respond differently to types and concentrations of various sugars. In addition, cases of phytotoxicity have been reported which demonstrate the sensitivity of these flowers to this postharvest disorder. Postharvest leaf blackening does not appear to be the product of one factor, but is rather the result of a complex chain of events that eventually culminate in the oxidation of phenolic compounds (Jones et al. 1995). Thus, due to the complex nature and limited understanding of how factors interact to cause leaf blackening, future studies should aim to delineate the phenomenon at a biochemical level. Patterns of gene sequence variation have been widely used to infer causes in expression at the phenotype level. Using RNAsequencing technology, it may be possible to identify genes or groups of genes responsible for leaf blackening. This, in turn, will be useful for hybrid selection of leaf blackeningresistant cultivars. As sea freight becomes the primary mode for transporting these cut flowers, better storage conditions during shipping become imperative. Controlled atmosphere technology during

transport with reduced O2 and increased CO2 concentrations, using the technologies of DCA and ILOS to slow down metabolic processes and prevent leaf blackening may be viable options to control the condition, especially if used in combination with postharvest pulsing practices. Protea as a cut flower holds many of the characteristics desired by international markets that are always searching for new, exciting floricultural products. Protea is available in an increasing number of cultivars displaying different forms, sizes, colors, and textures of the involucral bracts, with also recently having extended availability during the year. The flowers may have an exceptional extended vase life of 14 days or more. However, the one major challenge that still severely limits the reliability and marketability of this unique product, in all regions where Protea are grown, remains postharvest leaf blackening. As this everpresent disorder significantly diminishes the appeal of this unique flower to the consumer, and also limits the ability of the producer to deliver an unblemished, highquality product to discerning markets, interventions through research are urgently required to control – or at best eliminate – leaf blackening as a postharvest disorder in Protea.

LITERATURE CITED Akman M., J.E. Carlson, K.E. Holsinger and A.M. Latimer. 2016. Transcriptome sequencing reveals population differentiation in gene expression linked to functional traits and environmental gradients in the South African shrub Protea repens. New Phytol. 210: 295–309. Bieleski, R.L., J. Ripperda, J.P. Newman, and M.S. Reid. 1992. Carbohydrate changes and leaf blackening in cut flower stems of Protea eximia. J. Am. Soc. Hortic. Sci. 177:124–127. Blomerus, L., S. Joshua, and J. Williams. 2010. Breeding Proteaceae varieties for changing market trends. Acta Hortic. 869:173–182. Brink, J.A. 1987. The influence of the flower head on leaf browning of cut Protea neriifolia R. Br. Protea News 5:11–12. Brink, J.A. and G.H. De Swardt. 1986. The effect of sucrose in a vase solution on leaf browning of Protea neriifolia R.Br. Acta Hortic. 185:111–119. Cannon, A. and R. McConchie. 2001. Controlling leaf blackening in Protea. A report for the Rural Industries Research and Development Corporation, RIRDC publication No. 01/098, Project No. US89A, Australia. Coetzee, J.H. and G.M. Littlejohn. 2001. Protea: A floricultural crop from the Cape Floristic Kingdom. Hortic. Rev. 26:1–48. Coorts, G.D. 1973. Internal metabolic changes in cut flowers. HortScience 8:195–198. Cowling, R.M. and D.T. Mitchell. 1981. Sugar composition, total nitrogen and accumulation of C14 assimilates in floral nectaries for Protea species. S. Afr. J. Bot. 47:743–750.

Crick, S.G. and R. McConchie. 1999. Ethanol vapour reduces leaf blackening in cut flower Protea ‘Pink Ice’ stems. Postharvest Biol. Technol. 17:227–231. D’Andrea, S., C. Costa, M. Castellani, G. Burchi, and P. Menesatti. 2007. Influence of environmental factors on quality of Protea cynaroides and Leucospermum glabrum × tottum (Proteaceae) during midrange transportation. Acta Hortic. 755:277–282. Dai, J. 1993. Postharvest leaf blackening in Protea neriifolia R.Br. Ph.D. thesis. Univ. of Hawai at Manoa, Honolulu. Dai, J. and R.E. Paull. 1995. Source–sink relationship and Protea postharvest leaf blackening. J. Am. Soc. Hortic. Sci. 120:475–480. Dai, J. and R.E. Paull. 1997. Comparison of leaf susceptibility to enzymatic blackening in Protea neriifolia R.Br. and Leucospermum ‘Rachel’. Postharv. Biol. Technol. 11:101–106. De Swardt, G.H., J. Pretorius, and L. Burger. 1987. The browning of foliage leaves in Protea – a review. Protea News 6:4–9. De Swardt, G.H. 1979. Blackening of Protea leaves. SAPPEX research paper 13. Du Plessis, D.G.C. (1978). Leukoantosianienbinders as beheermaatreël vir die voorkoming van bruinwording in Protea loofblare. M.Sc. Thesis, Randse Afrikaanse University, Johannesburg, South Africa. Ekman, J., J. Eyre, and D. Joyce. 2008. Flowers by sea: improving market access for Australian wildflowers. Rural Industries Research and Development Corporation (RIRDC) Publication No 07/181, RIRDC Project No DAN235A. Export Standards and Requirements. 1997. Cut flowers and ornamental foliage. Part 3, Directorate food safety and quality assurance, pp. 30–51. http://www.daff.gov.za Ferreria, A. 2005. Further studies on leaf blackening in Protea. M.Sc. Agric. Thesis, Univ. Stellenbosch, South Africa. Ferreira, D.I. and G.H. De Swardt. 1980. Veranderinge in die respirasietempo, styselkonsentrasie, totale vry reduserende suikerkonsentrasie en totale vry aminosuurkonsentrasie in verouderende rose (cv. Sonia). Agroplantae 12:23–28. Ferreira, D.I. 1983. Prevention of browning of leaves of Protea neriifolia R. Br. Acta Hortic. 138:273–276. Ferreira, D.I. 1986. The influence of temperature on the respiration rate and browning of Protea neriifolia R. Br. Acta Hortic. 185:121–129. Ghahamani, F. and K.J. Scott. 1998. The action of ethanol in controlling scald of apples. Aust. J. Agric. Res. 49: 199–205.

Gollnow, B. and R. Worrall. 2010. Quality specifications for Australian wildflowers: King Protea. Rural Industries Research and Development Corporation (RIRDC) Publication No 10/ 050, RORDC Project No PRJ000331. Haasbroek, F.J., G.G. Rousseau, and J.P. De Villiers. 1973. Effect of gamma rays on cut blooms of Protea compacta R. Br., P. longifolia Lamarck and Leucospermum cordifolium Salisb. ex Knight. Agroplantae 5:53–62. Halevy, A.H. and S. Mayak. 1979. Senescence and postharvest physiology of cut flowers. Part 1. Hortic. Rev. 1:204–236. Hernández, M., M.F. González, C.E. Álvarez, M. FernándezFalcón, N. FernándezGarcía, and E. Olmos. 2012. Postharvest study of leaf blackening in Protea ‘Pink Ice’ and ‘Susara’. Acta Hortic. 937:769–776. Hernández, M., C.E. Álvarez, M. FernándezFalcón, N. FernándezGarcía, and E. Olmos. 2014. Ascorbic acid antioxidant activity against leaf blackening of Protea. Acta Hortic. 1031:135–140. Hettasch, H.B. 1999. Studies of the vegetative development of Protea cv. ‘Sylvia’ and ‘Cardinal’. M. Sc. Agric., Stellenbosch University, Stellenbosch, South Africa. Hettasch, H.B., K.I. Theron, and G. Jacobs. 2001. Dry mass accumulation and carbohydrate allocation in successive growth flushes of Protea cultivar Sylvia and Protea cultivar Cardinal shoots. Acta Hortic. 545:215–225. Hoffman, E.W. and M. Du Plessis. 2013. A preliminary study into protocols for the longterm cold storage of Leucospermum potted plants. Acta Hortic. 1007:161–170. Hoffman, E.W., N.E. Windell, and G. Jacobs. 2014. Optimum time of harvest for the control of leaf blackening in Protea ‘Sylvia'. Acta Hortic. 1031:125–133. Hoffman, E.W., M. Miller, and EL. Louw. 2015. The efficacy of LED lights and growth regulator sprays at controlling chilling injury in Leucospermum potted plants. Acta Hortic. 1097:47–54. Ingle, M. and M. D’Souza. 1989. Physiology and control of superficial scald of apples: A review. J. Food Sci. 24: 28–31. Jacobs, G. and H.R. Minnaar. 1977. Effect of temperature on the blackening of Protea leaves. SAPPEX Newsletter 17:20–24. Jackson, S. and S.W. Nicolson. 2002. Xylose as a nectar sugar: from biochemistry to ecology. Comp. Biochem. Physiol. Part B 131:613–620. Johnson F.T. and Y. Zhu. 2015. Transcriptome changes in apple peel tissues during CO injury symptom development under controlled atmosphere storage regimens. Hortic. Res. 2: 15061.doi:10.1038/hortres.2015.61

Jones, R. 1991. Understanding and controlling leaf blackening in Protea leaves: The use of high concentrations of sucrose. p. 313–322. In: Proc. Intl. Protea Assoc. Sixth Biennial Conference, Perth, Western Australia. Jones, R. and A. Cass. 1996. The possible involvement of a betaDglucosidase in postharvest leaf blackening in cut Protea flowers. p. 83–87. In: Proc. Australasian Postharvest Horticulture Conference ‘Science and technology for the fresh food revolution’. Melbourne, Australia, 18–22 September 1995. Jones, R.B. and K.A. ClaytonGreene. 1992. The role of photosynthesis and oxidative reactions in leaf blackening of Protea neriifolia. Scientia Hortic. 50:137–145. Jones, R.B., R. McConchie, W.G. van Doorn, and M.S. Reid. 1995. Leaf blackening in cut Protea flowers. Hortic. Rev. 17:173–202. Kader, A. 2002. Postharvest technology of horticultural crops. Postharvest handling systems: Ornamental crops (Reid, M.S.). Third ed. 25:315–325. Malan, G. 2012. Protea cultivation: from concept to carton. First ed. SUN Media, Stellenbosch, South Africa. Masie, W.E. 1979. Fenielmerkuriasetaat as voorkomingsmiddel vir loofblaarverbruining by Protea neriifolia R.Br. M. Sc. Thesis, Randse Afrikaanse University, Johannesburg, South Africa. McConchie, R. and N.S. Lang. 1993a. Postharvest leaf blackening and preharvest carbohydrate status in three Protea species. HortScience 28:313–316. McConchie, R. and N.S. Lang. 1993b. Carbohydrate metabolism and possible mechanisms of leaf blackening in Protea neriifolia under dark postharvest conditions. J. Am. Soc. Hortic. Sci. 118:355–361. McConchie, R., N.S. Lang, and K.C. Gross. 1991. Carbohydrate depletion and leaf blackening in Protea neriifolia. J. Am. Soc. Hortic. Sci. 116:1019–1024. McConchie, R., N.S. Lang, A.R. Lax, and G.A. Lang. 1994. Reexamining polyphenol oxidase, peroxidase, and leaf blackening activity in Protea. J. Am. Soc. Hortic. Sci. 119:1248–1254. Mellidou I., K. Buts, D. Hatoum, Q.T. Ho, J.W. Johnston, C.W. Watkins, R.J. Schaffer, N.E. Gapper, J.J. Giovannoni, D.R. Rudell, M.L. Hertog, and B.M. Nicolai. 2014. Transcriptomics events associated with internal browning of apple during postharvest storage. BMC Plant Biol. 14:328. doi:10.1186/s128700140328x Meyer, C. 2003. Carbohydrates and leaf blackening of Protea. M. Sc. Agric. Thesis. Univ. Stellenbosch, South Africa. Mostert, D.P., W.R. Siegfried, and G.N. Louw. 1980. Protea nectar and satellite fauna in

relation to the food requirements and pollinating role of the Cape sugarbird. S. Afr. J. Sci. 76: 409–412. Mulder, P.W.A. 1977. Primêre meganismes betrokke by die bruinwording van loofblare in Protea neriifolia. M.Sc. Thesis, Randse Afrikaanse University, Johannesburg, South Africa. Newman, J.P., W.G. van Doorn, and M.S. Reid. 1990. Carbohydrate stress causes leaf blackening in proteas. Acta Hortic. 264:103–108. Paull, R.E. and J.W. Dai. 1990. Protea postharvest black leaf: a problem in search of a solution. Acta Hortic. 264:93–101. Paull, R.E., T. Goo, R.A. Criley, and P.E. Parvin. 1980. Leaf blackening in cut Protea eximia; importance of water relations. Acta Hortic. 113:159–166. Paull, R.E., T. Jaroenkit, and N.J. Chen. 2003. Protea, a cut woody ornamental: diversity and variable postharvest responses. Acta Hortic. 602:127–132. Perold, G.W., P. Beylis, and A.S. Howards. 1973. Metabolites of Proteaceae. Part VII. Lacticolorin, a phenolic glucoside ester, and other metabolites of Protea lacticolor. J. Chem. Soc. Perkins Trans. 1:638–643. Perold, G.W., M.E.K. Rosenberg, A.S. Howards, and P.A. Huddle. 1979. Metabolites of Proteaceae. Part 9. Eximin (6OBenzoylarbutin) and the synthesis of aryl glucoside esters. J. Chem. Soc. Perkins Trans. 1:239–243. Perold, G.W. and L. Carlton. 1989. Neriifolin, an ester glucoside of benzene1,2,4triol. J. Chem. Soc. Perkins Trans. 1:638–643. Perold, G.W. 1993. Consistency and variation in metabolic patterns of South African Proteaceae: a chemical perspective. S. Afr. J. Sci. 89:90–93. PhilosophHadas, S., S. Droby, I. Rosenberger, and Y. Perzelan. 2007. Sea transport of ornamental branches: problems and solutions. Acta Hortic. 755:267–276. PhilosophHadas, S., Y. Perzelan, I. Rosenberger, S. Droby, and S. Meir. 2010. Leucadendron ‘Safari Sunset’: Postharvest treatments to improve quality of cut foliage during prolonged sea shipment. Acta Hortic. 869:207–217. Ramokanyane, T.M. 2016. Effects of dynamic controlled atmosphere and initial low oxygen stress on superficial scald of ‘Granny Smith’ apples and ‘Packham’s Triumph’ pears. M. Sc. Agric. Thesis. Univ. Stellenbosch, South Africa. Reid, M.S. 1992. Postharvest handling systems of ornamental crops. p. 201–209. In: Postharvest technology of horticultural crops. (ed.) A.A. Kader. Univ. California, Oakland, CA, USA. Reid, M.S., W.G. van Doorn, and J.P. Newman. 1989. Leaf blackening in proteas. Acta Hortic.

261:81–84. Reid, M.S. and CZ. Jiang. 2012. Postharvest biology and technology of cut flowers and potted plants. Hortic. Rev. 40:1–54. Robatscher, P., D. Eisenstecken, F. Sacco, H. Pohl, J. Berger, A. Zanella, and M. Oberhuber. 2012. Diphenylamine residues in apples caused by contamination in fruit storage facilities. J. Agr. Food Chem. 60: 2205–2211. RodríguezPérez, J.A., M.C. VeraBatista, A.M. de LeónHernández, and I. Rodríguez Hernández. 2012. Effect of sucrose pulsing and postharvest dehydration delay on vase life of Protea ‘Susara’. Acta Hortic. 937:837–845. Schmid, B., H. Nottebrock, K.J. Esler, J. Pagel, A. Pauw, K. BöhningGaese, F.M. Schurr, and M. Schleuning. 2015. Reward quality predicts effects of birdpollinators on the reproduction of African Protea shrubs. Persp. Plant Ecol. Evol. System. 17:209–217. Schuster S.C. 2008. Nextgeneration sequencing transforms today’s biology. Nat. Methods 5:16–18. Stephens, I.A. 2003. Leaf blackening in Proteas. Ph.D. Agric. Diss., Univ. Stellenbosch, South Africa. Stephens, I.A., D.M. Holcroft, and G. Jacobs. 2001. Low temperatures and girdling extend vase life of ‘Sylvia’ proteas. Acta Hortic. 545:205–214. Stephens, I.A., D.M. Holcroft, and G. Jacobs. 2003a. Postharvest treatments to extend vase life of selected Proteaceae cut flowers. Acta Hortic. 602:155–159. Stephens, I.A., D.M. Holcroft, and G. Jacobs. 2003b. Storage and vase life extension of ‘Sylvia’ protea flowers. Acta Hortic. 600:123–126. Stephens, I.A., C. Meyer, D.M. Holcroft, and G. Jacobs. 2005. Carbohydrates and postharvest leaf blackening of proteas. HortScience 40:181–184. Teixeira da Silva, J.A. 2003. The cut flower: postharvest considerations. J. Biol. Sci. 3:406– 422. van Doorn, W.G. 2001. Leaf blackening in protea flowers: recent developments. Acta Hortic. 545:197–204. Van Lieberg, M.J., W.G. van Doorn, and H. van Gelder. 1990. Prevention of phytochrome related postharvest losses in ornamentals through red light emitted by diodes. Acta Hortic. 272:347–351. Van Rheede van Oudtshoorn, M.C.B. 1963. Distribution of phenolic compounds in some South African Proteaceae: a contribution to the chemotaxonomy of the family. Planta Med. 4:399– 406.

Van Wyk, B.E. and S.W. Nicolson. 1995. Xylose is a major nectar sugar in Protea and Faurea. S. Afr. J. Sci. 91:151–152. Whitehead, C.S. and G.H. De Swardt. 1982. Extraction and activity of polyphenoloxidase and peroxidase from senescing leaves of Protea neriifolia. S. Afr. J. Bot. 1:127–130. Whitehead, C.S. 1979. Die verband tussen fenoliese verbruining en sekere oksidatiewe ensieme in loofblare van Protea neriifolia R. Br. Ph.D Diss., Randse Afrikaanse University, Johannesburg, South Africa. Windell, N.E. 2012. Leaf blackening and the control thereof in selected Protea species and cultivars. M. Sc. Agric. Thesis, Univ. Stellenbosch, South Africa. Windell, N.E., I.A. Stephens, G. Jacobs, and E.W. Hoffman. 2010. Effect of ethanol vapour on leaf blackening of selected Protea species. Acta Hortic. 869:135–139. Windell, N.E., G. Jacobs and E.W. Hoffman. 2015. Factors influencing the efficacy of glucose pulsing as a technique commercially employed to control leaf blackening in Protea ‘Sylvia’. Acta Hortic. 1097:133–141. Younis A., M.I. Siddique, C. Kim, and K. Lim. 2014. RNA interference (RNAi) induced gene silencing: A promising approach of hitech plant breeding. Intl. J. Biol. Sci. 10: 1150–1158. Zanella, A. and Strüz, S. 2013. Replacing DPA postharvest treatment by strategical application of novel storage technologies controls scald in 1/10th of EU’s apples producing area. Acta Hortic. 1012: 419–426.

4 Sapota (Manilkara achras Forb.): Factors Influencing Fresh and Processed Fruit Quality Babak Madani Tropical Fruit Research Center, Horticultural Science Research Institute, Agricultural Research Education and Extension Organization (AREEO), Hormozgan, Iran Amin Mirshekari Department of Agronomy and Plant Breeding, Faculty of Agriculture, University of Yasouj, Yasouj, Iran Elhadi Yahia Facultad de Ciencias Naturals, University of Queretaro, Queretaro, Mexico John B. Golding NSW Department of Primary Industries, Gosford, NSW, Australia

ABSTRACT Sapota (Manilkara achras Forb.) is an evergreen tropical tree, the fruit of which is used fresh and processed. Sapota, also known as sapodilla, contains high levels of ascorbic acid and phenolic compounds which contribute to its numerous purported human health benefits. The fruit is characterized by a climacteric ripening behavior with a short postharvest life at ambient temperature. The main limitation of postharvest shelf life is decay. Although low temperature storage prolongs the postharvest life of sapota fruit, chilling injury can develop if the storage temperature is less than 14 °C. The storage life of sapota fruit can also be extended with the use of modified and controlled atmospheres and the use of other postharvest treatments such as 1methyl cyclopropene and calcium application. Limited postharvest research has been conducted on sapota fruit over the last 30 years. This review discusses the preharvest and postharvest physiological and biochemical processes related to the maturation, ripening, and senescence of sapota fruit. In addition, the postharvest treatments, methods and technologies to maintain fruit quality and control of postharvest decay and physiological disorders are also reviewed. Future research is proposed, with an emphasis on proteomic and molecular approaches to investigate the mechanisms of sapota fruit ripening, senescence, and its susceptibility to chilling injury. The adoption of best postharvest handling practices and storage techniques developed from postharvest research and development will benefit the exploitation of sapota fruit onto world markets. KEYWORDS: Manilkara achras; postharvest treatments; storage life; quality;

ABBREVIATIONS ACC 1Aminocyclopropane1carboxylic acid AEAC Ascorbic acid equivalent antioxidant capacity CI Chilling injury HORAC Hydrophilic oxygen radical absorbance capacity LDPE Lowdensity polyethylene film MAP Modified atmosphere packaging 1MCP 1Methyl cyclopropene NMR Nuclear magnetic resonance PCA Principal component analysis PG Polygalacturonase PME Pectin methylesterase POD Peroxidase PPO Polyphenol oxidase SSC Soluble solids content TA Titratable acidity TAC Total antioxidant capacity TPC Total phenolics content I. INTRODUCTION A. Cultivars

B. Uses 1. Fresh Fruit 2. Cooking 3. Medicinal Uses II. NUTRITIVE VALUE A. Antioxidant Activity B. Anticancer Activity C. Potential Pharmaceutical Use III. PHYSIOLOGICAL AND BIOCHEMICAL CHANGES DURING FRUIT MATURATION AND RIPENING A. Respiration Rate and Ethylene Production B. Fruit Firmness C. Pigments D. Volatiles E. Ascorbic Acid F. Minerals G. Sugars H. Organic Acids I. Proteins and Enzymes J. Phenolic Compounds and Antioxidants IV. PREHARVEST EFFECTS ON POSTHARVEST QUALITY V. PHYSIOLOGICAL DISORDERS A. Chilling Injury B. Flesh Browning VI. POSTHARVEST DISEASES A. Control of Postharvest Diseases VII. POSTHARVEST TECHNOLOGY A. Grading and Packing B. Storage VIII. POSTHARVEST TREATMENTS

A. Ethylene Management and 1Methylcyclopropene B. Ethylene Management and LowEthylene Storage C. Plant Growth Regulators D. Calcium E. Essential Oils F. Edible Coatings G. Heat Treatments H. Irradiation I. ModifiedAtmosphere Packaging IX. NONDESTRUCTIVE METHODS FOR IDENTIFYING FRUIT MATURITY AND QUALITY X. PROCESSING XI. SUMMARY AND FUTURE PROSPECTS LITERATURE CITED

I. INTRODUCTION Sapota is from the Sapotaceae family with a diverse and significant range of 700 species and 35 or 40 poorly defined genera (Peiris 2007). These shrubs and trees are widely distributed throughout the tropics (Parle and Preeti 2015), and are recognized by the expression of milky latex and alternate leathery leaves with parallel secondary and tertiary veins (Parle and Preeti 2015). There are many synonyms for the scientific name of sapota, including Achras sapota L.; Achras zapota L. var. zapotilla Jacq.; Achras zapotilla Nutt.; Achras mammosa L.; Manilkara achras (Miller) Fosberg; Manilkara zapotilla (Jacq.) Gilly; Sapota zapotilla (Jacq.) Coville; Sapota achras Miller; and Sapota zapotilla (Coville) (Peiris 2007) (Plate 4.1). Sapota is a medium to large tree with a rounded canopy and many horizontal or drooping branches (Peiris 2007). The mature sapota fruit is classified as an ellipsoid or ovoid berry (about 12–15 cm diameter) with a thin skin, covered with a brown layer (Lim 2013). The immature fruit is hard, gummy and very astringent, and becomes soft and sweet as the fruit develops and ripens. Flesh colors range from a light yellow color to darkbrown or sometimes reddishbrown. Flesh texture may be coarse and grainy or smooth which becomes soft, very juicy, and sweet with a pleasant flavor (Plate 4.2) (Lakshminarayana and Subramaniyam 1966; Camargo et al. 2016). Some fruit are seedless, but normally there may be from three to 12 seeds (frequently five), which are easily removed as they are loosely held in the center of the fruit. The seeds contain a range of phytochemicals such as sapotin, saponin, and sapotinine. The seeds should be removed before eating because they contain the toxin hydrocyanic acid (Peiris 2007; Cortez et al. 2013).

Plate 4.1. Sapota tree

Plate 4.2. External and internal images of round and elongated sapota fruit Sapota is native to Central America, specifically in the adjacent areas of southern Mexico, along with northern Belize and northeastern Guatemala. However, due to its high adaptability to different soil and climatic conditions, relatively low production costs, high nutritive value, export potential and generally high economic returns, sapota is now widely grown across the world (Agrawal and Dikshit 2008; Kunyamee et al. 2009). It is extensively grown in the tropics, particularly in Central and South America but it is now widely grown commercially in Mexico, India and across Southeast Asia including the Philippines, Vietnam, Malaysia, Indonesia, Thailand, Sri Lanka and Bangladesh (Lim 2013; Parle and Preeti 2015). The largest producers of sapota in the world are India, Mexico, Guatemala, and Venezuela (Athmaselvi et al. 2014; Saxena 2014; Hiwale 2015). Sapota is an important fruit in India (Bhale et al. 2013), and is widely grown in the states of Gujarat, Karnataka, Maharashtra, Tamil Nadu, Andra Pradesh, and West Bengal. The state of Karnataka is the main producer of sapota in India, with 293 000 ha of plantations producing 360 000 MT of fruit each year

(Suhasini et al. 2012).

A. Cultivars Many cultivars of sapota are available which differ in branching, foliage color, fruit shape, fruit texture, and color and pulp quality (Hiwale 2015). A small number of seeds with sweet pulp are the main characteristics of an acceptable fruit. Inferior cultivars contain hard flesh with a “sandy” texture. Over 35 cultivars are commonly grown in India, from which ‘Kalipatti,’ ‘Kirthabarthi,’ and ‘PKM1’ are the most appreciated. ‘Kalipatti’ has oval shaped fruit with few seeds (one to four) and a sweet flesh. ‘Kirthabarthi’ has small to mediumsized fruit with a rough fruit skin, while ‘PKM1’ has thin skin, and a buttery and very sweet pulp. Moreover, crop improvement through clonal selection and planned hybridization has led to the introduction of many good new cultivars such as ‘DHS1,’ which has round to slightly oblong fruit, with a very sweet, soft, granular and mellow flesh (Hiwale 2015).

B. Uses 1. Fresh Fruit. Sapota is widely grown and is considered as one of the best fruits in Central America. Ripe sapota fruit are generally eaten by cutting the fruit in half and the flesh is eaten with a spoon. The flesh is also added to salads or fruit cups, sherbets, milk shakes, and ice cream. A sweet sauce can also made from ripe fruit by pressing the flesh, adding orange juice, and covering with whipped cream (Peiris 2007). 2. Cooking. Sapota flesh can also be used as an ingredient in baking when combined with egg custard mix before baking. However, sapota is generally not cooked or preserved, but is sometimes fried (Peiris 2007). Cooking the fruit alters the flesh to a red color (Lim 2013). 3. Medicinal Uses. Sapota fruit has been used as a traditional indigenous medicine in many cultures (Lim 2013). The immature fruit have a high tannin content and when boiled they can be used to treat diarrhea. Infusions of the young fruit have also been reported to relieve lung complaints (Kulkarni et al. 2007; Peiris 2007). Consumption of the fruit has been reported to decrease inflammation and pain in gastritis, boost immunity and prevent bacterial infection, which is thought to be due to the presence of ascorbic acid (Parle and Preeti 2015). It is thought to be good for the treatment of digestion and is used in the treatment of constipation. It is also useful for pregnancy where its high nutritional content can be used as a herbal remedy (Parle and Preeti 2015). Moreover, sapota has been used for treating pulmonary diseases in an Indian context because of its anti inflammatory, analgesic, and antimicrobial effects (Mund et al. 2016).

II. NUTRITIVE VALUE Sapota is a highly nutritious fruit containing a wide range of beneficial components such as dietary fiber, fructose, glucose, sucrose, vitamins, minerals and diverse phytochemicals, fatty acids and polyamines. The nutritional content of sapota fruit and juice is presented in Tables 4.1 and 4.2. Fruit contain high concentrations of minerals such as potassium, calcium, iron, copper and zinc and phenolic components (Kulkarni et al. 2007; Mund et al. 2016; Sumathi and Shivashankar 2017). The decrease in astringency during fruit development and ripening has been shown to be the result of polymeric changes, the interaction of other components such as sugars, and to a reduction in the concentration of polyphenols as fruit size increases (Lakshminarayana and Subramanyam 1966; Lim 2013). Table 4.1. Nutritional components of sapota (Manilkara zapota) fruit per 100 g edible portion. Constituent Water Protein Total lipid (fat) Ash Total dietary fiber Carbohydrate Calcium Magnesium Phosphorus Potassium Sodium Zinc Copper Iron Selenium Vitamin C Riboflavin Niacin Pantothenic acid Vitamin B6 Folate Vitamin A

Approximate value 78.00 g 0.44 g 1.10 g 0.50 g 5.3 g 19.96 g 21 mg 12 mg 12 mg 193 mg 12 mg 0.10 mg 0.086 mg 0.80 mg 0.6 μg 14.7 mg 0.020 mg 0.200 mg 0.2528 mg 0.037 mg 14 μg 60 IU

Palmitic acid Stearic acid

0.1 g 0.094 g

Oleic acid Linoleic acid Tryptophan Leucine

0.521 g 0.011 g 0.005 g 0.024 g

Lysine Arginine Aspartic acid

0.039 g 0.017 g 0.032 g

Lutamic acid Proline

0.038 g 0.036 g

Lycopene

41.93 (μg 100 g–1 DW)

Total phenolic

13.5(mg GAE 100 g–1)

(Almeida et al. 2011; Yahia and GutierrezOrozco 2011; USDA 2012; da Silva et al. 2014)

Table 4.2. Chemical composition of sapota fruit juice with pH 5.36 and titratable acidity 0.16% (% citric acid). Constituent Total soluble solids Reducing sugar Total sugar

Approximate value 20.68% 9.86% 11.06%

Protein

312.5 mg 100 g–1

Ascorbic acid

10.52 mg 100 g–1

Carotenoids

0.92 mg 100 g–1

Total phenolics

134.6 mg 100 g–1

Ca

20.67 mg l–1

K

6.15 mg l–1

Fe

0.11 mg l–1

Cu

0.09 mg l–1

Zn

0.5 mg l–1

Modified from Kulkarni et al. (2007)

A. Antioxidant Activity

Sapota fruit are considered to be a high source of antioxidants as they contain over 3000 mg l–1 ascorbic acid equivalent antioxidant capacity (AEAC) per 100 g fresh sample (Shui et al. 2004; Lim 2013). Polyphenolics, with basic blocks of catechin or gallocatechin, or both, attribute significantly to the antioxidant activity of the fruit (Lim 2013). It has been shown that sapota juice has high antioxidant activity when assayed with 2,2diphenyl1picrylhydrazyl (DPPH) as well as superoxide radical scavenging activity. The presence of carotenoids, phenolics and ascorbic acid has also been linked to the multiple radicalscavenging potential of sapota juice (Kulkarni et al. 2007; Lim 2013). Ma et al. (2003) isolated ten phenolic compounds from a methanol extract of sapota fruit, including methyl chlorogenate, dihydromyricetin, quercitrin, 4 Ogalloylchlorogenic acid, myricitrin, (+)catechin, (−) epicatechin, (+)gallocatechin, methyl 4Ogalloylchlorogenate, and gallic acid.

B. Anticancer Activity Phenolic compounds from sapota fruit have been shown to possess a range of anticancer activities. For example, methyl 4 Ogalloylchlorogenate isolated from sapota fruit has been shown to display cytotoxicity in the HCT116 and SW480 human colon cancer cell lines, with IC50 values of 190 μM and 160 μM, respectively (Ma et al. 2003; Lim 2013). Srivastava et al. (2014) considered the effects of methanolic extracts of sapota fruit (MESF) on NALM6 (preBcell leukemia) and K562 (chronic myelogenous leukemia) tumor cells of mice. This compound was found to induce cytotoxicity by activating proapoptotic and downregulating mitochondrial protective proteins. MESF induced depolarization and trans membrane potential in mitochondria and the translocation of phosphatidyl serine from the inner to the outer leaflet of the cell membrane, and significantly upregulated proapoptotic proteins and apoptotic markers which hallmarked apoptosis initiation. MESF treatment led to cell death of approximately 80% at 2 mg ml–1, and induced cytotoxicity in NALM6 and K562 cells with IC50 values of 0.9 mg ml–1 and 2.5 mg ml–1, respectively, after 72 h of treatment. Moreover, they showed that tumor progression inhibition resulted in a threefold increase in lifespan in 50% of mice, thus suggesting its potential role in preventing the genesis and progression of cancer cells (Srivastava et al. 2014).

C. Potential Pharmaceutical Use Sapota fruit pulp contains sapotin (a glucoside used in medicine to reduce fever), saponin, fixed oils, and other bitter alkaloids (Madhav and Khurana 2011). In formulating microemulsions of the antidepressant drug, escitalopram, a novel potent bioemulsifier extract from pulp was found to have potential. The biopolymer showed positive tests for the presence of proteins and carbohydrates. The polymer can serve potentially for formulating various drugs loaded into emulsions (Madhav and Khurana 2011; Lim 2013).

III. PHYSIOLOGICAL AND BIOCHEMICAL CHANGES DURING FRUIT MATURATION AND RIPENING

Mature sapota trees grown in semiarid rainfed conditions produce two crops per year. The summerharvested fruit are generally bigger in size and higher in yield compared to fruit harvested in the rainy season (Hiwale 2015). A single sigmoidal pattern of development characterizes growth of the fruit, where three unique stages of growth have been identified (Brito and Narain 2002; Yahia and GutierrezOrozco 2011). Cell division and maturation of the embryo occur during the first stage of growth following pollination, whilst the second stage occurs when growth is significantly decreased. The third and final stages of growth are characterized by cell enlargement which gives rise to full fruit size – the growth occurs within 5 to 7.5 months after fruit set (Lakshminarayana and Subramanyam 1966; Yahia and Gutierrez Orozco 2011). Identifying the best harvest indices of fruit is critical, as harvesting fruit after the optimum stage of maturity promotes rapid softening and more susceptibility to physical damage, whilst immature fruit are astringent and contain latex (Yahia 2004). There are significant maturity differences among fruit within a single tree, and even within a single cluster of fruit, although fruit maturity normally initiates at the base of the cluster and progresses to the tip. However, judging fruit maturity is difficult where cultivar differences play an important role in determining harvest time. Maturity for a specific sapota cultivar is often judged by appearance and size. Although some cultivars often shed the scruffy surface coat on their skin when mature (Love et al. 2014), rubbing the brown powdery “scruff” is not always a successful indicator of maturity index for all sapota fruit (Love et al. 2014). However, the shiny brown color and rounded stylar end of fruit are considered good maturity indices in some countries (Kute and Shete 1995; Yahia and GutierrezOrozco 2011). Waiting for the first ripe fruit to fall from the trees is commonly practiced commercially, but this is not an adequate method of determining maturity (Yahia and GutierrezOrozco 2011). Another good indicator of ripening is the lack of latex coming from the skin when it is scratched, as the latex content is reduced to almost zero when the fruit is fully mature (Yahia and GutierrezOrozco 2011; Love et al. 2014). The specific gravity of mature sapota fruit has been indicated to be between 1.025 and 1.10 depending on the cultivar; for example, for ‘Kalipatti’ it has been reported as being 1.10 (Awasarmal et al. 2011). Another maturity indicator is based on the days required from fruit set to maturity. Typically, it is between 120 and 245 days from fruit set to maturity depending on the climate and cultivar, but the erratic flowering habit and the presence of fruit at all stages of development on the tree makes it difficult to determine the optimum harvesting time of the fruit based on this indicator (Yahia and GutierrezOrozco 2011).

A. Respiration Rate and Ethylene Production Sapota is a climacteric fruit which can be harvested when fully mature and then ripened off the tree (Yahia 2004; Morrison et al. 2011). It is generally shown that the fruit is characterized by moderate respiration rates at 5 °C (10–20 mg CO2 kg–1 h–1), and high levels of ethylene production, that is, higher than 100 μl C2H4 kg–1 h–1 at 20 °C (Mangaraj and Goswami 2006). Vishwasrao and Ananthanarayan (2016b) showed that fruit from the cultivar ‘Kalipatti’ had a respiration peak 3 days after harvest, and later decreased, typical of other climacteric fruits.

High levels of ethylene (e.g., 100 μl l–1 for 24 h at 20 °C) have been shown to accelerate ripening (Kader 2009; Thompson 2014). Conversely, ripening can be delayed with the removal of ethylene from the storage atmosphere (Thompson 2014). Bhutia et al. (2011) showed the benefits of ethylene removal with ethyleneabsorbant sachets during sapota fruit storage (see also Section VIII B). Other postharvest treatments which have been shown to decrease respiration and ethylene production rates include hot air treatment and calcium chloride (CaCl2) dips (Thompson 2014). Chittham et al. (2002) showed with ‘MaKok’ fruit that treatment with hot air at 35 °C for 12 h followed by dipping in 5% CaCl2 for 30 min reduced the rates of respiration, ethylene production, and ACC oxidase activity.

B. Fruit Firmness Fruit texture is one of the most important quality attributes for the eating quality of sapota. Fruit texture is commonly quantified as fruit firmness or fruit hardness, which have been used as an indicator of fruit maturity (Raharjo et al. 1998). Fruit softening and ripening of sapota is associated with a reduction in fruit firmness, where firmness can decline from 60 N in early stages of ripening to 8 N in ripe fruit (Qiuping et al. 2006). Vishwasrao and Ananthanarayan (2016b) showed that flesh firmness decreased rapidly at the peak of the respiratory climacteric and then did not change significantly during further ambient storage.

C. Pigments Fruit color is an important consumer quality attribute, and is due to different pigments such as chlorophyll and carotenoids in the skin and flesh. Chlorophyll is responsible for the green background level in the skin of sapota fruit, and its concentration has been shown to continuously decrease during fruit ripening, denoting both a decrease in chlorophyll biosynthesis and an increase in chlorophyll degradation (Qiuping et al. 2006). Conversely, the concentrations of carotenoids have been shown to increase during fruit maturation and ripening (Kulkarni et al. 2007).

D. Volatiles Numerous studies have been conducted on the aroma of sapota fruit (Pino et al. 2003; Laohakunjit et al. 2007; Poonsawat et al. 2007; Uekane et al. 2017). These authors have used different extraction and identification methods, which have resulted in different results. Some of these differences may also be due, in part, to the cultivar studied and to the stage of fruit maturity at which the sample was taken for analysis. The use of headspace solidphase microextraction (SPME) with gas chromatographymass spectrometry (GC/MS) led to 23 volatile components being identified in ‘Kai’ sapota fruit (Laohakunjit et al. 2007). The major volatiles were ethyl acetate (29% of the total volatiles), acetaldehyde (22%), benzyl alcohol (12%), and 2butenyl benzene (7%). Pino et al. (2003) identified a total of 69 aroma volatiles using GC/MS, where the major components were

methanethiol (32%), hexadecanoic acid (26%), 3hydroxy2butanone (7%), ethyl acetate (6%), isoamyl alcohol (6%), and 2methyl1propanol (4%). Poonsawat et al. (2007), using a solidphase microextraction (SPME) technique, further showed that the aroma profile of freshcut sapota fruit changed during shelf life. Hexanal was a major volatile at the initial day of freshcut ‘Malay’ sapota during cold storage, but acetaldehyde and ethyl acetate became predominant by days 3 and 6, respectively (Poonsawat et al. 2007). Moreover, Uekane et al. (2017) identified 72 volatile compounds in the headspace of sapota fruit using headspace–solidphase microextractiongas chromatographytandem mass spectrometry (SHSPMEGCMS). These authors showed the main volatile classes in sapota fruit to be esters (33%), alcohols (27%), terpenes (18%), and others (21%). The authors further identified ethyl butyrate, isoamyl acetate, limonene, ethyl hexanoate, 6methyl5hepten 2one, nnonanal, ethyl octanoate, 2ethylhexanol, linalool, ethyl decanoate, borneol, a terpineol, ndodecanal, 1decanol, phenethyl acetate, 1 dodecanol and ethyl linoleate for the first time in sapota fruit (Uekane et al. 2017).

E. Ascorbic Acid Sapota fruit is a good source of ascorbic acid (Bose and Mitra 1990; Camargo et al. 2016), although the concentration of ascorbic acid rapidly decreases during ambient storage. For example, Vishwasrao and Ananthanarayan (2016b) showed that after 3 days of storage the fruit retained just 5.5% of their original ascorbic acid content. This rapid degradation of ascorbic acid indicates a fast catabolism during storage, which is a prerequisite for increased respiratory and antioxidant requirements during ripening (Mehta et al. 1980; Vishwasrao and Ananthanarayan 2016b). Ascorbic acid has been shown to be degraded in these later stages of storage as a result of the activities of phenol oxidase and ascorbic acid oxidase enzymes, which may be utilized for the production of other organic acids (Kostman et al. 2001; Vishwasrao and Ananthanarayan 2016b).

F. Minerals Sapota fruit contain relatively high concentrations of potassium and calcium (Kulkarni et al. 2007; Sumathi and Shivashankar 2017), with potassium being the most abundant mineral (7.2  mg g–1 DW). However, the concentrations of minerals in the fruit change during fruit growth. For example, Rastegar (2015) showed that the concentration of iron decreased from about 4.5  μg g–1 DW in the initial phases of fruit development to 1.0 μg g–1 DW in the full ripe stage. The concentrations of manganese have been shown to decline from about 7.0 μg g–1 DW in the initial phase of fruit development to a minimum of 0.5 μg g–1 DW in the full ripe stage (Rastegar 2015). Hamza et al. (2013) showed that sapota fruit contained high concentrations of iron (14.2 μg g–1 DW), manganese (1.5 μg g–1 DW), copper (1.7 μg g–1 DW), and zinc (1.0 μg  g–1 DW). It should be noted that there are considerable differences in the concentrations of minerals between different nutrition studies, but these may be the result of differences in the environment and other growing conditions.

G. Sugars Carbohydrates are the main constituents of sapota fruit, where the content of free sugars is high whilst starch is almost absent in the fully ripened fruit (Yahia and GutierrezOrozco 2011; Sumathi and Shivashankar 2017). The concentrations of individual sugars change during ripening, where levels of sucrose show the highest increase during ripening, followed by glucose and fructose (Yahia and GutierrezOrozco 2011; Sumathi and Shivashankar 2017). However, overripe fruit usually contain lower levels of sucrose than glucose and fructose (Yahia and GutierrezOrozco 2011). The levels of total sugars contained in mature green to the halfgreen fruit stage have been shown to increase from 11% to 15% because of the increase in nonreducing sugars, which increased from 2% to 6% (Brito and Narain 2002).

H. Organic Acids The major individual organic acids in mature sapota fruit are malic (18.25 mg g–1 FW), citric (8.30 mg g–1 FW) and tartaric (2.69 mg g–1 FW) (Sumathi and Shivashankar 2017). Other less significant organic acids such as fumaric, gluconic and oxalic acids have only been identified at specific stages of development and ripening (Das and De 2015). The flavor of sapota fruit is a balance between sugars and organic acids. Levels of titratable acidity (TA) in fruit have been shown to decrease from around 0.48% to 1.36% in the initial periods of fruit development to lower levels (0.11–0.41%) when the fruit is completely ripe (Yahia and GutierrezOrozco 2011). This decrease in TA during ripening has been shown to be irrespective of cultivar (Yahia and GutierrezOrozco 2011). The low levels of acidity and high soluble solids content (SSC) (Brix 15.8%) lead to an elevated SSC/TA ratio, which is characteristic of ripe sapota fruit (Brito and Narain 2002).

I. Proteins and Enzymes The protein content of sapota fruit is generally low, with the protein level being shown to vary between 0.52% and 0.76% FW at the eatingripe stage (Yahia and GutierrezOrozco 2011). The concentrations of proteins and soluble amino acids generally decline during ripening, but this is accompanied by an increased activity of several enzymes such as inulase, invertase, amylase, and phosphatase (Selvaraj and Pal 1984;Yahia and GutierrezOrozco 2011; Vishwasrao and Ananthanarayan 2016b). The concentration of soluble proteins has been shown to increase during storage. Camargo (2016) showed that fruit at harvest had an initial protein content of 1700 mg kg–1, and this reached a peak of 2506 mg kg–1 on the fourth day of storage at 27 °C. However, in senescent fruit the protein concentration was 2129 mg kg–1 (Camargo et al. 2016). These different levels are thought to be due to changes in metabolic pathways of maturation that involve the synthesis or activation of hydrolytic enzymes, such as αamylase, βamylase and starch phosphorylase, that lead to an accumulation of inverted sugars and polygalacturonase involved in softening of the tissue and accelerating changes during ripening (Vishwasrao and Ananthanarayan 2016b). The activities of other specific ripeningrelated enzymes such as catalase, pectin

methylesterase (PME), peroxidase and adenosine triphosphatase have been shown to increase during the ripening of sapota fruit (Rao and Chundawat 1988; Yahia and GutierrezOrozco 2011). Polygalacturonase (PG) and PME are softeningrelated enzymes which have been shown to increase during ripening (BautistaReyes et al. 2005). Even at 1 day after harvest the PG activity in the flesh of mature ‘MakokYai’ and ‘KraSuay’ fruit has been shown to increase (Kunyamee et al. 2010). Ethylene was also shown to hasten the increase in PG activity in both ‘MakokYai’ and ‘KraSuay’ cultivars during ripening (Kunyameea et al. 2010). Moreover, BautistaReyes et al. (2005) reported that PME activity in sapota fruit during ripening increased, and reached its maximum levels when the fruit was fully ripe. Levels of βgalacturonase activity are not detectable in fruit at harvest, but have been shown to increase as the fruit starts to ripen (Morais et al. 2008). Polyphenol oxidase (PPO) and peroxidases (POD) are involved in enzymatic browning (TomásBarberán and Espin 2001; Vishwasrao and Ananthanarayan 2016b) in fruit and vegetables, and their activities have been shown to increase during ripening and then to gradually decrease with senescence (Malhotra et al. 2009). Moreover, further studies have shown that sapota fruit possess maximum activities of PPO and POD coinciding with the respiratory climacteric, and then to subsequently decrease (Vishwasrao and Ananthanarayan 2016b).

J. Phenolic Compounds and Antioxidants Phenolics are an important class of phytochemicals contributing to a range of physiological functions, including antioxidant activity. The concentrations of phenolic compounds generally decline during ripening (Lim 2013; Camargo et al. 2016). The total phenolic content, as measured using Folin–Ciocalteu reagent, has been shown to range from 2.7 mg of gallic acid equivalents g–1 FW at the early stages of ripening down to 1.0 mg gallic acid equivalents g–1 FW in the final stages of ripening (Rastegar 2015). Moreover, Vishwasrao and Ananthanarayan 2016b) reported that total phenolic levels decreased by up to 90% after one week of storage at ambient temperature. As some phenolic compounds are often astringent in taste, the reduction of phenolic content makes the sapota fruit more suitable for consumption (Wang et al. 2012; Vishwasrao and Ananthanarayan 2016b). The decrease in the concentration of total phenolic compounds during storage is related to different activities, including those of some degradative enzymes (Fawole et al. 2013). The major individual phenolic compounds in unripe sapota fruit include catechin, epicatechin, gallocatechin, 4hydroxycatechins, chlorogenic acid, and gallic acid (Ma et al. 2003; Lim 2013). Proanthocyanidins have been identified in sapota fruit and have been suggested to be responsible for its high antioxidant activity (Shui et al. 2004; Lim 2013). Isabelle et al. (2010) showed that, among 38 different fruit, sapota fruit had the highest total phenolic content and hydrophilic oxygen radical absorbance capacity (HORAC). Sapota fruit have also been stated to contain leucoanthocyanidins, triterpenoids, and tannins which may relate to its antioxidant properties (Ma et al. 2003; Anand et al. 2004).

IV. PREHARVEST EFFECTS ON POSTHARVEST QUALITY Relatively few studies have been conducted on the effects of preharvest treatments on the postharvest quality of sapota fruit. Bhalerao et al. (2010) examined the influence of different preharvest orchard calcium treatments on the quality and storage life of ‘Kalipatti’ fruit. They showed that the longest shelf life was observed following a preharvest spray of 1% calcium chloride, whilst the level of postharvest decay was lowest following a preharvest treatment of 0.5% calcium chloride. They further showed that weight loss was lowest and firmness was highest with a preharvest treatment of 1% calcium chloride. These results demonstrated the positive effects of calcium chloride for shelflife extension and quality maintenance of sapota fruit. The postharvest quality of sapota fruit grown under organic and conventional agricultural systems in northeast Brazil was evaluated by Oliveira et al. (2010). No significant differences were determined in fruit soluble solids concentration (SSC) between the management systems, but organically grown fruit had higher titratable acidity and vitamin C concentration compared to fruit from the conventional system. However these comparisons were very limited in scope. A preharvest treatment of gibberellic acid (GA3) (200 μl l–1) has been shown to prolong the shelf life of sapota fruit by decreasing weight loss and decay (Choudhury et al. 2003). Moreover, Chavan et al. (2009) showed that 150 μl l–1 GA3 increased the total sugar content of the fruit at harvest.

V. PHYSIOLOGICAL DISORDERS A. Chilling Injury Chilling injury (CI) is characterized by welldefined symptoms that are a consequence of exposure to low temperature. Sapota fruit is susceptible to CI (MooHuchin et al. 2013), and this is a major storage problem. Symptoms of CI in sapota fruit include darkbrown spots and pitting of the skin, localized dark spots, uneven hardness, failure to ripen, increased decay after transfer to higher temperatures, and poor taste and flavor (Mohamed et al. 1996; Yahia and GutierrezOrozco 2011). The development of CI symptoms can occur when fruit is stored at less than 5 °C within 10 days, but can also occur within about 3 weeks when stored at 6–10 °C (Kader 2009). Postharvest treatments such as the use of controlled atmosphere (CA) storage, waxing, film packaging, or applications with polyamines, methyl jasmonate, methyl salicylate, or other natural compounds have been shown to alleviate CI in tropical fruits (Wang 2004). However, little research on these options has been conducted on sapota. Yahia (2004) reported that different edible coatings applied after harvest, such as ‘SemperfreshTM’ and ‘Stafresh TM’, maintained fruit quality for 40 days at 10 °C without showing CI symptoms. In addition,

postharvest hot air treatment (35 °C for 12 h) followed by dipping in 5% CaCl2 for 30 min, has been shown to lower CI symptoms in ‘MaKok’ fruit (Chittham et al. 2002). CI symptoms have also been shown to be prevented following treatment with 1methylcyclopropene (1 MCP) (MooHuchin et al. 2013).

B. Flesh Browning Flesh browning is a major quality constraint affecting the cut surfaces of sapota fruit used in minimal processing. The browning is mainly triggered by the action of PPO, which catalyzes the oxidation of polyphenols in the vacuoles producing oxidized products responsible for the brown pigments and associated sensory and nutritional changes (Rosenthal et al. 2002). Ascorbic acid has been shown to be the most effective postharvest treatment to control the action of PPO in sapota fruit, and browning can be managed by using this treatment during processing (Cortez et al. 2013). The latter authors further showed that sodium azide, acetic acid, sodium metabisulfite and honey also inhibited PPO activity, but tartaric, citric and oxalic acids did not show any inhibitory action against PPO in sapota. Edible coatings utilizing methyl cellulose and palm oil have been shown to decrease the browning of sapota fruit during storage at ambient temperature. Skin darkening was delayed by slowing the browning process caused by PPO and POD (Vishwasrao and Ananthananrayan 2016b).

VI. POSTHARVEST DISEASES Postharvest diseases are one of the main causes of postharvest losses in sapota fruit which, being soft textured, are highly sensitive to exogenous microorganisms, especially fungi. The main postharvest diseases of sapota fruit are black mold rot (Aspergillus niger), sour rot (Geotrichum candidum), blue mold rot (Penicillium itallicum), and anthracnose (Colletotrichum gloeosporioides) (Kader 2009). Other disease organisms include Phytophthora palmivora and species of Pestalotiopsis and Phomopsis (Snowdon 2010). Among these postharvest diseases affecting sapota fruit, black mold rot causes significant losses during storage (Wagh and Bhale 2012b). Fruit infected with anthracnose at the mature fruit stage leads to considerable losses during storage, transit, and marketing. When the anthracnose fungus invades mature fruit the resultant lesion is small and relatively cosmetic, with a shallow area of hardened tissue (Mossler and Crane 2002; Ahuja and Chattopadhyay 2015). Soft rot caused by the fungus Pestalotiopsis mangiferae is another major disease of sapota fruit in many parts of the world. The disease appears as watersoaked spots covering the entire fruit within 3–4 days (Srivastava 2014). The rotted fruit then become soft and dark brown, with numerous acervuli being visible in the rotted zones. Fruit rots caused by Phytophthora palmivora also exhibit watersoaked lesions which become brown within 2–3 days. Subsequently, the whole fruit is covered with tufts of mycelium (Srivastava 2014). Postharvest diseases significantly affect fruit physiology, morphology and biochemistry, and cause significant quantitative and qualitative losses (Bhale et al. 2013). For example,

postharvest fungal diseases interact with the physiology of the sapota fruit to enhance senescence and infection (Gadgile et al. 2010). Wagh and Bhale (2012a) showed that the levels of endogenous phenolic compounds in the flesh of sapota fruit decreased after infection with Aspergillus niger.

A. Control of Postharvest Diseases The control of postharvest diseases in sapota fruit is essential for quality retention and successful marketing. Geotrichum candidum, causing sour rot, can be controlled with postharvest dipping in carbendazim ranging from 2100 μg ml–1 to 4000 μg ml–1, and in mancozeb ranging from 80 μg ml–1 to 300 μg ml–1 (Wagh and Bhale 2012b). Active ingredients used for the control of selected diseases include azoxystrobin, mefenoxam, myclobutanil, hydrogen dioxide, carbonic acid, and phosphite, and Bacillus subtilis (Mossler and Crane 2002; Crane and Mossler 2003). A preharvest spray with 50 μl l–1 GA3 has also been shown to reduce postharvest Fusarium sp. infection in ‘PKM 1’ sapota, where no infection was observed for up to 6 days of storage (Sudha et al. 2007). Moreover, a preharvest spray with 50 μl l–1 GA3, in combination with a postharvest treatment of GA3 and carbendazim, reduced the development of postharvest spoilage for up to 12 days (Sudha et al. 2007). Tsomu and Patel (2014) reported that a postharvest spray with calcium chloride (5000 mg l–1) for 5 min could decrease postharvest spoilage of ‘Kalipatti’ sapota fruit for up to 12 days at ambient temperatures. They suggested that this may be due to the presence of higher calcium in the peel and the pulp, resulting in a stronger intracellular organization and rigidified cell walls (Tsomu and Patel 2014). Research with biological control agents such as Trichoderma sp., including T. koningii and T. pseudokoningii, has shown the potential of these biological control agents to inhibit and control postharvest diseases in sapota fruit (Bhale et al. 2013). Further research in this area should focus on in vivo studies on the efficiency of the Trichoderma species as biocontrol agents, where the fruit is dipped into a suspension of Trichoderma after harvest (Bhale et al. 2013).

VII. POSTHARVEST TECHNOLOGY The shelf life of sapota fruit is very short (2–3 days) at ambient temperature, and postharvest losses can commonly reach up to 30% (Polara 2013). Thus, there is a need to improve postharvest systems to preserve quality and prolong the postharvest life of the fruit. Sapota fruit are harvested from the tree by cutting the peduncle, using a knife or secateurs, where fruit selection is usually based on size and maturity stage (Polania Trujillo 1986; Yahia 2004). Fruit are usually carried to local market in bamboo baskets, using banana leaves as a lining material to protect the fruit from abrasion and excessive damage. A foldable transportation container of 10 kg capacity has been designed to minimize transportation losses of sapota fruit (Gontia 2016). This has enclosed conditions for protecting the produce from

adverse climate and microorganisms. Maximum firmness, marketable fruit and sensory score, and minimum weight loss of sapota fruit have been observed when using this container (Gontia 2016). In addition, bruising and impact damage and decay were substantially decreased with the use of the foldable plastic container as compared to a gunny bag, perforated polypropylene bag, corrugated fiber board (CFB) carton, egg tray in CFB carton, or plastic crate (Gontia 2016). Mature fruit are picked with or without the stalk (peduncle). Immediately after harvest a white latex exudes from the stalk, and should be removed because if it remains inside the fruit it coagulates and downgrades quality. The process of removing the latex is called “bleeding,” where the freshly harvested fruit are placed into a container with water. Bleeding is stimulated by scraping the stalk end with the thumbnail or with a sharp object. In addition, the fruit are scrubbed with a piece of cloth to remove the bloom, and then allowed to dry by placing them with their stalk ends down. A continuous mechanical system has been developed to clean sapota fruit in India. This is optimized to process up to 550 kg of fruit per hour with 100% cleaning efficiency and without causing damage to the fruit (Gontia 2016). Precooling is one of the primary postharvest operations required to maintain quality and to prolong postharvest life. The precooling of ‘Kalipatti’ fruit at 8 °C for 8 h has been shown effective in prolonging the postharvest life by lowering the percentage of spoilage, lowering losses from physiological disorders, and maintaining fruit firmness. Precooling has also been shown to reduce the increase in soluble solids concentration (SSC) and to increase the organoleptic rating with regards to color, taste, and overall acceptability (Polara 2013).

A. Grading and Packing Grading classifies fruit into uniform sizes and grades for the market, and is a vital prerequisite step in marketing sapota fruit. Grading can be conducted either manually (Varshney et al. 2005) or mechanically (Ukey and Unde 2010), where size and maturity can be used as grading indicators. Efficient grading operation, on the basis of physical dimensions of sapota fruit, can be made with the help of a mechanical sapota grader. The result is a product that is consistent in volume and shape and packs easily (Ali and Mandhar 2012). Fruit which have decay or mechanical injury are discarded (Yahia and GutierrezOrozco 2011). Most countries use standardized grading systems for sapota fruit where in all cases the fruit must be fresh in appearance and free from any foreign matter, pests, and diseases. Fruit must be mature, of similar varietal characteristics, clean, wellformed, of similar color, and free from fruit fly, decay, anthracnose, mechanical and chilling injury. Sapota are normally classified, based on the quality of the fruit, as Extra Class, Class I, and Class II (Philippine National Standard 2011; Thai Agricultural Standard 2011). In the Extra Class, the fruit must be of superior and uniform quality. Fruit in Class I must be of good quality, free from slight defects in shape, color, and skin defects (i.e., bruises) which should not exceed 5% of the surface area. In Class II, fruit do not meet the quality requirements for inclusion in the higher classes but satisfy the minimum requirements specified in this class, namely defects in shape, color, and skin defects (i.e., bruises) not to exceed 10% of the surface area. Fruit can also be graded into

six categories depending on their size, viz., jumbo (more than 221 g), extra class (181–220 g), large (141–180 g), medium (101–140 g), small (60–100 g), and very small (5.0 mm diameter (70 g weight); color/texture – corky brown, smooth surface without any latex deposit on skin; taste – sweet; packaging – corrugated cardboard boxes with 16 fruits (4 × 4) including an ethyleneabsorbant material. Each market has different specifications and uses different packing materials but, regardless, sapota fruit must be packed properly in a suitable container. The materials used inside the package must be new, clean, and of good quality to avoid any external or internal damage to the produce. In many markets such as India, after grading, fruit are packed in singlelayer trays with rice straw as padding material between fruit in fiberboard or wood flats (25–49 counts and 4.5 kg capacity). To avoid injury, the fruit weight should not exceed 20 kg per box (Yahia and GutierrezOrozco 2011; Love et al. 2014). Kenghel et al. (2015) compared a range of different filling materials and showed that coconut coir performed best and was the most economical filling material for the storage of ‘Kalipatti’ sapota fruit. However, all containers must meet the quality, hygiene, ventilation and strength characteristics to ensure optimum handling, shipping and preservation of the fruit (Philippine National Standard 2011; Thai Agricultural Standard 2011).

B. Storage Like many other tropical fruits, sapota fruit have a short storage and shelf life. The optimum storage conditions are 14 °C and 90–95% RH (Kader 2009), where fruit can be kept for 14–28 days (Yahia 2004). CI can be induced under storage at lower temperatures, whilst storage at higher temperatures hastens quality deterioration. Yahia (1998) reported that the storage life of sapota fruit was increased from 13 to 18 days at room temperature when stored in an atmosphere of 5 kPa CO2, 21 days in 10 kPa CO2, and up to 29 days in 20 kPa CO2. Freshly harvested ‘Cricket Ball’ fruit stored under 2 kPa O2 combined with 10 kPa CO2 has been shown to have an extended postharvest life about four to fivefold that of control fruit under ambient conditions, by minimizing changes in physicochemical characteristics and slowing down the process of ripening (Emerald et al. 2001). It has generally been shown that the storage life of sapota fruit can be improved beyond 2–4 weeks with optimal storage conditions, including the use of modified and controlled atmospheres. This is discussed in more detail below in the section on modifiedatmosphere packaging (MAP). Hypobaric storage involves the storage of produce under reducedpressure conditions. Hawa (2005) showed that storage of sapota fruit at 10 °C under a pressure level of 30 kPa reduced respiration rate and weight loss.

VIII. POSTHARVEST TREATMENTS

A. Ethylene Management and 1Methylcyclopropene 1Methylcyclopropene (1MCP) is widely used to delay ripening and softening of climacteric fruits such as apple (Morais et al. 2008). MooHuchin et al. (2013) indicated that treating sapota fruit with 1MCP (1 μl l –1) followed by storage at 16 °C prolonged the shelf life for up to 28 days. Furthermore, treatment with 1MCP considerably decreased respiration rates. MooHuchin et al. (2013) showed that the treatment of fruit with 1MCP resulted in firmer fruit, while Morais et al. (2008) showed that 1MCP treatment of ‘Itapirema31’ sapota fruit significantly delayed the rate of fruit softening, with less extensive solubilization of polyuronides, hemicellulose and of free neutral sugars when compared with control fruit. Moreover, Qiuping et al. (2006) showed that treatment with 1MCP (at 40 or 80 nl l –1 for 24  h at 20 °C) markedly inhibited the rates of respiration and ethylene production, and ethylene induced ripening processes such as softening and chlorophyll degradation were delayed. The same authors showed that 1MCP treatment delayed the appearance of the PG activity peak by an additional 6 days, and 1MCPtreated fruit had higher concentrations of SSC, TA, and ascorbic acid at the end of storage. These studies have shown that 1MCP is a potent ethylene action inhibitor in sapota fruit, and the application of 1MCP is a feasible technology for ripening inhibition and quality maintenance of harvested sapota fruit to allow extended storage and marketing.

B. Ethylene Management and LowEthylene Storage Storage in lowethylene atmospheres has been shown to increase the storage life of other climacteric tropical fruit such as banana (Wills et al. 2014). Bhutia et al. (2011) showed that the application of ethyleneabsorbent materials (e.g., sachets containing KMnO 4) could be used to manage sapota fruit ripening. They showed the use of inpackage ethylene absorbent to be dependent on the degree of ripening of sapota fruit, where the response was maximum in mature fruit, and delayed the ripening of fruit and prolonged marketable life for up to 13 days at ambient temperature (25–27 °C). A major reduction in postharvest decay (from approximately 40% to around 12–15%) was also observed (Bhutia et al. 2011). Thus, the use of lowethylene storage, such as when using ethylene absorbers in packed boxes of sapota fruit during longdistance transportation, could be a useful postharvest technology for delaying softening by removing the deleterious effects of ethylene around the fruit, delaying ripening, and further slowing down metabolic activities. This would have significant benefits where refrigeration is not available (Wills and Golding 2015).

C. Plant Growth Regulators Plant growth regulators are an integral component of tree fruit production, and have also been studied as postharvest treatments to enhance fruit quality. Gibberellic acid (GA3) is known for its general antisenescing properties which have been shown to delay fruit ripening (Tsomu et al. 2015). Combined treatments of GA3 (150 mg l–1) dips and storage at 12 °C have been shown to increase the storage life of ‘PKM1’ and ‘Kalipatti’ sapota (Patel et al. 2010).

Studies on the effects of various concentrations of ethephon (2chloroethylphosphonic acid) on ripening of fruit were carried out on ‘Kalipatti’ and ‘Cricket Ball’ cultivars. Fruit dipped twice in ethephon solution (1000 μl l–1) resulted in superior ripening after 3 days, with pleasant flavor, high SSC, lower acidity, and acceptable sensory quality (Bons et al. 2016).

D. Calcium Calcium plays a significant role in postharvest quality, due to its role in plant metabolism and membrane stability (Bahmani et al. 2015). Dipping sapota fruit in calcium chloride (CaCl2) has been shown to be effective in reducing weight loss, spoilage, maintaining fruit firmness, and increasing shelf life (by up to 12 days) of ‘Kalipatti’ fruit stored at room temperature (Tsomu and Patel 2014). Vacuum infiltration has been shown to improve the efficiency of calcium dips. Dhua et al. (2006) showed that sapota fruit could be stored for 16 days at 8 °C without deterioration in quality following vacuum infiltration with CaCl2 for 5 min. Fruit infiltrated with higher concentrations of calcium remained firmer and had higher SSC and sugars compared with those infiltrated with lower concentrations of calcium, and also untreated fruit. Moreover, gummy spot formation on the fruit surface due to milky latex was alleviated by dipping sapota fruit immediately after harvest in 1% calcium hydroxide for 5 min followed by wet rubbing, which improved the appearance of the fruit and extended its postharvest life (Tandel and Patel 2011).

E. Essential Oils Essential oils are a wide range of secondary metabolites, which have a broad spectrum of antimicrobial, antioxidant, and regulatory roles (Bahmani et al. 2015). The application of thyme (Thymus vulgaris) essential oils in combination with CaCl2 dips was evaluated on the quality of ‘Oval’ sapota fruit (Bahmani et al. 2015). Thyme essential oil alone or in combination with CaCl2 completely inhibited the growth of fungi on fruit during storage. The use of thyme essential oil at low concentration (250 μl l–1) with CaCl2 also had a positive effect on maintaining fruit firmness, reducing weight loss, and enhancing fruit appearance. CaCl2 at 2% and a high concentration of essential oil (500 μl l–1) maintained the highest fruit firmness, but the use of essential oil at this concentration also caused skin and flesh discoloration and an unpleasant odor and taste (Bahmani et al. 2015). It is suggested that thyme essential oil with or without low concentrations of CaCl2 could replace the use of chemical fungicides for the control of decaycausing fungi on sapota fruit, and increase its postharvest life.

F. Edible Coatings Edible coatings have been reported to play a major role in maintaining the quality of a wide range of fruits (Dey et al. 2014). These authors studied the effect of corn starch on the storage life and quality of sapota fruit, and showed that the coating delayed disease occurrence and physiological and biochemical losses (e.g., change in weight, ascorbic acid content, phenolic

content, and color). Chitosan coatings on sapota fruit have also been shown to maintain fruit quality, with lower weight losses and respiration rates (Ahlawat et al. 2015). The same authors further showed that after 2 weeks’ storage the 1.5% chitosancoated sapota had no deterioration in fruit quality. The use of other coatings, such as methyl cellulose and palm oil, have been shown to lower the activities of PPO, POD and PME, and delay the loss in phenolic compounds in fruit (Vishwasrao and Ananthanarayan 2016a). Moreover, pectinbased edible coatings have been shown to prolong the shelf life of sapota fruit for up to 11 days by delaying changes in physicochemical factors such as weight loss, SSC, pH, total acidity, ascorbic acid, firmness and color (Menezes and Athmaselvi 2016). Aloe vera is used extensively in the food and cosmetics industries, and has been shown to have beneficial effects on the quality of fresh fruits (MartínezRomaro et al. 2006). Padmaja et al. (2015) showed that a dip treatment in an aloe vera coating was effective in maintaining fruit quality, with an extension of postharvest life by up to 20 days (Padmaja et al. 2015). Palm oil coatings have also been tested and have shown benefits in fruit quality after storage, such as lower rates of weight loss, firmness, total soluble solids and increased shelf life for up to 15 days (Binti Arifin 2009). These results suggest that edible coatings could be novel and commercial means for maintaining sapota fruit quality.

G. Heat Treatments Heat treatments have been widely used as postharvest quarantine treatments, and also maintain fruit quality during storage. For example, the postharvest life of ‘Co1’ and ‘Co2’ sapota fruit was extended for up to 14 days when fruit were dipped in hot water at 50 °C for 5 min after harvest (Vijayalakshmi et al. 2004). The hot water treatment maintained quality parameters such as SSC, total sugars, and SSC/TA ratio. Chittham et al. (2002) also showed that sapota fruit quality was maintained for up to 40 days using fruit treatment with hot air (35  °C for 12 h), followed by dipping the fruit in 5% CaCl2 for 30 min. As sapota fruit can host quarantine pests such as fruit flies, postharvest disinfestation treatments are required to kill quarantine pests in the fruit to allow market access. The efficiency of the thermal treatment to kill the larvae of the quarantine pest Mediterranean fruit fly (Ceratitis capitata) was evaluated, and showed that vapor, hot water and hot air at 50 °C for 75–90 min were all efficient in killing larvae (Lima et al. 2006).

H. Irradiation Lowdose irradiation is used as a phytosanitary treatment to allow market access, and can be used to maintain fruit quality. Srinu (2014) showed that ‘Kalipatti’ fruit treated with 200 Gray (Gy) retained higher firmness and lower sugar content when compared with higher treatment doses and nontreated control fruit. Srinu et al. (2015) further showed that the irradiation of sapota fruit with 200 Gy irradiation, followed by storage at 15 °C, increased postharvest life to 26 days compared to 15 days for untreated fruit. However, fruit treated with doses higher than 400 Gy were softer and had a shorter shelf life, and developed injury and spoilage, compared to fruit irradiated at lower doses (200 Gy). Higher treatment doses (800 Gy) caused brownish

spots on the surface of the fruit within 3 days of storage. Srinivas et al. (2012) also reported that 200 Gy treatment retained a higher firmness of ‘Kalipatti’ fruit and that higher treatment doses resulted in fruit damage.

I. ModifiedAtmosphere Packaging Modifiedatmosphere packaging (MAP) involves modification of the atmosphere inside produce packaging to provide lower levels of oxygen (O2) and higher levels of carbon dioxide (CO2), which can delay fruit ripening, reduce respiration rate, and maintain fruit quality (Antala et al. 2014). MAP, using lowdensity polyethylene (LDPE) bags, has been shown to generate 5 kPa O2 + 5 kPa CO2 or 5 kPa O2 + 10 kPa CO2 inside modified atmosphere bags, which prolonged the postharvest life of ‘Kalipatti’ fruit for up to 49 days at 6 °C (Antala et al. 2014). Dash et al. (2012) proposed a model for sapota fruit packed in modified atmosphere using different packaging materials and based on the respiration rate and permeability of the packaging material. The most suitable packaging films for sapota fruit were determined to be LDPE, polyvinyl chloride, polypropylene and polystyrene films, whilst saran (a number of polymers made from vinylidene chloride, especially polyvinylidene chloride along with other monomers) and polyester films were found to be unsuitable (Dash et al. 2012). A comparison of the storage effects of LDPE bags with different gauges (100, 200, 300gauge) and three different ventilations (0.8, 1.2, and 1.6%) was conducted on mature ‘Kalipatti’ fruit. Fruit stored in LDPE bags had better quality and longer shelf life, whilst fruit stored in 200gauge LDPE bags with 1.2% ventilation had a maximum shelf life of 32 days and maintained the highest quality (Bindu Praveena et al. 2013). The longer shelf life and better quality of fruit in LDPE bags was due to the reduced permeability of the LDPE bags for O2 and an accumulation of CO2, coupled with low temperature. This resulted in lower respiration rates, a delayed onset of ripening, and slowing of the activity of celldegrading enzymes (Bindu Praveena et al. 2013). A range of other wrapping materials such as newspaper, cling film and polyethylene of 0.05 and 0.1 mm thickness, and two wrapping methods (i.e., individual wrapping and lining), have been assessed, and showed individual wrapping of fruit in clingfilm to be the best method for reducing water loss and decay (Pahel et al. 2015). Fruit packed in 100gauge polypropylene bags with 0.1% perforation resulted in a maximum total storability of 21 days (15 days at 15 °C and 6 days at room temperature) (Srinu et al. 2014). Other coating treatments such as bees wax have been shown to provide a good barrier to water loss and also to help maintain the texture of fruit under both refrigerated and ambient conditions (Morrison et al. 2011). For example, a bees wax coating was found to be superior in maintaining quality compared to cling wrap (Morrison et al. 2011). These results clearly demonstrate the potential of MAP to be an effective aid to prolong the shelf life of sapota fruit.

IX. NONDESTRUCTIVE METHODS FOR IDENTIFYING FRUIT MATURITY AND QUALITY Identifying the maturity stage of sapota at harvest is very difficult as no external color changes

occur during fruit development and ripening processes (Aziz et al. 2001). As sapota flowers and fruit form in clusters of two to four, there is a range of maturities within each flowering/fruit cluster. This variability in fruit maturity causes significant problems for packers and consumers. A number of nondestructive methods can be used to measure the internal quality of fruit and vegetables, and these have been used to sort fruit into different fruit maturity and quality classes. This segregation improves the consistency of fruit quality in each batch. A nondestructive technique using the “Kiwifirm” device (Plant and Food Research, New Zealand) has been shown to be useful in identifying the maturity stage of sapota fruit, predicting the quality at ripening stage and also the time required for the fruit to ripen (Aziz et al. 2001). “Kiwifirm” measures the deceleration of a tiny hammer that impacts the surface of the fruit via a small nonbruising and nonpenetrating tip. An inbuilt processor records the resulting collision, analyzes the waveform, and displays a value on a digital display. This technique is nondestructive and allows the testing of individual fruits, as well as repeated testing of the same fruit. This could be applied to predict which fruit are going to ripen during storage and to reduce inconsistent ripening within a batch (Aziz et al. 2001). However, more research is needed to develop this technique. Moreover, there is a need for more advanced grading techniques to improve the sorting of highquality attributes of the fruit. For example, the use of a nondestructive technique, such as applying impact response using Fruit Firmness 1 (FF1), has been used to assess and predict internal quality and storage life of ‘Subang’ sapota (Ibrahim 2007). Sapota fruit could be successfully sorted into classes with different storage periods, and fruit for immediate sale (eating ripesoft) could be identified separately from those that needed additional ripening/storage (e.g., 1–3 more days storage at 12 °C) to ensure consistent eating quality. The prediction and classification of fruit using non destructive methods will guarantee fruit quality in different markets and limit postharvest losses in the supply chain (Ibrahim 2007). The use of an “electronic nose” had been successful in detecting the aroma of sapota fruit, using four sensors during the ripening process, with the aim of classifying fruit for further sorting processes (Karyadi et al. 2012). Principal component analysis showed that it was possible to classify fruit into unripe, ripe, and overripe maturity stages (Karyadi et al. 2012). Other nondestructive measures, such as estimates of the composition of sugars in immature and mature sapota fruit, using nuclear magnetic resonance techniques, have been demonstrated (Chaughule et al. 2002). Sugars accumulated in sapota fruit during development could be clearly evaluated, but this technology is highly complex and expensive.

X. PROCESSING As the consumption of processed fruit products is becoming popular, the role of processing industries is becoming increasingly important in horticulture. Processing allows for the utilization of otherwise waste sapota fruit which do not meet fresh fruit consumer or market expectations. Of all the sapota fruit produced in India, nearly 30% is wasted due to incorrect postharvest handling, transportation, and storage (Ganjyal et al. 2005). The use of processing technologies can potentially utilize this waste and add value.

Minimally processed sapota is becoming popular, but rapid fruit softening and discoloration are significant problems which reduce shelflife. Postcutting applications of CaCl 2 have been shown to maintain fresh weight and firmness and to delay color changes of treated ‘Malay’ sapota. However, at high concentrations of CaCl2 there was evidence of the development of a deleterious white scale covering the cut surfaces (Poonsawat et al. 2007). Moreover, Purba (2011) showed that coating with 0.5% glucomannan, and additional modified atmosphere packaging using stretch film, was the optimum treatment for the storage and quality maintenance of freshcut ‘Sukatali ST1’ fruit. Milky white latex that is produced from unripe sapota fruit forms the base for making chicle, which is commercially produced in South East Mexico, Guatemala, and Belize. Chicle is also obtained from latex from the bark of sapota trees, and is the main ingredient of chewing gum (Ahmed et al. 2011). Mixed jams are also produced from mature sapota fruit and provide a valuable source of material for the manufacture of industrial natural fruit jellies (Ganjyal et al. 2003; Hiwale 2015). Dried sapota powder is also rich in vitamins, carbohydrates, fibers, and proteins (Jangam et al. 2008), and can be used as a fiber supplement. Dehydration has been identified as a costeffective and viable method for producing a wide range of products for the food industry (Ganjyal et al. 2005). Drying removes the moisture from fresh sapota fruit and can be used to preserve sapota slices for long periods of time (Divya et al. 2014). Several drying methods have been examined for fresh sapota fruit, including convection and vacuum oven drying, solar, forcedair, and forcedhot air drying. The highest recovery of soluble solids content, acidity, ascorbic acid and overall quality was obtained from sapota fruit when dried at 65 °C (Ganjyal et al. 2005). The dried powder from vacuum drying was shown to be of superior quality to that obtained from convectional drying (Ganjyal et al. 2005). Padmini et al. (2005) also showed that the moisture content of sapota fruit can be decreased from 76% to 10% within 76 h by solar drying, with time savings of 27% compared to passive sun drying. They further showed that the drying rate of unpeeled and whole fruit was lower than in peeled and thinner sliced fruit (Padmini et al. 2005). Other studies have indicated that the drying of whole fruit with peel was not possible (Jangam et al. 2008). Kumar et al. (2016) dried sapota fruit using a solar tray dryer and hotair cabinet dryer at drying air temperatures of 80 °C, 100 °C and 120 °C, and showed that the best overall results were obtained using hotair drying at 80 °C. Osmotic dehydration is a commonly accepted method for the partial removal of water by submersing fruit in sugar/salt solution (Patil et al. 2014). The osmotic dehydration process of mature sapota fruit was optimized for water loss and sugar gain (Patil et al. 2014), where the optimum conditions were found to be a treatment of 47 °C with sugar concentration of 55.5 °Brix for a total immersion processing time of 167 min. Rodrigues et al. (2009) studied the effects of ultrasound and ultrasoundassisted osmotic dehydration on fresh sapota tissue, where samples were immersed in distilled water and subjected to ultrasonic waves for 10, 20 and 30 min. The osmotic solution used was prepared by mixing foodgrade sucrose with distilled water to give concentrations of 35 and 70 °Brix. Changes in sapota cell structure could be induced by both methods, where the formation of microscopic channels following ultrasound treatment occurred chiefly through the breakdown of dense cells, which helped in

the dehydration process. The application of ultrasoundassisted osmotic dehydration resulted in major changes in the tissue structure of the fruit, with consequent increase in effective water diffusivity, because of the formation of microscopic channels and cell rupture (Rodrigues et al. 2009). Fruit bars have become increasingly popular for their convenience. The preparation of fruit bars includes fruit purées and other ingredients, along with additives such as pectins to increase physicochemical and sensory characteristics (Salleh et al. 2016). Increasing hardness, adhesiveness, and chewiness of sapota fruit bars was correlated positively with pectin concentration. The color (hue) of the fruit bar also increased as a result of high pectin content (3%), which suggests that pectin can be added as a stabilizer to extend shelf life, and to improve the texture of sapota fruit bars (Salleh et al. 2016). Nectar is another valueadded product which can be prepared from sapota fruit. Relekar et al. (2016) showed that nectar prepared with 25% juice and 18 °Brix SSC had the highest overall acceptability. Sapota wine is another alternative use for sapota fruit. Pawar et al. (2012) showed that the clarified juice of half ripe, ripe and overripe fresh sapota fruit could be used for preparing commercially acceptable wines that may be prepared from pulp with further pressing (Pawar et al. 2012).

XI. SUMMARY AND FUTURE PROSPECTS Sapota is a nutrientrich fruit which is widely grown in tropical areas. The fruit can be eaten fresh or processed. While there are numerous commercial cultivars, further improvements in fruit quality attributes, such as reduced softening and resistance to postharvest pathogens, would improve the production and quality of sapota fruit available to farmers and consumers around the world. Successful research and development in India has led to new sapota cultivars with improved storage, transport, and marketing characteristics. This example has significant applications to other growing regions with different agronomic and postharvest challenges. Sapota fruit are very perishable, with a strong climacteric ripening behavior that requires careful handling after harvest to allow transport to distant markets at optimum quality. Both pre and postharvest factors such as plant nutrition, storage conditions, and packaging all influence the postharvest quality and shelf life of sapota fruit. However, there needs to be a systematic investigation of different pre and postharvest treatments and management practices to effectively manage postharvest physiological disorders and diseases of sapota fruit, and to establish correct treatments to ameliorate postharvest handling problems. Numerous maturity indices have been considered, including color and days from full bloom to maturity, but more robust techniques need to be developed. Although sapota fruit ripening has been physiologically described, few studies have been performed of genetic differences and enzymatic changes that occur during ripening, such as cell wall pathogenesisrelated enzymes and antioxidant enzymes. There are also significant opportunities to study the associated genes, proteomics and other molecular approaches to investigate the mechanisms of ripening, senescence and resistance to chilling injury and

diseases in sapota fruit. Cold storage is effective in extending postharvest life, but more systematic research is needed to confirm the ideal temperature for different cultivars and different growing conditions. In addition, complementary treatments such as the use of 1MCP and calcium application should be developed to prolong the postharvest life of the fruit. The use of modified and controlled atmospheres, especially the use of MAP under refrigeration, have shown promising results for postharvest handling and storage, but further research is needed to establish appropriate commercial handling conditions. The establishment of practical postharvest handling and storage techniques will benefit the distribution of sapota fruit onto world markets, where robust export protocols and handling techniques are required to meet market access requirements and retail demands. Sapota fruit can also be processed into various types of food products. The development and application of new food processing technologies, such as highpressure hydrostatic processing, will assist new product development and the utilization of otherwise waste fresh sapota. Although sapota fruit is popular in many tropical countries, it is still only a specialty fruit in many parts of the world. Nevertheless, because of its sensory appeal and high nutritional status, sapota has the potential to increase in popularity, with the continuous supply of nutritious, highquality fruit onto world markets. Consumer awareness of the fruit and more detailed clinical trials into human health benefits would also increase its awareness and consumption.

LITERATURE CITED Agrawal, S. and S.N. Dikshit. 2008. Studies on the effect of plant growth regulators on growth and yield of sapota (Achras sapota L.) cv. Cricket Ball. Indian J. Agric. Res. 42:207–211. Ahlawat, A., A. Saha, Y.K. Tyagi, and R.K. Gupta. 2015. Development of chitosan based edible coatings to study sapota (Manilkara zapota) fruit shelf life. J. Chem. Pharm. Res. 7:879–885. Ahmed, T., M. Burhanuddin, M.A. Haque, and M.A. Hossain. 2011. Preparation of jam from sapota (Achras zapota). The Agriculturists 9:1–7. Ahuja, D.B. and C. Chattopadhyay. 2015. Pests of fruit trees (citrus, banana, mango, pomegranate and sapota) EPest surveillance and pest management advisory, ICARNational Research Centre for Integrated Pest Management, New Delhi and State Department of Horticulture, Commissionerate of Agriculture, Pune (Maharashtra): 124. Ali, S.M. and S.C. Mandhar. 2012. Development of electronic weight grader for sapota. AMA Agric. Mech. Asia Afr. 43:78–81. Almeida, M.M.B., P.H.M. de Sousa, Â.M.C. Arriaga, G.M. do Prado, C.E. de Carvalho Magalhães, G.A. Maia, and T.L.G. de Lemos. 2011. Bioactive compounds and antioxidant activity of fresh exotic fruits from northeastern Brazil. Food Res. Intl. 44:2155–2159.

Anand, K.P., R.S. Policegoudra, and S.M. Aradhya. 2004. Chemical composition and antioxidant activity of sapota fruit. J. Food Biochem. 31:399–414. Antala, D.K., R.M. Satasiya, P.D. Akabari, J.V. Bhuva, R.A. Gupta, and P.M. Chauhan. 2014. Effect of modified atmosphere packaging on shelf life of sapota fruit. Intl. J. Agric. Sci. Technol. 2:32–38. Athmaselvi, K.A., P. Jenney, C. Pavithra, and I. Roy. 2014. Physical and biochemical properties of selected tropical fruits. Int. Agrophys. 28:383–388. Awasarmal, A.B., S.B. Soni, and S.P. Divekar. 2011. Effect of different packaging materials on shelf life of sapota fruit. Intl. J. Proc. Post Harvest Technol. 2:125–128. Aziz, I.A., I.M. Nordin, A.A. Shukor, and A. Izham. 2001. Predicting storage life of sapodilla (Manilkara zapota L.) by nondestructive technique. J. Trop. Agr. Food Sci. 29:93–98. Bahmani, M.S., M. Yusefzadi, and M. Hamed Hasanzadeh. 2015. The effect of thyme essential oil, calcium chloride and storage time on quantity and quality of sapodilla fruit (Manilkara zapota L.) var. Oval. Intl. J. Agric. Crop Sci. 8:221–226. BautistaReyes, B., M.L. ArévaloGalarza, C. SaucedoVeloz, and M.T. MartínezDamián. 2005. Proceso de maduración de frutos de chicozapote (Manilkara zapota (L.). P. Royen. TipoFino. Revista Chapingo Serie Horticultura 11:387–391. Bhale, U.N., P.M. Wagh, and J.N. Rajkonda. 2013. Antagonistic confrontation of Trichoderma spp. against fruit rot pathogens on sapodilla (Manilkara zapota L.). J. Yeast Fungal Res. 4:5– 11. Bhalerao, R.R., B.R. Parmar, B.V. Padhiar, and P.P. Bhalerao. 2010. Preharvest spray of different sources of calcium to improve the bio chemical qualities of sapota fruits (Manilkara achras (Mill) Fosberg) cv. Kallipatti. Asian J. Hortic. 5:93–95. Bhutia, W., R.K. Pal, S. Sen, and S.K. Jha. 2011. Response of different maturity stages of sapota (Manilkara achras Mill.) cv. Kallipatti to inpackage ethylene absorbent. J. Food Sci. Technol. 48:763–768. Bindu Praveena, R., V. Sudha Vani, and M. Rajasekhar. 2013. Influence of low temperature on shelf life and quality of sapota [Manilkara achras (Mill.) Fosberg] fruits packed in polybags. Acta Hortic. 1012:873–879. Binti Arifin, N. 2009. Effect of palm oil coating on sapodilla [Manilkara achras (Mill.) Fosberg] in different concentrations at ambient temperature. FASM, Universiti Malaysia Terengganu. Bons, H.K., R.S. Boora, and H.S. Rattanpal. 2016. Studies on effect of various levels of ethephon on ripening of sapodilla fruits [Manilkara achras (Mill) Fosberg]. Intl. J. Agric. Environ. Biotechnol. 9:563.

Bose, T. and S.K. Mitra. 1990. Fruits: tropical and subtropical. Naya Prakash, Calcutta. Brito, E.S.D. and N. Narain. 2002. Physical and chemical characteristics of sapota fruit at different stages of maturation. Pesquisa Agropecuária Brasileira 37:567–572. Camargo, J.M., A.T. Dunoyer, and L.A. GarcíaZapateiro. 2016. The effect of storage temperature and time on total phenolics and enzymatic activity of sapodilla (Achras sapota L.). Revista Facultad Nacional de Agronomía, Medellín. 69:7955–7963. Chaughule, R.S., P.C. Mali, R.S. Patil, G.D. Joshi, and Y.M. Lo. 2002. Magnetic resonance spectroscopy study of sapota fruits at various growth stages. Innov. Food Sci. Emerg. Technol. 3:185–190. Chavan, S.R., M.B. Patil, G.N. Phad, and A.B. Suryawanshi. 2009. Effect of growth regulators on yield attributes and quality of sapota [Manilkara achras (Mill.) Forsberg] cv. Kalipatti. Asian J. Hortic. 4:176–177. Chittham, A., S. Kanlayanarat, and C. Wongsaree. 2002. Effect of heat treatment and calcium chloride on chilling injury of Archras sapota Linn. Agric. Sci. J. 33:122–126. Choudhury, S., D.P. Ray, B.K. Das, and G.S. Sahu. 2003. Effect of pre and postharvest chemical treatments on ripening, quality and storage life of sapota (Manilkara achras) cv. Pala. Orissa J. Hortic. 31:54–57. Cortez, J.T., C.H.H. Méndez, E.S. Duch, M.D.L.V. y Vargas, and S.S. Pereira. 2013. Purification and partial characterization of polyphenol oxidase from sapodilla plum (Achras sapota). Food Nutr. Sci. 4:727. Crane, J.H. and M.A. Mossler. 2003. Pesticides registered for tropical fruit crops in Florida. University of Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, EDIS. CIR 1414. da Silva, L.M.R., E.A.T. de Figueiredo, N.M.P.S. Ricardo, I.G.P. Vieira, R.W. de Figueiredo, I.M. Brasil, and C.L. Gomes. 2014. Quantification of bioactive compounds in pulps and by products of tropical fruits from Brazil. Food Chem. 143:398–404. Das, S. and B. De. 2015. Analyzing changes in metabolite profile during postharvest ripening in Achras sapota fruits: GCMS based metabolomics approach. Intl. Food Res. J. 22:2288– 2293. Dash, K.K., T.K. Goswami, and S.C. Deka. 2012. Modeling for the modified atmosphere packaging of sapota fruit. Agric. Eng. Int. CIGR J. 14:102–109. Dey, K., A. Ghosh, and N.B. Howmick. 2014. Physicochemical properties of sapota [Manilkara achras (Mill) Fosb.] fruits coated with corn starch. J. Crop Weed. 10:43–49. Dhua, R.S., S. Nath, and J. Kabir. 2006. Vacuum infiltration of calcium delayed sapota fruit ripening. p. 254–261. In: S.N. Ghosh, S.K. Mitra, B.C. Banik, M.A. Hasan, S.K. Sarkar, and

J.K. Hore (eds.), Proc. Natl Symp. on Production, Utilization and Export of Underutilized Fruits with Commercial Potentialities. Kalyani, Nadia, West Bengal, India. Divya, A.R., S. Jayashree, and B. Bhogi. 2014. Effect of storage methods on the nutritional quality of sapota candy. Asian J. Dairy Food Res. 33:104–108. Emerald, F.M.E., V.V. Sreenarayanan, and R. Parvathy. 2001. Physicochemical responses of sapota packed under modified atmosphere. Madras Agric. J. 88:271–273. Fawole, O.A. and U.L. Opara. 2013. Effects of storage temperature and duration on physiological responses of pomegranate fruit. Ind. Crops Prod. 47:300–309. Gadgile, D.P., R.B. Kakde, G.M. Rathod, and A.M. Chavan. 2010. Postharvest fungal diseases of some tropical fruits. Biosci. Disc. 1:7–10. Ganjyal, G.M., M.A. Hanna, and D.S.K. Devadattam. 2003. Processing of sapota (Sapodilla): drying. J. Food Sci. 68:517–520. Ganjyal, G.M., M.A. Hanna, and D.S.K. Devadattam. 2005. Processing of sapota (sapodilla): powdering. J. Food Technol. 3:326–330. Gontia, N.K. 2016. All India coordinated agronomic research project (ICAR). Postharvest engineering and technology. Annual Report 2015–2016. Hamza, S., S. Naseem, E. Bashir, G.H. Rizwani, and B. Hina. 2013. Trace element geochemistry of Manilkara zapota (L.) P. Royen, fruit from Winder, Balochistan, Pakistan in perspective of medical geology. Pakistan J. Pharmaceut. Sci. 26:805–811. Hawa, L.C. 2005. Study on weight loss and development of a respiration rate model of sapote fruit (Achras sapota L) under hypobaric storage. Jurnal Teknologi Pertanian. 2. Hiwale, S. 2015. Sustainable horticulture in semiarid dry lands. Springer. Jangam, S.V., V.S. Joshi, A.S. Mujumdar, and B.N. Thorat. 2008. Studies on dehydration of sapota (Achras zapota). Drying Technol. 26:369–377. Ibrahim, A.A. 2007. Predicting and monitoring the internal quality of stored sapodilla. N.Z. J. Crop Hortic. Sci. 35:275–302. Isabelle, M., B.L. Lee, M.T. Lim, W. Koh, D. Hwang, and C.N. Ong. 2010. Antioxidant activity and profiles of common fruit in Singapore. Food Chem. 123:77–84. Kader, A.A. 2009. Sapotes (sapodilla and mamey sapote): recommendations for maintaining postharvest quality. Department of Plant Sciences, University of California, Davis, CA. Available: http://postharvest.ucdavis.edu/Produce/ProduceFacts/Fruit/sapotes.shtml (accessed 15 January 2010). Karyadi, J.N.W., S. Rahayoe, and A.A. Oka. 2012. Applications electronic nose to determine

the maturity of sapodilla using pattern recognition system. In: Proc. Second Asia Pacific Symp. on Postharvest Research, Education and Extension. Acta Hortic. 1011:87–93. Kenghe1, R., F. Nilesh, and K. Kalyani. 2015. Design, development and performance evaluation of an onfarm evaporative cooler. Int. J. Sci. Tech. Soc. 2:1–5. Kostman, T., N. Tarlyn, F. Loewus, and V. Franceschi. 2001. Biosynthesis of Lascorbic acid and conversion of carbons 1 and 2 of Lascorbic acid to oxalic acid occurs within individual calcium oxalate crystal idioblasts. Plant Physiol. 125:634–640. Kulkarni, A.P., R.S. Policegoudra, and S.M. Aradhya. 2007. Chemical composition and antioxidant activity of sapota (Achras sapota Linn.) fruit. J. Food Biochem. 31:399–414. Kumar, A., A.S. Chauhana, R. Ravib, and V.B. Kudachikara. 2016. Influence of pretreatments and storage condition on the quality of sapota (Achras zapota) fruit and on its processed product marmalade. Intl. J. Curr. Sci. 19:133–145. Kunyamee, S., S. Ketsa, W. Imsabai, and W.G. van Doorn. 2009. The transcript abundance of an expansin gene in ripe sapodilla (Manilkara zapota) fruit is negatively regulated by ethylene. Funct. Plant Biol. 12:1205–1211. Kunyamee, S., S. Ketsa and W.G. van Doorn. 2010. Gene expression of cellwall degrading enzymes in sapodilla (Manilkara zapota) fruit. ScienceAsia 36:18–25. Kute, L.S. and M.B. Shete. 1995. Sapota (sapodilla). p. 475–484. In: D.K. Salunkhe and S.S. Kadam (eds.), Handbook of fruit science and technology production, composition, storage and processing. Marcel Dekker, Inc. New York. Lakshminarayana, S. and H. Subramaniyam. 1966. Physical, chemical and physiological changes in sapota fruit (Achras sapota L.) during development and ripening. J. Food Sci. Technol. (Mysore). 3:151–154. Laohakunjit, N., O. Kerdchoechuen, F.B. Matta, J.L. Silva, and W.E. Holmes. 2007. Postharvest survey of volatile compounds in five tropical fruits using headspacesolid phase microextraction (HSSPME). HortScience 42:309–314. Lim, T.K. 2013. Manilkara zapota. Volume 6, Edible medicinal and nonmedicinal plants, Springer. p.110–118. Lima, A.N., J.L. Batista, N.P. Costa, W.E. Pereira, C.H. Brito, A.B. Silva, and J.V.M.L. Marinus. 2006. Influence of thermal treatments in the control of the larval phase of Ceratitis capitata Wied. (Diptera: Tephritidae) in fruits of sapota (Achras sapota L.). Ecossistema 31:89–94. Love, K., N. Chen, and R. Paull. 2014. Quick harvest and postharvest tips for better quality and longer postharvest life. Fruit, Nut, Bev. Crops 36:1–13. Ma, J., X.D. Luo, P. Protiva, H. Yang, C. Ma, M.J. Basile, I.B. Weinstein, and E.J. Kennelly.

2003. Bioactive novel polyphenols from the fruit of Manilkara zapota (Sapodilla). J. Nat. Prod. 66:983–986. Madhav, N.V.S. and S. Khurana. 2011. A novel, potent, bioemulsifier from the fruit of Manilkara zapota for formulating escitalopram microemulsions. Intl. J. Pharmtech. Res. 3:929–933. Malhotra, S., K. Mondal, V. Jain, and R. Singh. 2009. Oxidative stress and antioxidant systems in guava (Psidium guajava L.) fruits during ripening. Physiol. Mol. Biol. Plants 15:327–334. Mangaraj, S. and T.K. Goswami. 2009. Modified atmosphere packaging of fruits and vegetables for extending shelflife – a review. Fresh Produce 3:1–31. MartínezRomero, D., N. Alburquerque, J.M. Valverde, F. Guillén, S. Castillo, D. Valero, and M. Serrano. 2006. Postharvest sweet cherry quality and safety maintenance by aloe vera treatment: a new edible coating. Postharvest Biol. Technol. 1:93–100. Mehta, P., T. Suma, and T. Prasad. 1980. Effect of ethrel treatment on the postharvest changes in the turnover of ascorbic acid and reducing sugar in Achras sapota fruits. J. Biosci. 2:305– 310. Menezes, J. and K.A. Athmaselvi. 2016. Study on effect of pectin based edible coating on the shelf life of sapota fruits. Biosci. Biotechnol. Res. Asia 13:1195–1199. Mossler, M.A. and J. Crane. 2002. Florida crop/pest management profile: mamey sapote and sapodilla. CIR 1414, Institute of Food and Agricultural Sciences, University of Florida. Mund, M.D., S. Alam, U.H. Khan, U. Tahir, M.S. Zubair, T. Younas, and B.E. Mustafa. 2016. Phytochemicals as complementary and alternative therapeutic formulations with potential pro apoptotic effects on various cancerous cell lines: a literature survey. Focus on Sci. 2: DOI: 10.20286/focsci020220. Mohamed, S., B. Taufik, and M.N.A. Karim. 1996. Effects of modified atmosphere packaging on the physicochemical characteristics of chiku (Achras sapota) at various storage temperatures. J. Sci. Food Agric. 70:231–240. MooHuchin, V.M., R. EstradaLeón, I.A. EstradaMota, L.F. CuevasGlory, and E. SauriDuch. 2013. Ripening of sapodilla fruits [ Manilkara zapota (L.) P. Royen] treated with 1methylcyclopropene after refrigeration. Afr. J. Plant Sci. 7:561–570. Morais, P.L.D., M.R.A. Miranda, L.C.O. Lima, J.D. Alves, R.E. Alves, and J.D. Silva. 2008. Cell biochemistry of sapodilla (Manilkara sapota) submitted to 1methylcyclopropene. Brazilian J. Plant Physiol. 20:85–94. Morrison, D., O. Ellis, and E. Nelson. 2011. Effect of beeswax and modified atmosphere packaging on the storage life of sapodilla (Manilkara zapota) in refrigerated and ambient conditions. Acta Hortic. 894:201–209.

Oliveira, A., C. Moura, E. Gomes Filho, R. Ness, and M. Miranda. 2010. Postharvest quality and chlorophyll content in sapodilla cv. Sapota Tropical grown under organic and conventional systems (abstr.). Intl. Hortic. Congress. Lisboa, Portugal. Padmaja, N., S.J.D. Bosco, and J.S. Rao. 2015. Physicochemical analysis of sapota (Manilkara zapota) coated by edible aloe vera gel. Int. J. Appl. Sci. Biotechnol. 3:20–25. Padmini, T., G.S.V. Raghavan, and T. Thangaraj. 2005. Dehydration of sapota flakes. Proc. 4th AsiaPacific Drying Conference; Kolkata, India, 13–15 December, 2:1161–1165. Pahel, P., D. Singh, S.S. Kundu, and M.K. Rana. 2015. Effect of various packaging materials and methods on physiological parameters and organoleptic rating of sapota. Environ. Ecol. 33:1815–1820. Parle, M. and Preeti. 2015. Chickoo: a wonderful gift from nature. Int. J. Res. Ayurveda Pharm. 6:544–550. Patel, R.G., B.R. Parmar, and S.J. Patil. 2010. A study on biochemical changes of sapota varieties in cold storage with application of GA3. Asian J. Hortic. 5:127–130. Patil, B.N., S.V. Gupta, and V. Wankhade. 2014. Response surface methodology of osmotic dehydration for sapota slices. Int. J. Agr. Food Sci. Technol. 5:249–260. Pawar, C.D., A.A. Patil, and G.D. Joshi. 2012. Quality of sapota wine from fruits of differential maturity. Karnataka J. Agric. Sci. 24:501–505. Peiris, K. 2007. Sapodilla Manilkara zapota L. p.183–224. In: D.K.N.G. Pushpakumara, H.P.M. Gunasena, and V.P. Singh (eds), Underutilized fruit trees in Sri Lanka. van Royen Publisher: World Agroforestry Centre, South Asia Office, New Delhi, India. Philippine National Standard, National Bureau of Agricultural Commodity and Food Standards, Ministry of Agriculture and Cooperatives Fresh Fruit – Chico (Sapodilla) – Classification and Grading. 2011. ICS 67.080.10. Pino, J.A., R. Marbot, and J. Aguero. 2003. Volatile components of sapodilla fruit (Manilkara achras L.). J. Essential Oil Res. 15:374–375. Polania Trujillo, H. 1986. Recomendación para el cultivo del níspero. ESSO Agrícola 33:16– 18. Polara, N. 2013. Effect of precooling on post harvest life of sapota [ Manilkara achras (Mill) Fosberg] cv. Kalipatti. Grin Velag, India. Poonsawat, P., H. Nimitkeatkai, A. Uthairatanakij, S. Kanlayanarat, and C. WongsAree. 2007. CaCl2 affecting quality and aroma volatiles of freshcut ‘Malay’ sapodilla. Acta Hort. 746:309–315. Purba, I.J.B. 2011. Shelf life and quality of edible coated fresh cut sapota (Achras zapota, L)

cultivar Sukatali ST1 stored in modified atmosphere packaging. Di bawah bimbingan Prof. Dr Ir. Hadi K. Purwadaria, M. Sc. Thesis. Faculty of Agricultural Technology Bogor Agricultural University, Indonesia. Qiuping, Z., X. Wenshui, and Y. Jiang. 2006.Effects of 1methylcyclopropene treatments on ripening and quality of harvested sapodilla fruit. Food Technol. Biotechnol. 44:535–539. Raharjo, B., T. Tranggono, and J.S. Dethan. 1998. Textural behavior of ripening sapota fruits. Indonesian Food Nutr. Prog. 5:43–48. Rao, D.V.R. and B.S. Chundawat.1988. Ripening changes in sapota cv. Kalipatti at ambient temperature. Indian J. Plant Physiol. 31:205–208. Rastegar, S. 2015. Physical, biochemical and mineral evaluation of sapota fruits during growth, development and ripening. Agric. Commun. 3:14–19. Relekar, P.P., A.G. Naik, and B.V. Padhiar. 2016. Studies on standardization of recipe for sapota nectar and its storage. Indian J. Hortic. 73:149–152. Rodrigues, S., M.C. Gomes, M.I. Gallão, and F.A. Fernandes. 2009. Effect of ultrasound assisted osmotic dehydration on cell structure of sapotas. J. Sci. Food Agric. 89:665–670. Rosenthal, A., D. Ledward, A. Defaye, S. Gilmour, and L. Trinca. 2002. Effect of pressure, temperature, time and storage on peroxidase and polyphenol oxidase from pineapple. Prog. Biotechnol. 19:525–532. Salleh, R.M., T.L. Ying, and L. Mousavi. 2016. Development of fruit bar using sapodilla (Manilkara zapota L.). J. Food Process. Preserv. DOI: 10.1111/jfpp.12806. Saxena, M. 2014. Handbook on horticulture statistics. Government of India Ministry of Agriculture Department of Agriculture and Cooperation: New Delhi. Selvaraj, Y. and D.K. Pal.1984. Changes in chemical composition and enzyme activity of the sapodilla (Manilkara zapota L.) cultivars during development and ripening. J. Hortic. Sci. 59:275–281. Shui, G., S.P. Wong, and L.P. Leong. 2004. Characterization of antioxidants and change of antioxidant levels during storage of Manilkara zapota L. J. Agr. Food Chem. 52:7834–7841. Snowdon, A.L. 2010. A colour atlas of postharvest diseases and disorders of fruits and vegetables. Volume 1: General introduction and fruits. Manson Publishing Ltd, London. Srinivas, B.A., J. Dilip Babu, and D. Vishnuvardhan Reddy. 2012. Synergic effect of MAP and irradiation on storability of Achras sapota cv. Kalapati at low temperatures. Abstracts of 7th Intl. Postharvest Symp. (IPS2012). 25–29 June, 2012. Putra World Trade Centre (PWTC), Kuala Lumpur, Malaysia. Srinu, B. 2014. Studies on modified atmospheric packaging and irradiation on shelf life and

quality of sapota [Manilkara achras (Mill) Fosberg] cv. Kalipatti. p. 1420. In: 2nd Intl. Conf. on Agricultural and Horticultural Sciences, Hyderabad, India. Srinu, B., A. Bhagwan, and J.D. Babu. 2014. Studies on modified atmospheric packaging on shelf life and quality of sapota (Manilkara zapota L.) cv. Kalipatti. Plant Archives 2:757–761. Srinu, B., A. Bhagwan, and J.D. Babu. 2015. Effect of combination of polypropylene packaging and irradiation on shelf life of sapota fruit stored at low temperature. Environ. Ecol. 33:1672–1675. Srivastava, A.K. 2014. Aesa based IPM package sapota. National Institute of Plant Health Management, India. Srivastava, M., M. Hegde, K.K. Chiruvella, J. Koroth, S. Bhattacharya, B. Choudhary, and S.C. Ragavan. 2014. Sapodilla plum (Achras sapota) induces apoptosis in cancer cell L lines and inhibits tumor progression in mice. Sci. Reports.4:6147.82. doi: 10.1038/srep06147. Sudha, R., R. Amutha, S. Muthulaksmi, W. Baby Rani, K. Indira, and P. Mareeswari. 2007. Influence of pre and postharvest chemical treatments on physical characteristics of sapota (Achras sapota L.) var. PKM1. Res. J. Agric. Biol. Sci. 3:450–452. Suhasini, J., V.C. Kanamadi, G.S.K. Swamy, A.M. Shirol, and M. Chavan. 2012. Performance of sapota (Achras sapota L.) varieties and hybrids under Ghataprabha command area. Karnataka J. Agric. Sci. 25:548–551. Sumathi, M. and S. Shivashankar. 2017. Metabolic profiling of sapota fruit cv. Cricket Ball grown under foliar nutrition, irrigation and water deficit stress. Scientia Hortic. 215:1–8. Tandel, Y.N. and N.L. Patel. 2011. Alleviation of SAP injury on sapota fruit through calcium hydroxide. Asian J. Hortic. 6:525–527. Thai Agricultural Standard, National Bureau of Agricultural Commodity and Food Standards. 2011. Ministry of Agriculture and Cooperatives, Sapodilla. ICS 67.080.10. Thompson, A.K. 2014. Specific recommendations for herbs, in fruit and vegetables: harvesting, handling and storage, John Wiley & Sons, Ltd, Chichester, UK. doi: 10.1002/9781118653975.ch17. TomásBarberán, A. and J. Espin. 2001. Phenolic compounds and related enzymes as determinants of quality in fruits and vegetables. J. Sci. Food Agric. 81:853–876. Tsomu, T. and H.C. Patel. 2014. Effect of postharvest treatments of chemical and plant growth regulators on physical parameters of sapota fruit cv. Kalipatti. J. Food Process. Technol. 5:347. Tsomu, T., H.C. Patel, R.M. Thakkar, M.A. Jang, and R.P. Vasara. 2015. Response of post harvest treatments of chemical and plant growth regulators on biochemical characteristics of sapota fruit cv. Kalipatti. The Bioscan 10:33–36.

Uekane, T.M., L. Nicolotti, A. Griglione, H.R. Bizzo, P. Rubiolo, C. Bicchi, M.H.M. Rocha Leao, and C.M. Rezende. 2017. Studies on the volatile fraction composition of three native AmazonianBrazilian fruits: murici ( Byrsonima crassifolia L., Malpighiaceae), bacuri (Platonia insignis M., Clusiaceae), and sapodilla (Manilkara sapota L., Sapotaceae). Food Chem. 219:13–22. U.S. Department of Agriculture, Agricultural Research Service (USDA). 2012. USDA National Nutrient Database for Standard Reference, Release 25. Nutrient Data Laboratory Home Page, http://www.ars.usda.gov/ba/bhnrc/ndl. Ukey, P.D. and P.A. Unde. 2010. Design and development of sapota fruit grader. Int. J. Agric. Eng. 3:35–39. Varshney, A.C., S. Mangaraj, and S.L. Birla. 2005. Grading, packaging and marketing of sapota in south Gujarat. Agric. Eng. Today 29:40–47. Vijayalakshmi, D., U. Bangarusamy, and R.M. Vijayakumar. 2004. Postharvest treatments to extend the shelf life of sapota fruit cultivars Co 1 and Co 2. Crop Res. (Hisar) 27:219–226. Vishwasrao, C. and L. Ananthanarayan. 2016a. Postharvest shelflife extension of pink guavas (Psidium guajava L.) using HPMCbased edible surface coatings. J. Food Sci. Technol. 4:1966–1974. Vishwasrao, C. and L. Ananthanarayan. 2016b. Delayed postharvest ripeningassociated changes in Manilkara zapota L. var. Kalipatti with composite edible coating. J. Sci. Food Agric. doi:10.1002/jsfa.7758. Wagh, P.M. and U.N. Bhale. 2012a. Changes in phenolic contents of sapota pulp (Achras sapota) due to different isolates of Aspergillus niger. Biosci. Disc. 3:263–265. Wagh, P.M. and U.N. Bhale. 2012b. Efficacy of chemical control against Geotrichum candidum Link ex Fries on postharvest Manilkara achras Mill. Sci. Res. Rep. 2:91–93. Wang, C.Y. 2004. Alleviation of chilling injury in tropical and subtropical fruits. In: III Intl. Symp. on Tropical and Subtropical Fruits. Acta Hortic. 864:267–273. Wang, H., T. Liu, L. Song, and D. Huang. 2012. Profiles and αamylase inhibition activity of proanthocyanidins in unripe Manilkara zapota (Chiku). J. Agr. Food Chem. 60:3098–3104. Wills, R.B.H. and J.B. Golding. 2015. Reduction of energy usage in postharvest horticulture through management of ethylene. J. Sci. Food Agric. 95:1379–1384. Wills, R.B.H., D.R. Harris, L.J. Spohr, and J.B. Golding. 2014. Reduction of energy usage during storage and transport of bananas by management of exogenous ethylene levels. Postharvest Biol. Technol. 89:7–10. Yahia, E.M. 1998. Modified/controlled atmospheres for tropical fruits. Hortic. Rev. 22:123–

183. Yahia, E.M. 2004. Sapodilla and related fruits. U.S. Department of Agriculture, Handbook 66, (http://www.ba.ars.usda.gov/hb66/index.html). Yahia, E.M. and F. GutierrezOrozco. 2011. Sapodilla ( Manilkara achras (Mill) Fosb., syn Achras sapota L.). p.351–362. In: Postharvest biology and technology of tropical and subtropical fruits. Mangosteen to white sapote. Woodhead Publishing Ltd, Cambridge, UK.

5 The Citron (Citrus medica L.) in China David Karp University of California, Riverside, CA, USA Xulan Hu Yunnan Provincial Department of Agriculture, Kunming, Yunnan, China

ABSTRACT Citron (Citrus medica L.) is the type species of the genus Citrus, one of the primary species of cultivated citrus, and a parent or ancestor of the commercially important acid citrus fruits. Southwestern China is a crucial center of origin and diversity for citron. The diversity of this species in China is extensive, but poorly documented until recently. Prized for its fragrance, citron played an important role in Chinese art and culture. Today, citron is grown in most of the warmer citrusproducing areas of China, and is used in Chinese traditional medicine, ornamental pot culture, and for human consumption as both fresh and processed products. In Chinese traditional medicine, dried citron is used as a tonic, to regulate Qi, the life force. Citrons are classified by fruit shape as either common (non fingered) or fingered. In most of China, fingered citrons predominate because they are more suited to cultivation for medicinal and ornamental uses. Typically, common citrons grown in China have locules in which juice vesicles are absent, or are very scanty and rudimentary. ‘Ning’er Giant’, typically weighing 3–5 kg, sometimes reaching 8–10 kg, may be the world’s largest citrus cultivar. Fingered citron trees vary in stature, cold hardiness, flower color, and other significant horticultural characteristics; the fruits vary in size, shape, number and thickness of fingers, proportion of fruits that are open or closed, and the point on the fruit at which the carpels become distinct. The distinctive forms of many named cultivars reflect different genotypes. There also exist in China genotypes that are intermediate in morphology between common and fingered citron. Putatively “wild” citron trees producing small fruits grow in forests, in semiwild margins between natural and cultivated areas, and in private gardens in the subtropical forests of western and southern Yunnan. Scientists in China and elsewhere have paid increasing attention to citron germplasm resources during recent years. KEYWORDS: citron, Citrus medica L., China; citrus taxonomy, medicinal plants, traditional medicine I. INTRODUCTION II. HISTORY AND CULTURE III. NOMENCLATURE

IV. CURRENT CITRON CULTIVATION IN CHINA A. Propagation 1. Cuttings 2. Grafting 3. Seed 4. Tissue culture B. Diseases, Pests, and Freeze Damage C. Medicinal Uses and Properties D. Ornamental Cultivation E. Culinary Uses and Processing V. MAJOR CULTIVARS OF CHINESE CITRON AND SELECT CITRON HYBRIDS A. Common (NonFingered) Citron Cultivars 1. ‘Dog Head’ 2. ‘Jianshui Round’ 3. ‘Jinggu’ 4. ‘Jinghong Water’ 5. ‘Ning’er Giant’ 6. ‘Persistent Stigma’ 7. ‘Pumpkin’ 8. ‘Weishan Bullet’ 9. ‘Weishan Sour’ 10. ‘Weishan Sweet’ B. Fingered Citron Cultivars 1. ‘Aihua’ 2. ‘Chuan’ 3. ‘Guang’ 4. ‘Octopus’ 5. ‘Qingpi’ 6. ‘Yun’ 7. ‘Zhaocai’

C. Types Intermediate Between Common and Fingered Citron 1. ‘Half and Half’ 2. ‘Muli’ D. “Wild” Citrons 1. Wild Citron, Type 1 2. Wild Citron, Type 2 E. Citron Hybrids 1. Goucheng (Yunnanese) 2. ‘Yunmao Oval’ VI. GERMPLASM STATUS; REGIONAL AND GLOBAL PERSPECTIVE LITERATURE CITED

I. INTRODUCTION The diversity of citron (Citrus medica L.) in China is extensive. Citron trees grow wild in the forests, and dozens of cultivars, varying greatly in size and including round, oblong, oblate, fingered, and intermediate shapes, exist either in the wild or in cultivation. China recently surpassed Brazil to become the world’s largest producer of citrus fruits (FAO 2016 ). China is a primary center of origin and diversity for many species and types of citrus, including mandarin (Citrus reticulata), kumquat (Fortunella spp.), and trifoliate orange (Poncirus trifoliata) (Swingle and Reece 1967 ; Gmitter and Hu 1990 ). Although its citron cultivation and resources are little known in the rest of the world, China is a crucial center of diversity for this fruit, and very likely the world’s largest producer. Citron cultivation and studies are paltry compared to those of more commercial citrus such as oranges, mandarins and pummelos, but more scientific publications on citron appear in China than in the rest of the world combined. Books on citron cultivation have been published in China, but these are chiefly practical horticultural treatises (Xu and Deng 2002 ; Shi and Shao 2008 ). In order to better understand Chinese citron germplasm resources and their uses, the authors began a project in 2008 to visit the country’s major citrongrowing areas, with a particular focus on the southwestern province of Yunnan. The resulting observations, integrated with a survey of the published literature, are presented in this chapter.

II. HISTORY AND CULTURE The earliest references to citron appear in the Vajasaneyi, a collection of sacred Hindu texts written in India around the 8th century B.C., more than a millennium before the first description

in China (Simoons 1991 ). De Candolle ( 1886 ) and Hodgson ( 1967 ) both believed that citron originated in northeastern India and adjacent regions, because there it grows wild and natural hybrids showing citron characteristics are abundant. Bonavia ( 1888 ) was not sure whether citron originated in India and spread to China, or vice versa. Since the first Chinese reference to citrons appeared in the 4th century A.D., long after references to other forms of citrus such as pummelo and mandarin, Tolkowsky ( 1938 ) thought that citron was not native to China, but a foreign importation. Gmitter and Hu ( 1990 ) described some of the diversity of citrus, and particularly citron, in Yunnan; they maintained that this region played a more important role than previously supposed in the origin and distribution of Citrus species. Guo ( 1993 ) proposed, on the basis of geographical distribution, that western and southwestern Yunnan might be considered to be the center of origin in China, with wild citron from Cangyuan Va Autonomous County in western Yunnan being the progenitor species from which cultivated citron and lemon evolved. The earliest mention of citron in China is in the flora of Chi Han (304 A.D.), who called it “kouyuan ,” the name by which it was known until the Song Dynasty. Chi Han describes how the fruit, which had a nonfingered shape familiar from most citrus fruits, was highly fragrant, and prized by foreigners. In the previous century, he wrote, western countries sent 10 jars of citron to the emperor as a tribute (Simoons 1991 ). In the early 10th century Liu Xun, a T’ang official, elaborated in his “Register of Strange Things Beyond the Ranges” on how citrons were prized in southern China: “Its skin resembles that of an orange, but is the color of gold, so that men make much of it, and they love its fragrant aroma. The elite and nobility close to the royal seat in the capital display them in their households on platters and mats, attracted to them as rare fruits from a distant quarter. The flesh is thick and white, like that of a radish. Female artisans in the south vie at taking its flesh to carve and chase into flowers and birds, which they steep in bee’s honey and touch up with rouge. These hold a unique place for their wonderful artistry…” (Schafer 1967 ). In the fingered citron the carpels grow separately, starting at the end opposite the stem (i.e., the distal end), so that the fruit looks like a fist or a hand. This type is often assumed to have been brought to China from India by Buddhist monks, and has long been called “foshou,” meaning “Buddha’s Hand”, but there is no clear evidence that this happened, or when. It was already important in the 10th century in the Min Empire, in modernday Fujian Province. Chinese artists depicted fingered citrons in jade (Plate 5.1) and ivory carvings, bronze teapots, and lacquered wood panels (Tolkowsky 1938 ; Simoons 1991 ).

Plate 5.1. Sculpture of fingered citron carved in carnelian in 18th century China. (Photo credit: Metropolitan Museum of Art, New York.)

Paintings and drawings from the Ming and Qing eras often depict household scenes featuring bowls holding one to a dozen citrons, common or fingered, sitting on a table, by themselves or with different citrus or other fruit. In various images, lovers disport themselves, scholars and bureaucrats read, and wealthy homeowners repose, next to a bowl of citrons. For Chinese artists of erotic images, the fingered citron was emblematic of female sexuality. In the Qing era, fingered citron was the fruit most frequently carved in bamboo by Chinese artists (Rawson et al. 2015 ). In China, the citron was grown from the Yangtze Valley southward, and the fingered form was the most common. In northern regions citrons had to be brought at considerable expense from the south, and thus were affordable only to the wealthy. The Chinese considered the citron to be a symbol of happiness, wealth, and longevity. They prized it for New Year’s gifts, and placed it as an offering at the shrines of household gods in their homes, or on altars at their temples, and as an alternative to incense, to perfume the air (Tolkowsky 1938 ; Simoons 1991 ; Xu and Deng 2002 ).

III. NOMENCLATURE Nonfingered citrons, with an oblong to spherical shape familiar from many other citrus types, are called “common citrons” in this review. The use of this term to distinguish nonfingered from fingered citrons dates back at least to Firminger ( 1874 ) and was endorsed by Bailey ( 1919 ). The name for the nonfingered citron in Chinese is “ xiāng yuán” ( in Chinese ideograms). The Chinese name for fingered citron is “fo&c.acute; shoˇu” ( in Chinese characters), meaning “Buddha’s Hand.” This is the name by which it is known, in closely related or cognate versions, in many languages around the world. The botanical name is Citrus medica var. sarcodactylis (Hoola van Nooten) Swingle. (This name came from the Greek words “sarkos” meaning “fleshy” and “dactylos” meaning “finger.”) Curiously, Buddha’s Hand or fingered citrons are frequently called “bergamot” in Chinese scientific literature in English (Zhou et al. 2005b ; Lu and Wu 2006 ; Mei et al. 2006 ; Zhang and Xu 2007 ). The true bergamot (C. bergamia Risso) is a very different acid citrus fruit, originating in Italy, that most closely resembles sour orange. It is unclear when, how, or why this erroneous terminology came to be employed; a similar misnaming occurs in Korea, where yuzu (C. junos) is often erroneously called “citron” in English translation. Fortunately, this misuse of “bergamot” appears to be diminishing and in many recent Chinese publications on citron the proper term, “fingered citron,” is used (Guo et al. 2009a , b , c , d ; Chen et al. 2010 ; Cao et al. 2011 ; Liao et al. 2013 ). Several Chinese citron cultivars, notably those commonly cultivated in Zhejiang Province,

have been described in the scientific literature, but no comprehensive description or taxonomy of Chinese citron cultivars has yet been published. As a result, in order to refer to particular types, it has been found necessary in this chapter (as in a previous paper; Ramadugu et al. 2015 ) to identify putatively distinctive genotypes of citron by choosing names, usually based on morphological and/or geographical characteristics, for example, ‘Ning’er Giant,’ ‘Yunmao Oval,’ ‘Octopus,’ and ‘Weishan Bullet.’

IV. CURRENT CITRON CULTIVATION IN CHINA Citron is grown in most of the warmer citrusproducing areas of China, from Yunnan and Sichuan provinces in the west to Zhejiang and Jiangsu provinces in the east and to Guangdong Province and Guangxi Zhuang Autonomous Region in the south. Citron is also grown in Chongqing Municipality, Hunan, Hubei, and Henan provinces (Deng 2008 ); in southwest Guizhou Province, the lower elevations of eastern Tibet Autonomous Region (Xizang), and Hainan Province (Lim 2012 ); and in Anhui Province (Anon. 2017 ) (Fig. 5.1a and b).

Fig. 5.1. (a) Map of Citron cultivation in China. (b) Map of Citron cultivation in Yunnan and Southern Sichuan. (Credit: Seth Karp.)

The main uses of citron are for processing for Chinese traditional medicine, for ornamental pot culture, and for human consumption of fresh or processed fruits. The fruits are also sold for ornamental and decorative purposes. In some areas, citron is used as a rootstock for other citrus, such as mandarin, sweet orange, and citron itself, or as a hedgerow (author’s personal observations, 2008–2017). Other uses include extraction of essential oils, and additives for the flavoring of cigarettes (Li. et al. 2000 ; Liang et al. 2006a ; Guo et al. 2009a ; Lü et al. 2011 ). Citron is grown mostly on a small scale, often just a few trees in a rural yard. There are some larger plantings, such as a 133 ha planting of common citron in Yunnan, and substantial facilities for producing potted fingered citron plants in Jinhua, a prefecturelevel city in central Zhejiang. Because citron is a relatively minor crop in China, reliable statistics regarding cultivation and production are not available. It has been estimated that 2000 ha of citron were cultivated in China in 2008 (Xiuxin Deng, personal communication, 2008).

A. Propagation Citron trees are propagated in China using four different methods: 1. Cuttings. 

Both hardwood and softwood cuttings are widely used for both common and fingered citrons. This method is easy and quick, but has the disadvantage that cold hardiness and drought resistance are not as good as for grafted trees. Citron cuttings, which are easy to root, are usually made in spring, and sometimes in summer or fall. Usually, each cutting is 20–30 cm long, with a few buds. The rooting media used are sand, peat moss, or even common soil. Sometimes rooting hormones such as indole acetic acid (IAA), indole3butyric acid (IBA), or 1naphthaleneacetic acid (NAA) are applied, while in other cases no root induction agent is used. 2. Grafting.  In some areas grafting is used for both common and fingered citrons. The major rootstocks are trifoliate (Poncirus trifoliata (L.) Raf.) and local citron seedlings. Citron seedlings are preferred because they result in a uniform graft union and vigorous growth. Trifoliate rootstock is also used, although the graft union is typically uneven, with the rootstock larger than the scion. In Jinhua, Huyou (C. × changshanhuyou Y.B. Chang, a natural hybrid of C. sinensis, C. grandis and possibly other species of Citrus) is used as an interstock for bonsai fingered citron (Shi and Shao 2008 ). Grafted trees live longer and perform better than trees grown from cuttings. 3. Seed.  Citron is sometimes grown directly from seed (Liu et al. 2014 ). 4. Tissue culture.  Researchers have developed methods to use tissue culture to propagate citron trees. Tissue culture has the potential to multiply citron plants by 1 to 3 million a year (Zhou et al. 2005a ; Zhang et al. 2009a ).

B. Diseases, Pests, and Freeze Damage Huanglongbing (HLB, citrus greening), a highly destructive bacterial disease associated with the fastidious Gramnegative, phloemlimited, αProteobacteria , Candidatus Liberibacter sp., does not spare citron (He et al. 2005 ; Deng et al. 2008 ). In some parts of China where citron is grown, HLB has not yet arrived, but when it does spread it often becomes a serious problem. A survey in Guangdong, where HLB is endemic, found that fingered citron was the most infected of 16 citrus cultivars, with all samples diseased (Deng et al. 2012 ). In contrast, Ramadugu et al. ( 2016 ) found that citron, along with its progeny, lemons and limes, performed better than other Citrus types in withstanding decline from HLB infection. In contrast, researchers have observed that citron has never demonstrated canker disease (Xanthomonas axonopodis pv. citri Hasse) in the field, and inoculated plants of ‘Chinese’ citron did not develop lesions (Deng et al. 2010 ). Other significant diseases of citron in China include anthracnose (Colletotrichum gloeosporioides), citrus scab (Elsinoë fawcettii Bitanc. & Jenkins), and green mold

(Penicillium digitatum [Pers.:Fr.] Sacc.) (Li et al. 2011 ). Major insect pests include citrus red mite (Panonchus citri McGregor), citrus leafminer (Phyllocnistis citrella Stainton), chaff scale (Parlatoria pergandii Comstock), spirea aphid (Aphis citricola van der Goot), and orange spiny whitefly (Aleurocanthus spiniferus Quaintance). Premature leaf abscission and sunburn are other common problems. Since citron is among the most coldtender of citrus crops, damage from freezing weather is a significant problem in districts with cool winters. In Jinhua (Zhejiang Province), where temperatures in winter can fall as low as –3 °C to –5 °C, many producers of bonsai fingered citron trees use protected greenhouses and plastic tunnels, and spray water on the trees during freezes. Researchers there have evaluated the available fingered citron germplasm for cold hardiness (Guo et al. 2009d ; Jiang et al. 2012a , b ), and elucidated the molecular mechanisms underlying cold regulation of plant defense and stress responses (Guo et al. 2009c ; Chen et al. 2010 ; Cao et al. 2011 ; Yang et al. 2012 ).

C. Medicinal Uses and Properties The primary use for which citron is grown in China is as a medicinal plant. In Chinese traditional medicine, dried slices (Plate 5.2) and strips of both common and fingered citron are used as a tonic, to regulate Qi, the life force or energy flow (Zhang 2015 ). The fruit are often sliced thin and dried in the sun or on trays in ovens (Plate 5.3). As is the case for other citrus grown for medicinal use, citrons are typically harvested somewhat immature, when the rind is still greenish, because the medicinal properties are said to be more potent at this stage. Liang et al. ( 2006a ) found that the essential oil content of individual fruits reaches a maximum when they are a light yellow or golden color, but Wu et al. ( 2013 ) determined that the antioxidant activity of fingered citron essential oils does indeed decrease with maturity at harvest, declining by approximately 20% from the immature to the fully mature stage.

Plate 5.2. Dried ‘Chuan’ fingered citrons, for use in Chinese traditional medicine, in Xiaogu, Sichuan. (Photo credit: David Karp; 7 Nov. 2008.)

Plate 5.3. Drying fingered citron slices for Chinese traditional medicine near Huaning, Yunnan. (Photo credit: David Karp; 22 Oct. 2011.)

According to Jiao ( 2001 ), common citron (xiang yuan) is “acrid, sour, and bitter in flavor and warm in nature. [It] regulates Qi, loosens the chest, and transforms phlegm. Xiang yuan is suitable for liver Qi depression that causes ribside pain, stomach duct pain, fullness and oppression in the stomach duct and abdomen, belching, and vomiting. … [It also] enhances the appetite during the initial stage of pregnancy.” Similarly, Jiao ( 2001 ) characterises the fingered citron (foshou) as follows: “…rectifies Qi and harmonizes the center, and soothes the liver and resolves depression. [It] is suitable for liverstomach disharmony, Qi stagnation stomach pain, oppression in the chest, ribside distention, poor appetite, and vomiting.” Common citron is said to have a stronger phlegmtransforming effect than fingered citron, but the latter to be more effective in checking retching. Some practitioners consider common citron to be an inexpensive substitute for fingered citron (Jiao 2001 ). In addition to fruit products, the leaves of citron trees are used to treat cough and respiratory ailments (personal observation, 2008). Modern medical researchers are finding evidence that citron does offer beneficial health

properties. Fingered citron has been reported to contain various bioactive coumarins, flavonoids, tetranortriterpenoids, monoterpenoids, and acridone alkaloids (Chan et al. 2010 ; Li et al. 2013 ). A review of medical literature by Panara et al. ( 2012 ) found that citron possesses analgesic, hypoglycemic, anticholinesterase, anticancer, antidiabetic, hypocholesterolemic, hypolipidemic, insulin secretagogue, anthelmintic, antimicrobial antiulcer, and estrogenic properties. Among dozens of studies on the medical effects of citron, Guo et al. ( 2009a ) observed that the essential oil extracted from fingered citron fruits has a strong antibacterial effect, with good thermal stability and longlasting effectiveness. Peng et al. ( 2009 ) confirmed that fingered citron had an insulin secretagogue effect in rats, and stated that it would be very beneficial to patients with type 2 diabetes. Both Lü et al. ( 2011 ) and Shao et al. ( 2011 ) found that fingered citron essential oil had an antiproliferative effect on melanoma cells in vitro. Kim et al. (2013) confirmed that fingered citron essential oil possesses antiinflammatory properties while AlKalifawi ( 2015 ) found that essential oils of fingered citron could be used as a therapeutic agent for human microbial infections.

D. Ornamental Cultivation The second most important use of citron in China is as an ornamental dwarf, or bonsai, tree, to be sold in a pot or basket (Plate 5.4). The center of ornamental citron production is the city of Jinhua, in Zhejiang Province, which annually produces about 500 000 fingered citron trees in pots, with a value of approximately $US 5 million.

Plate 5.4. Bonsai fingered citron plants in pots for sale at a street market along a road in Jinhua, Zhejiang. (Photo credit: David Karp; 17 Oct. 2008.)

There, citron trees are cultivated both in greenhouses with removable plastic covers, to protect them against rain and winter cold, and in openair plantings. Fingered citron plants are raised in black plastic pots initially before transplanting into more decorative pots. The growing medium comprises perlite + peat + mushroom bran (1:1:1) with incorporated slowrelease fertilizers (Lu and Wu 2006 ). Nearinfrared reflectance spectroscopy can be used to diagnose nutritional deficiencies (Liao et al. 2012 ). Producers use trellises, stakes and ties to ensure that the limbs and fruit maintain an attractive upright growing habit (Plate 5.5).

Plate 5.5. In Jinhua, fingered citron growers use trellises, stakes, and ties to ensure that the limbs and fruit maintain an attractive upright growing habit. (Photo credit: David Karp; 17 Oct. 2008.)

Local growers and researchers have selected and developed several dwarfed varieties, such as ‘Aihua’ and ‘Qingpi,’ which are naturally suited to bonsai production. Fruits are often grafted onto small plants, giving them the look of an ornamental flowering plant. Jinhua growers also harvest about 500 tonnes of fingered citrons annually for fresh sales and for processing for use in Chinese traditional medicine.

E. Culinary Uses and Processing The authors have observed in citrongrowing areas of China that culinary uses and processing are the third most important use of citron in China overall, but quite important for common citron. The albedo from select common citron cultivars is sweet and juicy, and the Chinese often enjoy eating it fresh (Plate 5.6). The rind and albedo are also used in savory dishes, but recipes in Chinese cookbooks are not particularly common. Karp ( 1998 ) mentions three recipes using fingered citron: chicken rolls steamed in lotus leaf; stewed fruit and white fungus; and crispy mock Buddha’s Hand (with pork, cornstarch, and eggs).

Plate 5.6. Mr. PeiYong Zhang eats a slice of common citron at his home in Dehong Prefecture, Yunnan. (Photo credit: David Karp; 4 Nov. 2008.)

One common use is candied rind, which is similar to products made in Brazil, the Mediterranean area (primarily Italy, Greece, and Corsica), and Puerto Rico (Klein 2014 ). Those areas produce candied citron mostly for export, but virtually all of the Chinese production is consumed domestically. It seems possible that China might export this product in the future; equally, producers in other countries might well consider using elite Chinese common citron cultivars such as ‘Ning’er Giant’ for their large size and naturally sweet albedo. Citronprocessing factories visited by the authors are small and family owned, making from a few hundred kilograms to 20 or 30 tonnes of candied citron annually. Manufacturers are often very secretive about the particulars of their processing methods. Both common and fingered citron are used, but mostly common citron is processed. The harvest season for culinary citron is August to October. Fruit are harvested at full size, although the rind may be light green or yellow. The basic technique, similar to the traditional method used in the Mediterranean, is to scoop out the central portion (locules and seeds, and juice vesicles, if any), then to cut the rind and albedo into thin strips which are soaked in water to remove the bitterness and then boiled in sugar syrup. Tonghai, in Yunnan (elevation 1800 m), has a long history of making candied citron peel, which is a famous local product. Local factories use some 80–100 tonnes of citrons annually from nearby growing areas, particularly Huaning County. Other products made from citron in China include soft drinks, wine, preserves, juice concentrate, tea, and essential oils. For extracting juice from fingered citron, Jie ( 2007 ) found that steam distillation and supercritical carbon dioxide extraction were the best methods. In Jingjiang, in Jiangsu Province, citron trees have been planted on a large scale as street trees, because of their attractive shape, appearance, and aroma. Since 2007 these trees have produced roughly 2000 tonnes of fruit each year, most of which have been discarded. To make use of this fruit, local authorities have encouraged processors to make citron soft drinks.

V. MAJOR CULTIVARS OF CHINESE CITRON AND SELECT CITRON HYBRIDS Citrons are classified by fruit shape as either common (nonfingered) or fingered; there are a few rare types that are intermediate in character between these two groups. It is also useful to distinguish pure C. medica specimens from citron hybrids, using both morphological and genetic analyses. Wild and wildtype are smallfruited common citrons, and might include both pure and hybrid forms.

A. Common (NonFingered) Citron Cultivars The common (nonfingered) citron is the most commonly cultivated type of citron in many

areas of the world. However, in most of China fingered citrons predominate, because they are more suited to cultivation for medicinal and ornamental uses. In China, only in Yunnan and in adjacent western regions is common citron grown on a significant scale. The majority of the plantings are of small size, but accurate statistics of annual yield and planted area are not available. Nonetheless, it seems clear that both production and planted surfaces are substantially less for common citron than for fingered citron. There are two basic types of common citrons in China: 1. Common citrons with segments and juice vesicles, roughly similar to the citrons such as ‘Diamante’ that prevail in the Western world. (Smallfruited wildtype citrons, which do have juice vesicles, belong in this category morphologically, but are treated in a separate section below, as are citron hybrids.) 2. Common citrons with locules in which juice vesicles are absent, or are very scanty and rudimentary. China is notably one of the few citrongrowing countries in which citrons of the second category (no or very few juice vesicles) outnumber those of the first, both in number of cultivars and areas grown. (The ‘Temoni’ cultivar, traditionally cultivated in Yemen, and now grown in Israel for Jewish ritual, is the bestknown example of this type in the Western world. It would be tempting to suggest that ‘Temoni’ might have been derived somehow from Chinese citrons of similar type, but in Ramadugu et al. ( 2015 ) ‘Temoni’ was clustered with Western rather than Chinese citrons.) The reason for this prevalence in China of common citrons without juice vesicles probably relates to the primary economic importance of the albedo, which is consumed either fresh or cooked. To those unfamiliar with the cultivars used this might seem surprising, since the albedo of most citrus fruits is unattractively bitter, and citron rind used for candying in other countries undergoes lengthy processing to remove this bitterness. However, the albedo of at least some forms of common citron grown in China is nonbitter, moist but firm, and reasonably pleasant to eat fresh, almost like a bland apple. Common citron rind is also candied, and dried chips of the rind are used in Chinese traditional medicine. Most nonfingered citrons probably had juice vesicles like all other Citrus species, before the selective process of domestication. It would be speculative to surmise just how thick the albedo was on these ancestral citrons. However, because the albedo has long been the fruit part of primary importance, it is not surprising that farmers have over time selected cultivars in which a large part of the fruit is albedo, the segments or locules are small, and the juice vesicles are scanty or (most frequently) absent altogether. One can speculate that since the juice vesicles did not serve a useful purpose, it did not serve the interests of growers to cultivate varieties in which they were prominent. In Yunnan, citrons with juice vesicles are sometimes called “sour” citrons, while cultivars without juice vesicles are called “sweet.” These words have a different meaning in this context than in Western citrus terminology, in which “sweet” refers to the flavor of the pulp, not the rind. All of the Chinese cultivars with juice vesicles encountered by the authors have acidic

pulp, not acidless pulp like ‘Corsican.’ 1. ‘Dog Head’ [] (

Plate 5.7).

Plate 5.7. ‘Dog Head’ citron in Mojiang, Yunnan. (Photo credit: Xulan Hu; 3 Nov. 2010.)

Distribution: Mojiang and Ning’er counties and Simao District in southern Yunnan. Elevation: 1000–1800 m. Botanical characteristics: trees medium to large; young flush purpletinted; shoots green, leaves large; flower buds purpletinted; fruits large, typically 3–5 kg; fruit shape oblong, cylindrical; rind wrinkled, warty and ridged; albedo thick, sweet, and relatively juicy; c. 12 locules, juice vesicles absent; seeds numerous, monoembryonic. Uses: fresh; processing (candied citron rind); rootstock for fingered citrons; traditional Chinese medicine; religious and decorative purposes. Note: this is the major citron cultivar in Mojiang, a village near Pu’er that grows about 100 mu (7 ha) of citron, yielding about 40–50 tonnes of fruit annually.

2. ‘Jianshui Round’ (Plate 5.8).

Plate 5.8. ‘Jianshui Round’ citron, Bai Shi Yan village, Xizhuang township, Jianshui County, Yunnan. (Photo credit: David Karp; 23 Oct. 2011.)

Distribution: Jianshui and Shiping counties in southern Yunnan. Elevation: 1300–1500 m. Botanical characteristics: trees short or medium; young flushes purpletinted, shoots darkgreen, leaves large, flower buds purple to pink; fruits medium, typically 1.5–3 kg; fruit round (spherical); rind surface smooth; albedo thick, sweet, and juicy; 10–15 locules, juice vesicles absent; seeds numerous, monoembryonic. Uses: rootstock for fingered citrons; consumed fresh; traditional Chinese medicine; religious and decorative purposes. 3. ‘Jinggu’ (Plate 5.9).

Plate 5.9. ‘Jinggu’ citron, Jinggu Dai and Yi Autonomous County, Yunnan. (Photo credit: Xulan Hu; 6 Jan. 2017.)

Distribution: Jinggu and Zhenyuan counties in southern Yunnan. Elevation: 900–1300 m. Botanical characteristics: thorns long and thin; fruits small, 300–800 g, obovate to ellipsoid, rind smooth with prominent longitudinal ridges; albedo thick and sweet; locules 11–13; juice vesicles absent; seeds numerous, monoembryonic. Uses: consumed fresh; traditional Chinese medicine; trees are used as hedges to protect

from animals; leaves are used to cook meat and treat respiratory disorders. Note: according to local farmers, fruits floated down a river called Weiyuan Jiang from nearby uplands, and its seeds grew into trees along the riverbank. It has characteristics typical of wild citrons, such as long, thin thorns, small fruit size, and numerous seeds. This may in fact be a form of “Wild citron, Type 2” (see below) that has entered cultivation. 4. ‘Jinghong Water’ (Plate 5.10).

Plate 5.10. ‘Jinghong Water’ citron, Jinghong, Yunnan, with large central cavity and small locules with numerous seeds and scanty juice vesicles. (Photo credit: Xulan Hu; 1 Nov. 2010.)

Distribution: Jinghong and Mengla counties in southern Yunnan. Elevation: 400–1000 m. Botanical characteristics: tree short or medium in stature, foliage dense, leaves large, relatively pointed for a citron; fruits medium, typically 1.5–3 kg, oval to pearshaped; rind pale whitishyellow, surface rough, ridged; albedo thick, sweet, and juicy; 12–14 locules, which are small, while the central cavity is enlarged; juice vesicles almost absent, although some rudimentary juice vesicles are present; seeds numerous, monoembryonic.

Uses: rootstock for fingered citrons; consumed fresh; traditional Chinese medicine; religious and decorative purposes. Note: in microsatellite analysis performed by Ramadugu et al. ( 2015 ), this cultivar was the only one of 26 Chinese citron accession in which 0% heterozygosity was observed. 5. ‘Ning’er Giant’ (Plate 5.11).

Plate 5.11. ‘Ning’er Giant’ citron weighing ca. 8 kg, held by the farmer, Li Hua Zhong, near Ning’er, Yunnan. (Photo credit: David Karp; 18 Oct. 2011.)

Distribution: Ning’er, Simao, and Mojiang counties in southern Yunnan. Elevation: 800–1700 m. Botanical characteristics: flower buds light purple, young flushes purple; thorns medium long; leaves oval or obovateelliptic with round tip and fine saw margin; fruits very large, typically weighing 3–5 kg, sometimes reaching 8–10 kg; shape oblong, blocky; rind often very wrinkled or warty; albedo thick, sweet and juicy; locules c. 13, with numerous monoembryonic seeds; juice vesicles absent, although a few rudimentary vesicles can be found. Fruits on the same tree can vary considerably in size and rind texture. Uses: rootstock for fingered citrons; consumed fresh and processed (candied rind); traditional Chinese medicine; religious and decorative purposes. Note: at its largest, this cultivar may be the largest citrus fruit in the world. Ning’er is one of the major citron production areas in Yunnan, possibly producing over 1000 tonnes annually. The ‘Ning’er Giant’ variety, sometimes called “zhou pi xiangyuan” (“wrinkled citron”) because of its wrinkled rind, has a long growing history and is well known in southern Yunnan. The total growing acreage of citron grown in the area is about 700 mu (50 ha). The fruit can be harvested yearround, with a peak season in fall. 6. ‘Persistent Stigma’ (Citrus medica var. stigmapersistens Liang S.H. and Pen G.N., ) (Plate 5.12).

Plate 5.12. ‘Persistent Stigma’ citron, Huaning, Yunnan. (Photo credit David Karp; 22 Oct. 2011.)

Distribution: Huaning County and Mile City in central Yunnan. Elevation: 1100–1700 m. Botanical characteristics: fruits small to medium, 1–2 kg, blocky oval to pearshapes; rind moderately rough, sometimes ridged; 12–14 segments; juice vesicles present, sour; seeds numerous, monoembryonic; style persists on most but not all fruits. Uses: rootstock for fingered citrons; consumed fresh and processed (candied rind); traditional Chinese medicine; religious and decorative purposes. Notes: described, based on a type specimen found at Honguang Farm, Yuangjiang county, Yunnan, in an article (in Chinese only) by S.H. Liang and G.N. Peng in 1980. This form appears to have vanished from Yuangjiang, but a similar form that typically retains its style was recently found in Huaning (Yuxi City), Yunnan (it is not clear that they are genetically the same or even related). Open pollinated seedling (OPS) material from this genotype clustered with the pure C. medica accessions in Ramadugu et al. ( 2015 ). 7. ‘Pumpkin’ (Plate 5.13).

Plate 5.13. ‘Pumpkin’ citron in Xiaowan, Fengqing County, Yunnan. (Photo credit: Xulan Hu; 28 Oct. 2015.)

Distribution: Majie village, Xiaowan township, Fengqing County, Lincang City, in southwestern Yunnan; introduced from Shuangjiang County, Yunnan. Elevation: 1400 m. Botanical characteristics: tree upright and vigorous; young flushes purple; fruit medium to large, 1.5–2.0 kg, flat, prominently lobed, uneven in shape (one side is often larger than the other), indented at the stem and stylar ends; rind rough; albedo thick; locules c. 18, with numerous monoembryonic seeds; juice vesicles absent. Use: the albedo is eaten as fresh fruit. Note: recently discovered by Xulan Hu; grown on a small scale in southwestern Yunnan. 8. ‘Weishan Bullet’ (Plate 5.14).

Plate 5.14. ‘Weishan Bullet’ citron near Weishan, Yunnan. (Photo credit: David Karp; 12 Oct. 2011.)

Distribution: Weishan and Fengqing counties in western Yunnan. Elevation: 1200–1500 m. Botanical characteristics: flower buds pink; fruit small to medium, 0.4–0.8 kg, oblong, pointed at the stylar end (some specimens could be said, with a bit of imagination, to be bulletshaped); rind moderately rough, ridged in some specimens; albedo moderately thick; c. 12 segments, juice vesicles present, sour; seeds numerous, monoembryonic. Uses: rootstock for fingered citrons; consumed fresh; traditional Chinese medicine; religious and decorative purposes. Note: in microsatellite analysis performed by Ramadugu et al. ( 2015 ), the heterozygosity observed in this cultivar was 8.70%. It clustered with other Chinese pure C. medica genotypes. 9. ‘Weishan Sour’ (Plate 5.15).

Plate 5.15. ‘Weishan Sour’ citron near Weishan, Yunnan, China. (Photo credit: David Karp; 22 Oct. 2011.)

Distribution: Weishan County in western Yunnan. Elevation: 1400–2100 m. Botanical characteristics: trees medium size; young flushes light green, shoots dark green; leaves large, flower buds white; fruits large, typically 3–6 kg; fruits oblong, cylindrical, to spherical with a pointed stylar end; albedo thick, somewhat sour, and juicy; 12–15 segments, juice vesicles present, sour; seeds numerous, monoembryonic (Li et al. 1983 ). Uses: rootstock for fingered citrons; consumed fresh; traditional Chinese medicine; religious and decorative purposes. Note: this cultivar is unusual among Yunnanese citrons in having juice vesicles, which is normal for virtually all other citrus, but not typical of pure Chinese common citrons. The population structure analysis in Ramadugu et al. ( 2015 ) shows this cultivar to be intermediate between the cluster of pure Chinese citrons and the cluster of hybrid citrons. 10. ‘Weishan Sweet’ (Plate 5.16).

Plate 5.16. ‘Weishan Sweet’ citron, Tuanshan village, Nanzhao township, near Weishan, Yunnan. (Photo credit: David Karp; 12 Oct. 2011.)

Distribution: Weishan County in western Yunnan. Elevation: 1400–2100 m. Botanical characteristics: trees medium size; young flushes light green, shoots dark green, leaves large, and flower buds white; fruits large, typically 3–5 kg; fruit oblong, cylindrical, to spherical with a pointed stylar end; rind texture rough, ridged; albedo thick, sweet, and juicy; 15–20 locules, juice vesicles absent; seeds few, monoembryonic. Uses: rootstock for fingered citrons; consumed fresh; traditional Chinese medicine; religious and decorative purposes.

B. Fingered Citron Cultivars Fingered citrons are split apically into fingerlike sections, resembling a human hand. Some find the unusual appearance of the fruit bizarre (Hesser 1998 ), but others find it beautiful and exotic. In addition, fingered citron has a strong, very pleasant fragrance, reminiscent of osmanthus (which in turn is described as having a fruity apricot aroma) (Lim 2012 ). The aroma volatile

component betaionone is responsible for this characteristic osmanthuslike odor (Shiota 1990 ). Fingered citron oil is primarily composed of terpenoids, particularly limonene, beta myrcene and terpinolene (Huang et al. 1998 ). Fingered citrons generally have no seeds or juice vesicles, consisting exclusively of flavedo on the surface and albedo internally. The authors have observed a specimen of ‘Yun’ fingered citron that had a few sections with juice vesicles (Plate 5.17), but such fruit appear to be rare. Hodgson ( 1967 ) mentioned a clone containing “seeds hanging free in the locules,” but no such types appear in more recent literature.

Plate 5.17. ‘Yun’ fingered citron, showing how one fruit actually has some juice vesicles. (Photo credit: David Karp; 8 Nov. 2008.)

Chen ( 2014 ) emasculated and bagged fingered citron flowers before they opened, found that fruit growth did not change significantly compared to controls, and concluded that fingered citron is parthenocarpic. Fingered citrons are more commonly cultivated in China than nonfingered types, because they are more suited for the two citron uses of primary economic importance: (1) they are more suited for Chinese traditional medicine because they have a higher ratio of flavedo, in which many medicinally significant compounds occur, to fruit weight; and (2) because of the unusual flowerlike shape of fingered citron fruits they are better suited than common citrons for the

production of dwarf (bonsai) plants for the ornamental trade. In China, fingered citrons are grown in Sichuan, Yunnan, Guangdong, Fujian and Guangxi provinces, primarily for use in traditional Chinese medicine. In addition, they are cultivated for primarily ornamental use (bonsai potted plants) in Jinhua, Zhejiang Province; Zhaoqing, Guangdong Province and in Yanling, Henan Province (Deng 2008 ). Historically, Western citrus literature has regarded ‘Buddha’s Hand’ citron as a single horticultural variety, C. medica L. var. sarcodactylis Swingle (Swingle and Reece 1967 ), much in the same way that citrus scientists wrote as if there were only one cultivar of ‘Etrog’ used for Jewish ritual (Hodgson 1967 ; Saunt 2000 ). Hodgson ( 1967 ) recognized two clones, with and without seeds. However, based on Chinese publications, morphology and genetic analysis, it is clear that more than six cultivars of fingered citron exist in China alone. Fingered citron trees vary in stature, cold hardiness, flower color, and other significant horticultural characteristics; the fruits vary in size, shape, number and thickness of fingers, proportion of fruits that are open or closed, and the point on the fruit at which the carpels become distinct, as well as rind density, color, and aroma. Because a single tree can produce fruits of widely varying shapes – for example, depending on the time of year of flowering, the proportion of fruits with fingers that are spread or closed can vary considerably, and other characteristics such as rind color can vary depending on the growing environment – it has not always been clear when varying phenotype reflects varying genotype. Indeed, it appears that fingered citron fruit is more variable in morphology than most other types of citrus. From a purely morphological standpoint, it is essential to observe many fruits on multiple trees over time. It is possible that in an orchard of ‘Yun’ fingered citron one might find a few fruits with atypically slender, gnarled fingers resembling ‘Octopus’ fingered citron grown in the same area, but it would be highly unlikely that all the fruits on a ‘Yun’ tree would look like that. During the past 15 years multiple studies have confirmed that the distinctive forms of many named cultivars reflect different genotypes (Chen et al. 2002 ; Ma et al. 2002 ; Gao et al. 2007a , b ; Zhang et al. 2008 ; Hao 2009 ; Sang 2011 ; Ramadugu et al. 2015 ). Ramadugu et al. ( 2015 ) was able to distinguish seven different genotypes of fingered citron, and Ma et al. ( 2002 ) could distinguish two closely related Jinhua cultivars, ‘Qingpi’ and ‘Aihua,’ using random amplification of polymorphic DNA (RAPD) analysis. In China, several types of fingered citrons are classified into cultivars or cultivar families named after the province or area where they primarily are grown: ‘Guang’ for Guangdong, ‘Jin’ for Jinhua, ‘Yun’ for Yunnan, and ‘Chuan’ for Sichuan. RAPD analysis (Zhang 2008) and simple sequence repeat (SSR) markers (Hao 2009 ) have found that genotypes within each of these geographical categories clustered in clades, some of greater diversity (‘Guang’) than others (‘Jin’). Sang ( 2011 ) found a significant correlation between genetic and geographic distances. It seems clear that specimens of these geographical cultivar families are not necessarily clones identical to each other; it would not be surprising if some citrons identified by the same family name were not in fact closely related. In addition to area of origin, fingered citron cultivars are distinguished by fruit shape, tree

habit and flower color (Chen et al. 2008 ). Since all other forms of citrus have whole fruits, segments and juice vesicles, it can be assumed that the original form of C. medica was the common (nonfingered) citron. On at least one occasion in the past, a fingered mutation occurred naturally. Since fingered citrons generally have no seeds, and have not been found growing wild, it seems likely that this mutation survived because humans found it interesting or useful, and propagated it by cuttings or grafting. Did all fingered citrons arise from one original mutation from a common citron parent, or did such mutations occur in multiple instances, presumably from different parents? The neighbor joining tree and maximum parsimony tree investigated by Ramadugu et al. ( 2015 ) showed common and fingered citrons interspersed, implying that fingered citrons as a group are polyphyletic – that is, they do not all descend from a common ancestor of similar type. Anecdotally, scientists in China say that they do suspect that certain fingered selections mutated from common citron parents that are similar in rind aroma and flesh texture (Duowen Li, personal communication, 2008). Because fingered citron does not have seeds and is vegetatively propagated in cultivation, it has not developed as many different forms and cultivars as might be the case if diverse genotypes had long been recombining genes through zygotic seed propagation. However, over the many centuries during which the Chinese have grown fingered citron, they have noticed spontaneous mutations, and selected genotypes best adapted to different geographical areas and uses. More recently, researchers in Zhejiang, the center of the ornamental fingered citron industry, have developed a dwarf variety, ‘Aihua,’ through mutation breeding (Chen and Shentu 2003 ; Guo et al. 2006 ; Wang et al. 2007 ; Chen et al. 2008 ; Sang, 2011 ; Zhang et al. 2009b ). In order to increase the range of fingered citron varieties available to growers, researchers in Zhejiang have started to experiment with genetic transformation methods for citron, using particle bombardment, but none of the cultivars created using this technology have yet been introduced (Guo et al. 2005 ; Zhou et al. 2005b , 2006 ). 1. ‘Aihua’ (‘Qianzhi Baitai’) (Plate 5.18).

Plate 5.18. ‘Aihua’ fingered citrons. (Photo credit: WenRong Chen; 5 Dec. 2008.)

Distribution: Jinhua and Lanxi districts of Zhejiang Province. Elevation: 50–800 m. Botanical characteristics: a recently developed dwarfed mutation of ‘Qingpi’ obtained by gamma irradiation (Guo et al. 2006 ), introduced in 2009. Crown dwarf and compact; average height 45 cm; leaves small and thick; thorns few; length between nodes short; fruit small, average weight 48 g.; rind golden yellow. Produces numerous fruits per tree (average on 3yearold plants 17.3 fruits); suitable for ornamental pot culture (Wang et al. 2010 ). Uses: bonsai plants (Chen and Shentu 2003 ); dried fruit chips, flowers and leaves for Chinese traditional medicine; may be used for citron wine, tea, fresh fruit, and essential oil. Note: ‘Aihua’ (the name means “dwarf” in Chinese) is (with ‘Qingpi’) one of the two main cultivars of the ‘Jin’ family of fingered citrons grown in the Jinhua area of Zhejiang. In an analysis of SSR markers, Wang et al. ( 2007 ) and Zhang et al. ( 2009b ) confirmed that the dwarfing of this cultivar has a genetic basis. 2. ‘Chuan’ (Plate 5.19).

Plate 5.19. ‘Chuan’ fingered citrons with typical closed fingers, Shuangshi village, Xinfan township, Muchuan County, Sichuan. (Photo credit: David Karp; 7 Nov. 2008.)

Distribution: Qianwei, Muchuan, Luzhou, Leshan, Yibin, and Hejiang districts of Sichuan Province and Jiangjin, Nanchuan, and Shizhu districts of Chongqing Municipality. Elevation: 300–800 m. Botanical characteristics: tree canopy extensive; shoots light brown, hanging down loosely; young flush and flowers dark purple; leaves large, deep green; fruits relatively small, 0.3–0.7 kg; shaped like an oblong or cylindrical fist, with fingers typically closed; rind ridged, light yellow. Sensitive to cold. Yields can reach 37 500 kg of fresh fruit per hectare, which can be dried to 3600 kg. Uses: traditional Chinese medicine; considered a highquality product, strongly aromatic, and potent in promoting Qi; also used to make juice, candy, preserves, tea, and essential oil, etc., and to adorn altars. Notes: a molecular marker study (Hao 2009 ) found that ‘Chuan’ clustered with ‘Guang’, and may have derived from it; the neighbor joining tree of Ramadugu et al. ( 2015 ) also depicts them as close. ‘Chuan’ dried citron chips are mostly sold in Beijing, Tiangjin, and other northern China areas. 3. ‘Guang’

(Plate 5.20).

Plate 5.20. ‘Guang’ fingered citron in Guangdong. (Photo credit: Xulan Hu; 14 Oct. 2012.)

Distribution: Zhaoqing and Gaoyao districts of Guangdong Province and Guilin, Liuzhou, and Wuzhou districts of Guangxi Zhuang Autonomous Region. Elevation: 40–200 m. Botanical characteristics: tree relatively tall, with sparse canopy. Shoots light brown, hanging down loosely; young flushes and flowers purplish red; leaves larger and deep green; fruits large, 0.8–1.5 kg, fistshaped; rind light yellow. Uses: traditional Chinese medicine; dried chips are lightly aromatic, weak in promoting Qi, but considered high quality. Also used to make juice, candy, preserves, tea, and essential oil, etc.; widely used as an adornment or offering at altars. Notes: RAPD analysis by Zhang et al. ( 2008 ) identified three distinct Guang cultivars originating from different districts of Guangdong. Guangdong produced 120–150 tonnes of

dried fingered citron chips in 2007. Plantings in Guangxi were less than 600 ha, producing 300 tonnes of dried chips. In 1999 ‘Guang’ dried chips accounted for half of China’s production. 4. ‘Octopus’ (Plate 5.21).

Plate 5.21. Typical specimens of ‘Octopus’ fingered citron grown near Weishan, Yunnan. (Photo credit: David Karp; 12 Oct. 2011.)

Distribution: Weishan Yi and Hui Autonomous County, Yunnan Province. Elevation: 1200–1500 m. Botanical characteristics: trees medium size; shoots green; young flushes and flowers purplishred; leaves large and deep green; fruits medium to large, 0.5–1.2 kg, with thinner and longer fingers than are typical in other forms of citron, with distinct carpels all the way from the stem end, giving it a shape reminiscent of an octopus; rind yellow. Relatively sensitive to cold. Uses: traditional Chinese medicine; processed (candied rind); religious and decorative purposes.

Note: occasionally fruit on trees of other fingered genotypes might also have the open multifingered look of ‘Octopus’, but only on ‘Octopus’ do the great majority of the fruits have this distinctive shape. 5. ‘Qingpi’ (‘Jinhua Qingpi’, ‘Baiyixiushi') (Plate 5.22).

Plate 5.22. ‘Qingpi’ fingered citron plants in bonsai cultivation in Jinhua, Zhejiang. (Photo credit: Xulan Hu; 31 Oct. 2013.)

Distribution: Jinhua and Lanxi districts of Zhejiang Province, and Jiangxi Province. Elevation: 50–800 m. Botanical characteristics: tree medium in stature with compact canopy. Trunk surface light grayish green, young flushes light green, leaves small; flower buds and petals white; fruits medium to large with mostly open fingers. Fingers thin and long, golden, with a slightly wrinkled surface. Produces 10–15 fruit per plant, average fruit weight 0.5 kg, 11 250–22 500 kg ha–1 annually; potted bonsai plants produces 4–6 fruits each; average fruit weight 0.2–0.3 kg. Uses: suited both for ornamental bonsai pot culture, and for production of fruit (Chen and Shentu, 2003 ). Chinese traditional medicine; also for citron wine, tea, fresh fruit, and

essential oil. Note: ‘Qingpi’ is (with ‘Aihua’, q.v.) one of the two main cultivars of the ‘Jin’ family of fingered citrons grown in the Jinhua area of Zhejiang. The name refers to the greenish color of the trunk, in Chinese. It is also called also called ‘Qingyitongzi,’ meaning “white flower/green flush fingered citron.” Some observers consider ‘Baiyixiushi’ to be a cultivar distinct from ‘Qingpi,’ with a white skinned tree trunk; other authorities say that the two putative varieties are the same (Dr WeiDong Guo, personal communication, 2015). Yang et al. ( 2012 ) found that ‘Qingpi’ is less coldtolerant than ‘Aihua.’ ‘Qingpi’ is a mutation of ‘Aihua.’ In an analysis with 23 pairs of microsatellites, Ramadugu et al. ( 2015 ) could not distinguish between ‘Qingpi’ and ‘Aihua,’ but Ma et al. ( 2002 ) could distinguish between them using RAPD analysis. 6. ‘Yun’ (Plate 5.23).

Plate 5.23. ‘Yun’ fingered citrons sold at a produce stand at the airport in Kunming, Yunnan. (Photo credit: David Karp; 6 Nov. 2008.)

Distribution: Pu’er, Yuxi, Dali, and Baoshan districts, Yunnan Province. Elevation: 1000–1800 m. Botanical characteristics: tree relatively tall; shoots green, young flushes and flowers purplishred; leaves large and deep green; fruit large, 1.5–3 kg, with numerous thick, short, stubby fingers and open shape; rind yellow. Sensitive to cold. Uses: mainly for the production of dried chips, for traditional Chinese medicine. Also used to make juice, candy, preserves, tea, and essential oil, etc.; for religious and decorative purposes. Note: Sang ( 2011 ) found that ‘Yun’ had the highest concentration of starch and lowest concentration of amino acids among 20 citron genotypes. The vitamin C concentration in the mesocarp and total antioxidant capacity in the exocarp of ‘Yun’ were significantly higher than in other fingered citrons. 7. ‘Zhaocai’ (‘Red Prince,’ ‘Chijinwangzi,’ ‘Red Flower’) (Plate 5.24).

Plate 5.24. ‘Zhaocai’ fingered citron, Jinhua, Zhejiang. (Photo credit: Xulan Hu; 31 Oct. 2013.)

Distribution: Jinhua and Lanxi districts of Zhejiang Province. Elevation: 50–800 m. Botanical characteristics: young leaves, young fruit and flowers purple; fruit attractive, aromatic, golden yellow when ripe; produces 10–15 fruit per plant annually; average fruit weight 0.5 kg; fruit production is 11 250–22 500 kg ha–1. Uses: Mainly grown for producing fruit, but also for ornamental pot culture (Chen and Shentu 2003 ). Chinese traditional medicine; also used for citron wine, tea, fresh fruit, and essential oil. Note: Considered to be a distinct cultivar by some growers in Jinhua, but a leading authority there, Dr WeiDong Guo, considers it to be “similar to Guang foshou from Guangdong” (personal communication, October 2015).

C. Types Intermediate Between Common and Fingered Citron The occurrence of forms that are intermediate between the common and fingered types is emblematic of the role of western China as a center of diversity for citron. It is unclear whether they originated as a mutation from common citron, or as a hybrid of common and fingered types (presumably with the common citron being the maternal parent, since fingered citrons rarely produce seeds). The most distinctive and fully documented of these intermediate types is ‘Muli,’ first described by Siejie Yang in 1980 (Yang 1980 ), in which a relatively conventional round to flat fruit encloses fingerlike projections that sometimes protrude from a navellike stylar orifice. More common, if less spectacular, are several local genotypes sometimes called “xiang yuan foshou” = “citron/Buddha’s Hand.” One such type, found in Chuanjie, Lufeng County, Yunnan (west of Kunming), has a smooth, round, bulbous stem end, and fingers developed along the rest of the fruit. Local growers clearly view it as a type intermediate between common and fingered citron, but it is not clear, without further research, that the shape is genetically based, and consistent or unusual enough to merit designation as a cultivar. In contrast, the appearance of the fruit that we have named ‘Half and Half’ is so remarkable – the bottom half of the fruit looks like a smoothskinned common citron, while the top half resembles a fingered citron – that it seems worthy of describing and illustrating here, despite the lack of definitive information about it. 1. ‘Half and Half’ (“xiang yuan foshou” = “citron/Buddha’s Hand”) (Plate 5.25).

Plate 5.25. ‘Half and Half’ fingered citron at a market in Jinghong, Yunnan. (Photo credit: David Karp; 17 Oct. 2011.)

Distribution: Jinghong, southern Yunnan. Elevation: 550 m. Botanical characteristics: fruit small to medium in size, 0.5–1 kg; typically round at the bottom, with a smooth rind, and little or no indication of the development of fingers, which typically starts more than halfway toward the stylar end of the fruit. In some specimens the fingers are thick and stubby, and closed in like a fist; in others, the fingers are irregular, from thin to thick, and curve outward. The interior of the fruit resembles that of fingered citrons, with no locules, seeds, or juice vesicles, just albedo straight through, aside from a narrow axial central cavity; the albedo is dense, moist and sweet, as in many common citrons in which the albedo is the primary part consumed. Uses: for traditional Chinese medicine; consumption as fresh fruit; religious and decorative purposes. Note: this description is based solely on observation of fruits for sale at a market in Jinghong on October 17, 2011, so it is not certain that these fruits derive from a genetically stable trait, that there was more than one tree of this type, and if so that this genotype is extant. Despite these limitations, the fruit seems of great intrinsic interest, because its

morphology is so different from that of any other genotype, and seems truly intermediate between the fingered and nonfingered classes of citron. 2. ‘Muli’ (Citrus medica L. var. muliensis W.D. et Y.; , “ xiangyangguo,” ) ( Plates 5.26 and 5.27).

Plate 5.26. ‘Muli’ citron in Baidiao, east of Muli, Sichuan. (Photo credit: David Karp; 10 Nov. 2008.)

Plate 5.27. Characteristic form of ‘Muli’ citron, Muli, Sichuan. (Photo credit: SieJie Yang; 1980 .)

Distribution: Southern Sichuan Province and northern Yunnan Province, particularly the area in and around Muli Tibetan Autonomous County (Sichuan). Elevation: 1600–2200 m, mostly along rivers such as the Yalong, Muli, and Shuiluo. Botanical characteristics: small evergreen shrub or tree, height 2–3 m; leaves oval, elongated oval, or upsidedown egg shape, with leaf tip rounded, leaf edge scalloped; leaves 6.8 × 4.8 cm; petioles wingless, 0.2–0.3 cm long. New shoots green and slightly triangularshaped, becoming rounder and solider when mature; thorns medium in length. Short racemes with 2–13 florets. The top flower is larger and normally bears fruit. Flower buds light lavender, 1.3 × 0.8–1.1 cm; flowers 3.6 cm in diameter, with five petals; stamens 55–60. Segments 5–6 (Yang 1980 ; Wang et al. 1983 ). Fruit intermediate in morphology between common and fingered citron: the external carpel wall is whole and undivided, but the internal carpel is divided into several fingerlike sections. Weight 1360–2378 g; dimensions 8.5–12.5 × 13.5 × 18 cm; strong citron

fragrance. In the same plant, fruits may be found in various shapes, such as flattened spheres, palms with three cracks, and lotus throne shapes. The outside carpels develop into the outer peel, and inner carpel group then develops into the fingershaped parts. In some fruits, fingershaped parts protrude through the stylar orifice and spread out from the fruit (Plate 5.26). Pulp scanty, juice vesicles white and sour. Seed coats yellow white; inner part of the seed is purple red, with deep purple speckles. Cotyledons milky white and monoembryonic (Yang 1980 ; Wang et al. 1983 ). ‘Muli’ was traditionally regarded as a citron, and molecular analysis shows its similarity with citron, but also that its chloroplast gene might have come from pummelo, so it is likely that it is a hybrid (Yang et al. 2015 ). Uses: consumed as fresh fruit or preserves, as a preserved ingredient in a type of wine, and in herbal medicine. It can be used as a dwarfing stock (Yang 1980 ; Wang et al. 1983 ; Wen et al. 1986 ; Gmitter and Hu 1990 ). Note: called “Muli” or “xiangyanggo” (meaning, “fragrant sun fruit”) by the local people. Muli Tibetan Autonomous County is a remote area of southern Sichuan, where Tibetan and Yi people are the largest ethnicities. The American naturalist Joseph F. Rock, one of the first Westerners to visit Muli, wrote about his celebrated journey in “The Land of the Yellow Lama” (Rock 1925 ); Muli served as the inspiration for ShangriLa in James Hilton’s novel Lost Horizons.

D. “Wild” Citrons Putatively “wild” citron producing small fruits grow in wild forests, in semiwild margins between wild and cultivated areas (on the borders between forest and field, or in inhabited forest areas), and in private gardens in the subtropical forests of western and southern Yunnan. They do not seem to be cultivated on a commercial scale, at least in China. Whether these forms are truly wild, in that trees of this type existed in the area before local cultivation of citron, it is difficult, or perhaps impossible, to judge. As Bonavia noted in 1888: “When a botanist says he found a certain plant wild in a certain place, he probably does not mean that it has been there from the beginning of time, or even from the very commencement of the citrus family on this earth. … We can rarely know to what extent man and other animals may have aided in carrying seeds of plants from place to place. … in most cases, it is next to impossible to decide, whenever a citrus is found wild anywhere, whether it was there, or in the vicinity, from the beginning of its family history, or whether it became naturalized there through the aid of birds, &c” (Bonavia 1888 ). From their botanical characteristics these wildtype trees appear to be pure C. medica, with typical growing habit and leaves, and local inhabitants refer to them simply as “xiang yuan” citron. However, most of the trees currently accessible (type 1 – see below) produce fruits with thinner skin and juicier flesh than is typical of pure citron, and in a recent study (Ramadugu et al. 2015 ) these forms appeared in a cluster with known citron hybrids. Another form which is rare and poorly documented, has thick albedo and locules devoid of juice vesicles, and thus appears to be pure C. medica, but the authors have not yet been able to

find specimens for genetic testing to confirm this hypothesis. The habitat for “wild” citrons in western Yunnan is quickly disappearing as forests are cleared. The preservation and study of such forms, assuming that they remain extant, is critical. 1. Wild Citron, Type 1 (with thin rind, segments and juice vesicles) (Plates 5.28 and 5.29).

Plate 5.28. Wild citron, type 1 (with thin rind, segments and juice vesicles), from a wild forest in a valley in the mountains 2 km west of Hexinchang village, Mangshi township, Luxi City, Yunnan. Elevation: 1190 m. (Photo credit: Xulan Hu; 17 Sept. 2010.)

Plate 5.29. Wild citron, type 1, growing wild in Mangshi township, Yunnan. Top: view of the valley where wild citron grows as an understory plant. Bottom: trees of wild citron, type 1. (Photo credit: Xulan Hu; 6 Sept. 2010.)

Distribution: Subtropical forests of western and southern Yunnan Province, including Dehong, Baoshan, Lincang, and Pu’er prefectures, particularly in the Mangshi and Ruili districts of Dehong Prefecture. Also in northern Myanmar. Elevation: 900–1800 m. Botanical characteristics: trees medium size, growing under the lower forest canopy. Young leaves light green, leaf blades ellipticovate or oblong with tip blunt, petioles wingless, margin entire or shallowly serrated; young twigs angular or rounded, shoots with medium thorns. Flowers white, fruits small with a size of 12–18 × 7–10 cm and a spindle like shape with a sharp tip; rind may be rough, light to deep yellow/brown; albedo relatively thin for a citron, closer to the thickness typical for lemon (C. limon), albedo flavor sour; segments 9–13, juice vesicles sour and bitter, seeds numerous and monoembryonic. Uses: peel oil extract is used in perfumery; traditional Chinese medicine. Note: the fruits are small, fitting easily in the palm; some have the shape and bumpy rind typical of Etrogtype citrons used in Jewish ritual celebrations, and from the outside of the fruits it might seem as if southwestern China could have served as the primordial origin for such citrons. In fruit that are cut open, however, it is clear that the albedo is thinner, and the pulp much juicier, than is typical for known pure C. medica. That still leaves open the possibility that the original wild citrons more closely resembled this type than the kind with thick albedo that is now accepted as the standard of purity. However, in Ramadugu et al. ( 2015 ), four wildtype specimens – three from Ruili and one from Mangshi – all clustered in a clade that also included ‘India Lemon,’ a known hybrid, as well as ‘Suanmaliu,’ a lessknown but clearly hybrid sample; both of these two contained noncitron alleles. No such hybrids appeared in the other clusters, which included Chinese and Westerntype citrons. More research needs to be done to verify this result, but the preliminary conclusion is that the socalled “wild citrons” of this type (thin rind) are hybrids of citron and noncitron genotypes. 2. Wild Citron, Type 2 (with thick albedo, no segments or juice vesicles) (Plate 5.30).

Plate 5.30. Wild citron, type 2 (with thick albedo, no segments or juice vesicles). (Photo credit: Xulan Hu; 1997.)

Distribution: Subtropical forests of western and southern Yunnan Province, including Dehong, Lincang, and Pu’er prefectures, particularly in Ruili County (Dehong Prefecture). Elevation: 800–1800 m. Botanical characteristics: trees medium size, growing under the lower forest canopy. Young leaves light green/yellow color, mature leaf blades ellipticovate or oblong with tip blunt, petioles wingless, margin entire or shallowly serrate; young twigs angular or rounded; shoots with medium thorns. Flower buds and flushes purpletinted; fruit size 14– 20 × 5–9 cm with a long spindle shape with a sharp tip; skin light to deep yellow with some rough surface; thick albedo with sour flavor, segments degenerated, locules empty except for seeds, numerous and monoembryonic. Uses: peel oil extract is used in perfumery; traditional Chinese medicine. Note: one of the authors (Xulan Hu) encountered and photographed this citron type in Dehong prefecture during the 1990s (Plate 5.29), but despite numerous attempts we have not been able to find similar specimens growing in the wild in more recent years (2008– 2016). (As noted above, the recently discovered ‘Jinggu’ citron may well be a form of Wild citron, Type 2 that has entered cultivation.) It would be of great scientific interest if such specimens could be found and tested with modern molecular marker methods, to determine whether they are of pure C. medica or hybrid ancestry. We do not know of any citrus other than pure C. medica with thick albedo and empty locules (no juice vesicles). It could be that such a genotype represents an early form of C. medica, from which modern largefruited forms evolved through selection over the centuries.

E. Citron Hybrids Included in this category are forms that are regarded in their growing area as citrons, but for which morphological and/or genetic evidence indicates that they are actually hybrids. Compared to other citrongrowing areas such as India and Italy, citron hybrids play a lesser role in cultivation than one might expect. Probably this is because the two types most important in commerce, fingered citrons (for medicinal use and bonsai) and common citrons with very thick albedo (for candying or other consumption of the albedo) are both likely to be pure C. medica. Given the proximity of Yunnan to northeastern India, where a wide range of hybrid citron and other citrus exists, and the rich native citrus germplasm resources in China, it is not surprising that hybrid citrons have developed there. 1. Goucheng (Yunnanese) (Citrus medica L. var. yunnanensis S.Q. Ding ex Huang) () (

Plate 5.31).

Plate 5.31. ‘Goucheng’ citron hybrid, Binchuan, Yunnan. (Photo credit: David Karp; 2 Nov. 2008.)

Distribution: mainly western Yunnan Province, including Bingchuan, Yangbi, Weishan, and Yongping counties. Elevation: 1100–1700 m. Botanical characteristics: trees small or shrublike, height of 2–3 m; young shoots

purple, hairless, and thorns longer in each node; leaves oval or obovateelliptic with round tip and fine saw margin; fruits broad ovate with sharp tip; small, 0.3–1.2 kg; rind yellow, rough; albedo relatively thin; segments 10–12; juice vesicles present, pulp very sour; seeds numerous, monoembryonic, occasionally polyembryonic. Uses: rootstock for mandarins and sweet orange in Yunnan. Dried seed and fruit are a traditional Chinese medicine to treat certain stomach disorders. Because of its thorniness, used as a hedge for orchards and gardens. Note: “Found at Binchuan district in Yunnan … in 1976–78. … The flowers are rather large, each with stamens cohering in two or three whorls, ovaries different from the ordinary forms, carpels of fruits appendaged, pulpvesicles well developed. According to the traits of flowers and fruits, they give evidence of Yunnan citron as an intermediate variety evolved from fingered citron (C. medica L. var. sarcodactylis) to citron (C. medica L.)” (Ding 1979 ). Clustered with citron hybrids in Ramadugu et al. ( 2015 ). In 1980, soon after the initial publication of this putative subspecies, Professor Tsuin Shen of Peking Agriculture University sent seeds of it to the University of California, Riverside Citrus Variety Collection. The accession derived from these seeds (CRC 3798 / PI 600630) produces fruits that are small to medium in size; pearshaped; seedless; rind moderately thin to moderately thick; pulp juicy, fruity, very tart, and very fragrant. 2. ‘Yunmao Oval’ (Plate 5.32).

Plate 5.32. ‘Yunmao Oval’ citron hybrid near Mengwan village, Dehong prefecture, Yunnan. (Photo credit: David Karp; 4 Nov. 2008.)

Distribution: Mangshi and Longling counties in western Yunnan Province. Elevation: 900–1300 m. Botanical characteristics: tree vigorous, upright; young flushes light green, shoots dark green, leaves large, and flower buds pink to purple; fruits medium, typically 1–3 kg; fruit oval to topshaped, bluntly pointed at stylar end; rind texture smooth, sometimes slightly rough; albedo thick; 12–14 locules, small; juice vesicles absent, although a few rudimentary ones can be found; seeds few, monoembryonic. Uses: rootstock for fingered citrons; fruit consumed fresh; dried chips in traditional Chinese medicine; religious and decorative purposes. Notes: named after Yunmao village in Dehong prefecture. Ramadugu et al. ( 2015 ) found that ‘Yunmao Oval’ clustered with hybrid citrons.

VI. GERMPLASM STATUS; REGIONAL AND GLOBAL PERSPECTIVE Until recently, Chinese citrus scientists paid relatively little attention to citron germplasm

resources, both wild and cultivated, and made few attempts to document the diversity of available forms. Wild citron germplasm resources, in particular, may have been significantly eroded by development and neglect. More attention has been paid to citron germplasm during recent years (Chen and Shentu 2003 ; Zhou and Guo 2005a,b; Chen et al. 2006 ; Liang et al. 2006a , b ; Wang et al. 2007 ; Zhang and Xu 2007 ; Chen et al. 2008 ; Deng 2008 ; Zhang et al. 2008 ; Hao 2009 ; Wang et al. 2010 ; Sang 2011 ). In 2010, the present authors established an informal Chinese Citron Germplasm Collection in Jianshui, 175 km south of Kunming, which within a few years contained approximately 20 distinct accessions. The objectives were to collect, characterize, preserve, and distribute citron germplasm. The collection is no longer in existence onsite, but the accessions were shared with Chinese citrus germplasm collections, local and national. In addition, the material was used in a wideranging genetic study (Ramadugu et al. 2015 ) which analyzed the genetic diversity of 47 citrons – 32 from China and 15 of Mediterranean origin. These citrons clustered into three distinct groups: 1. From China, wild and hybrid citrons with pulp and seeds. Three of these, ‘Suanmaliu,’ ‘India Lemon,’ and CRC 3819 appear to be hybrids involving citron and noncitron ancestors. It seems likely that most or all of this group are of hybrid ancestry, but more work will need to be done to confirm this. 2. From China, both common and fingered citrons. These appear to be “pure” C. medica. The common citrons have thick albedos; some have juice vesicles, others do not. 3. From the Mediterranean and Western areas, common citrons, “pure” C. medica, both with juice vesicles and without. This cluster includes a representative selection of citrons used for Jewish ritual (‘Assads,’ ‘Braverman,’ ‘Halperin,’ ‘Kivelevitz,’ ‘Temoni,’ and an unspecified ‘Etrog’ which clusters with ‘Kivelevitz’). There are also typical Mediterranean cultivars, ‘Diamante,’ ‘Italian,’ ‘Corsican,’ and ‘Citron of Commerce,’ but no Chinese cultivars. The C. medica group appeared to be monophyletic in this study, which also indicated that some citrons have a higher level of heterozygosity than previously estimated. It is notable that Ramadugu et al. ( 2015 ) found that pure C. medica specimens cluster into two separate groups depending on their geographical origin, from China and from the Mediterranean area. Phenotypically there are no characteristics of tree or fruit that would serve to distinguish the groups, but a distinction is evident genetically. Assuming that citron originated in northeastern India, southwestern China or adjacent regions, it may be that the citrons from the Mediterranean and Western areas derive from ancestors of Indian origin, which may differ genetically from the citrons now growing in China. Indian scientists have started to survey the citron germplasm of northeastern India and adjacent regions (Ray and Deka 1999 ; Kumar et al. 2014 ; Barbhuiya et al. 2016 ). A study of how Indian citrons relate genetically to populations in China and the Mediterranean area would do much to define the spread of pure and hybrid citrons across the globe.

LITERATURE CITED AlKalifawi, E.J. 2015. The antimicrobial activity of essential oils of AlAbbas’s (AS) hand fruit peel (Citrus medica) var. sarcodactylis Swingle. J. Nat. Sci. Res. 5(12):19–27 (in English, with Arabic abstract). Anon. 2017. Chinese medicine. , accessed 13 Feb., 2017 (in Chinese). Bailey, L.H. 1919. Citrus medica. In: The standard cyclopedia of horticulture. London, Macmillan, p. 781. Barbhuiya, A.R., M.L. Khan, and S. Dayanandan. 2016. Genetic structure and diversity of natural and domesticated populations of Citrus medica L. in the Eastern Himalayan region of Northeast India, Ecol. Evol. 6(12):3898–3911. Bonavia, E. 1888–90. The cultivated oranges and lemons of India and Ceylon. London, W.H. Allen, 2 vol. Cao, Y.B., R. Shi, W.R. Chen, and W.D. Guo. 2011. Comparative analysis of expression changes of ethylene response factor 6 (ERF6) in fingered citron and trifoliate orange under cold stress. Acta Hortic. Sin. 38(10):1873–1882 (in Chinese, with English abstract). Chan, Y.Y., C.H. Li, Y.C. Shen, and T.S. Wu. 2010. Antiinflammatory principles from the stem and root barks of Citrus medica. Chem. Pharm. Bull. 58(1):61–65. Chen, B.C., T.Q. Zhao, B.J. Ma, and C.R. Shao. 2002. Type delimitation and RAPD analysis of Jinhua fingered citron (Circodactylitrus medica var. sarcodactylis Swingle). J. Fruit Sci. 19(4):278–280 (in Chinese, with English abstract). Chen, B.C. and W.Y. Shentu. 2003. Compared on biological characteristics in 3 types of Jinhua fingered citron. For. Res. Beijing 16(6):774–776 (in Chinese, with English abstract). Chen, H.L., Y. Wang, and Y. He. 2008. Bergamot research 1: origin distribution resource types ornamental characteristics and varieties. South China Agric. 2(6):16–20 (in Chinese). Chen, M.G. 2014. Preliminary studies on fruit set and fruitshaped development of fingered citron. Dissertation in botany at Zhejiang Normal Univ. (in Chinese, with English abstract). Chen, S.Y., Y.J. Peng, Q. Huang, and F. Yang. 2006. Study of the decorative and medicinal values of two varieties of Citrus medica. For. Byproduct Spec. China 6(85):70–72 (in Chinese, with English abstract). Chen, W.R., Z.Z. Zhang, D.D. Xin, and W.D. Guo. 2010. Identification and expression analysis of coldregulated genes in fingered citron ( Citrus medica var. sarcodactylis Swingle). In Vitro Cell. Dev. Biol. Animal 46(Suppl. 1):S93–S211 (poster abstract).

De Candolle, A. 1886. Origin of cultivated plants, 2nd ed. London, Paul, Trench. Deng, X.L., Y.D. Gao, J.C. Chen, X.L. Pu, W.W. Kong, and H.P. Li. 2012. Current situation of ‘Candidatus Liberibacter asiaticus’ in Guangdong, China, where citrus huanglongbing was first described. J. Integr. Agric. 11(3):424–429. Deng, X.X. 2008. Citrus varieties in China. Beijing, China Agric. Press, 203 pp. (in English and Chinese). Deng, X., J. Chen, Z. Shan, G. Zhou, H. Li, and E.L. Civerolo. 2008. Identification of ‘Candidatus Liberibacter asiaticus’ from foshou (Citrus medica) in China. Plant Pathol. 57:365. Deng, Z.N., L. Xu, D.Z. Li, G.Y. Long, L.P. Liu, F. Fang, and G.P. Shu. 2010. Screening citrus genotypes for resistance to canker disease (Xanthomonas axonopodis pv. citri). Plant Breed. 129:341–345. Ding, S.Q. 1979. Yunnan citron – a new variety of Citrus medica L. from Yunnan. Acta Hort. Sin. 6(2):85–86,133 (in Chinese, with English abstract). FAO (Food and Agriculture Organization of the United Nations). 2016. Citrus fruit statistics 2015. Rome. Firminger, T.A.C. 1874. A manual of gardening for Bengal and Upper India (3rd ed.). Bombay, Thacker, Spink and Co. Gao, X.X., X.Y. Chen, and Y.S. Luo. 2007a. Preliminary study on rDNA ITS sequencing and characteristics of Citrus medica L. var. sarcodactylis (Noot.) Swingle. J. Chin. Med. Mater. 30(3):268–70 (in Chinese). Gao, X.X., X.Y. Chen, and Y.S. Luo. 2007b. Preliminary study on rDNA ITS sequencing and characteristics of Citrus medica L. var. sarcodactylis (Noot.) Swingle. Lishizhen Med. Mater. Med. Res. 18(1):162–163 (in Chinese, with English abstract). Gmitter, F.G., Jr., and X. Hu. 1990. The possible role of Yunnan, China, in the origin of contemporary citrus species (Rutaceae). Econ. Bot. 44(2): 267–277. Guo, T.C. 1993. Study on the citron in China. China Citrus 22(4):3–6 (in Chinese). Guo, W.D., C.L. Zhou, M. Lu, and X.S. Xie. 2005. Biolistic mediated transformation and transient expression of Gus gene in leaves of Citrus medica L. var. sarcodactylis (Noot) Swingle. J. Zhejiang Normal Univ. (Nat. Sci.) 28(4):425–428 (in Chinese). Guo, W.D., M. Lu, C.L. Zhou, F.L. Li, and Y.F. Pan. 2006. Irradiation effects of 137Csγray on growth of cuttings of Citrus medica var. sarcodactylis. J. Zhejiang Univ. (Agric. & Life Sci.) 1:31–33 (in Chinese). Guo, W.D., J.S. Zheng, G. Deng, W.R. Chen, and J. Sun. 2009a. Antibacterial effects of

essential oil from fingered citrons. J. Chin. Cereals and Oils Assn. 24(8):103–107 (in Chinese). Guo, W.D., D. Sang, J.S. Zheng, H.B Liao, and W.R. Chen. 2009b. Effects of nitrogen deficiency on the gas exchange, chlorophyll fluorescence and chloroplast ultrastructure in fingered citron. J. Zhejiang Univ. (Agric. and Life Sci.) 35(3):307–314 (in Chinese, with English abstract). Guo, W.D., J.S. Zheng, Z.Z. Zhang, W.R. Chen, and Y.P. Guo. 2009c. Effects of short term chilling stress on the photosynthetic physiology of fingered citrons (Citrus medica var. sarcodactylis Swingle). Acta Ecologica Sin. 29(5):2286–2293 (in Chinese, with English abstract). Guo, W.D., Z.Z. Zhang, X.W. Jiang, M.G. Chen, J.S. Zheng, and W.R. Chen. 2009d. Semilethal temperature of fingered citron (Citrus medica var. sarcodactylis Swingle) under low temperature stress and evaluation on their cold resistance. Acta Hortic. Sin. 36(1):81–86 (in Chinese, with English abstract). Hao, D.Q. 2009. Study on germplasm resources and genetic diversity of Citrus medica var. sarcodactylus (Noot.) Swingle. Master’s thesis in botany at Sun YatSen Univ. (in Chinese, with English abstract). He, H., G.F. Zhang, C.M. Pan, H. Huang, and X.H. Lin. 2005. New record of Huanglongbing on Citrus medica var. sarcodactylis (Noot.) Swingle in Guangdong. Acta Phytopathol. Sin. 35(2):190–192 (in Chinese). Hesser, A. 1998. A long and happy life of ugliness. The New York Times, December 9, 1998. Hodgson, R.W. 1967. Horticultural varieties of citrus. p. 431–587. In: W. Reuther, H.J. Webber, and L.D. Batchelor (eds), The citrus industry, vol. 1. Berkeley, Calif., Univ. of Calif. Huang, X.Y., X.Y. Zhong, and Y. Su. 1998. Extraction and identification of volatile oil in fingered citron. J. South China Agric. Univ. 19(3):101–106 (in Chinese). Jiang, H, Y.C. Xu, Y.R. Li, and, M. Zhai. 2012a. The study on cold resistance of different Citrus medica strains. Acta Hortic. Sin. 39(3):525–532 (in Chinese, with English abstract). Jiang, H., Y.C. Xu, Y.R. Li, M. Zhai, and C.L. Dong. 2012b. The study on cold resistance of the filial individual of halfsib provenance of Citrus medica L. J. Jiangsu For. Sci. Tech. 39(1):1–6 (in Chinese, with English abstract). Jiao, S.D. 2001. Ten lectures on the use of medicinals from the personal experience of Jiao ShuDe. Brookline, MA: Paradigm Publ. (in Chinese and English). Jie, Z.X. 2007. Study on extraction and components analysis of bergamot oil and process of juice beverage. Master’s thesis in food science, China Agric. Univ. (in Chinese, with English abstract).

Karp, D. 1998. Buddha’s hand citron. Flavor & Fortune 5(4):5–6. Kim, K.N, Y.J. Ko, H.M. Yang, Y.M. Ham, S.W. Roh, Y.J. Jeon, G. Ahn, M.C. Kang, W.J. Yoon, D. Kim, and T. Oda. 2013. Antiinflammatory effect of essential oil and its constituents from fingered citron (Citrus medica L. var. sarcodactylis) through blocking JNK, ERK and NFκB signaling pathways in LPSactivated RAW 264.7 cells. Food Chem. Toxicol. 57:126–131. Klein, J.D. 2014. Citron cultivation, production and uses in the Mediterranean region. p. 199– 214. In: Z. Yaniv, and N. Dudai (eds), Medicinal and aromatic plants of the MiddleEast. Dordrecht, Netherlands, Springer. Kumar, S., S.K. Malik, A. Uchoi, R. Chaudhury, and S.R. Bhat. 2014. A new wild type of citron (Citrus medica L., Rutaceae) identified through morphology and psbMtrnD spacer region of chloroplast DNA. Trees 28:1115–1124. Li, D., C. Wan, and S. Liang. 1983. An investigation on fingered citron in Weishan County. Yunnan’s Sci. Tech. 2:60–61 (in Chinese). Li, X.M., W.J. Liu, J. Zhou, and Z.Y. Mao. 2000. Volatile components in Citrus medica L. leaves and its application in tobacco flavoring. Tobacco Sci. Technol. 5:24–25 (in Chinese with English abstract). Li, J.X., H.C. Wang, M.H. Gao, L.W. Yang, and H. Li. 2013. Coumarins and flavonoids from the fruit of Citrus medica. J. Pharm. Res. 32(4):187–191 (in Chinese with English abstract). Li, Y.Q., J.N. Yang, W.R. Chen, M. Lu, and W.D. Guo. 2011. Identification of Penicillium species causing postharvest diseases of finger citron. Acta Phytopathol. Sin. 41(5):534–537 (in Chinese, with English abstract). Liang, S. and G.N. Peng. 1980. Preliminary investigation on Citrus medica var. stigmapersistens. S.H. Liang and G.N. Peng, privately published journal: p. 58–59 (in Chinese). Liang, S.J., H. Wu, X. Lun, and D.W. Lu. 2006a. Secretory cavity development and its relationship with the accumulation of essential oil in fruits of Citrus medica L. var. sarcodactylis (Noot.) Swingle. J. Integrative Plant Biol. 48(5):573–583. Liang, Y.H., X.W. Chu, and D. Kai. 2006b. Review of progress in fingered citron. Modern Chin. Med. 8(5):24–27 (in Chinese). Liao, H.B., J.G. Wu, W.R. Chen, W.D. Guo, and C.H. Shi. 2012. Rapid diagnosis of nutrient elements in fingered citron leaf using near infrared reflectance spectroscopy. J. Plant Nutr. 35(11):1725–1734. Liao, F.L., M.G. Chen, D. Sang, W.R. Chen, and W.D. Guo. 2013. Genetic diversity analysis of fingered citron germplasm by ISSR. Acta Hort. Sin. 40(11):2222–2228 (in Chinese, with

English abstract). Lim, T.K. 2012. Citrus medica and Citrus medica var. sarcodactylis. p. 683–694. In: Edible medicinal and nonmedicinal plants. Dordrecht, Springer. Liu, C.Q., D.Q. Li, L.C. Niu, M. Yu, and J.F. Song. 2014. Research progress of citron exploitation and usage. J. Jiangsu Agric. Sci. 42(7):1–5 (in Chinese). Lu, L.G. and J.L. Wu. 2006. The cultural techniques for raising potted plant of Jinhua bergamot. South China Fruits (No.1):16–17 (in Chinese). Lü, X.W., L.X. Shao, J.P. Zhang, Y.F. Ma, G. Deng, L.L. Xu, and M.L. Mei. 2011. Inhibitory effect of fingered citron essential oil on proliferation of B16 melanoma cells in vitro. J. Chin. Cereals Oils Assoc. 26(8):50–54 (in Chinese, with English abstract). Ma, B.J., B.Y. Zhang, B. Chen, and B.C. Chen. 2002. RAPD analysis of genetic polymorphism and molecular identification of Citrus medica cv. sarcodactylis from Jinhua. Chinese Trad. Herbal Drugs 33(5)460–462 (in Chinese, with English abstract). Mei, S.F., H. Zhao, X.L. Liu, Z. Mei, L.X. Shi, and G.Q. Lu. 2006. Development of bergamot resources and its utility post development. China Seed Ind. 10, 68–69 (in Chinese). Panara, K., K. Joshi, and K. Nishteswar. 2012. A review on phytochemical and pharmacological properties of Citrus medica Linn. Int. J. Pharm. Biol. Arch. 3(6):1292–1297. Peng, C.H., Y.B. Ker, C.F. Weng, C.C. Peng, C.N. Huang, L.Y. Lin, and R.Y. Peng. 2009. Insulin secretagogue bioactivity of finger citron fruit (Citrus medica l. var. sarcodactylis hort, Rutaceae). J. Agr. Food Chem. 57(19):8812–8819. Ramadugu, C., M.L. Keremane, X. Hu, D. Karp, C.T. Federici, T. Kahn, M.L. Roose, and R.F. Lee. 2015. Genetic analysis of citron (Citrus medica L.) using simple sequence repeats and single nucleotide polymorphisms. Scientia Hortic. 195:124–137. Ramadugu, C., M.L. Keremane, S.E. Halbert, Y.P. Duan, M.L. Roose, E. Stover, and R.F. Lee. 2016. Longterm field evaluation reveals huanglongbing resistance in Citrus relatives. Plant Dis. 100:1858–1869. Rawson, J., et al. 2015. China. Grove Art Online. Oxford Art Online. Oxford University Press. Web. 28 Dec. 2016. . Ray, B.K. and P.C. Deka. 1999. Characterization of citrons growing in North East India using hierarchical cluster analysis. Indian J. Plant Genet. Resour. 12(1):50–55. Rock, J.F. 1925. The land of the yellow lama. Nat. Geogr. 47(4):447–492. Saunt, J. 2000. Citrus varieties of the world: an illustrated guide, 2nd ed. Norwich, England, Sinclair International Limited.

Sang, D. 2011. Analyzed on genetic diversity, collection, and evaluation of fingered citron. Master’s thesis in botany, Zhejiang Normal Univ. (in Chinese, with English abstract). Schafer, E.H. 1967. The vermillion bird: T’ang images of the south. Berkeley and Los Angeles, University of Calif. Shao, L.X., X.W. Lu, G. Deng, J.P. Zhang, L.L. Xu, Y.F. Ma, and J.Q. Bi. 2011. Effects of fingered citron essential oil on the proliferation and tyrosinase activity of B16 melanoma cells in vitro. J. Zhejiang Normal Univ. (Nat. Sci.) 34(2):202–206 (in Chinese, with English abstract). Shi, X.W. and J.A. Shao. 2008. Jinhua fingered citron cultivation using new technologies. Beijing, China Agricultural Science and Technology Press (in Chinese). Shiota, H. 1990. Volatile components in the peel oil from fingered citron (Citrus medica L. var. sarcodactylis Swingle). Flavour Frag. J. 5(1):33–37. Simoons, F.J. 1991. Food in China: a cultural and historical inquiry. Boca Raton, Florida, CRC Press. Swingle, W.T. and P.C. Reece. 1967. The botany of citrus and its wild relatives. p. 190–430. In: W. Reuther, H.J. Webber, and L.D. Batchelor (eds.), The citrus industry, vol. 1. Berkeley, Calif., University of California. Tolkowsky, S. 1938. Hesperides: a history of the culture and use of citrus fruits. John Bale, Sons and Curnow, London. Wang, H.K., Y.X. Ding, and S.J. Yang. 1983. Muli citron – a variety of Citrus medica L. Acta Hortic. Sin. 10(3):181–182 (in Chinese). Wang, H.X., J.L. Zhang, W.D. Guo, J.S. Shen, and Y.F. Chen. 2010. A new dwarf type of fingered citron cultivar ‘Qianzhi Baitai’. Acta Hortic. Sin. 37(6):1029–1030 (in Chinese, with English abstract). Wang, J.P., W.D. Guo, C. Nikai, and Z.Z. Zhang. 2007. Morphological characters and genetic background on dwarf Citrus medica var. sarcodactylis Swingle. J. Zhejiang Agric. Sci. 2007(4):407–410 (in Chinese). Wen, M.Z., Y.Z. Huang, S.C. Xiao, W.J. Ren, H. Zhao, and S.J. Yang. 1986. A study on the chemical components of the essential oil from the peel of Citrus medica L. var. muliensis W.D. et Y. Acta Bot. Sin. 28(5):511–516 (in Chinese). Wu, Z., H. Li, Y. Yang, Y. Zhan, and D.W. Tu. 2013. Variation in the components and antioxidant activity of Citrus medica L. var. sarcodactylis essential oils at different stages of maturity. Ind. Crops Prod. 46:311–316 (in English). Xu, G.L. and B.Z. Deng. 2002. Pollutionfree cultivation and application of fingered citron. Beijing, China Agriculture Press (in Chinese).

Yang, S.J. 1980. Cultivation history and distribution of Muli citron. Liangshan Horticulture (suppl.):3–11 (in Chinese). Yang, L., J. Ye, W.D. Guo, C.C. Wang, and H.T. Hu. 2012. Differences in cold tolerance and expression of two fatty acid desaturase genes in the leaves between fingered citron and its dwarf mutant. Trees 26:1193–1201. Yang, X.M., H. Li, M. Liang, Q. Xu, L.J. Chai, and X.X. Deng. 2015. Genetic diversity and phylogenetic relationships of citron (Citrus medica L.) and its relatives in southwest China. Tree Genet. Genomes 11:129. Zhang, G.F. and H.H. Xu. 2007. Bergamot germplasm resources research overview. J. Guangzhou Univ. Traditional Chin. Med. 1:69–72 (in Chinese). Zhang, G.F., H.H. Xu, and H. He. 2008. An RAPD analysis on genetic diversity of Citrus medica L. var. sarcodactylis (Noot.) Swingle germplasms. Lishizhen Medi. Mater. Medica Res. 20(3):604–606 (in Chinese) Zhang, G.F., H. He, and H.H. Xu. 2009a. Factors affecting on the survival rate for shoottip micrografting of Citrus medica L. var. sarcodactylis Swingle. J. Guangzhou Univ. Traditional Chin. Med. 5:480–484 (in Chinese, with English abstract). Zhang, Q.W. 2015. Citrus medica L. var. sarcodactylis Swingle: (Foshou, Finger Citron). p. 327–331. In: Liu, Y., Z.M. Wang, and J.Z. Zhang (eds), Dietary Chinese Herbs Chemistry, Pharmacology and Clinical Evidence. Vienna, Austria: SpringerVerlag. Zhang, Z.Z., W.D. Guo, and J.P. Wang. 2009b. SSR marker analysis of genetic difference in dwarf Citrus medica var. sarcodactylis Swingle. Zhejiang Agric. Sci. (1):67–71 (in Chinese). Zhou, C.L. and W.D. Guo. 2005a. Current studies on the plasm of Citrus medica L. var. sarcodactylis Agric. Sci. 3:89–110 (in Chinese). Zhou, C.L. and W.D. Guo. 2005b. Progress in fingered citron germplasm. Shaanxi J. Agric. Sci. 3:89–92 (in Chinese). Zhou, C.L., W.D. Guo, and Y.Y. Zhang. 2005a. Selective test of sensitivity to kanamycin and tissue culture of Citrus medica. var. sarcodactylis. J. Northwest For. Univ. 20(4):76–79 (in Chinese). Zhou, C.L., W.D. Guo, D.J. Wang, C.L. Yu, M. Lu, and Y.P. Li. 2005b. Exploration of the transformation parameters of the leaf discs of bergamot by particle bombardment in virtue of transient expression of GUS gene. Acta Botan. BorealiOccidentalia Sin. 25(11):2145–2150 (in Chinese). Zhou, C.L., W.D. Guo, M. Lu, J. Chen, and Y.P. Li. 2006. Factors effecting the transformation of Citrus medica L. var. sarcodactlis mediated by Agrobacterium. J. Trop. Subtrop. Bot. 14(5):374–381 (in Chinese).

6 Apple Rootstocks: History, Physiology, Management, and Breeding Richard P. Marini Department of Plant Science, The Pennsylvania State University, University Park, PA, USA Gennaro Fazio USDA/ARS Plant Genetic Resources Unit, Geneva, NY, USA

ABSTRACT For more than two millennia, superior fruit tree genotypes have been grafted onto rootstocks to maintain the genetic identity of the desirable scions. Until the 20th century most fruit trees were grafted onto seedling rootstocks. Following the classification, evaluation, and propagation of clonal rootstocks during the early 1900s, dwarfing rootstocks became important to the commercial apple industries. Although trees on dwarfing rootstocks are more economical to maintain, and are more precocious and productive than trees on seedling rootstocks, there remains a need for dwarfing rootstocks to be adapted to different growing conditions. During the past 100 years, considerable effort has been made to understand the physiological changes in the scion induced by rootstocks. More recently, molecular techniques have been utilized to identify the genes that control interactions between scion and rootstock. Modern rootstock breeding programs are combining molecular and traditional techniques to develop rootstocks that are dwarfing, productive, and tolerant to biotic and abiotic stresses. In this chapter, the history, development and current use of apple rootstocks, the current understanding of rootstock–scion interactions, and current efforts to develop and evaluate superior rootstocks are discussed. KEYWORDS: Malus × domestica, dwarfing, Malling, roots, rootstock–scion interaction I. INTRODUCTION II. HISTORY A. Europe B. North America III. ROOTSTOCK–SCION INTERACTIONS A. Influence of Rootstock Roots versus Rootstock Stems B. Influence of Rootstock on Scion Growth C. Mechanisms of Stock–Scion Interaction

1. Rootstock Anatomy and Morphology 2. Assimilate Production and Distribution 3. Uptake and Distribution of Nutrients 4. Influence on Leaf Nutrient Concentrations 5. Hormones 6. Graft Union Failure and Tree Anchorage 7. Burrknots IV. STRESSES INFLUENCING ROOTSTOCK PERFORMANCE A. Abiotic Stresses 1. Temperature 2. Water Relations 3. Root Characteristics Affecting Water Relations 4. Soil pH and Salinity B. Biotic Stresses 1. Fire Blight 2. Viruses 3. Soil Microorganisms and Apple Replant Disease 4. Apple Rootstocks for Organic Management V. INTERSTEMS VI. INFLUENCE OF ROOTSTOCK ON FRUIT CHARACTERISTICS A. Fruit Mineral Content B. Fruit Quality and Maturity C. Fruit Size VII. GENETICS AND BREEDING A. Genetic Factors Controlling Important Rootstocks Traits 1. Induced Traits 2. Inherent Traits B. Apple Rootstock Breeding Programs 1. European Apple Rootstock Breeding Programs 2. North American Apple Rootstock Breeding Programs

3. Asian and Pacific Islands Apple Rootstock Breeding Programs VIII. ROOTSTOCK EVALUATION LITERATURE CITED

I. INTRODUCTION Apple (Malus × domestica) originated in central Asia and is the most widely grown tree fruit. Apples are heterozygous and do not come true from seed, and apple cuttings are difficult to root. Therefore, to maintain genotypes with desirable characteristics, rootstocks have been used as a means of propagating fruit trees for more than 2000 years (Roach 1985). Until about 80 years ago, most trees in commercial apple orchards around the world were propagated on seedling rootstocks. Trees on seedling rootstocks were easy to produce, relatively inexpensive, freestanding and performed fairly well on a wide range of soil types and in most climates. Although dwarfing rootstocks have been available for at least 2000 years, attempts to grow trees commercially on these rootstocks usually failed. During the 20th century, the economics of the fruit industry demanded uniform, relatively small, and productive trees. Dwarfing clonal rootstocks offered the potential to satisfy these demands, but these were not widely adopted until researchers developed orchard systems that could take advantage of the beneficial characteristics while minimizing the weaknesses of dwarf rootstocks. The precocity of dwarfing rootstocks caused the leader on young trees to bend and trees to fall over or break. This prevented trees from developing canopies large enough to sustain high yields as the orchard matured. The development of tree support systems, coupled with new tree training techniques, has allowed commercial fruit growers to take advantage of the superior productivity of dwarfing rootstocks, and is responsible for yields per hectare that are more than double those of just 25 years ago. All currently available dwarfing rootstocks have weaknesses, however, such as weak graft unions, propagation difficulties, and susceptibility to biotic and abiotic stresses. The aim of this chapter is to provide an overview of the voluminous work on apple rootstocks that has been conducted and published during the past century, and to identify potential areas for future research.

II. HISTORY A. Europe Apple cultivars are difficult to root and do not come true from seed; consequently, for more than two millennia superior apple genotypes have been maintained by grafting or budding onto rootstocks. Until the mid20th century, seedling rootstocks were most common. It is thought that in the 4th century B.C., Alexander the Great sent to Greece a dwarf selfrooting apple, which later became known as Paradise rootstock (Tukey 1964). For more than 2000 years the use of Paradise was limited to gardens, and it was mentioned in various books as a method for growing small trees that were precocious. Tukey (1964) also described another rootstock called Doucin or English dwarfing stock as only slightly dwarfing and no more precocious than

seedling. During the first half of the 19th century, Dutch Paradise and Doucin were being recommended in England (Loudon 1822; Lindley 1846; Barry 1851) and America (Thomas 1859) for certain cultivars, and especially for espalier training of apple trees to reduce tree size and to induce annual bearing. The rootstock Paradise was described as causing early bearing (Cole 1858), but in his book American Pomology, Warder (1867) described the French Paradise dwarfing rootstock and recommended it for gardens only, because it was “… wholly unsuited for orchard planting”. By 1914 there was an account of an orchard in eastern Massachusetts, USA with 600 dwarf trees on Doucin stock (Anonymous 1915). During the late 1800s, San Jose scale (Quadraspidiotus perniosus) was a serious pest in North American orchards. Controlling the insect required spraying insecticides or fumigating trees with cyanide gas, and large trees were difficult to treat. Beach (1902) suggested growing dwarf trees on Doucin or Paradise in commercial orchards to facilitate scale control. He noted that these rootstocks were shallowrooted and required fertile soil, but such trees were more precocious than trees on seedling rootstocks. Beach cited S.T. Wright, of the Royal Horticultural Society, as indicating that commercial production of Paradise was increasing in England, and that the most profitable systems of apple culture utilized dwarf trees because they obtain early returns, orchard work can be performed from the ground, trees are less injured by wind, and interior fruit can be thinned out. It was recommended that dwarf trees be planted 3  m × 3 m (1074 trees per hectare) and such mature orchards could be expected to produce 4700–9500 kg ha–1, whereas standard trees would produce 13 000 kg ha–1. However, lime sulfur sprays soon became available to effectively control scale and American interest in dwarfing rootstocks diminished. In Europe during the late 1800s and early 1900s, nonuniformity of trees was a problem, with variable tree performance having a negative economic impact on commercial fruit production and also hampering research efforts (Hatton 1917). At that time, seedling rootstocks were referred to as Crab regardless of origin, and those obtained by vegetative propagation (layers) were called Paradise. A second clonal rootstock had been described in France as Doucin, but commercially this became known as English Paradise. An 1869 article in the Gardeners Chronicle indicated there was confusion about the identity and nomenclature of various forms of Paradise rootstocks, and new seedling “Paradise” stocks were being grown and tested by several nurserymen (cited by Hatton 1917). In England there were at least five rootstocks with the name Paradise, but with different characteristics. By the late 1800s, names for these dwarf rootstocks included “Paradise,” “French Paradise,” ‘English Paradise,” and “Doucin.” Bunyard (1920) attempted to trace the history of Paradise rootstock and explained the confusion. He felt that there were several dwarfing rootstocks that were collectively known as “Paradise.” He divided this collection into two groups: “French Paradise” was the true dwarf and “Broadleaf Paradise” was not the true dwarf, but was probably a dwarf seedling that rooted easily from cuttings. “Doucin” was one of these dwarfing rootstocks that rooted from cuttings. As rootstock mixtures was suspected to be at least one of the factors contributing to tree variation in the orchard, in 1912 R. Wellington and R.G. Hatton developed a rootstock collection at East Malling, England for evaluation and categorization. Some 71 collections of Paradise were obtained from 35 nurseries in England, Holland, France, and Germany. In most

cases, Wellington and Hatton received 12 plants from each nursery. Nursery observations of plant phenology, leaf and bark characteristics, and shoot length during the first year indicated that there was variation within and between nurseries. Over the next few years, additional vegetative characteristics, such as the presence of burrknots and root suckers, date of leaf abscission, tree size, fruiting, fruit characteristics and growth habit allowed segregation of the samples into nine types, designated as Type I to Type IX, later designated as (East Malling) EM I–EM IX. It was suspected that the confusion in the trade had resulted from mislabeling in the nursery and the development of sports from mutations. There was also variation among Crab rootstocks, probably due to different parentage (Hatton 1920a). To evaluate the influence of these rootstocks on scion growth and productivity, the first apple rootstock trial was established at East Malling in 1919, with the scion cultivar ‘Lane’s Prince Albert’ grafted onto 16 clonal rootstocks. In 1920, additional trials were established on four different soil types. Over the next 15 years Hatton (1927, 1935) summarized the results of these trials. First, he noted that the rootstock exhibited a greater influence on the scion than had previously been thought. Trees on EM IX and EM VIII were very small, but the other 14 rootstocks created a continuum of tree sizes, and trees on EM XII were the largest. In addition to tree size, rootstocks also influenced flowering and fruit set, and there was a negative relationship between tree vigor and cropping. However, two rootstocks of similar vigor sometimes differed in cropping. Hatton also noted that trees on EM IX produced larger fruit than trees on EM I. Although the science of statistics was quite young, Hatton used standard deviations to estimate the number of singletreereplicates that would be required in future experiments to detect differences between two rootstocks. He estimated that nine replications would be needed to detect a 30% difference in total wood growth, which is similar to recent estimates. Using data from multilocation rootstock trials, Marini (unpublished results) estimated that 10 replications are needed to detect a 23% difference in trunk crosssectional area at the 5% level of significance. The identification of a set of rootstocks capable of inducing such a wide range of characters on the scion allowed researchers to use dwarfing rootstocks as a tool to study various aspects of apple tree growth and physiology. In the 1923 East Malling Annual Report there was mention of 17 trialacres (6.9 ha) of apple trees of 11 cultivars on selected rootstocks. They learned that the rootstock effect was similar with different cultivars on a given soil. Recently, results from a multilocation trial in North America verified the lack of a rootstock × cultivar interaction (Autio et al. 2001). During the 1970s the nomenclature for EM rootstocks changed to M, and Arabic numbers replaced Roman numerals. For example, over the years Type IX changed to EM IX, to M IX, and is currently M.9. During the 1950s to 1970s, the Malling stocks were heattreated to eliminate known viruses, with the designation East MallingLong Ashton (EMLA), so the virusfree M.9 is designated M.9 EMLA. The nontreated rootstocks sometimes referred to as “dirty,” were usually slightly less vigorous. To eliminate confusion, in this chapter the designations M. (Malling) and MM. (MallingMerton) will be used while discussing rootstocks released from the first two programs at East Malling. After classifying the known rootstocks, a joint breeding program was initiated by M.B. Crane

of the John Innes Horticultural Institute at Merton and H.M. Tydeman at East Malling (Preston 1956). The resulting MallingMerton (MM) rootstocks were bred for resistance to woolly apple aphid (WAA; Eriosoma lanigerum), which was a common problem in North America and especially in New Zealand. The East Malling clones were crossed with the WAA resistant ‘Northern Spy’ and evaluated for WAAresistance and propagation characteristics (Le Pelley 1927). Crossing ‘Northern Spy’ with susceptible cultivars produced resistant progeny, and ‘Northern Spy’ was heterozygous for the trait. This was one of the first truly interdisciplinary fruit research programs, involving both pomologists and entomologists. Of the resulting 3500 seedlings, 15 were selected and numbered MallingMerton 101–115, but later were renamed MM.101–115. The resulting MM series was described by Preston (1953), and details of their behavior in the nursery, along with photographs of leaves, buds and stems, were published by Tydeman (1952). The clones varied in vigor, but none was as dwarfing as M.9. The clones were then tested abroad (Preston 1955). A second set of new clones was bred by Tydeman at East Malling to increase the range of vigor and productivity (Tydeman 1933, 1934). The seedlings had M.9 as a common parent and were crossed with other Malling rootstocks. Of the 1000 seedlings, 18 clones were selected for field trial, and an inverse relationship was reported between rootstock vigor and degree of root suckering. Another group of rootstocks was developed by the English Cider Institute (later the Long Ashton Research Station) as desirable rootstocks for cider apples. Little information is available on these rootstocks, but LA G6, LA G7, and LA G8 were imported by American researchers for evaluation (Ritter and Tukey 1959).

B. North America By the 1930s there was interest in the United States for smaller trees for economic reasons and because the turnover rate of commercial apple cultivars was becoming more rapid (Tukey and Brase 1939b; Zeiger and Tukey 1960). Tree variability also hindered apple research in North America. During the 1920s, most rootstock trials in North America utilized clonal material imported from the East Malling Research Station (Shaw 1935). In 1920, R.D. Anthony attempted to reduce treetotree variation in his research orchard by importing M.12 for a nutrition study in Pennsylvania (Anthony and Thomas 1923; Anthony 1927). Several years later, additional Malling stocks were imported for orchard trials in Massachusetts, West Virginia, Pennsylvania, Vermont, and New York. In the 1920s, several test orchards on vegetatively propagated roots were established in Pennsylvania (Anthony 1927). In 1927, five apple cultivars on four Malling stocks were planted in trials in Ontario, Canada (Ferree and Carlson 1987). Federal quarantine stopped the importation of rootstocks in 1931, which altered the sources of commercial rootstocks for American nurseries. Seedlings from domestic cultivars replaced French crab rootstocks (Anthony 1946). When the importation of plant material was terminated, there was interest in producing clonal rootstocks in North America, and Tukey and Brase (1935) reported on methods for propagating clonal rootstocks and also described nursery characteristics of M.1–M.16; M.10–M.16 were unnamed (Tukey and Brase 1939a). Based on observations of sixyearold trees, Sudds (1939) concluded that M.2 was worthy of further trial, but M.9 had no commercial potential except for certain special

purposes. In 1959, Ritter and Tukey (1959) summarized American experiences with Malling and Malling Merton rootstocks and with 75 scion–rootstock combinations, including some Long Ashton rootstocks. Dwarfing rootstocks were not commercially important in North America until the mid1990s. In fact, pomology text books written before 1960 did not even include the term “rootstock” in their indices. During the 1960s through the 1980s, researchers and commercial growers attempted to grow apples on dwarfing rootstocks with varying degrees of success. Commercial growers had experience with seedling or vigorous clonal rootstocks (mostly M.4, M.7, MM.111, and MM.106) that did not require tree support. They wanted to grow freestanding or nonsupported trees because most felt that tree support could not be economically justified. When trees on dwarfing rootstocks were not supported, tree loss was high due to trees falling over, and the short trees caused by leader bending produced low yields. Commercial use of dwarf rootstocks was limited to lowtrellis systems, primarily for pickyourown operations. By the early 1980s, M.9 had been an important rootstock in Europe for more than 40 years (Vyvyan 1955; Webster 1984), whereas most North American apple growers preferred free standing central leader trees on vigorous seedling rootstocks or the semidwarf M.7, MM.106, and MM.111. Resistance of North American growers to adopt dwarfing rootstocks probably occurred because, unlike in Europe, land and labor were relatively plentiful and inexpensive. Although M.9 was the most popular rootstock in Europe, it had weaknesses, including a lack of winter hardiness, it was difficult to propagate, and the trees required support. In an attempt to overcome some of these problems, European nurserymen and researchers selected clones of M.9, with varying effects on the scion. In North America, dwarfing rootstocks did not become important until the mid1990s when research from a multilocation trial coordinated by the NC140 project showed that the vertical axis system (Lespinasse 1980), utilizing dwarf rootstocks, was more profitable than the traditional central leader using semidwarf rootstocks (Marini et al. 2001c; Marini and Barden 2004). By 2005, the North American apple industry was well on the way to transitioning to intensive orchards utilizing the principles of the vertical axis system with dwarfing rootstocks, such as M.9, B.9, O.3, and M.26.

III. ROOTSTOCK–SCION INTERACTIONS Interaction between the scion and rootstock genotypes is complex because the root system provides the scion with water, hormones and nutrients, whereas the root requires assimilates, hormones, and other compounds from the scion. Rootstock can have profound effects on the scion, such as precociousness, flowering intensity and cropping of young trees, as well as tree size, performance in various soils, and tolerance to biotic and abiotic stresses. Recently, grafting has been used to improve productivity in herbaceous plants, and this has stimulated research activity to understand mechanisms underlying the phenotypic variability resulting from rootstock × scion × environment

interactions (Albacete et al. 2015). During the past decade considerable research has been carried out on the interaction of scion and rootstock in vegetable crops. The physical and biochemical (MarínezBallestra et al. 2010), hormonal crosstalk between the two plant parts (Aloni et al. 2010), and the molecular aspects of these interactions (Kanehira et al. 2010; Koepke and Dhingra 2013) have been reviewed for herbaceous plants. At the molecular level, scion–rootstock interactions may be similar for woody and herbaceous plants, but the discussion here will concentrate on research with apple.

A. Influence of Rootstock Roots versus Rootstock Stems Before rootstock research was stimulated by the classification of the Malling rootstocks, the scion was thought to have a greater influence than the rootstock on the growth and productivity of the tree. Hatton (1920a) was the first to report that the rootstock influenced tree vigor and precocity, and these observations were confirmed by Tydeman (1928). Hatton also found a negative relationship between vegetative vigor and cropping, but rootstocks of a similar vigor sometimes varied in their effects on cropping, flowering, and fruit set. There was disagreement on how much the scion affected the roots, and how much of the scion response was due to the roots versus the rootstock stem piece. Much of this confusion was because American researchers were working with seedling rootstocks and English researchers were working with the newly classified clonal Malling rootstocks. Roberts (1929, 1931a,b) found that the scion greatly influenced the seedling roots when grafted onto pieces of 1yearold root, but there was little scion effect when budded onto the stem of the seedling rootstock. Roberts concluded that the rootstock influence on the scion came from the stem of the rootstock rather than the roots, and this was verified by Grubb (1939), who reported that the length of a dwarfing interstem increased the dwarfing effect on the scion. However, dwarfing rootstocks influenced scion growth even when the scion was grafted onto a piece of root in the absence of the rootstock stem (Hatton 1931). The scion cultivar had little influence on root depth, and the proportion of fine roots versus coarse roots was affected primarily by rootstock (Hatton et al. 1924). Shallowplanted trees on dwarf rootstocks grew slower than deepplanted trees (Hatton et al. 1924). Using dwarfing rootstocks as interstems to study dwarfing, Vyvyan (1938) and Knight (1927) found that roots were most important, but the rootstock stem also contributed to dwarfing and the rootstock had a greater effect on the scion when used as a rootstock rather than as an interstem. The interstem did not affect the morphology of the root system. Inarching of vigorous rootstocks to trees on dwarfing rootstocks also enhanced the vigor of the trees (Hearman et al. 1937). As late as the 1950s the relative importance of the scion and rootstock on scion growth characteristics was still being debated. To study the relative influence of the rootstock and scion, Vyvyan (1955) used three rootstock genotypes as the scion and rootstock in reciprocal grafts to obtain nine scion/rootstock combinations. He periodically harvested trees over three years. The ratio of the mean weight of trees with reciprocal “unlike” unions (M.9/M.4 and M.4/M.9) to that of “like” unions (M.9/M.9 and M.4/M.4) was consistently about 0.9, and deviation from unity (1.0) was never significant, indicating that the relative growth rates of composite trees are primarily controlled by the scion genotype. However, rootstock had a

greater influence on scion genotype than the scion had on the rootstock, because the size of the composite tree was more similar to the genotype used as the rootstock. Studying the influence of the scion on the root system is difficult and requires the complete excavations of mature trees (Rogers and Vyvyan 1928; Rogers and Vyvyan 1934; Beakbane and De Wet 1935). In general, when budded to scion cultivars with different growth characteristics, the morphology of the root systems of Malling clones remained constant. Additionally, for rootstocks varying in vigor, the stem:root weight ratio remained constant for a given soil type. After nearly 35 years of work, the general conclusion was that for clonal rootstocks, the rootstock root has the greatest influence on scion growth and precocity, but the rootstock stem enhanced the influence of the root and they can also be slightly modified by the scion. Using dwarfing rootstocks as interstems or bark rings can also influence the scion, but to a lesser degree than when the scion is budded onto the rootstock.

B. Influence of Rootstock on Scion Growth Before there was much research on rootstocks, dwarfing rootstocks were thought to influence tree vigor by starving the scion of mineral nutrients obtained from the soil, or by partially inhibiting the translocation of photosynthates (Beakbane 1956). Root systems of dwarf rootstocks were thought to be shallower and to occupy a smaller volume of soil relative to their canopy size. Rogers and Vyvyan (1928, 1934) found that trees on the vigorous M.1 had more upright branches and a root system with greater spread than trees on M.9, while M.9 tended to have a more onesided root system. However, trees on M.9 had roots that were deeper, with about 70% of the roots below 33 cm, compared to only 27% for M.1. The total root mass was about twice as great for M.1, the fruit:root ratio was twice as high for M.9, but the ratio of stem:root was similar for both rootstocks, as was the proportion of fine roots compared to coarse roots. For a given scion cultivar on a given soil, the stem:root ratio was similar for rootstocks varying in vigor, and the root systems were similar regardless of the scion for trees that were rootgrafted onto pieceroots or stemgrafted (Vyvyan 1934; Rogers and Vyvyan 1934). Later, Rao and Berry (1940) found that carbohydrate concentrations were higher in scions on dwarfing than on vigorous rootstocks. Rootstock can also influence the growth habit of the scion. Seleznyova et al. (2003) found that ‘Royal Gala’ grafted on M.9 had fewer nodes per shoot, and that the initiation of new metamers in scions grafted onto M.9 ceased soon after bud break; the reduced number of nodes resulted in fewer axillary annual shoots during the year, and this growth pattern accumulated. Although dwarfing rootstocks have been used in commercial apple production for decades, the mechanisms by which they affect various aspects of tree development are still being studied. Until recently, pomologists have used trunk crosssectional area and branch crosssectional area (Westwood and Roberts 1970; Moore 1978) to characterize the influence of rootstock on tree growth because they are related to shoot length (Hirst and Ferree 1995) and above ground tree mass (Barden and Marini 2001b). As trees and branches aged, the number of nodes and length of annual shoots declined, but a rootstock may affect some cultivars more than

others (Costes and Garcia 2001). As trees aged over a 5year period, the number of laterals per annual shoot declined but was similar for trees on M.7 and M.9. However, the number of long laterals and the crosssectional area of annual shoots were always greater for M.7, and the number of short laterals was always greater for M.9. During the first year after grafting, dwarfing rootstocks suppressed tree size by altering the development of axillary meristems, which affected the type of growth that occurred the following season. Compared to the vigorous M.793 during the first season, M.9 decreased the number, length and node number of sylleptic shoots and/or increased the proportion of floral buds the second season. In another study, trees with many flower buds had fewer and shorter vegetative shoots the following year (Seleznyova et al. 2008; Foster et al. 2014). Plant architecture is a relatively new tool for evaluating genetically controlled plant growth (Hallé et al. 1978; Kenis and Keulemans 2007). This tool utilizes the metamer as the basic unit of plant structure, consisting of a node, a leaf, an axillary bud, and subtending internodes (White 1979). The addition of metamers results in shoot (axis) extension. A group in New Zealand is using architectural analysis to study the influence of rootstocks and interstems on tree structural development and vigor (Selenznyova et al. 2003, 2008). M.9 reduced vigor of vegetative annual shoots and induced an earlier transition to flowering by increasing the number of less vigorous floral axillary annual shoots, resulting in fewer annual shoots with extension growth units along the trunks. The vegetative growth units on the M.9 rootstock had fewer nodes and shorter internodes than trees on MM.106. The number of extension growth units, vegetative spurs and fruiting spurs per annual shoot changed over a 3year period, but they were not affected by the rootstock (Selenznyova et al. 2003). Information from these studies may be useful for rootstock breeders to evaluate the dwarfing potential and precociousness of rootstock selections during the first one or two seasons after grafting. Recent research has demonstrated that effects of rootstock on tree architecture are also influenced by environmental conditions. Shoot length, and the numbers of nodes and sylleptic shoots of 1 yearold trees of a given rootstock–scion combination varied when grown at two locations in New Zealand (Seleznyova et al. 2008). In another study, ‘Royal Gala’ was grafted onto M.27, M.9 and M.793, and grown in containers at three locations, with the experiment being repeated for two years to obtain a range of environmental conditions (Foster et al. 2016). Total growth and aspects of the canopy architecture were influenced by rootstock, temperature, and wind, but there was little interaction between rootstock and location. It was concluded that early growth cessation on dwarfing rootstocks was primarily under genetic control, whereas other traits, such as sylleptic shoot growth, flowering and final dry weight, were controlled by the combination of rootstock and growing conditions. The type of scion bud (vegetative versus floral) strongly affected growth; trees developing from vegetative buds had final dry weight 15–45% greater than for trees developing from a floral scion bud. The rootstockinduced effect that was most consistent across locations and years was that dwarfing rootstocks induced an earlier cessation of the primary growth axis. Data reported by various researchers are often difficult to compare and interpret due to differences in methods. Care must be taken to standardize controls, propagation methods, and rootstock and interstem lengths, planting depth and environmental conditions, because these

factors can influence the growth of the scion (Wertheim 1998). Some researchers compared ownrooted trees to trees on dwarfing rootstocks (Lauri et al. 2006), but the lack of a graft union may have altered the growth of the ownrooted trees. Many rootstock experiments aiming to compare rootstock effects on scion architecture, flowering and cropping do not use appropriate controls. The preferred control would be an ownrooted scion genotype grafted with the same genotype as a scion. Since dwarfing rootstocks initially modify scion architecture in either the first (Rao and Berry 1940; van Hooijdonk et al. 2010) or the second (Tukey and Brase 1943; Seleznyova et al. 2008) year of growth, it would be advantageous for rootstock physiologists and breeders to develop protocols for testing new rootstock selections for size control and precocity.

C. Mechanisms of Stock–Scion Interaction Over the years, several theories have been suggested to explain the influence of rootstocks on the growth and development of the scion. These theories involve root system size, hormones and other compounds, carbohydrate distribution, hydraulic conductivity and water relations, the anatomy of the graft union, and mineral uptake and allocation. Many of these proposals have been reviewed (Rogers and Beakbane 1957; Tubbs 1973; Lockard and Schneider 1981; Jones 1986), but no hypothesis has totally explained the observed rootstock effects on scion growth and cropping. In a recent review, Koepke and Dhingra (2013) cited a review by Webster (2004), where the influences of rootstock on apple tree growth were discussed along with the potential mechanisms that were essentially the same as proposed 40 years earlier, indicating that little progress had been made towards understanding how rootstocks influence the scion. In many of the studies intended to test these hypotheses, variables such as water relations, assimilates or hormones were measured above and below the graft union, but tree growth was often not measured. In future experiments, some aspects of growth should be measured. One growth variable that is consistently affected by dwarfing rootstocks is early shoot growth cessation; hence, repeated shoot length measurements to determine the date of growth cessation may be a good indicator of dwarfing. 1. Rootstock Anatomy and Morphology. Beakbane (1952) reviewed the literature on anatomic differences of apple rootstocks and roots. In general, the bark:root ratio of young apple trees was related to mature tree size. The morphology and anatomy of both seedling and Malling clones were not affected by the scion (Vyvyan 1930). Dwarfing rootstocks have higher ratios of bark:wood, but similar ratios of stem:root, have highly parenchymatous xylem and phloem, and contain more living tissue per unit volume of stem and root than vigorous rootstocks (Beakbane 1941; Beakbane and Thompson 1939, 1945). Therefore, the metabolic activity of dwarfing rootstocks per unit volume of tissue is likely greater than for vigorous rootstocks (Rogers and Beakbane 1957). The latter authors hypothesized that a relatively greater proportion of total assimilates would be utilized by the roots of dwarfing rootstocks than vigorous rootstocks, but Hassan (1953, cited by Rogers and Beakbane 1957) found that the rate of respiration per unit of living tissue was higher for vigorous than for dwarfing rootstocks.

The relative proportions of storage to conducting and strengthening tissues in stems and roots were related to the precocity and productivity of the tree. Compared to vigorous rootstocks, dwarfing rootstocks had more living and less conducting and strengthening tissue per unit of root volume. The size of conducting elements was usually positively related to vigor (Beakbane 1941). Therefore, different rootstocks likely vary in their capacity to absorb, store, and utilize nutrients. To study phloem transport in dwarf apple trees, Dickson and Samuels (1956) supplied radioactive phosphorus through a leaf petiole of 3yearold trees that had been dwarfed by inverting two rings of bark the previous year. The isotope accumulated above the bark inversion, and phloem transport to the roots was reduced. The authors then performed a similar experiment with 5yearold trees with a M.9 interstem, and the isotope accumulated in the interstem. It was hypothesized that both phloem and auxin transport are retarded in the dwarfing interstem. It is difficult to interpret these results because the authors did not report at what point during the growing season the experiment was conducted. Translocation of the isotope would likely be influenced by the growth stage of the tree (early versus lateseason). In general, vigorous rootstocks have larger xylem vessels, and this supports the concept that dwarf rootstocks may reduce the flow of assimilates and other compounds between the roots and scion. The effect of the xylem on water flow is discussed in the section on water relations (see below). Based on the research with inverted bark rings and phloemblocking viruses, Rogers and Beakbane (1957) concluded that it seems more likely that phloem transport, rather than xylem transport, would be a limiting factor in rootstocks and interstems. Simons and Chu (1980) reported that the outer bark of dwarfing rootstocks was thicker than that of semidwarfing rootstocks, and 60% of this bark was nonfunctional phloem. Also, vascular tissues that developed between the rootstock and scion were arranged in a swirling pattern and became necrotic during growth of the tree. More recently, Soumelidou et al. (1994) found that the xylem of M.9 linking the bud to the rootstock contained fewer and smaller vessels than for MM.106, and this difference could reduce supplies of water and dissolved nutrients to the scion. These authors hypothesized that hormones, possibly auxin, might be involved in the differentiation of callus cells into vessel elements in the callus between the bud and rootstock. More recent research with herbaceous plants seems to support the role of auxin in graft union formation. At the scion–rootstock interface of a grafted plant, cells expand and divide to form callus, which is undifferentiated stemcelllike tissue (Sugimoto 2010). Callus tissues surrounding the cut then differentiate to phloem and xylem, and vascular strands connect the scion and rootstock. Eventually, a common cell wall forms between the scion and rootstock and plasmodesmata form across the cell wall. Callus proliferation and cell differentiation continue such that eventually vascular connections are reestablished (Melnyk 2016). Yin et al. (2012) proposed a model for graftunion developmental stages, and suggested that there is communication between the scion and rootstock resulting in auxin accumulation; the auxin then regulates cell division and differentiation. For apple, Soumelidou et al. (1994) reported that the vessels in M.9 in the first year were larger than for MM.106, but smaller the following year. The authors suggested that the difference was due to the level of auxin reaching the bud. Auxin had difficulty crossing the graft interface, and the local accumulation of auxin resulted in highly parenchymatous, abnormal xylem in the rootstock

below the bud. The level of auxin was inadequate for true xylogenesis, but the basipetal flow of auxin is unhindered. 2. Assimilate Production and Distribution. During autumn, photoassimilates are translocated out of the leaves to the roots and other woody organs, where they are converted to starch; these reserve carbohydrates are utilized for spring growth (Priestley 1960). Lateseason modifications in the carbohydrate status of the tree can affect growth the following spring (Abusrewil et al. 1983). Based on experiments with containergrown fruiting and nonfruiting trees, Avery (1970) hypothesized that trees on dwarfing rootstocks are smaller because they produce fewer growing points which continue extension for a shorter period, and are unable to fully utilize photosynthates for growth because of a limiting rate of growth by the root system. The hypothesis that lateseason differences in stored assimilates influence vegetative growth the following season may not be valid, because summer pruning that was severe enough to suppress lateseason trunk and root growth did not influence the shoot growth of fieldgrown or containergrown trees the following season (Marini and Barden 1982a,b). Research results on the influence of rootstocks on photosynthesis are conflicting. Gregory (1957) reported higher net assimilation rates for vigorous rootstocks over the whole season compared to dwarfing rootstocks, primarily because photosynthetic activity continued later in the season for vigorous stocks. Ruck and Bolas (1956) grew four rootstocks of varying vigor in three types of soilless media with varying pH and found that the vigorous Crab C had higher photosynthetic rates than M.9 at low nitrogen, but at higher nitrogen levels the differences were less pronounced. Ferree and Barden (1971) reported that net photosynthesis (Pn) of containergrown trees on seedling rootstocks was higher than for trees on MM.106. However, in a similar study, Barden and Ferree (1979) found that Pn and dark respiration were unaffected by rootstocks. Schechter et al. (1991) found that Pn was higher for fieldgrown cropping trees on vigorous than on dwarfing rootstocks. Baugher et al. (1994) reported that Pn was greater for mature trees on M.7 and MM.111 than on M.9, but Fallahi et al. (2002) reported that mature trees on M.9 had higher Pn and transpiration rates than trees on M.7. These results are difficult to interpret, because crop density was not reported and photosynthesis is related to crop load (Palmer et al. 1997). Working with 25 rootstocks, Tworkoski and Fazio (2011) reported that photosynthesis and transpiration of 1year ‘Fuji’ on dwarfing rootstocks were lower than for trees on semivigorous and vigorous rootstocks. Šabajevienė et al. (2006) measured leaf pigments of the cultivar ‘Auksis’ on 12 rootstocks growing in the field. Although the results varied somewhat from year to year, chlorophyll and carotenoid concentrations were higher for trees on M.9 and York 9 than for trees on M.26 and B.9. These results may have varied due to differences in crop load or water relations, but the methods used were reported in toolittle detail to interpret the data. Since gas exchange results were so variable, and even when differences were significant they were not very large, at this point there are too few supporting data available to conclude that rootstock vigor can be explained by differences in carbon assimilation. A detailed experiment comparing wholetree gas exchange measurements for trees on rootstocks of varying vigor is still needed to determine

if vegetative vigor may be related to carbon assimilation. Gas exchange measurements for this type of experiment should be made periodically throughout the season to take into account any early cessation of shoot growth on dwarfing rootstocks. The careful management of crop load and water relations would also be critical for determining the influence of rootstock on gas exchange. Several studies tested the hypothesis that the greater volume of living tissue in dwarf rootstocks may lead to greater utilization of total metabolites in those tissues. The season following a heavy crop, the vegetative scion growth of ‘Lane’s Prince Albert’ on the vigorous M.4 was 91% that of the previous season compared to only 47% on M.9 (Rogers and Booth 1964). Stutte et al. (1994) reported that rootstock affected concentrations of starch, sucrose, and sorbitol at harvest, and concentrations of starch and sorbitol at leaf fall. Roots of trees on MM.111 had greater starch and less sorbitol and sucrose at harvest than roots of trees on M.9. Starch concentrations in young and mature roots of trees on MM.111 were higher than for trees on M.9, and trees on MM.111 had greater shoot and branch starch and less root sorbitol and sucrose at leaf fall. Brown et al. (1985) used two scion cultivars on two rootstocks to measure carbohydrates of above and belowground (roots plus rootstock shank) tissues of young containergrown apple trees throughout the season. For both rootstocks and cultivars, the dry weight and sorbitol content of above and belowground sections of the tree increased until bud set, and declined during the dormant period. For the aboveground portion of the tree, MM.111 had higher starch, sorbitol and soluble sugar concentrations than trees on M.9 from midseason until budbreak the following spring; however differences in the belowground portion of the tree occurred only during the early winter. Dwarf rootstocks had higher root sucrose, glucose, and fructose concentrations during dormancy than semidwarfing rootstocks under greenhouse conditions, but not when trees were grown in the field (Saavedra 1991). When M.9 was used as an interstock bridge, starch and soluble sugar concentrations increased in the bark above the graft, but declined in the bark below the graft; however, these changes may have been due to the girdling effect of grafting (Samad et al. 1999). Taken together, these reports support the concept that carbohydrate concentrations are higher in the aboveground organs of trees on vigorous rootstocks during the late fall and during dormancy, and availability of reserves may influence vegetative growth the following spring. The influence of stored carbohydrates on spring growth is still poorly understood. Most researchers have reported the concentration of various carbohydrates, but the total amount of carbohydrate may be more important than the concentration at any one time. As carbohydrates are converted to sugars and utilized for spring growth, additional starch may be converted to sugars to maintain some critical concentration, above which growth is not limited. Additional research is needed to determine changes in carbohydrates in various parts of the tree throughout the season. 3. Uptake and Distribution of Nutrients. Since the roots are responsible for absorbing minerals and nutrients for the plant, early rootstock researchers hypothesized that one mechanism by which dwarfing rootstocks exerted their influence on the scion might be by altering the uptake and/or translocation of organic or inorganic materials. It seems likely that the movement of metabolites in the phloem may depend

on the size of the sieve tubes, and the relative capacity of the phloem for conducting and storing may be involved in plant vigor (Beakbane 1956). Modifications of the sieve tube size or number occurring at the graft union might hinder the downward flow of organic nutrients and lead to carbohydrate accumulations in the scion. Since dwarfing rootstocks have smaller vessels, movement of water and metabolites in the xylem and phloem would likely be impeded. Work by Rogers and Vyvyan (1934) and Pearse (1940) showed that the size of the root system and uptake of nutrients were not limiting factors in dwarf trees. Trees on M.9 had higher bark:wood ratio and higher Ca, but lower K in bark, wood and leaves compared to more vigorous rootstocks, while P and Mg concentrations were positively correlated with rootstock vigor (Vaidya 1938). Ruck and Bolas (1956) grew nongrafted Crab C and M.9 rootstocks with a range of nitrogen (N) levels, and found that Crab C produced the largest trees. At the lowest N concentration, the mean dry weight of Crab C trees was 31% higher than for M.9, but at the highest N concentration the difference between rootstocks was only 5%. These authors concluded that Crab C absorbed N more efficiently than M.9, especially when N is limited. However, greater amounts of carbohydrates, N, and total amino acids were found in the bark tissue of nongrafted cuttings of M.9 than the more vigorous M.16 (Martin and Williams 1967). Bukovac et al. (1958) and Jones (1976) suggested that dwarfing was caused by reduced solute translocation through the rootstock to the scion. Dana et al. (1963) also found that trees with dwarfing interstems accumulated N more slowly in scion leaves than vigorous rootstocks, but water loss was not influenced by interstems. Jones (1971) collected sap from the xylem of decapitated mature apple trees, and concluded that nutrient concentrations in the scion alone would not explain the dwarfing effect because the volume of exudate was similar for different rootstocks, and transpiration of the scions was not affected by rootstock (Knight 1925). Concentrations of N, P, and K were higher in the sap of scions on vigorous rootstocks, and there was a greater reduction of these nutrients above the interstem piece for dwarfing rootstocks. Young containergrown ‘Smoothee Golden Delicious’ trees on M.9 had higher leaf concentrations of N, Ca, Mg, P, and Zn, but lower concentrations of B and Fe than trees on MM.111 (Zeller et al. 1991). Jones (1971) hypothesized that interstocks may simply restrict the movement of nutrients, causing an accumulation near the basal end, or that solutes may be removed from the sap as it moves through the xylem. Jones (1971, 1974) proposed that M.9 interstems diminished the vigor of scions by reducing nutrient concentrations in the xylem sap stream. The same author later found that sap flowing from above an interstock had lower concentrations of N, P, K, Ca, and Mg than from the rootstock levels, and these differences increased with the dwarfing effect of the interstocks. Nutrient concentrations were higher below the interstocks, but were reduced in the sap flowing from the interstocks into the scion. Analyses of sap from above, midway, and below interstems indicated that the changes in concentration were produced in, or close to, the graft union between scion and interstem (Jones 1976). However, the concentration of amino acids was similar above and below the interstem (Jones and Pate 1976). Contradictions in the literature may be due to the fact that some researchers used nongrafted rootstocks, whereas others used grafted trees or interstock trees, and the rootstocks being compared often differed. In addition, experiments with containergrown trees may not reflect conditions in the field. Roots of fieldgrown trees are not confined, they are exposed to moderate temperatures, and

soilborne communities of microorganisms may interact with the roots and influence the uptake of water and nutrients. An area of research where there is a dearth of information for fruit trees, but is currently being investigated by agronomists, is the role of root architecture in water and nutrient acquisition. In maize and beans, the depth and spread of roots can affect the uptake of nitrogen, phosphorus and water; cultivar performance in varying soil types appears to be related to the characteristics of the root system (Lynch and Brown 2012). Although root systems of some of the Malling clones were described during the first half of the 20th century, similar information is lacking for newer rootstocks. Detailed characterization of rootstock root systems and the influence of these characteristics on uptake of water and nutrients may be a productive area for future research. 4. Influence on Leaf Nutrient Concentrations. During the 1940s and 1950s, apple orchard nutritional programs were developed for widely spaced trees on vigorous rootstocks. As commercial growers in North America started adopting dwarfing rootstocks, there was a need to establish nutritional requirements for trees on dwarfing rootstocks and interstems. Tukey et al. (1962) was among the first to report differences in leaf nutrient concentrations of apple leaves due to rootstocks and interstems. However, results from different cultivar/orchard combinations were not consistent. Whitfield (1963) reported higher leaf concentrations of Ca and Mg for trees on M.9 than on M.7, and trees on M.2 had higher levels of Ca, Mg, and P than trees on M.16. However, other researchers found no influence of rootstock (Dzamic et al. 1980) or interstems (Bould and Campbell 1970) on leaf nutrient levels. In the presence of viruses, trees on MM.106 had higher leaf Mg and Ca than on other rootstocks. High leaf concentrations of Mg for trees on M.26 were reported from several rootstock experiments (Fallahi et al. 1984, 1998; Rom et al. 1991; Fallahi and Simons 1993; Fallahi and Mohan 2000). One of the challenges associated with determining the influence of rootstock on apple tree mineral nutrition is that rootstocks influence crop load, and crop load can influence leaf nutrient levels (Sadowski et al. 1995). In most cases, crop load was not carefully controlled or adequately accounted for in the statistical analyses used in the research outlined above. In future experiments, researchers must carefully control crop load to compare rootstock effects on various aspects of tree growth and physiology. A number of investigators have reported leaf tissue analyses from rootstock trials, but these results are difficult to interpret and compare because they used different scion cultivars and rootstocks, reported results from varying numbers of seasons (Abdella et al. 1982), and crop loads varied or sometimes were not reported. Sometimes, data for two or more seasons were averaged (Poling and Oberly 1970). Over a 4year period, Fallahi et al. (2002) reported that the effect of rootstock (M.7, M.26, and M.9) on ‘Fuji’ leaf nutrient concentrations varied with season, but trees on M.7 frequently had greater K and trees on M.26 always had greater leaf Mg than trees on other rootstocks. In another study with ‘Gala,’ results were also inconsistent over years, but trees on M.9 tended to have higher leaf concentrations of Ca than trees on MM.111 (Fallahi and Mohan 2000). Results from a multilocation rootstock trial for 8 years

were not very consistent over locations or years within locations, but leaf Ca levels tended to be lower for trees on M.7 than on M.9 or M.26 (Rom et al. 1991). In general, it seems that rootstocks have little influence on N and K, and have a greater effect on Ca and Mg, and these two mineral elements have been implicated in bitter pit development. Taken as a whole, results from various experiments support the conclusions of Lockard and Schnieder (1981), that rootstock effects on tree size and precocity are not due to differences in leaf mineral concentration. Although rootstocks do influence leaf nutrient levels in some years, at present results are too variable to adjust orchard nutritional programs for different rootstocks. Less information is available on fruit nutrient concentrations, which may influence fruit storability. Rootstocks and crop load can affect fruit nutritional levels, fruit quality, and storability. An accurate evaluation of the effect of rootstock on leaf and fruit nutrient levels will require the comparison of rootstocks with a range of crop loads over more than one season, and possibly with more than one cultivar and at multiple locations. Such an experiment would be particularly useful with a cultivar such as ‘Honeycrisp,’ which is very susceptible to bitter pit. 5. Hormones. Trees tend to maintain a constant top:root ratio. Each scion/rootstock combination has a specific top:root ratio, and attempts to alter this ratio by pruning the top or roots result in the plant changing its growth pattern until the ratio is reestablished. Maintenance of the ratio requires interaction between the scion and rootstock. During the 1920s, nurseries had difficulty in producing uniform trees on seedling rootstocks, and research was directed at determining the influence of seed source and scion source on growth of the root system and scion (Maney 1930). The latter author found that not only did the rootstock affect scion growth, but the scion also influenced root growth. There is substantial evidence that rootstock effects on apple tree vigor are mediated by endogenous plant hormones. Some of the first evidence for this concept was reported by Martin and Stahly (1967), who found that the bark of actively growing non grafted shoots of M.9 had less growthpromoting compounds, and more growthinhibiting compounds than bark of M.14 shoots. After reviewing the literature on the dwarfing effects of rootstocks, Lockard and Schneider (1981) proposed that the dwarfing effect of apple rootstocks was due to hormones in both the scion and rootstock. They presented evidence that the primary signal from the shoot to the root was phloemtransported auxin, and the primary signal from the root to the shoot was xylem translocated cytokinin. Based on a series of experiments involving bark grafts, it was proposed that auxin produced in the shoot tip is translocated down the phloem, and the amount reaching the root influences root metabolism and controls the amount and kind of cytokinins synthesized and translocated to the shoot through the xylem. However, few data were available to support this hypothesis and the suggestion was based primarily on observations of bark grafts and work with callus tissue. This hypothesis now appears to be an oversimplification of the role of hormones in explaining the effect of rootstocks on scion growth. Kamboj et al. (1997) reported higher cytokinin concentrations in shoot xylem sap and root pressure exudate from scions on vigorous rootstocks compared to dwarfing rootstocks, greater [3H]indole acetic acid (IAA) movement from the scion to roots of vigorous rootstocks, and the ratio of abscisic acid (ABA)

to IAA in the rootstock bark was negatively related to rootstock vigor. Unfortunately only the radiolabel, rather than IAA, was measured. Kamboj et al. (1999a) found that zeatin was the predominant cytokinin in xylem sap from the dwarfing M.9 and M.27, whereas zeatin riboside was the predominant cytokinin in xylem sap of the semidwarfing MM.106. Cytokinin concentrations from sap collected above and below the graft union in composite trees increased with increasing rootstock vigor. Cytokinin concentrations in the shoot sap of non grafted rootstocks were also positively related to rootstock vigor. Shoot bark of nongrafted M.27 and M.9 had higher concentrations of ABA and higher ratios of ABA:IAA than stems of the more vigorous MM.106 and MM.111 (Kamboj et al. 1999b). Since low concentrations of IAA in the cambial region stimulates the differentiation of cambium to phloem (Aloni 1987), the higher ABA:IAA ratios in dwarfing rootstocks may explain the higher bark:wood ratios that have been reported for dwarfing rootstocks. Researchers in New Zealand studied the role of plant hormones in rootstockinduced scion architecture modification. After grafting ‘Royal Gala’ onto three rootstocks varying in vigor, van Hooijdonk et al. (2006) hypothesized that a reduced root supply of cytokinin to the scion likely controlled bud break, lateral production, and the allocation of growth between plant axes. At the end of the first season, the length and node number of the primary shoot were similar for scions on M.9 and on an ownrooted ‘Royal Gala’ rootstock control, but trees on M.9 had fewer secondary shoots and fewer grew late in the season. In addition, the dry mass of trees on M.9 was less than on the ‘Royal Gala’ rootstock control. During the late season, M.9 had greater concentrations of zeatin riboside and lower concentrations of giberillin (GA)19 in the xylem sap compared to ownrooted trees. It was concluded that dwarfing apple rootstocks may limit rootproduced GA 19 supplied to shoot apices of the scion, where GA19 may be a precursor of bioactive GA1 required for shoot extension growth (van Hooijdonk et al. 2010, 2011). Determining the role of cytokinins and gibberellins in rootstocks is difficult because cytokinin concentrations can change rapidly, especially during the first 15 days after bud break (Tromp and Ovaa 1994). Endogenous cytokinins increase in the spring just before budbreak in shoot xylem sap, and this peak may be responsible for budbreak and branching habit (Cook et al. 2001). Gibberellins are difficult to study because there are several bioactive forms and a number of precursors, and they can be converted back and forth; moreover, bioactive forms are often at concentrations that can be difficult to measure (Yamaguchi 2008). Gibberellic acid insensitive mRNA can move in both upward and downward directions via the graft union within 5 days of grafting (Xu et al. 2010). Michalczuk (2002) measured free and conjugated IAA levels in wood, bark and cambial sap of several mature nongrafted rootstocks varying in vigor. Conjugated IAA in bark and wood tissues was higher than the free form and was not affected by rootstock. The level of free IAA in cambial sap was much higher than in the bark and wood tissues, and was lower for M.9 than for M.26 and MM.106, while IAA in cambial sap tended to decline in M.9 later in the season. Using mature trees, Tworkoski and Miller (2007) found that auxin concentrations were a little lower and ABA was a little higher in apical buds of trees on M.9 and M.7 compared to seedling, but the auxin:cytokinin ratio was nearly twice as high in seedling than the other rootstocks. This suggested that the ratio may be a factor regulating sylleptic budbreak and the growth habit in apple scions, and that rootstock

modified the hormone concentrations in shoot tips. Tworkoski and Fazio (2011) found that concentrations and fluxes of IAA and cytokinin in xylem exudates of containergrown ‘Fuji’ were not affected by rootstock when measured after 30 days of growth, but the concentration of ABA per unit volume of exudate per hour (ABA flux) was greater in dwarfing rootstocks than in more vigorous rootstocks. ABA flux was also negatively correlated with vessel cross sectional area and xylem flow. Scions on dwarfing rootstocks had smaller vessel diameters but greater vessel density, and the resulting total lumen area of vessels in the scion was about 14% greater for vigorous than for dwarfing rootstocks. Hydraulic conductivity was lowest in the most dwarfing rootstocks, suggesting that reduced hydraulic conductivity, caused by an undetermined factor, may lead to water stress and the induction of ABA synthesis. ABA moving up the stems may exert a growthinhibitory effect, but the role of ABA in size controlling effects of apple rootstocks remains to be elucidated. The trees used by Tworkoski and Miller (2007) had diverse growth habits, and were budded onto four rootstocks with a wide range of vigor. Although not stated, the treatment structure appeared to be a factorial of six scion genotypes and four rootstocks, and the presence of interaction was not tested. However, there appears to be an interaction between scion genotype and rootstock because the influence of rootstock on trunk diameter was not consistent for all scion genotypes. If tree vigor was controlled solely by rootstock, then one would expect rootstocks to have a similar effect on tree size, regardless of scion genotype. To elucidate the relative importance of rootstock and scion on tree vigor, additional research is needed, involving reciprocal grafts, and the controls should be rootstocks grafted onto the same rootstock. Lockard and Schneider (1981) concluded that there was little evidence to support a role for gibberellins in apple dwarfing. However, evidence supporting a role for GA was later presented by Richards et al. (1986), who found that M.9 interstems can greatly alter the distribution and metabolism of xylemapplied [ 3H]GA4, with an overall effect of reducing the levels of [3H]GA metabolites arriving in the upper shoot and leaves. These authors suggested that endogenous GAs, possibly associated with other hormones, probably contribute to the dwarfing mechanism. ABAlike activity was negatively related to rootstock vigor in tissues of nongrafted rootstocks and levels of GA 4+7like compounds were positively related to rootstock vigor (Yadava and Lockard 1977). Concentrations of GA3like compounds were positively related to vigor in aboveground tissues, but negatively related to vigor in the roots. It was hypothesized that rootstock vigor may be related to the ratios of cytokinins and gibberellins in the roots and shoots. Tworkoski and Fazio (2016) used rootstocks of varying vigor as scions on rootstocks of varying vigor. They found ABA and its conjugate, ABA glucose ester, were higher in the root and rootstock shank below the graft union, in the scion above the graft, and in the xylem exudate of M.9 than MM.111 rootstock. Elevated ABA and reduced GA were associated with the more dwarfing rootstocks. Since results for greenhouse and fieldgrown trees sometimes differed, it is possible that tree age and environmental stress may interact to affect hormone signals and other sizecontrolling factors. Additionally, containergrown trees in greenhouses, where roots may be restricted and soil

temperatures may be excessive, may not be representative of trees growing in the field. In addition to GA and ABA, other hormones may be involved in sizecontrol. Tworkoski and Fazio (2016) found higher IAA in the rootstock below the graft and in the exudate of MM.111 than for M.9. In reciprocally grafted trees, IAA was higher in the scion above the graft for trees on M.9 than MM.111, but did not differ in rootstock below the graft or the root. It was hypothesized that understanding gene expression associated with hormone metabolism may help explain sizecontrol in rootstocks and assist the selection of rootstocks for size control. Although results are not totally consistent, most of the research supports the concept that rootstock vigor is at least partially mediated by hormones. Since hormones affect tree growth and architecture, it seems likely that rootstockinduced differences in tree growth are related to the interaction of auxin, GA, and cytokinins. High ratios of auxin and possibly GA to cytokinin may enhance scion vigor directly – or possibly indirectly – by altering cell differentiation, leading to smaller vessels that suppress the movement of water and other growthpromoting compounds through the graft union. 6. Graft Union Failure and Tree Anchorage. Some scion/rootstock combinations have weak unions that break in the nursery or orchard during wind storms. For example, ‘Golden Delicious,’ ‘Gala,’ ‘Honeycrisp,’ and ‘Granny Smith’ on M.26 sometimes break at the union. Shaw and Southwick (1944a,b) were among the first to report that several apple cultivars performed poorly in the nursery when grafted onto the clonal rootstock ‘Spy 227.’ In a review of stock–scion interactions as related to dwarfing, Simons (1987) discussed characteristics associated with incompatibility, which included poor growth and unsatisfactory unions, often resulting in breakage. Graft combinations with mechanically weak unions, which often break, were sometimes associated with mechanical obstruction at the union or with abnormal starch distribution. In a review of graft incompatibility, Andrews and Serrano Marquez (1993) defined incompatibility as “…the failure of the graft combination to form a strong union and to remain healthy due to cellular, physiological intolerance resulting from metabolic, developmental, and/or anatomical differences”. They also suggested that incompatibility reduced vascular continuity and the transport of carbohydrates and nutrients across the union, causing these materials to accumulate on either side of the union. Mosse (1962) and Simons (1987) suggested that incompatibility is the primary cause of graft union failure. Some apple cultivars have brittle wood, possibly due to a high proportion of fiber cells that influence flexibility (Simons 1975), and cultivars or rootstocks with brittle wood may be prone to union breakage. Rootstock can also influence the number of fiber cells, the thickness of fiber cell walls, and the proportions of fiber and parenchyma cells (Beakbane and Thompson 1947; Doley 1974). Basedow (2015) compared the xylem tissues of eight scion/rootstock combinations that varied in their tendencies to break, and found that fiber cell wall thickness varied with combination. Weak combinations also had higher percentages of parenchymatous tissue than strong combinations. In addition to anatomic characteristics, biochemical factors may also contribute to weak graft unions. Compounds, such as peroxidases (Feucht et al. 1983), isoperoxidases (Gulen et al.

2002), the glycoside prunasin (Gur et al. 1968), polyphenols (Dos Santos Pereira et al. 2014) and plant hormones, such as auxin and cytokinin (Aloni 1982; Aloni et al. 2010), may be involved in the division, differentiation, and function of cells in the graft union. Recent nursery and field observations indicate that unions of some newer cultivars, such as ‘Cripps Pink’ and ‘Scilate’ on the new rootstocks G.41 and G.935, are also brittle and trees may break in wind storms and while digging trees in the nursery or planting trees in the orchard. Since tree breakage can have significant economic consequences for nurserymen and orchardists, researchers have evaluated methods for determining union strength and flexibility. Ermel et al. (1997, 1999) evaluated histological traits related to different aspects of graft union formation in pear, but there was considerable variation in most traits. Using multivariate analyses, it was possible to discriminate between compatible and incompatible combinations before differences between graft combinations became evident using macroscopic or qualitative microscopic examination. The histological variables responsible for the discrimination were related to bark discontinuity, cambial dysfunction, and starch accumulation in the scion (Ermel et al. 1999). Unions with different vascular connectivity could be distinguished using magnetic resonance imaging (Warmund et al. 1993) and Xray computed tomography (Milien et al. 2012). Laser ablation tomography (Basedow 2015; Chimungu et al. 2015) may be used on newly budded trees to develop threedimensional models of the unions. Biochemical gene expression and activity may also be used to evaluate graft combinations (Gulen et al. 2002; Dos Santos Pereira et al. 2014). The amount of force required to break graft unions has been evaluated for some scion/rootstock combinations. The graft union of ‘Gala’/G.30 was more brittle than ‘Gala’/M.26, and ‘Empire’ on G.30 and CG.41 also had some tree breakage (Robinson et al. 2003). Rehkugler et al. (1979) used a universal testing machine to measure the bending strength of graft unions, and found that 18yearold ‘Golden Delicious’ on M.9 and G.30 could withstand only onethird of the force that caused breakage on vigorous rootstocks. Adams et al. (2017) applied various plant growth regulators to graft unions in the nursery in an attempt to enhance the flexibility of the union. Using a bench testing machine to measure fracture strength, it was found that foliar applications of prohexadionecalcium and benzyl adenine applied to the union in latex paint increased the flexural strength per scion crosssectional area, and the flexibility of the union. To avoid tree breakage problems in the future, routine measurement of graft union strength will likely precede the introduction of new rootstocks into commerce. For scion/rootstock combinations with known union weaknesses, many pomologists recommend support systems with at least three wires so that branches can be tied to wires to prevent trees twisting in the wind. In addition to union breakage, some rootstocks have long been known to be poorly anchored and tend to lean. The term “anchorage” refers to the resistance of trees to lean or fall over. There is a wide range of anchorage among apple rootstocks. Preston (1955) reported that M.9 and M.4 were weak, M.2 was fair, and M.14 had excellent anchorage. MM.104 was better anchored than either M.6 or MM.111. The cause of poor anchorage for dwarfing rootstocks is likely due to asymmetric root distribution or to brittle root systems, possibly due to short fibers (Rogers and Beakbane 1957). Following wind storms, it became obvious that tree leaning was

due to the scion/rootstock combination rather than to the rootstock itself (Marini et al. 2001a). In Maine, the angle of leaning from vertical was greater than 30° for ‘Starkspur Supreme Delicious’ trees on MAC39, B.9, P.1, P.2, P.22, M.26, 16° for M.7, and less than 10° for trees on MAC1 seedling, B.490, and Ant.313. In Massachusetts, no tree leaning was observed for ‘Puritan’ on six rootstocks, whereas the percentage of leaning trees was 0%, 57%, and 100% for ‘Delicious’ trees on MM.106, M.2, and M.7, respectively. Following a severe storm in Virginia, less than 10% of ‘Starkspur Supreme Delicious’ on B.490 and P.22, ‘Golden Delicious’ on M.26 and ‘McIntosh’ on M.26 were leaning, whereas more than 80% of the ‘Starkspur Supreme Delicious’ on M.4 and C.6 were leaning. No ‘Golden Delicious’ or ‘Empire’ trees on M.7 exhibited leaning, whereas 88% of the ‘Triple Red Delicious’ trees on M.7 were leaning. In modern orchards, trees are always supported, so rootstock anchorage will likely be less important (Marini 2001). In Virginia (R.P. Marini, personal observations) and in Michigan (R. Perry, personal communication), vigorous cultivars on M.7 rootstock growing on heavy or clay soils often lean. Investigations in Michigan determined that the loss of anchorage was associated with an asymmetric pattern of roots around the stem shank. Heavy soils appear to exacerbate the situation. 7. Burrknots. Clonal apple rootstocks differ in their tendency to form burrknots (Rom 1973). Burrknots are areas of partially developed adventitious root initials originating in apple tree stem tissue (Rom and Brown 1979). Root primordia form from nondifferentiated parenchyma tissue in the bud and leaf gap areas at each node (Swingle 1927). Burrknots form on the aboveground portion of many apple rootstocks, and on some scion cultivars, such as ‘Gala’ and ‘Empire.’ These areas can enlarge as the tree grows and cause trunk fluting or partial girdling, which can interfere with vascular transport and stunt the tree (Rom and Carlson 1987). Burrknots can also be sites for ovipositioning of dogwood borer (Synthanthedon scitula Harris), feeding for ambrosia beetles (Xylosandrus germanus Blandford), and infection for fire blight (Erwinia amylovora Burrill). Since burrknots are undesirable, rootstocks with good rooting characteristics with little burrknot formation would be preferred. Liners of MM.111 produced more large (>20 mm diameter) burrknots than liners of M.26, M.7, and MM.106 (Rom and Brown 1979). In multilocation NC140 rootstock trials, burrknot severity is usually quantified as the percentage of the rootstock circumference affected. When comparing burrknot data from various rootstock trials, it is obvious that burrknot development varies with site. Soil and climatic conditions, as well as orchard management practices are all confounded in site, so it is impossible to determine which of these factors promote burrknot development. With ‘Gala’ as the scion, burrknot severity on 5year old trees was significantly influenced by rootstock at 11 of 20 locations (Marini et al. 2000). After 10 years, burrknot development was more severe on B.491, M.27 EMLA, MARK, and M.26 EMLA than on six M.9 clones (Marini et al. 2006). In a trial with ‘Golden Delicious’ as the scion, trees on PiAu 514, B.10 and M.26 EMLA had more burrknot development than trees on M.9 NAKBT337, G.11, G.41 and G.935 (Marini et al. 2014). In another trial with ‘Gala’ the severity of burrknots was ranked: M.26 NAKB > M.26 EMLA > B.9 > M.9 NAKBT337 > M.9 RN29 (Autio et al. 2013). In addition, there was a poor relationship between burrknot severity after 5 years and that after

10 years (Marini et al. 2003, 2014). The number of differentsized burrknots and the burrknot density (burrknots per cm2 trunk crosssectional area; TCA) on the ‘Gala’ scion were recorded for eight dwarf rootstocks in North Carolina and Virginia (Marini et al. 2003). For unknown reasons, burrknot development on the scion was greater in North Carolina than in Virginia, but the rankings for the rootstocks were similar except for MARK. At both locations, burrknot density was highest for trees on O.3 and lowest for trees on B.9 and M.27 EMLA, and there was a poor correlation between tree vigor and scion burrknot severity. It is interesting that M.27 EMLA induced few burrknots on the scion, because it usually ranks high for burrknot development on the rootstock shank (Marini et al. 2003). When reviewing burrknot data from several NC140 trials, burrknot development tended to be greatest in British Columbia, North Carolina, and Virginia. Conditions conducive to burrknot development are not known, but may be related to orchard practices, environmental conditions, or soil type. Future experiments aimed at identifying the factors involved in burrknot development are needed.

IV. STRESSES INFLUENCING ROOTSTOCK PERFORMANCE Unlike the scion cultivar, apple rootstocks are exposed to both aboveground and below ground stresses. Until fairly recently, rootstocks were not bred to tolerate stresses other than WAA. In general, the adaptation of rootstocks to various stresses was usually identified during field trials, but occasionally rootstocks were subjected to stresses in controlled experiments. More recently, rootstock breeders have been challenging new rootstock genotypes with various stresses, and also studying plant responses to stresses at the wholeplant and the molecular level.

A. Abiotic Stresses 1. Temperature Cold Stress. Lowtemperature injury is one of the most important factors limiting apple production in northern areas, while in other regions high temperatures can adversely affect tree growth. Root cold injury is common in cold applegrowing regions. Increased frequency of temperature extremes associated with climate change may influence tree survival and performance (Quamme et al. 2010). Therefore, the industry requires rootstocks that can tolerate extreme temperatures, and it would also be advantageous if the rootstock can impart these characteristics to the scion. Terminology has long been problematic in the literature related to lowtemperature injury in plants. In this chapter, the term “frost” will refer to temperatures below 0 °C during the spring or fall, when trees have leaves or blossoms. “Freeze” will refer to belowfreezing temperatures occurring from late fall (after leaf abscission), to early spring before green tissues appear in the buds. The term “hardy” will refer to the plant’s ability to tolerate cold

stress. Most of the information concerning rootstock cold hardiness comes from field observations following a cold event, and from controlled freezing experiments in the laboratory. Because it is difficult to evaluate root responses to cold stress, information on apple rootstock hardiness is limited. Sublethal effects of root injury on subsequent tree growth are also difficult to evaluate. The most common methods for evaluating rootstock cold hardiness include: 1) observing tree survival and growth following test winters; 2) exposing young grafted or nongrafted rootstocks to freezing temperatures in the laboratory, followed by the evaluation of specific tissues or plant survival; and 3) subjecting pieces of rootstock roots or shoots to controlled freezing conditions and evaluating injury. Injury may be evaluated by recording various aspects of growth the following season, or in the case of stem and root pieces by observing tissue oxidative browning, performing differential thermal analysis, or by measuring electrolyte leakage. Interpreting the literature on cold hardiness from field trials is challenging because the severity of lowtemperature injury is influenced by a number of factors such as: 1) time of year when the cold event occurred and when observations were made; 2) temperatures preceding the freezing event; 3) characteristics of the cold event (the rate of temperature drop, the minimum temperature experienced, the length of time at the minimum temperature, and the rate of thaw); 4) the relative health and vigor of the trees before cold stress; 5) tree age; and 6) the amount of insulating snow cover on the ground at the time of the cold event. Although laboratory freezing tests eliminate some of these sources of variation, differences exist in sampling time, storage and preparation of plant material for freezing, and methods used to assess injury. For these reasons, results from different studies sometimes do not agree. The aboveground parts of woody plants acclimate to low temperatures in autumn in response to short photoperiods and declining temperatures (Weiser 1970). Plants deacclimate in the late winter in response to warm temperatures, and can reacclimate to some extent upon exposure to lower temperatures (Arora and Rowland 2011). Compared to aboveground parts of the tree, apple roots are less coldhardy, and roots are slower to harden in the fall and slower to deharden in the spring (Chandler 1957; Wildung et al. 1973a). The critical midwinter temperature for apple stem tissue is below –30 °C for hardy cultivars, and the critical temperature for roots is likely between –10 and –18 °C, depending on a number of factors. Trees with root winter injury usually leaf out in the spring, but new shoots grow slowly and often wilt in hot weather. Tissues of shallow roots and the belowground portion of the rootstock may exhibit oxidative browning. Roots at lower depths, which were not exposed to lethal temperatures, may be uninjured, but may eventually die. Trees often continue to die for at least two years after the cold event (Czynczyk 1979). Trees on sandy or gravelly soils are most prone to winter injury. During a 2year study, Wildung et al. (1973b) reported a root killing temperature below –12 °C in one year and –10 °C the next year, despite a warmer soil temperature the first year. It was suggested that a low soil moisture the second year may have been responsible for the decreased hardiness. It was felt that the scion and roots developed hardiness independently of each other because root hardiness followed changes in soil temperature. It was also found that 2yearold roots were hardier than 1yearold roots.

Several researchers have studied the acclimation and deacclimation of apple rootstocks. Quamme (1990) used 4 or 5yearold trees budded to two cultivars for controlled freezing tests. The roots of three rootstocks increased in hardiness from late November to mid January, and then hardiness decreased in March. It was suggested that the capacity to acclimate varies with rootstock. Roots of M.26, MM.106 and M.7 in January were hardy to –14, –10, and –9 °C, respectively, whereas shoots of the same rootstocks were injured at –30 °C. For most of the test period the hardiness rank was M.26 > MM.106 > M.7, and the minimum survival temperature was estimated at –10, –7.2, and –6.7 °C, respectively. Root hardiness responded to changes in soil temperature. Czynczyk and Holubowicz (1984) reported that a winter soil temperature of –11 °C caused root injury and tree death in the orchard, with tree mortality being greater on M.9 than on M.26 or B.9. In another study, roots were exposed to a range of temperatures by sweeping snow and/or covering the soil with mulch to obtain a minimum temperature of –12 °C. Less root injury occurred at –10 °C than at –11.5 °C. To study deacclimation in a 2year study, Malling and Polish rootstocks were exposed to low temperatures after exposing the trees to 5 or 10 days of temperatures ranging from 1 to 10 °C. In January, 10 days, but not 5 days, of abovefreezing temperatures resulted in a loss of hardiness. B.9 lost the most hardiness, followed by M.9 and M.26, whereas P.2 and P.22 deacclimated the least. In midFebruary, P.2 and P.22 deacclimated less than B.9 and M.26. In late March, B.9, M.9 and M.26 deacclimated fairly quickly after being exposed to –7 °C. Shoots of Robusta 5 deacclimated earlier in the spring than most apple cultivars (Forsline 1983). The cold temperature tolerance of the Malling rootstocks has been questioned almost since they were imported to North America. Stuart (1941) was among the first to report on experiments involving controlled freezing of root pieces and stems of stools of Malling rootstocks. Based on electrolyte leakage, the roots of M.3 were most hardy and M.1 and M.9 were the least hardy, whereas the stems of M.3 and M.7 were most hardy and M.8 and M.2 were least hardy. Later, Filinger and Zeigler (1951) reported that all Malling stocks being tested at Kansas were killed during the winter of 1947, and all rootstocks of the Kseries of French Crab rootstocks survived. They also exposed rootstock shoots to controlled freezes and found that M.9 was hardier than 16 of the 17 clones of French Crab rootstocks. A number of researchers have evaluated the cold hardiness of various rootstocks. Lapins (1963) summarized the information in the literature and, of the commercially important rootstocks, he classified M.7, M.2, M.4, M.9 and MM.106 as tender, MM.111, Alnarp 2 (A.2) and Antonovka seedling as moderately hardy, and Beautiful Arcade as hardy. In Poland, Antonovka seedlings suffered more damage than A.2, and Polish 1 (P.1) was more hardy than M.7 or MM.106, but less hardy than MM.111. Of the dwarfing rootstocks, P.2, P.22 and M.26 were most hardy. When M.9 or B.9 were used as interstems on Antonovka seedlings, scions were more hardy when B.9 was the interstem (Czynczyk and Holubowicz 1984). Zurawicz and Lewandowski (2014) tested stool shoots of nine Polish (P), three Malling (M), one Malling Merton (MM) rootstock and Antonovka seedlings at three temperatures in late winter for 2 years. All plants survived –12 °C and, based on regrowth, P.59, P.60, P.66, P.67, P.68, M.7, and MM.106 recovered best, whereas subsequent growth was poorer for P.2, P.14, P.16, P.22,

M.9 and M.26. In British Columbia, P.2 was most hardy, followed by O.3 and then A.2, Jork 9 (J.9), B.9, O.3 and P.2, while M.7 was least hardy (Quamme and Brownlee 1997). When Embree (1988) froze 2yearold containergrown trees at a range of root temperatures, root hardiness was greater for B.118, B.490, M.26, O.3, A.2 and P.1 than for M.7 or M.2. Privé and Embree (1997a) exposed root tissue to a range of temperatures and, based on electrolyte leakage, reported that results did not agree with regrowth. Kentville Stock Clone 28 (KSC.28) and M.26 had the best regrowth and survival; B.118 and B.490 had fair to good survival and regrowth; O.3, MM.106 and MM.111 had good survival and poor regrowth, whereas M.7 and M.9 had poor survival and regrowth. Roots from 2yearold rootstocks had less mortality and regrowth than those from 1yearold rootstocks. Moran et al. (2011) froze nongrafted M.26, Geneva 41 (G.41), G.30, B.9, G.11, P.2 and G.935 rootstocks in plastic bags to –8 to – 16 °C. Based on regrowth in the greenhouse, G.41, G.11, G.30, B.9, P.2 and M.26 had similar hardiness, whereas G.935 had greater root hardiness than M.26. Wildung et al. (1973b) compared nongrafted layers of M.7, M.9, M.26, M.104 and MM.106 for two winters under various mulch treatments in Minnesota, and also used stem and root tissues for controlled freezes. During the first winter, the January soil temperature was –17.8 °C for 5 hours with bare ground, and the same temperature was recorded for only 2 hours with snow cover. In February, the soil temperature was below –12.2 °C on 10 days in bare ground, and –10.6 °C with snow. Trees growing in bare ground had more injury than mulched trees, and M.26 was less injured than the other rootstocks. Survival was 88% for M.26, 50% for M.9, 38% for MM.104, 12% for MM.106, and 0% for M.7. There was more snow cover the second year, and minimum soil temperatures were –18.3 and –7.8 °C for bare ground and snowcovered ground, respectively. Again, M.7 had most injury, M.26 had the least injury, and the others were intermediate. Freezing tests in November showed no differences in stem hardiness, and slight injury occurred on all clones between –23.3 and –26.1 °C, but all survived –28.9 °C. In November, roots of M.26 and M.7 were killed at –10.6 and –6.7 °C, respectively, and the other rootstocks were intermediate. Freezing in May caused the same trend, with all clones sustaining severe injury near –6.7 °C. Rootstock influenced tree survival following the mid winter cold event of 2004 in a rootstock trial in the Champlain region of New York State. With ‘Honeycrisp’ and ‘McIntosh’ as the scions, O.3, V.1, V.3, G.16, G.30 and Mark had the greatest survival, followed by B.118, M.9 T337, B.9, M.9 Nic 29 and Supporter 4. M.26, MM.111, M.7 and MM.106 had very poor survival (Robinson et al. 2006). Based on his research and reports in the literature, Quamme (1990) classified rootstocks as follows: very tender (M.7), tender (M.2, M.4, M.9, MM.106, and P.16) moderately hardy (M.26, MM.111, MM.104, P.1 and J.9) and hardy (Antonovka seedling, A.2, Beautiful Arcade, O.3, O.8, B.9, P.2, P.22, and P.18). Freezing trees in pots requires a large freezer and quite a long time because the soil in the pot buffers the temperature. Privé and Embree (1997b) froze nonbudded rootstocks at –12 °C in pots containing various types of media. The different media had different insulating properties and provided a range of temperature drop, but root injury and subsequent plant growth were similar for roots in plastic bags, soilless mix, and sawdust. The authors suggested placing the root systems in plastic bags for future wholeplant cold hardiness studies because

temperatures throughout the root system were more uniform and less time was required to attain the desired temperature. There are three aspects of rootstock cold hardiness to consider: 1) the hardiness of the belowground portion of the rootstock; 2) the hardiness of the aboveground portion of the rootstock; and 3) the influence that the rootstock and scion may have on each other. Stuart (1937) and Quamme (1990) reported that there was no effect of rootstock on scion hardiness. However, Rollins et al. (1962) found that when several crabapple cultivars were used as rootstocks they seemed to induce early scion deacclimation. Lapins (1963) also reported that M.7 induced early acclimation of the scion. Quamme and Brownlee (1997) later reported that Robusta 5 imparted about 2–3 °C more hardiness to the scion than other rootstocks. The range in scion hardiness for all rootstocks was 3.2 °C. Robusta 5 was the earliest rootstock to leaf out in the spring, followed by B.9, M.7, and M.26. Embree and McRae (1991) grafted the hardy ‘Wealthy’ and the tender ‘Gravenstein’ on five rootstocks. The trees were grown in pots and the pots were moved to storage in November until freezing in December or February. Injury was assessed by browning of the xylem and cambium of the rootstock and scion as well as the roots, and they also measured regrowth. When exposed to –8 or –11 °C, root survival was not affected by scion cultivar. However, when exposed to −25 °C, trunk tissue survival was consistently lowest for scions on M.7 EMLA, and regrowth was less than for scions on M.26 and MM.111. When Embree (1988) froze the tree tops to a range of temperatures without freezing the roots and trees on M.26, A.2, the Beautiful Arcade seedlings and O.3 had less injury than trees on M.7, MM.106 and MM.111. It was concluded that the rootstock may have an effect on trunk hardiness. Peach flower bud survival was influenced by rootstock in two test winters in New Jersey (Durner 1988), but rootstock did not greatly influence cold hardiness of 1yearold shoots of peach on several rootstocks in Michigan (Gucci et al. 1988). Ranking of rootstock hardiness is remarkably consistent considering the methods used to obtain the information. Field testing is important because trees varying in age and condition are exposed to lowtemperature stress. However, each cold event is fairly unique and injury depends on a combination of factors, such as the minimum temperature, the length of time at the minimum temperature, the rate of temperature drop, time of year, the temperatures proceeding the cold event, snow cover, soil moisture, nutritional status of the tree and cropping history. Although field testing provides “real world” tests, controlled freezing is important for screening new rootstocks and should be incorporated into rootstock breeding programs. Heat Stress. Some appleproducing regions experience temperatures exceeding 40 °C, and one consequence of global climate change is that these regions will likely experience more frequent hightemperature events that may seriously affect apple tree growth and production. Rootstocks respond differently to rootzone temperatures. Top growth of nongrafted MM.106 and MM.111 was greatest when roots were held at 18 °C and least at 7 °C, whereas top growth of MM.104 and MM.109 responded little to root temperatures. Root growth of MM.104, MM.109 and MM.111 was greatest at 13–18 °C, whereas root growth of seedlings and MM.106 did not respond to temperature (Carlson 1965a).

Plants have evolved physiological and molecular mechanisms to resist heat stress. “Acquired thermotolerance” is the ability to tolerate otherwise lethal heat stress, and is induced by a short acclimation period at moderately high, but survivable, temperatures. “Basal thermotolerance” is the ability to survive exposure to temperatures above the optimal for growth (Larkindale et al. 2005). Stress physiologists are using transcriptomics (signaling components, such as protein kinases and transcription factors) and proteomics (functional genes, such as heat shock protein and catalase) to identify heat stressresponsive genes and proteins in plants. Unfortunately, most of the research on the mechanisms involved in thermotolerance has been conducted with Arabidopsis or seedlings (Guo et al. 2016). Heat shock proteins (HSP) are likely the major control proteins during heat stress response, but the functions of these proteins are poorly understood (Qu et al. 2013). HSPs are common to many organisms, and tend to increase in response to heat stress in plants, insects (Key et al. 1981; cited by Murthy and Ravishankar 2016), and bacteria (Ritossa 1962). Heat acclimation induces the transcription and translation of HSPs, and can be regulated by hormones. HSPs were proposed to act as chaperones to protect cellular proteins against irreversible heat induced denaturation, to facilitate refolding of heatdamaged proteins, and may facilitate the translocation of unstable proteins for degradation to lysosomes or proteasomes (Boston et al. 1996). Larkindale et al. (2005) tested Arabidopsis mutants that were defective for acquired thermotolerance at five growth stages and found that, in addition to HSP induction, ABA, active oxygen species and salicylic acid pathways were involved in acquired thermotolerance, and that the uvh6 gene plays an important role in temperature responses. It was concluded that thermotolerance is a complex multigenic process, with different gene sets involved in acquired and basal thermotolerance at different plant growth stages. Although the apple genome has been fully sequenced, the heat shock transcriptional factor (Hsf) gene family has not been characterized in detail. Jensen et al. (2003) reported differences in gene expression due to rootstock in ‘Gala’ trees that were grafted on M.7 and M.9. Twice as many genes with homology to stressrelated genes were identified in ‘Gala’/M.7 scions as in ‘Gala’/M.9 scions. The same group also identified a clone from ‘Gala’/M.7 with homology to HVA22, which is a member of a family of stressregulated genes (ABA, drought, salt, cold) believed to play a role in stress tolerance (Shen et al. 2001). Another clone had homology to SP1/POP3, a protein implicated in drought tolerance and Hsp20, a heatshock protein. Using leaves, blossoms and fruit from mature ‘Golden Delicious’ trees on M.9 rootstock, Giorno et al. (2012) identified five Malus domestica heat shock families (MdHsfs), classified in three main groups (class A, B, and C) according to structural characteristics and to phylogenetic comparisons with Arabidopsis thaliana and Populus trichocarpa. The apple genome comprises 25 fulllength Hsf genes, and these are important regulators in the sensing and signaling of different environmental stresses. Rootstock breeders should be able to use this information to facilitate the selection of rootstock candidates with increased heat stress tolerance. Zhou et al. (2016) exposed six nongrafted apple rootstocks to heat stress, and rootstocks of the Shao series (SH series, native to China) showed better heat stress resistance than B.9, CG.24, and M.26. SH1 and SH6 had higher heat stress resistance than SH40. Based on

photosynthesis and leaf conductivity, M.26 was least tolerant to heat stress. It was suggested that a high temperature tolerance of the SH series rootstocks may be related to greater osmotic adjustment, because there were smaller reductions in leaf relative water content, higher turgor potentials and higher leaf gas exchange rates compared with the other rootstocks. Following heatacclimation and exposure to 42 °C, Brestic et al. (2011) found that leaves of JTEF (from the Czech Republic) had more thermostable photosystem II (PSII) than leaves of MM.106 because ABA stimulated total peroxidases activity and they suggested that JTEF may be better adapted in heatstress regions. While studying the effect of anthocyanins on heat resistance, Trutneva (2011) found that the scion cultivars, ‘Korichnoe Polosatoe’ and ‘Antonovka Obyknovennaya’ grafted onto the red leaf rootstock Paradizka Budagovskogo had greater resistance to high temperature than the same scions grafted onto a greenleaf mutant Paradizka Budagovskogo. The nongrafted redleaf type also was more heatresistant than the nongrafted greenleaf type. It was suggested that the ability to resist heat stress was due to a higher water content of leaf tissues induced by anthocyanins, and also that anthocyanins have antioxidant properties and may act as antistress factors. These studies showed that leaves of rootstocks vary in heat tolerance, and that heat tolerance of the scion can be influenced by rootstock. However, there is a need to include these rootstocks in field trials in regions that are prone to heat stress to verify that rootstock can influence growth and productivity of trees in hot regions. 2. Water Relations. Plants respond to water stress at the physiological, biochemical, and genetic levels. Fruit trees can suffer from too much or too little water. Water relations in a grafted tree may be influenced by the ability of the roots to absorb water, or by the ability to conduct water from the roots through the rootstock stem piece, across the graft union, and through the scion stem to the leaves. The ability of rootstocks to absorb water at the root surface has not been studied, and information is lacking on the importance of young fine roots. Although recent research suggests that first and secondorder fine roots are most important for water and mineral acquisition, they have typically been ignored in apple root studies. Excess Water. Many orchards in nonarid regions periodically experience wet soils and flooding. In general, apple trees tolerate wet soils better than many woody species. Apple trees are usually expected to survive flooding for about 6 weeks during the dormant season, but they tolerate flooding less well during the growing season (Rom and Brown 1979). Childers and White (1942) found that 2 to 7 days of apple seedling root submergence reduced transpiration, photosynthesis and leaf respiration. Some roots died after 18 days of submergence, but new roots were observed 8 days after the water was drained. Rom and Brown (1979) submerged the roots of 1yearold trees on five rootstocks for 6 weeks at different times of the year, and measured regrowth and tree mortality. The order of flooding tolerance was M.26 > M.7 >  MM.106 > MM.111, regardless of flooding time from November to April. It was suggested that

the fibrous root system of MM.111 may be advantageous under drought, but not under flood conditions. However, these results contradicted field observations where trees on M.26 and MM.106 were sensitive to wet soils (Ferree and Carlson 1987). Makariev (1977) ranked rootstock tolerance to asphyxia as MM.106 (most tolerant), and M.7, MM.111, ‘Golden Pearmain’ seedling, M.2, M.4, MM.104, A.2, M.26, MM.109 and M.9 (least tolerant). Other researchers reported that MM.106 tolerated flooding better than many other rootstocks, and M.1, M.13 and M.16 had field tolerance to wet soil conditions (Ferree and Carlson 1987). Although MM.106 tolerates wet soils fairly well, tree mortality is often high on wet soils because MM.106 is quite susceptible to Phytophthora (Browne and Mircetich 1993). In China, Malus sieversii and M. hupehensis are commonly used as apple rootstocks. M. sieversii is native to the semiarid region in northwestern China, and M. hupehensis is native to a wet climate region in Eastern China. M. sieversii is more vigorous and shallowrooted than M. sieversii. Young M. hupehensis seedlings tolerated hypoxia better than M. sieversii, and the ability to tolerate hypoxia appeared to be related to changes in multiple hormones (Bai et al. 2011). Drought Stress. Many important fruitproducing regions are arid, where rootstocks that perform well under relatively dry conditions are beneficial. In addition, rootstock water relations have been hypothesized to influence rootstock vigor. Research on drought stress is more voluminous than for root flooding. Drought stress research has been performed with grafted and nongrafted rootstocks, and with containergrown and fieldgrown trees. Drought tolerance is actually a compromise between plant survival and vegetative growth and cropping. Stomatal closure effectively conserves moisture, but at the cost of reduced carbon assimilation. Water stress can induce physiological and biochemical changes in trees, and rootstock can influence leaf function of the scion. Apple rootstocks influenced leaf size, specific leaf weight and net photosynthesis in the scion (Ferree and Barden 1971) and carbon partitioning to various parts of the tree (Zhou et al. 2015). Stomatal closure is related to root physiology, possibly due to chemical signals, such as ABA from roots to the shoot in the transport stream. In grape, the same scion cultivar had different stomatal density and pore sizes when grafted to different rootstocks (Serra et al. 2013). Total plant dry weight gain was usually suppressed by increasing water stress. Based on the dry weight of containergrown nongrafted rootstocks, Sakalauskaite et al. (2006) classified MM.106, M.26, B.118, M.9, P.60, P.59, P.2 and B.396 as droughtsensitive, whereas seedling, P.22 and M.9 were droughttolerant. Based on shoot length and dry weight measurements of droughtstressed, nongrafted rootstocks, Preston and Rogers (1961; cited by Carlson 1967) concluded that MM.106, MM.109 and M.9 were more droughttolerant than M.2, Robusta 5 and A.2. In another study, based on transpiration and shoot diameter increase, P.22 was more droughttolerant than P.16 (Klamkowski and Treder 2002). Atkinson et al. (1999) exposed containergrown, nongrafted rootstocks to drought stress and after 6 months the dwarfing rootstocks had less root dry matter, but drought affected the production of coarse and fine roots differently depending on the rootstock. For AR2956, AR36019 and

AR6282, the production of fine and coarse roots declined with increasing soil water deficit, whereas for AR697 and M.26, root production increased slightly. For M.9, coarse root dry matter declined while fine root production increased, and this may enhance the capacity of a root system to extract more water. For M.26, the coarsetofine root ratio increased with reduced irrigation. The authors suggested that plants with large root systems may be more droughttolerant (Higgs and Jones 1990), but the dwarfing rootstocks AR6283, AR2956 and AR4861 produced more root mass than the more vigorous rootstocks, especially M.26, and this response was associated with rootstocks that have O.3 as a parent. Unfortunately, O.3 has not been included in any rootstock experiments involving drought stress. Psarras and Merwin (2000) found a slight shift towards finer root diameter with water stress of M.9 and MM.111, and total root dry matter production declined whereas root respiration increased with soil moisture. Hydraulic conductivity through the graft union of apple trees was generally less than that through adjacent scion and rootstock stem tissue (Knight 1926; Warne and Raby 1939). The effect was greater for trees on M.9 than for trees on more vigorous rootstocks, and the differences in conductivity between the union and adjacent stem declined as trees aged. Gur and Blum (1975) found no difference between hydraulic conductivity of compatible unions and adjacent stem tissue for 4 to 10yearold apple, peach, and plum trees. Olien and Lakso (1984) felt it was unlikely that resistance at the graft union could account for differences in midday stem potential. Since root distribution and amounts of roots varied with rootstock (Rogers and Vyvyan 1934), but the shoot:root ratio was similar for all rootstocks in a given soil environment, Olien and Lakso (1984) also felt that the size of root system probably does not affect root resistance among rootstocks, though they postulated that there are differences in root resistance among rootstocks. Results from experiments using grafted trees may be more relevant to field situations; however, results from such experiments have been inconsistent. Most researchers using grafted trees did not report tree dry weights, but usually reported shoot length, leaf gas exchange, or plant water status responses to drought. Based on field observations, Tukey and Brase (1939c) classified M.1, M.7 and M.16 as tolerant of high soil moisture, M.2, M.4, M.12 were intolerant of low soil moisture, whereas M.7 and M.13 tolerated low soil moisture. Fernandez et al. (1997) reported that 1yearold ‘Gala’/Mark trees were most sensitive to drought, MM.111 was intermediate, and M.9 EMLA was least sensitive. Leaf ABA concentrations increased after drought stress and were highest for M.9 EMLA and lowest for Mark. These results were supported by others. ‘Granny Smith’ on M.9 was most droughttolerant, followed by seedling and MM.106 (Kaynas et al. 1995), and for ‘Golden Delicious’, M.9 was most drought tolerant, followed by MM.106 and M.7 (Giulivo et al. 1985). However, Chandel and Chauhan (1990) using ‘Delicious’ as a scion, found that of 11 rootstocks, trees on M.9 grew most poorly when droughtstressed. The same authors later reported that drought tolerance was positively related to levels of proline, ABA, and carbohydrates in the scion leaves (Chandel and Chauhan 1991). Apple rootstock root systems have low root length per volume of soil, roots are not uniformly distributed under the tree, and rootstocks respond differently to soil structure (Rogers 1939;

Fernandez et al. 1995). The root system can influence water uptake and transport, and may also detect soil water deficits and send signals that regulate stomatal functioning. Fernandez et al. (1997) reported a positive association between the drought tolerance of three rootstocks and ABA levels in scion leaves. Loveys and During (1984) found that leaf ABA concentrations increased during drought stress, but ABA levels alone could not explain the higher leaf stomatal conductance and photosynthesis measured on grafted scions compared to ownrooted cuttings. Following drought stress, nongrafted M.9, Gami almasi (a dwarf apple rootstock native to Iran) and its seedlings had more free proline and soluble sugars, and were more droughttolerant than MM.106 and trees of the local apple ‘Azayesh’ (Alizadeh et al. 2011a). In Prunus rootstocks, droughtinduced accumulations of sorbitol, raffinose and proline, conferred drought tolerance (Jimenez et al. 2013). Because initial molecular responses were related to the biochemical responses, it was proposed that the accumulation of leaf sorbitol, root raffinose, and root and leaf proline could be used as drought tolerance markers for the early selection of Prunus rootstocks. The differential expression of PSC5 in roots could also be used as a drought tolerance marker. The peach–almond hybrid rootstocks GF 677 and ROOTPAC® R performed better under drought stress than the dwarfing rootstock ROOTPAC 20. Differential rootstock performance could be related to differences in genetic background and vigor. Jimenez et al. (2013) also suggested that additional research is needed to determine if these metabolic compounds participate in osmotic adjustments in plants. Jones (2012) felt there is a need to identify genes that are involved in drought tolerance, and stressed the importance of evaluating fruiting trees on different rootstocks in different hydraulic environments. Marguerit et al. (2012) investigated the architecture of the rootstock control of scion transpirationrelated traits in grapes by waterstressing the plants. It was concluded that the scion transpiration rate was controlled by the rootstock through different genetic architectures. Genes have recently been identified that may be involved in the response of rootstocks to abiotic stresses. In droughtstressed grapevines, scion transpiration was controlled by a small number of loci, each accounting for less than 10% of the phenotypic variance (Marguerit et al. 2012). The genetic control of transpiration rate and water extraction capacity was independent of the genetic control of transpiration rate acclimation. There were eight genomic regions associated with scion transpirationrelated traits suggesting hormonal (ABA), and hydraulic signaling between the scion and rootstock plays an important role in responses to water deficit. Scion transpiration rate and its acclimation to water deficit are controlled genetically by the rootstock, through different genetic architectures. Using ‘Gala’ as the scion, Jensen (2010) found twice as many genes with homology to stressrelated genes in trees on M.7 than trees on M.9, and they suggested that this may account for the observations that tees on M.7 tolerate disease, drought and cold better than trees on M.9 (Carlson 1970; Ferree et al. 1995; Wertheim 1998; Cline et al. 2001). Liu et al. (2012a) grafted ‘Gale Gala’ onto seedlings of M. sieversii and M. hupehensis grown in pots and exposed them to two irrigation treatments. M. sieversii was more resistant to drought and had smaller reductions in growth, photosynthesis, leaf area, chlorophyll and relative water content. Following drought stress, leaves and roots from trees grafted onto M. sieversii had greater synthesis of ascorbic acid and glutathione, as well as higher activities of

superoxide dismutase, catalase, ascorbate peroxidase, monodehydroascorbate reductase, dehydroascorbate reductase, and glutathione reductase. These results suggest that the rootstock can enhance drought resistance by improving the antioxidant system in a plant. Irrigation systems have become an integral component of apple production in dry regions and are becoming more common in humid regions. However, since periodic drought will likely be more common in many applegrowing regions as the climate changes, breeding for drought tolerance should be considered an important aspect of rootstock breeding and evaluation in the future. 3. Root Characteristics Affecting Water Relations. The effects of soil characteristics, rootstocks, fruiting, and abiotic stresses on the gross structure and growth of apple root systems have been reported, but less is known about the function of apple roots. A better understanding of apple root growth and function is required before the interactive effects of scion and rootstock responses to abiotic stress can be adequately evaluated and to allow rootstock breeders to select for traits that enhance tree performance. The amount of water available to a tree depends on the volume and distribution of roots under the tree, which may depend on how a rootstock grows in a given soil. Therefore, rootstock response to drought is, in part, likely independent of the physiological characteristics of the rootstock. Because entire root systems of mature trees in the field are difficult to study, there is little information on these aspects of water relations as affected by rootstock. Additionally, during root sampling, the young fine roots – which are most important for the absorption of water and nutrients – are typically not measured. Rootstock interactions with scion growth and cropping (Marini et al. 2013a,b) likely also complicate such studies. Studying the effect of rootstocks on water relations with mature trees is complicated by several factors that may be confounded; vigorous rootstocks have larger root systems that can explore greater soil volume, but also have more leaf area and tend to use greater amounts of water than trees on dwarfing rootstocks (Higgs and Jones 1990). For some, but not all, experiments involving drought stress, the soil and plant water status were measured. Future studies should focus on tree responses to estimates of soil or plant water status, such as predawn water potential or midday stem potential (Lakso 2003). In most experiments using nongrafted and grafted trees, M.9 was among the more droughttolerant and MM.106 among the least droughttolerant rootstocks. However, comparing data from various experiments is difficult because different rootstocks were compared, and the severity and duration of drought stress varied. Additionally, at least in field experiments, drought stress may have been confounded with ambient air and soil temperatures. Although such research is difficult and expensive to perform, more information is needed concerning the drought tolerance of cropping trees on various rootstocks, not only in the year of the stress but also in the following season. During the past 30 years considerable progress has been made in understanding how water moves from the soil, into the roots, and eventually to the leaves of grafted apple trees. Under dry conditions, leaf water potential, leaf conductance and water uptake were lower for

dwarfing rootstocks (Olien and Lakso 1984, 1986; Higgs and Jones 1990; Hussein and McFarland 1994), which was attributed to hydraulic conductance (Olien and Lakso 1984, 1986; Higgs and Jones 1990). Mature ‘Smoothee’ trees on M.9 used less water than MM.106 (Cohen and Naor 2002), and differences in water use were attributed to differences in leaf conductance because conductivity of the wood and the maximum sap velocities were similar for the two rootstocks. The hydraulic conductivity of 1.5 mmdiameter roots from dwarfing rootstocks was lower than similar sized roots from semivigorous rootstocks. Water movement into the root from the soil has been considered the major component of root resistance in healthy plants, but the ability of rootstocks to absorb water at the root surface has not been evaluated (Landsberg and Jones 1981). Total root length relative to tree size is more limited in fruit trees than in other species (Atkinson et al. 1980). Thus, fruit trees require higher rates of water absorption per unit of root length, increasing the potential for resistance to water absorption and to limit the water transport system. Olien and Lakso (1984) suggested that differences in root resistance among rootstocks may involve water absorption at the root surface and the radii and number of xylem vessels for water transport. Atkinson et al. (1980) described the growth and aging of apple roots, but the function of roots of different diameters and age were not known. Pomologists studying roots have generally classified roots on the basis of root diameter, and usually the smallest classification is for fine roots (