ALTERNATE BEARING 129 b. Cool Weather 151 c. Low Air Humidity 151 2. Edaphic Stresses 151 3. Other Environmental Stresses 152 B. Endogenous Factors of Alternation 152 1. Inhibition of Flower Initiation by Growing Fruits 152 2. Fruit Set as Conditioned by Pollination 153 3. The Effect of Seeds on Prevention of Fruit Drop 154 Alternate Bearing in Fruit Trees1 4. Contribution of Leaves to Reproductive Growth 155 5. Competition between Vegetative and Reproductive Sinks 156 6. The Effect of Fruit Overload 157 S.P. Monselise and E.E. Goldschmidt VI. Horticultural Control of Alternation 159 Department of Horticulture, The Hebrew University of Jerusalem, A. Control of Flower Formation 161 Rehovot, Israel B. Thinning Fruits 161 C. Reducing Crop Influence by Early Harvest 161 D. Pruning Techniques 161 I. Introduction 129 VII. Conclusions 162 II. Definition of Alternation 130 VIII. Literature Cited 164 A. Biennial and Other Cycles 130 B. Quantitative Evaluation of Alternation 131 C. Alternation in Whole Areas of Individual Trees and Individual I. INTRODUCTION Branches 132 Problems of alternate (or biennial) bearing in fruit trees have been III. Representative Cases of Alternation 135 A. Apple 136 investigated and reviewed on many occasions in the last decades. Only B. Pistachio Nut 137 review papers with extensive bibliographic lists will be quoted here C. Pecan 139 (Singh 1949; Davis 1957; Singh 1971; Williams and Edgerton 1974; D. Olive 140 Jonkers 1979). Most of these reviews deal mainly with one sort of tree. E. Citrus 141 Alternation is, however, a very widely spread phenomenon, occurring in F. Avocado 144 both deciduous and evergreen trees. G. Mango 145 The fact that different types of fruit trees show alternate bearing­ IV. Horticultural Traits 146 notwithstanding wide differences in dormancy, time of flower forma­ A. Genetic Determination 146 tion, flowering habits, set-abscission relationships, length of fruit de­ 1. Families, Genera, Species 146 velopment stages as compared with vegetative activities, and many Cultivar Differences 14 7 2. other aspects-indicates that alternation is somehow inherent to the B. Effect of Rootstocks 148 C. Effect of Tree and Clonal Age 148 nature of the polycarpic plant and calls for a unified theory or at least a D. Flowering Habit 149 logical set of alternative hypotheses explaining the alternating behav­ E. Autonomy of Branches with Regard to Alternation 149 ior in a unified way. V. Causes of Alternation 150 The purpose of this review is to scrutinize available knowledge with A. Environmental Triggers 150 the aim of eventually presenting an integrative interpretation of the 1. Climatic Stresses 151 alternate bearing phenomenon valid for different types of fruit trees. a. Frost 151 An alternate bearing tree (or branch) is one that does not bear a regular crop year after year; rather, heavy yields are followed by ex­ 'This review was partly supported by a grant from the United States-Israel tremely light ones and vice-versa. The term "biennial bearing," which is (Bination~)8 Agriculture Research and Development Fund (BARD). We acknowledge wi~h than the contribution of 30 specialists from 5 continents who kindly provided published and customary (e.g., in the annual indexes of Horticultural Abstracts), does unpublished material and useful suggestions. not necessarily imply a regular biennial pattern (Pearce and Dobersek- 128 130 HORTICULTURAL REVIEWS ALTERNATE BEARING 131 Urbane 1967l. For this reason we prefer the term "alternate bearing"· grapefruit in Florida, 'Washington Navel' orange in California, and the term "irregular bearing" is sometimes used, but this term is not 'Redblush' grapefruit in Texas. In 'Marsh' grapefruit in Florida 660 of sufficiently precise. the variance was imputable to peaks at 2 and 2.5 years; very little The fact that the development of scientific horticulture began earlier cyclical bearing was found for 'Navel' orange in California; and 430 of in the temperate zone of the northern hemisphere than anywhere else is variance was imputable to a 4-year cycle in 'Redblush' grapefruit in probably the main cause for the large emphasis on alternation of decid­ Texas. In the period studied for this last cultivar and location, occur­ uous trees (especially apples and pears) in the early literature. Never­ rences influencing yield at 4-year distances were a very severe freeze, a theless, alternation is very common in widely different families and hurricane, and a severe pruning treatment. Only the Florida grapefruit species of trees (see Table 5.4l. showed sufficiently clear biennial patterns. It is anticipated that more An early reference to the almost universality in fruit trees may be studies along these lines will be forthcoming. found in Jewish sources from the fourth century (Talmud Yerushalmi Cycles of about 3 years in apple production, influenced by environ­ Sheviith, Chapter 1 ): "All trees bear fruit one year and leave one yea~ ment and heavy crops, have been pointed out by L.D. Tukey (personal off, but the fig bears regularly every year." More recent quotes (six­ communication, 1980l. teenth century) are cited in Sparks ( 19751 for pecan and Singh and Khan (1940) for mango. It should be remembered that most of these trees were B. Quantitative Evaluation of Alternation grown without irrigation. There is an obvious interest in developing reliable parameters to evaluate bienniality of alternation, its intensity (maximum and aver­ II. DEFINITION OF ALTERNATION age of deviation from pluriannual mean values), synchrony in different locations, and other features (see Table 5.ll. A. Biennial and Other Cycles Accepted and proposed indexes of alternation have been recently Each fruit tree in commercial groves does not bear equal crops year reevaluated by Pearce and Dobersek-Urbanc ( 1967l. Two parameters to after year. Small yearly variations can easily be accounted for by cli­ evaluate alternation had been proposed by Hoblyn et al. (1936l. The matic variations in particular years, as well as by natural or pathologi­ first, B, expresses percentage ofbienniality while amplitude of fluctua­ cal changes occurring in trees with the progress of time. Our main tion is expressed by I, a measure of intensity (see Table 5.1). After concern is cyclical changes in cropping. A biennial cycle is very usual, so reevaluating Band I and proposing some other indexes with the help of that an "on-year" !large yields) is followed by an "off-year" (little or no rather extensive simulation experiments, Pearce and Dobersek-U rbanc yield) and so on for a sequence of several years. Even with a rather (1967) concluded that B is very insensitive, as an alternation of positive regular biennial sequence, however, two or more on- or off-years can be and negative signs can continue by chance for 8 years so that a 1000 found in sequence; this would still be considered a biennial behavior by value would become significant at the 5rlr level only if obtained after 9 most workers (Pearce and Dobersek-Urbanc 1967). consecutive years. Moreover, a full bienniality cannot be expected in Cycles other than biennial also occur; according to Gardner (1966), practice. Nevertheless, B remains an easy and rapid way to assess quoting a classical German forestry textbook, such cycles are common bienniality over a long period of years. On the other hand, I is more in forest trees. Cycles of2 to 3 years in seed production are common with satisfactory than other parameters tried; K~ can be obtained by sweet chestnuts, hazelnuts, and elms; of 3 to 5 with pines and oaks; of 5 squaring the terms instead of arbitrarily ignoring their signs; however, to 7 with spruces; and of 10 to 15 years with beeches. it minimizes small and maximizes large differences in yield. A better Two recent studies using power spectral analysis have attempted to statistical distribution can be obtained by extracting the square root of detect all possible periodic patterns in yield data. Gemoets et al. (1976) K2 to get K. These authors conclude that there is no reason for abandon­ have attempted to describe periodicity of U.S. pecan production (sub­ ing I, the accepted index of evaluation. divided into "native" and improved cultivars) and detected, in addition Another problem of evaluation, usefui when discussing alternation, to the expected 2-year periodicity, strong peaks in the power spectra at seems to have not been quantitatively approached: synchronization of 2.5 years (unexplained but highly significant), at 4 years, and at 11 to 16 different plants within a single orchard or of different orchards within a years. J.F. Fucik and J.E. Chance (unpublished) have studied cyclical single region. Such an attempt would provide a basis to evaluate to what bearing problems for 10 to 12 individual trees of 'Marsh' seedless extent external factors (common to a grove, an area, etc.) are dominant ALTERNATE BEARING 133 132 HORTICULTCRAL REVIEWS TABLE 5.1. DEFINITIONS AND WAYS TO CALCULATE SOME ALTERNATION PA­ country-wide, regional, and provincial (=county I oil production during RAMETERS 14 successive years. ill B ian evaluation ofbiennialityl =percentage of occasions I pairs of successive years) As expected, B is rather insensitive. Country-wide production, obvi­ where trends of increase or decrease in yield are reversed. 100';( of occasions 1n- 21. 100'( full to o1; lack of biennialitv. ously more buffered than regional or provincial productions, has a n = no. of years of observation. · B-value only slightly lower than the regional; provinces present irregu­ 121 I ian evaluation of intensity of deviation in vield in successive years I. lar B values. RP and I, on the other hand, present a clear picture of the stronger alternation evinced by regions than by the country as a whole. 1 ( a:2 -- a 1 a: 1 - a:2 a,n 1·- an) I=--~--- I ----r As to the regions, we can distinguish between two groups (fully con­ n-1 a:2 ~ a a.