California Avocado Society 1996 Yearbook 80:133-149




D. N. Zamet

ACCO Experimental Station, Ministry of Agriculture, Acco, Israel


The text of this article was published in the 1995 Yearbook. Through misadventure, the tables and figures intended for inclusion in the paper were omitted, regrettably diminishing its usefulness. The paper is here presented in full, inclusive of the graphic accompaniments provided by the author.



Data are presented to show that avocado fruit, both parthenocarpic and normal seeded fruit, show multimodular fruit size distribution. Attention is drawn to the fact that with regard to the time of the flowering period of a cultivar the peaks of fruit size and minimum temperature are very similar in number. It is suggested that the minimum temperature controls the time and amount of fruit set. It is also suggested that parthenocarpic fruit are those fruitlets which though not having completed the conditions for normal fruit development can, due to the in-tree hormone status, still manage to "hang on" and grow.



Avocado fruit vary enormously in size (1) from less than 10 grams to well over one kilogram (personal observation). This remarkable size variation is usually between different cultivars; however, it also occurs between fruit of the same cultivar, usually to a smaller extent; e.g., Fuerte, where the size can range from 20 grams to over 500. Although some cultivars can give normal seeded fruit of very small size, most of the smaller fruits are seedless and usually shaped like—and called—"cucumbers" or "cukes" (in a few cultivars these "cukes" are spherical). Many people have labored very hard to explain the reason for these seedless fruits.

On the way, they have produced excellent diagrams describing the differences between seedless and normal fruit (3,12). Trees have been identified which give parthenocarpic fruit every year, and here differences in floral behavior have been noted (11). Steyn et al. have suggested (11) "that the disturbance of the polarity along the micropylar-chalazal axis of the nucleus primordium sets in motion a series of aberrant phenomena in the developing ovule." This prevents the seed developing and can result in a "cuke." This explanation, however, leaves open the question of "what causes the disturbance?"

It is very difficult to know when a particular flower sets and begins its way to becoming a fruit, normal or parthenocarpic. On a medium-size tree there will be 20—30,000 flowers open on average each day during the flowering period according to counts made in 1976 (5), and of these a very small percentage become fruit.

The occurrence of parthenocarpic fruit varies from year to year, just as can the levels of normal fruit. It is the author's opinion that for any particular cultivar, as the number of normal fruit rises, that of the seedless goes down. Years of relatively low yields seem to be the best years for cuke development. Steyn et al. (11) found that if normal fruit did occur on "parthenocarpic" trees", then they did not occur on the same branch. This author has found them growing in the same inflorescence cluster on parthenocarpic trees and many cases of Siamese twins—one normal and one seedless fruit growing on the same fruit stalk (a two-ovuled flower). This would tend to belay the idea that there is always something special about particular flowers, branches, or trees; and that this speciality, if present, can give rise to parthenocarpic fruit.

Zilka and Klein have shown that larger Hass fruit set before smaller fruit (14). Sedgely and Annells have shown in the laboratory the importance of temperature on avocado fruit set (14). Lomas showed that heat stress during flowering and fruit set in the orchard played a role in yield levels6. This author has shown that minimum temperature below 10°C affects the crop level (13). Working together, Lomas and Zamet have shown that crop level in avocado depends on heat stress, minimum temperature, March soil temperature, and winter rainfall (17). These four factors can be explained to work in the following manner. Excess winter rainfall leads to death of many roots, so that in early spring, during the flowering and fruit set period, the tree has difficulty in absorbing sufficient moisture. Low soil temperature in March prevents root growth immediately previous to this same critical period with similar effect of insufficient moisture absorption. Low minimum air temperature prevents or slows down the fruit set process, or causes the "disturbance" as suggested by Steyn et al. (11) Heat stress removes all those fruitlets which for one reason or another are "weak", or in extreme cases will simply burn off the fruitlet. Total winter rainfall and soil temperature are factors which do not change significantly from day to day. Temperature, however, can fluctuate considerably in Israel during the avocado flowering period (Figs. 1, 2). If, as put forward by Zilka and Klein (14), first set fruit is larger, then with waves of fluctuating minimum temperature we should get waves of fruit set. As flowering slowly builds up to a maximum and then falls off until it ceases altogether, it should be expected that the "waves of fruit set" will be higher in percentage of the total crop in the middle of the distribution curve, and that the waves of larger and smaller fruit will be smaller. This paper attempts to show that these waves of fruit set exist, that their intensity varies as just suggested, and what can be learned from fruit size distribution curves. This first part deals mainly with parthenocarpic fruit; the second will deal with normal fruit.


As part of an investigation into fruit size, fruit was picked from trees of several different cultivars, each bearing numerous parthenocarpic fruit. Some of the cultivars also bore normal seeded fruit which was also picked. Each fruit was weighed separately. Normal fruit were weighed on a spring balance to the nearest gram, while the parthenocarpic fruit were weighed on a Precisa electronic scale to the nearest tenth of a gram. The weights were grouped into 20 gram groups for normal fruit and into two gram groups for the parthenocarpic fruit. The percentage of fruit in each group was calculated and distribution curves drawn.

The fruit examined was from the 1994-95 crop from trees growing on the Acco station. In most cases, the total crop of the tree was picked within ten days, the two exceptions being two trees of the cultivar 4403, where 271 fruit were picked from a tree bearing an estimated 1,600 parthenocarpic fruit, and 211 were picked from a tree bearing an estimated 600 parthenocarpic fruit. Both trees had two normal fruit each.

The minimum temperature regime data were obtained from the records of the meteorological station close to the orchard. These data were drawn as a graph for each day from the end of March through till mid-May. In order to give an idea of how the temperature regime varies between years in the flowering period, data were extracted from the records and the information drawn as three day averages. Flowering data were also extracted from the records.





The minimum temperature regime for the flowering period of 1994 is shown in Fig. 1. The fact that the minimum temperature varied considerably from day to day can be clearly seen. It can also be seen clearly that the minimum temperature varies frequently above and below the 10°C level, with maximum range of over 20 degrees. Fig. 2 shows the variation in temperature regimes for selected years. The fact that the regimes vary from year to year can be seen, and also the overall level of the temperature. The flowering period for each cultivar is shown in Table 1. These dates are accurate to ±3 days. In Fig. 3 is given the progress of flowering for the commercial cultivars Hass and Ettinger, showing the relatively regular progress of flowering for 1994.






The numbers of fruits weighed from each tree are given in Table 2. A total of over 2,000 fruits were weighed for this part of the investigation.

In Figs. 4-11 are given the fruit weight distribution curves for the various cultivars. In three cases, distribution curves for normal fruit are also given.


Distribution curves are usually described as "unimodal" or "multimodal"2. When a multimodal curve is obtained, this indicates that more than one factor is controlling the curve. Flowering is basically a unimodal curve. It commences slowly, builds up to a maximum, and then falls off till complete cessation. It can show slowing downs or speeding ups with prolonged temperature variations7 and, as can be seen in Fig. 2 with the Ettinger cultivar; thus (except for a cultivar such as Pinkerton, which can begin flowering in the autumn and give bursts of activity as winter weather conditions permit), avocado flowering is basically unimodal.

If the work of Zilka and Klein (14) is true, then fruit size should show a unimodal distribution curve. The first set fruit will be the largest; however, as flowering is in the beginning weak, the numbers of fruit setting will be few. As flowering intensity increases, more fruit will be able to set, and this will raise the curve; and then as flowering subsides, fewer fruit will be able to set and the size distribution curve will fall.






The suggestion is often raised that citrus and/or wild flowers act as competitors for the services of the honeybee. As far as citrus is concerned, this frequently stops flowering quite suddenly following a heat wave. If this would affect the avocado fruit set, then it should be able to find a sudden single change in the distribution curve. In all the work carried out so far, both on parthenocarpic and normal fruit, this does not seem to be so. The trees used for this investigation grow quite close to citrus, with the bee hives closer to the citrus.

Temperature affects bee activity (4) and also the opening and closing of the various stages of the flower (4). This article does not purport to explain by what means minimum temperature affects fruit set, and therefore the fruit size distribution curve, but it tries to show that this is an actual possibility.



The minimum temperature regime during the 1994 flowering period (Fig.1) shows a typical situation for a Western Galilee spring. It is typical in that the minimum temperature varies considerably up and down, sometimes above 10°C, and sometimes below. The number and range of these ups and downs varies from year to year. It would seem, however, that they are always present. In spite of these variations, there is a general slow increase in the minimum temperature so that by the first ten days of May it is mainly above 10°C. Temperature has already been shown to be an extremely important factor in avocado fruit set (6, 7, 10, 13). It has also been shown that 10°C is critical for avocado production in the Western Galilee (13).




When we examine the fruit size distribution curves of the cultivars used in this work, they seem to fall into several different types. First, there are those which have a considerable crop of normal fruit in addition to the parthenocarpic ones. It is interesting to note that the two cultivars with considerable normal fruit carried on flowering into the month of May, while the others finished by the end of April. It has already been noted by other workers that late flowering cultivars are frequently good yielders; and, on this basis, work is being carried out to try to delay flowering in order to increase yield (9). Another grouping can be made with cultivar 3637 on one side as compared to all the others. The majority show multimodal distribution both for parthenocarpic and for normal fruit, while 3637 shows multimodal distribution only with the normal fruit, and these are very few. It would seem that under Western Galilee conditions multimodality is normal. This would seem to point to the possibility that size distribution both with seeded and with non-seeded fruit is controlled by the same factor. With 3637, an additional factor may be playing a function. A third grouping can be made with regard to the steepness of the distribution curve found with the parthenocarpic fruit of 3637 as compared to the shallower curve as in cultivar 647. It is intended to discuss this grouping when dealing with normal fruit distribution curves.

When we compare the curves of the two 4403 trees, we can see—although of different intensity—the same five peaks. Four occur as the curve rises to a maximum, and then one which is smaller. There is a slight indication of a sixth peak at the large fruit end, but this is less than one percent of the total.

With the cultivar 4209, we have only four clear peaks, but with a "bulge" around the 20 gram weight, and another smaller one at the 34 gram weight, both of which could be signs of two weak peaks giving a total of six peaks. In the case of cultivar 647, we have a wide spread of weights and six clear peaks with a possible seventh at 44 grams. As with 4403, we have one peak only on the downward run to smaller fruit. In the case of 122.16, we again have a wide distribution curve and five peaks as in 4403, but three of them are on the downward side as the fruit size drops.

With the two cultivars with a considerable crop of normal fruit, we have eight or nine peaks with parthenocarpic fruit and seven or nine with the normal fruit. In both cultivars, there is a gradual increase in fruit size before reaching the maximum percentage and several peaks on the downward run as the fruit gets smaller. There is a considerable similarity between the distribution curves of the normal and of the parthenocarpic fruit which again indicates the strong possibility that the same factor is controlling the fruit size distribution.





As noted previously, cultivar 3637 seems to be in a category all by itself. This would appear true for both parthenocarpic and for normal fruit. A knowledge of what the tree underwent during the previous winter can maybe explain the reason for the apparent difference in the distribution curve between 3637 and the remaining cultivars. During the previous winter, water was standing around the tree on three sides for a considerable period. This could have caused asphyxiation and death of many roots. This would lead during the early part of flowering to an insufficient moisture uptake and inhibition of fruit set. Thus, in the curve of the normal fruit there is a possible sign of a peak at the 400 gram mark, a rise to the maximum, and then three peaks as the fruit size drops. In the case of the parthenocarpic fruit, there are two very small peaks (at 38 and 32 grams) before the maximum, and then a steep drop to zero. Parthenocarpic fruit of less than two grams weight would most likely be too weak to survive—the end result being that some of the peaks have simply not developed at all.






The distribution curves can be also grouped thus: 3637 with few peaks, possibly with some which have not survived; at the other end, the two cultivars which bore numerous normal fruit, with up to nine peaks; and the remainder with five or six peaks. If we now examine the minimum temperature regime, we find that for the total flowering period of the trees examined (20th March-10th May), there are nine major peaks of temperature. There also two very tiny peaks (17th and 25th April) which it is felt can be ignored. In this way, we have the same number of minimum temperature peaks as we have size distribution peaks in cultivars 9.76 and 125.06. The remaining cultivars had finished flowering by the 30th of April, and therefore the last three minimum temperature peaks could not have affected their fruit set, and thus they have fewer peaks. This situation is what would be expected if minimum temperature13 and fruit set time14 affect fruit size.




As will be shown in the part on normal fruit, the number of peaks varies each year depending on the number of minimum temperature peaks during the flowering period. As has been suggested with cultivar 3637, one or more of the other three climatic factors controlling yield (heat stress, March soil temperature, and winter rainfall—this latter could also be called poor soil aeration7) can cause erasure of one or more of the distribution curve peaks due to bringing about the drop of the weakest fruitlets. Length of the flowering period will control the maximum possible peaks in the curve.

Steyn et al. have found that E-type trees have earlier flowering commencement (11). This author has noted that trees which have borne low crops tend to commence flowering earlier than trees of the same cultivar which have borne moderate to good crops, while in some cases trees which have borne very heavy crops do not flower at all in the following spring. A crop of parthenocarpic fruit could be considered as acting as a low level crop as regards the nutritional balance of the tree. Earlier and prolific flowering could give rise to in-tree competition which in itself could lead to weaker fruit set. Some cultivars give numerous parthenocarpic fruit, while others do so only rarely. Steyn et al. (11) point out the possibility that fruit development is controlled by phytohormones, and that possibly instead of their being supplied by the developing seed, the tree itself or the mesocarp of the young fruitlet can supply sufficient hormone for the fruitlet to "hold on" and develop into a seedless fruit. A cultivar which has a greater ability to supply the necessary amount of hormone from outside the seed will be more likely to produce a crop of parthenocarpic fruit than one which does not have the hormone producing capability. That these phytohormones (cytokinins and gibberellins) are present could be inferred from the following: Martin et al. (8) have found that these hormones can be used to cause elongation in apple fruit. According to the work of Zilka and Klein (14), later set Hass avocado fruits tend to be longer than early set fruit. Field observations by this author have indicated also that later set fruit tends to be longer and less round than early set fruit.

Whether too low minimum temperature causes a cessation of the reproductive process or if the fruitlet needs a certain length of time above a critical minimum temperature level in order to complete certain essential processes before development can proceed remains to be clarified. However, what does seem to be certain is that minimum temperature controls fruit set. The earlier a fruit sets, the larger it will be able to become, and this is true of normal as well as parthenocarpic fruit. Any condition, nutritional in the tree or external in the soil or air, causing a disturbance of the fragile reproductive process will reduce the "hanging on" power of the young, newly fertilized flower-fruitlet; but the minimum temperature level will control the number of flowers being productively fertilized. It is suggested that parthenocarpic fruit are those fruitlets which through meteorological conditions were unable to pass through the critical stage in the early development of the seed, but due to the hormone production of the tree managed to "hang on."

The fact that cultivars flowering after the first of May are in general good yielders (e.g.; 125.06 and 9.76) does not seem to concur with the distribution curves, for it would be natural to expect that these cultivars would have the largest peak at the end, or small fruit side, of the curve, and this was only found with the parthenocarpic fruit of 122.16 which finished flowering early. Could it be that later flowering and better yields are genetically linked?

Different parts of the tree have slightly different temperature regimes, depending on tree side (toward or away from the sun), height above ground, or climatic conditions. As a result, it is possible to find on the same tree slightly different flowering regimes, not only as to when a particular area of the tree begins to flower, but also the opening and closing of the various stages (this is based partly on unpublished data of Lomas and Zamet). Thus, each flower, depending on the cultivar, will have slightly different reaction to cold or warm spells. Each flower opens at a slightly different time, and is liable to be pollinated at a different time. Small differences such as these could cause the difference between normal or parthenocarpic fruit set, or simply total flower or fruitlet death.

One cannot ignore the fact that the actual size a fruit will grow to depends not only on the cultivar and the temperature regime, but also on its position on the tree, the number of fruit on the tree, the number of fruit on a particular branch, and the number and size of leaves supplying photosynthates to any particular fruit, and finally to the light regime in the area of the said leaves. Nevertheless, with the present total sample of over 2,000 fruits and seven cultivars, the evidence for waves of fruit set seems very strong; and their similarity to the number of minimum temperature peaks (with regard to the flowering period) is sufficient to indicate a strong possibility that it is these minimum temperature peaks which are the cause of the waves of fruit set. Avocado growers, at least in Israel, frequently talk about "waves of fruit set." This work clearly shows that these waves do exist.

Literature Cited

1. Anon. 1976. Research in progress. California Agriculture 30 (12) 3.

2. Babcock, E. B., R. E. Clausen. 1927. Genetics in Relation to Agriculture. McGraw-Hill Book Company. New York.

3. Blumenfeld, A., S. Gazit. 1974. Development of seeded and seedless avocado fruits. J. Am. Soc Hort. Sci. 99:442-448.

4. Ish-Am, G. 1985. Avocado Pollination by Honey Bees in the Fuerte, Ettinger and Hass Varieties. Thesis submitted toward the M.Sc. degree of Tel Aviv University.

5. Lahav E., D. N. Zamet. 1976. Flower, fruitlet, and fruit drop from avocado trees, pp. 57-63. In Research in Subtropical Fruit Trees. The Volcani Center for Agricultural Research. (In Hebrew.)

6. Lomas, J. 1988. An agrometeorological model for assessing the effect of heat stress during flowering and early fruit-set on avocado yields. J. Am. Soc. Hort. Sci. 113(1):172-176.

7. Lomas, J., D. N. Zamet. 1994. Long term analysis and modeling of agro-climatic effects on national avocado yields in Israel. Agricultural and Forest Meteorology. 71:315-336.

8. Martin G.C., D. S. Brown, M. M. Nelson. 1970. Apple shape changing possible with cytokinin and gibberellin sprays. California Agriculture. 24(4): 14.

9. Lovatt, C. 1994. Report on a research symposium in California. Translated by A. Ben-Yaacov. Alón Hanotea 4(11):510. (In Hebrew.)

10. Sedgley M., C. M. Annells. 1981. Flowering and fruit-set response to temperature in the avocado cultivar 'Hass'. Sci. Hortis. 14:27-33.

11. Steyn, E. M. A., P. J. Robbertse, D. Smith. 1993. An anatomical study of ovary-to-cuke development in consistently low producing trees of the 'Fuerte' avocado (Persea americana Mill.) with special reference to seed abortion. Sex. Plant Reprod. 6:87-97.

12. Tomer, E., S. Gazit, D. Eisenstein. 1980. Seedless fruit in 'Fuerte' and 'Ettinger' avocado. J. Am. Soc. Hort. Sci.105:341-346.

13. Zamet, D. N. 1990. The effect of minimum temperature on avocado yields. Calif. Avoc. Soc. Yrbk. 74:247-256.

14. Zilka, S., L. Klein. 1987. Growth kinetics and determination of shape and size of small and large avocado fruits of the cultivar 'Hass' on the tree. Scientia Hort. 32:195-202.