California Avocado Society 1991 Yearbook 75:125-137

 

New Insight into Avocado Flowering in Relation to Its Pollination

 

 

G. Ish-Am, and D. Eisikowitch

Tel Aviv University, Department of Botany, Tel Aviv 69978, Israel

 

Additional index words: Persea americana, cross-pollination, Honey-Bees, flower-imitation

 

Abstract

Flower morphology and flowering phenology were explored in five avocado (Persea americana Mill) cultivars in Israel. Ten morphological stages were described during the flower's development. Although both appearance and disappearance of each stage are synchronized within the cultivar, the individual flowers change their stages one by one, during 2-3 hrs. period. A daily bisexual period within each cultivar was observed, which could enable efficient intra-cultivar pollination. A significant linear correlation of flowering time and temperature was found, with slopes of -15 to -50 min. per 1°C. A similarity between the pistillate and the staminate stages was recognized, which could be significant for both bees' behavior and flower pollination.

Introduction

Since the beginning of the avocado pollination research, its synchronously protogynous flowering rhythm has been accepted as a typical adaptation to cross-pollination, which is performed by pollen transfer from B type cultivars to A type cultivars during the morning, and vice versa in the afternoon (Stout: 1923). Evidence of yield increase of several cultivars, after being exposed to cross-pollination, supports this hypothesis (Bergh: 1969; Bergh & Garber: 1964; Gustafson & Bergh: 1966; McGregor: 1976). However, some other phenomena indicate that there is a significant avocado intra-cultivar pollination as well. A reasonable yield of both isolated trees and solid blocks was explained as a product of a pollen flow within the cultivar, either through self-pollination of staminate flowers (Davenport: 1986, 1989) or during close-pollination of coincidental pistillate and staminate flowers within the cultivar (Davenport: 1986; Gustafson & Bergh: 1966, McGregor: 1976), and even by transferring the pollen through the hive on bees' bodies (Bergh: 1977; Gazit: 1977).

In this work, while exploring the optional routes of avocado pollination, a new understanding of its flowering rhythm was achieved which seems to give a better explanation to both its inter- and intra-cultivar pollination processes.

Terminology

Hereafter, inter-cultivar pollination is referred to as cross-pollination, intra-cultivar pollination as close-pollination, and pollination within a single flower as self-pollination. The flower parts will be called as specified in Fig. 1.

Fig. 1 - The flower parts

 

Materials and Methods

Observations were made during 1982-1984 and 1989 seasons in avocado orchards in the Western Gallilee hills of Israel, 10 km west of the Mediterranean coast, 300 m above sea level. The orchard consisted of the cultivars 'Fuerte', 'Ettinger', and 'Nabal' (flowering type B) and 'Hass', and 'Reed' (flowering type A), which were planted in 1975-1976.

Morphology and phenology of individual flowers were recorded by tagging and checking all open flowers on sample branches every 2-4 hours during four consecutive days. The flowers were measured, photographed, and described at different temperatures. The daily flowering course of the cultivar was monitored for 30 days and 6 nights by checking the trees at ½ - 2 hour intervals. A standard meteorological equipment was installed within the orchard, 1 m above ground, and correlation between the daily temperatures and the flowering course was noted. Time was measured according to Israeli Winter Standard Time. Statistical calculations were implemented according to Sokal & Rohlf (1981).

Results

The flower morphology progressive sequence was divided into 10 distinctive stages which were found to be identical in the checked cultivars (Fig. 2). The pistillate (female) opening, open, and closing flowers were indicated as B₁, B₂, and B₃, respectively; the interim closed flower as C; the staminate (male) stages before pollen shedding, during the two pollen shedding stages, and at closing as D₁, D₂, D₃, and D₄, respectively; and the final closed flower was indicated as E.

 

Fig. 2 - Morphological stages of the flower

 

 

The flower parts grow during stages B₁ (tepals, style, and staminodes) and D₁ (tepals, stamens, and nectaries); and therefore the stigma, which is prominent and exposed during the pistillate stages, is covered by the growing inner stamens by the end of stage D₁. The nectar is secreted by the staminodal nectaries during the pistillate stages and by the "true" nectaries during the staminate stages. Anther dehiscence occurs first in the flower lower valves (D₂) and later in the upper valves (D₃). The stigma and the staminodes turn brown during the staminate stages, although on cool days they may be still fresh during D₁ and D₂. The tepals' ends tend to bend towards the pedicle, especially during the staminate stages D₂-D₃ and on hot days. On the other hand, on cool days the pistillate flowers of type B cultivars may stay half-opened for the whole of their flowering period, or even fail to open.

The stage manifestations are found to be synchronized within each cultivar, although the individual flowers go through the stages one by one, during a period of 2-3 hours, and not simultaneously (Figs. 3, 4). Due to this nonsynchronization among the cultivar's flowers, several consecutive stages occur on the tree concurrently most of the day. Flowers that open early to the first stages in the morning continue their course to the other stages earlier than the "late flowers". The whole stages' sequence occurs earlier on hot days (Fig. 3), and much later on cool days, during which the last stages continue to bloom throughout the night and into the next morning (Fig. 4). B type cultivars' staminate flowers that are open continuously for a second day in a cool period look abnormal and do not secrete nectar, while the other stages are not changed significantly in these cases. An overlapping period of 1-3 hours between pistillate and staminate flowers within the cultivar was observed, except in 'Nabal' on hot days (Tmax>30°C).

Each cultivar presented an almost constant daily sequence of flower stages, which was usually identical among cultivars of equal flowering type (Table 1). Slight changes of this sequence were noticed during hot days; however, on cool days the whole sequence started very late, continued throughout the night and the next morning, and underwent more changes.

 

Fig. 3 - Flowering course on a warm day. (April 18, 1984, Temp. max. = 27.0°C, Temp. min. = 12.4°C)

 

Fig. 4 - Flowering course on a cool day. (April 20, 1984, Temp. max. = 21.7°C, Temp. min. = 9.7°C)

 

Significant linear correlations were found between the stages' flowering time and temperature, with negative regression slopes of -15 to —50 min. per 1°C. The morning stages' time was correlated best with the previous night's minimum (Table 2), the afternoon stages' time with the same day's maximum (Fig. 7), and the other stages' time with the daily average temperature (Figs. 5, 6). The bisexual flowering period existed throughout the checked temperature range (striped area in Figs. 5, 6), as an overlapping between the morning pistillate stages B₂-B₃ and the noon D₂-D₃ staminate stages of A type cultivars, and between the morning staminate D₃-D₄ stages and the afternoon B₁B₂ pistillate stages of B type cultivars (Table 1).

 

 

 

Fig. 5 - Flowering times of 'Fuerte' stages vs. average temperature of the day. (Striped area indicates self bisexual flowering period)

Discussion

The above definition of the avocado flower stages follows and elaborates Stout's (1923) classical description. This specification of the flower development has a biological meaning, since the various flower stages exhibit different structures and nectar displays, and are visited by bees in a different manner and frequency (Ish-Am, not published). It also gives a useful tool by which quantitative correlations may be drawn between flowering sequence and temperature (Table 2, Figs. 5, 6, 7, 8), nectar secretion regime, pollinator activity, and pollination course (Ish-Am: 1985).

Morphological similarity between the pistillate and the staminate stages may be recognized. The pistillate stages Bj and B, are similar to the staminate stage D₄, and B₂ is similar to D_1; D₂, and D₃ (Fig. 2). This resemblance is a result of similar location of the pistillate flower style and stigma and the staminate flower three inner stamens and anthers, respectively, as well as a similar spatial arrangement of the former staminodes and the mutually attached tepals and stamens, and the latter nectaries and tepals, respectively. This phenomenon was found to contribute to both bees' behavior and flower pollination (Ish-Am: 1985) and might be referred to as "undeceptive mimicry between pistillate and staminate flowers" (Vane-Wright: 1976).

 

Fig. 6. - Flowering times of 'Hass' stages vs. average temperature of the day (Striped area indicated self bisexual flowering period).

 

According to McGregor (1976), the nectar is secreted from the staminodes during the pistillate period and from the "true" nectaries during the staminate period. Sedgley (1985), on the other hand, reported that both staminodes and nectaries are active during the pistillate and the staminate stages. Our findings support the former description, although on cool days we observed some staminodal activity during the early staminate stages D_l and D₂.

The lack of synchronization among the flowers within the cultivar had been observed by several investigators (Papademetriou: 1976, Eisikoitch and Melamud: 1982; Sedgley and Grant: 1983; Sedgley: 1985), although others described a significant avocado flowering synchronization (Davenport: 1986; McGregor: 1976; Stout: 1923). The latter may have referred to the simultaneity of the stages' appearance and disappearance within the cultivar. This asynchronization seems to be of highly biological importance, since the gradual anther dehiscence during the two successive stages D₂-D₃ (Fig. 1) significantly lengthens the available pollen period, and the asynchronized "early" and "late" pistillate flowers are exposed to different pollination conditions (see below).

 

Fig. 7 - Last stages of the day (at 18:00) vs. maximum temperature

(1) Stages order according to Table 1.

 

Lesley & Bringhurst (1951) reported a strong correlation between flower opening time and temperature, but it was not quantitatively recorded. Our work shows a linear response of flowering time to temperature (Table 2, Figs. 5, 6, 7, 8). During the avocado flowering period in Israel, changes of 10-15 °C on successive days are common and may cause large changes in the daily flowering time. However, both stage sequence order and phase synchronization within the cultivar appear to be stable in wide range of temperatures.

It seems that the flowering time may be correlated best with a weighted average of temperatures, which gives a greater weight to more recent temperature. For instance, on the morning of April 6, 1984, after a hot night (minimum of 18.1°C), flowering started early, as was expected. At 10:00 a.m. the temperature was 30.5 °C, and then dropped to 28.0°C, 25.5 °C, and 23.8°C at 12:00, 14:00, and 16:00 hours, respectively. The flowering course slowed down, and at 16:00 hours was delayed by 1-2 hours more than had been expected according to this day's average temperature.

 

Fig. 8 - Cross-pollination period of 'Fuerte' and 'Hass' vs. average temperature of the day. (Vertically striped area indicates pollination of 'Fuerte' by 'Hass'. Diagonally striped area indicates pollination of 'Hass' by 'Fuerte')

 

The daily period of bisexual flowering within the cultivar (striped area on Figs. 5, 6) gives a good explanation to the close-pollination phenomena. This is in accordance with many works (Bringhurst: 1951; Lesley & Bringhurst: 1951; Gustafson and Bergh: 1966; Papademetriou: 1976; Sedgley: 1977; 1985; Stout: 1923), although for some cultivars it was reported that this period did not exist (Gustafson and Bergh: 1966; McGregor: 1976; Papademetriou: 1976). This important issue should be further studied with other cultivars and under different climatic conditions.

It seems that A type cultivars are subjected to better close-pollination conditions that are B types. The latter begin their close-pollination period at noon, while their staminate flowers are in D₃ stage, bearing no fresh pollen for 2 hours, and are closing by D₄ and E (Fig. 5). The A type cultivars, on the other hand, start this period 2 hours earlier and expose their B₂ and B₃ pistillate flowers to D₂ staminate ones, which shed fresh pollen and open progressively.

The avocado cross-pollination takes place during the relatively long overlapping period of opposite gender flowers of A type and B type cultivars (Fig. 8). It appears that B type cultivars may serve as good pollinizers for A type ones, since the former's staminate and the latter's pistillate flowering periods coincide for about 5 hours throughout all the pistillate period. On the other hand, the pollination of B type by A type is less efficient, since their overlapping period occurs after most of the latter's dehiscence process and lasts 1.5 - 3.5 hours only (see Papademetriou: 1976). Moreover, on cool days the pistillate flowering time of B type cultivars is delayed to the evening while the A type pistillate stages complete their course during the day even on cold days (Figs. 5, 6, 7, 8), having the same time period for cross-pollination (see also Davenport: 1986; Lesley & Bringhurst: 1951; Sedgley & Grant: 1983; Stout: 1923).

A different pollination course of "early" and "late" pistillate flowers is suggested. Type A "early" pistillate flowers are open coincidentally with type B dehiscence, and hence are exposed to efficient cross-pollination (Fig. 8). These flowers close early at noon, do not overlap with their cultivar's dehiscence, and are not exposed to close-pollination (Fig. 6). On the other hand, the "late" pistillate flowers of these cultivars are less exposed to cross-pollination during the morning, but gain all the close-pollination at noon. Type B "early" pistillate flowers are exposed to both efficient close- and cross-pollination (Figs. 5, 8), but their "late" pistillate ones are exposed to low, or zero, close-pollination and also to low cross-pollination. Evidence for these pollination routes was found by field study on stigmas (Ish-Am: 1985).

The pollination process of B type cultivars seems to be more sensitive to low temperatures than the A types', since both close- and cross-pollination periods of the former are shortened by low temperature, while the latter's periods do not seem to be dependent on temperature (Table 1, Figs. 5, 6, 8). Moreover, the appearance of B type pistillate flowers is delayed to the evening in cool periods, while the A type cultivars still complete their pistillate flowering during the day (Fig. 7).

Acknowledgments

This work was partially supported by the Fund of Tzvi Meriminki and his sister, Sonia Meriminki; by the Perpetuation Fund of the Israeli Beekeepers Organization; and by Tel Aviv University. We wish to thank Dr. A. Eshel for assistance in data handling, R. Director for editing the manuscript, and M. Ish-Am for helping in the observations.

References

1. Bergh, B.O. 1969. Avocado, Persea americana Miller. In: F. P. Ferwerda, F. Wit (eds.): Outlines of Perennial Crop Breeding in the Tropics, pp. 23-51.

2. Bergh, B. O. 1977. Factors affecting Avocado fruitfulness. In: J.W. Saults, R. L. Philips, and L. K. Jackson (eds.). Proc. First Intl. Trop. Fruit Short Course: the Avocado. Univ. of Florida, Gainesville, pp. 83-87.

3. Bergh, B. O., Garber, M. J. 1964. Avocado yields increased by interplanting different varieties. Calif. Avocado Soc. Yrbk. 48: 78-85.

4. Bringhurst, R. S. 1951. Influence of glasshouse conditions on flower behavior of Hass and Anaheim Avocados. Calif. Avocado Soc. Yrbk. 36: 164-168.

5. Davenport, T. L. 1986. Avocado flowering. Hort. Rev. Westport, Conn. 8: 257-289.

6. Davenport, T. L. 1989. Pollen deposition on avocado stigmas in southern Florida. Hort Sc. 24: '844-845.

7. Eisikowitch, D., Melamud, H. 1982. A preliminary report on the role of Honey Bees in avocado pollination (In Hebrew). Alón Hanotea. 37:19-29.

8. Gazit, S. 1977. Pollination and fruit set of Avocado. In: J. W. Saults, R. L. Philips, and L. K. Jackson (eds.). Proc. First Intl. Trop. Fruit Short Course: the Avocado. Univ. of Florida, Gainesville, pp. 88-92.

9. Gustafson, C. D., Bergh, B. O. 1966. History and review of studies of cross pollination of Avocados. Calif. Avocado Soc. Yrbk. 50: 39-49.

10. Ish-Am, G. 1985. Avocado pollination by Honey Bees in 'Fuerte', 'Ettinger' and 'Hass' cultivars (In Hebrew). MS Thesis, Dept. of Botany, Tel Aviv Univ. Tel Aviv, Israel, pp. 245.

11. Lesley, J. W., Bringhurst, R. S. 1951. Environmental conditions affecting pollination of Avocados. Calif. Avocado Soc. Yrbk. 36: 169-173.

12. McGregor, S. E. 1976. Insect pollination of cultivated crop-plants. U.S.D.A. Agriculture handbook No. 496: 93-98.

13. Papademetriou, M. K. 1976. Some aspects of the flower behavior, pollination and fruit set of Avocado (Persea americana Mill.) in Trinidad. Calif. Avocado Soc. Yrbk. 60: 106-153.

14. Sedgley, M. 1977. The effect of temperature on floral behaviour, pollen tube growth and fruit set in avocado. J. Hort. Sci. 52: 135-141.

15. Sedgley, M. 1985. Some effects of day length and flower manipulation on the floral cycle of two cultivars of Avocado (Persea americana Mill, Lauraceae), a species showing protogynous dichogamy. J. of Exper. Bot. 36: 823-832.

16. Sedgley, M., Grant, W. J. R. 1983. Effect of low temperatures during flowering on floral cycle and pollen tube growth in nine Avocado cultivars (Persea americana). Sci. Hortic. 18: 207-213.

17. Sokal, R. R., Rohlf, F. J. 1981. Biometry. State Univ. of N. Y., Stony Brook, W. H. Freeman and Comp.

18. Stout, A. B. 1923. A study in cross-pollination of avocado in southern California. Calif. Avocado Assn. Ann. Report 7: 29-45.

19. Vane-Wright, R. I. 1976. A unified classification of mimetic resemblances. Biological J. of the Linnean Soc. 8: 26-56.