The watermelon is our largest edible fruit. It is a native of dry parts of Southern Africa (Wein 1997). When ripe, the sweet juicy pulp is eaten fresh, and the rind is sometimes preserved (Dupree et al. 1953). Seeds are roasted as a snack or ground into an ingredient in oils or sauces. Spivey (1960) listed three types of melons - regular, icebox, and seedless. Juice from the red part of a watermelon contains 8 to 10 percent solids, of which 20 to 50 percent is sucrose. An edible sirup or fermented beverage can be made from the juice (Webster and Romshe 1951). The pulp of the relatively rare stockmelon, pie melon, or citron melon is used in pies. This melon, indistinguishable externally from the watermelon, can only be opened with great difficulty. It is inedible in the raw state.

Watermelons usually range in size from about 5 to 50 pounds, depending upon the cultivar and growing area. Isolated growers in southwest Arkansas and northeast Texas specialize in jumbo sizes that weigh in excess of 100 pounds (Kennerly 1960); one was produced near Hope, Ark., that weighed 195 pounds. The trend is for icebox size melons of 8 to 12 pounds. Market quality (sweetness and amount of edible fruit) is foremost in harvest decisions with sugar content of 10-14% ideal (Pierce 1987).

Since 1990, triploid or 'seedless' watermelons have been in great demand (Beste et al 1998, Sanford 1990, Mussen 1998). Over 50% of all watermelons grown on the west coast are now seedless partly due to market pressure from imports from Mexico (Sanford 1998). Although around for much longer (Kihara 1951, Mohr et al 1956, Watts 1962), they still were not widely utilized when this section was originally published (McGregor 1976).

Watermelons are grown in almost every State, but roughly two- thirds of the 175,900 acres grown in 1996 were in four States: Texas (56,000), Florida (38,000), Georgia (36,500), and California (20,000 acres). The value of the 1996 crop was $305 million. Worldwide, watermelons are grown in all major regions with Asian production (China) accounting for greater than 50% of tonnage (15,746 of world total 27,063 m. tons)(FAO 1994).

The watermelon plant is a slender, sprawling, slightly hairy, monoecious annual. The stems or runners may extend 0.3 to 5 m. The deeply lobed leaves are 2.5 to 15 cm wide and 5 to 25 cm long, on 2.5 to 12.5 cm stems. The fruit varies according to the cultivar; the shape from oblong to round; the rind, from light green to dark green or mottled light and dark green; the flesh, from red to yellow, rarely to light green or white; the seeds, from white to yellow, brown, black or reddish black; and the shipping quality, from a tender easily broken or bruised skin to a firm and tough rind (Whitaker and Davis 1962*).

The fruit and the vine are susceptible to frost. It is a warm season crop best adapted to mean temperatures above 20oC. Fruit maturity is best at higher temperatures (35-40oC) with warm nights. The plant is started from seed in rows about 2 m apart, the plants 0.3 to 2 m apart in the row. The root system is highly branched up to 2m deep. One to two marketable melons are harvested per plant -- seedless varieties produce more.

Because of the care necessary and the time consumed in harvesting the perishable ripe melons (virtually all harvesting is done by hand), vast acreages are seldom grown by individuals. Fields of 50 to 125 acres (20 to 50 h)are most prevalent in watermelon growing regions and some may be as large as 200 to 300 acres (5000-7500 h); much smaller fields are more common elsewhere.

All cultivars of watermelons and citron melons are monoecious and bear staminate and pistillate flowers, except for a few that bear hermaphrodite flowers instead of pistillate ones (Rosa 1925, Goff 1937). The pale yellow to greenish flowers, about 1 inch in diameter, are much less conspicuous than those of several other genera of the family Cucurbitaceae. The pews of the flower are united in a tiny tube, just as in the cucumber, and are deeply five-lobed, with three stamens around a short blunt style and a three-lobed stigma tightly crowded into the corolla tube.

The flowers are borne singly in the axils, the pistillate or hermaphroditic one occurring in every seventh axil in many cutivars, the staminate ones occupying the intervening axils. The basal main stem node has male flowers -- cultivars vary in ratio of male to female flowers which might also vary due to environmental conditions of temperature and light (Wein 1997). Nectar is secreted in the base of the corolla. All staminate and most of the pistillate flowers shed, and there does not seem to be a definite cycle to fruit setting. The fruit sets more or less irregularly throughout the season, as long as the plants are growing vigorously with adequate moisture. Developing fruit inhibit additional fruit set (Cunningham 1939; Hibbard 1939) but is less pronounced in seedless varieties.

The flowers open 1 to 2 hours after sunrise. The pistillate flower and the staminate flower just below it open the same day. The anthers have dehisced when the corolla expands (Seaton and Kremer 1939), but the pollen remains on the anthers in sticky masses. The stigma is receptive throughout the day although most pollination takes place in the forenoon (Poole and Porter 1933, Sedgley and Buttrose 1978). Large, sticky pollen grains and an adhesive stigma signal the necessity for active pollen transfer between flowers for pollination (Sedgley and Scholefield 1980).

Watermelon pollen is not windblown (Porter 1931). The flowers are almost exclusively insect pollinated. There is no self-sterility so far as the plant is concerned. Pollination is equally effective if the pollen is brought from the adjacent staminate flower on the same branch or from another plant. The watermelon style has no styler canal, but most pollen grains grow directly downward from their point of deposit. Mann (1943) found that 21 to 22 percent of the pollen tubes show some lateral movement, but if an insufficient amount of pollen is deposited on any lobe of the stigma, an asymmetrical melon results. It may be lopsided or it may be smaller on one end than the other. Pollination is required for the triploid "seedless" varieties (Ambrose et al 1995, Sanford 1990), perhaps in higher numbers, as well as in diploid seeded varieties.

Adlerz (1966) studied the relationship between time of day and set of flowers. He (and Parris 1949) found that the highest percentage of fruit set resulted from deposition of pollen on the stigma between 09:00 and 10:00 h. Porter (1933) and Poole and Porter (1933) concluded that fertilization after hand pollination was most likely between 07:00 and 11:00 h. Cordova (1990) found no difference in fruit quality (shape) and seed numbers in honey bee pollinations from 08:00 to 13:00 h. Goff (1937) reported that bees, in Florida melon fields, reached their greatest abundance around 08:30 to 09:00 h.

Adlerz (1966) also studied the relationship between fruit set and length of the ovary at time of pollination. He found that the longer the ovary the better the chance that a fruit would set. Only 22 percent of 103 ovaries 20 mm or less in length set fruit, whereas 67 percent of those over 28 mm set fruit. Cunningham (1939) concluded that both the physiological condition of the plant and the number of fruit already set on it seem to determine the number of pistillate flowers that set later. Hibbard (1939) showed the value of thinning and that the presence of a cull will inhibit setting of normal fruit for several weeks.

Seedless watermelon varieties, gaining in popularity, require special planting schemes for seed production. The seedless hybrids result from a cross of tetraploid (4X chromosomes) female with diploid (2X chromosome) male plants. Tetraploids result from treatment of diploid seedings with a mitotic-inhibiting chemical (Colchicine). A 'seedless' watermelon actually has seeds, but they are small, white seed coats that are eaten along with the fruit (Beste et al 1998).

Hybrid seed to grow a seedless variety is produced by planting diploid and tetraploid lines together in alternating rows (Beste et al 1998). The tetraploids are mostly self sterile, but a small (0-20%) amount of self pollination occurs. Honey bees are useful for essential pollination. Some seed producers prefer hand pollination to avoid tetraploid seed contamination. Since tetraploids produce one-tenth to one-fifth as much seed as diploid plants, seed production is more expensive.

Fruit enlargement in watermelon requires grow-promoting hormones that the developing seeds release -- in triploids, pollen provides these hormones. Since pollen is not abundant in the sterile triploids, commercial fruit production requires growers to interplant diploid (seeded) varieties with seedless. It is recommended that the diploid pollenizer fruit variety be easily distinguished (oblong shape, different color) from the seedless fruit. If an icebox variety is used, two plantings are recommended to ensure adequate pollen availability for the longer-flowering seedless variety (Beste et al 1998). Traditionally two rows of seedless are planted followed by one row of seeded (pollenizer) variety. Recent tests of 'CalSweet' variety in California (reported by Musser 1998) found plantings of pollenizers every third of fourth plant within the same row (vs. in every third row) resulted in the highest yields. The tests need to be repeated and duplicated as within row planting are more costly and make harvest more difficult. In Delaware planting a pollinizer every third versus every fourth row resulted in < 3 mature fruites/50ft of seedless row when pollen source was three rows distant improved to nearly 8 mature fruites/50 ft row when 2 rows distant. The row adjacent to the pollinizer yielded 9 to 14 fruits/50 ft row section (Kee, unpublished data)

The recognition of the need for insect pollination of watermelons is not new. Newell (1903) quoted the following statement made by P. J. Berckmans in August 1877, "Our watermelon growers would find their occupation gone if honey bees and other (pollinating) insects were out of existence."

Porter (1933) concluded that watermelon pollination is almost entirely by insects. Goff (1937, 1947) collected different species of bees, Apis mellifera L., Halictus spp. Augochlorella gratiosa Smith, Agapostemon splendens Lepeletier, and Augochloropsis caerulea Ashmead (listed in the order of their abundance) and concluded that the honey bee was by far the most abundant. Rosa (1925) and Jones and Rosa (1928*) concluded that pollination was chiefly by bees. Purseglove (1968*) stated that watermelon is pollinated by insects, particularly honey bees. Smith (1933) concluded that the lack of sufficient honey bees to pollinate early watermelon blooms in the Big Bend area of Oklahoma cost the growers in that district thousands of dollars annually. Adlerz (1966) showed that honey bees are highly effective as pollinators if they are sufficiently abundant in the field. Stanghellini et al (1998) concluded that bumble bees could serve as backup or alternate pollinators if honey bees were not available

In India, Bhambure (1958) recorded staminate flowers open at 08:30h and pistillate at 09:30h near Bombay. Apis Cerana, A. Florea, and Melipona spp. collected watermelon pollen from 08:30h to a peak at 10:30h; A. cerana deserted the crop by 12:00h each day, but A. florea continued working it until sunset. Near Pune, (Rao and Suryanarayana, 1988) flowers were fully open by 07:00h and most were closed by 14:00h. Peak pollen collection occurred at 09:00h and decreased thereafter. A. cerana comprised 87% of the pollinating insects with A. florea and Tigona iridipennis also found.

One study that failed to demonstrate the value of honey bee pollination was that of Brewer (1974) in Colorado. He documented the value of pollination with a strong positive correlation between watermelon fruit weight and number and weight of mature seeds per fruit but cages to exclude honey bees had populations of wild (native) bees (unidentified). His data on melon number and seed weight inside cages was not different from partially open cages and uncaged plots. In all other studies where all potential pollinators were excluded, flowers without bee visitation abort and do not produce fruit (Adlerz 1966, Spangler and Moffett, 1979, Rau and Suryanarayana 1988, Caron 1990, Stanghellini et al 1998).

Honey bee visits to melon flowers are primarily in the morning from 1 to 2 hours after sunrise when the flower first opens until mid-afternoon, depending on temperature and other weather conditions. The peak period of activity is usually mid-morning (Cordova 1990). Adlerz (1966) recorded the average time that honey bees spend on melon flowers: 5.7 seconds per female flower in 1959, and 8.0 seconds per female and 5.7 per male flower in 1960. He considered duration of the visit relatively unimportant. Honey bees visit the flowers for both nectar and pollen, but because of the scarcity of blooms they never store surplus amounts of either.

The effect of number of visits to the flower is of great importance to production of the mature melon. Adlerz (1966) found fruit set and yield after eight or more bee visits to the flower superior to four or fewer visits. Only two of 64 flowers receiving one bee visit and one of 72 receiving two bee visits developed fruit, which were small, badly shaped, and unmarketable. Fruit set after eight bee visits was significantly better than after two or four visits.

Cordova (1990) reported 8 visits necessary for 'All sweet' variety while 12 visits yielded a significantly higher percent of normal shaped and higher seed numbers in 'Jubilee' variety in studies that compared 0, 4, 8 and 12 visits of honey bees to open pollination. Stanghellini et al (1998) compared fruit abortion following 1,6,12 and 18 flower visits of honey and bumble bees with no visits and open pollination in 'Royal Jubilee' in North Carolina. Abortion was 100% for flowers receiving no visitation and then it decreased as pollination visit numbers increased. Only flowers receiving one honey bee visit had significantly fewer fruit abortions -- even one bumble bee visit led to about 50% fruit formation equal to the open pollination control. Because bees do not uniformly visit all flowers, some flowers will receive more than 8-12 visits if all are to receive the minimum number necessary for marketable fruit formation.

Adlerz (1966) concluded that distribution of the 1,000 pollen grains he found necessary for perfect fruit formation over the stigmatic surface depended more upon multiple visits than upon length of visits or movement of the bee on the flower. Mann (1943) showed that if adequate amounts of pollen are not deposited on every stigmal lobe, the melon will be misshapen - the most common reason for rejecting marketed melons from the number one(highest priced) category -- fruit of lower quality usually have little or no market value.

Peto (1951) reported that one to five hives of honey bees were used per acre (1 h = 2.5 ac)on cucumbers, cantaloupes, and watermelons grown for seed in relatively small fields. Wadlow (1970) used one colony per five acres of watermelons, the colonies placed in small groups in the field. Breece (1962) recommended one colony per acre, the bees on at least two sides of a 40-acre field. Adlerz (1966) made his studies in fields with one colony per acre and concluded that he had more visitors than necessary to provide eight visits per flower. Eckert (1959*) stated that one colony for each 2 acres may be enough. Atkins et al (1979) recommended 2-3 colonies/acre for California and Hughes et al (1982)indicated one strong hive/acre in North Carolina as did Beste et al (1998) for Delmarva. The Arizona Agricultural Experiment Station and Cooperative Extension Service (1970) and McGregor (1976) recommended a bee population that will provide one bee for each 100 flowers in all parts of the field. This recommendation seems to be the best considering the influence of various environmental factors on bee activity.

Ambrose et al 1995 examined both a feeding attractant/stimulant (BeeLine() and an attractant based on worker bee nasanov gland pheromone (Bee Scent() for their influence on yield and crop value in 'Royal Sweet' watermelon variety. Neither compound increased bee activity on the blossoms nor did they increase the monetary yield of the crop for growers. Elmstrom and Maynard (1990) had earlier found an increase in bee activity and seed content when Bee Scent was used on watermelons in Florida. Ambrose did not recommend use of bee attractants for increasing watermelon yield at the expense of increasing the number of honey bees available for crop pollination.

ALDERZ, W. C.
1966 HONEY BEE VISIT NUMBERS AND WATERMELON POLLINATION. Jour. Econ. Ent. 59: 28-30.

AMBROSE, J.T., J.R. SCHULTHEIS, S.B. BAMABARA, and W. MANGUM.
1995. AN EVALUATION OF SELECTED COMMERCIAL BEE ATTRACTANTS IN THE POLLINATION OF CUCUMBER AND WATERMELON. Amer. Bee Jour. 135:267-272.

ARIZONA AGRICULTURAL EXPERIMENT STATION and COOPERATIVE EXTENSION SERVICE.
1970. MELONS AND CUCUMBERS NEED BEES. Ariz. Agr. Expt. Sta. and Coop. Ext. Serv. Folder 90, leaflet.

ATKINS, E.L., E. MUSSEN, AND R. THORP.
1979. HONEY BEE POLLINATION OF CANTALOUPE, CUCUMBER AND WATERMELON. Univ. California Div. Agric. Sci., leaflet 2253.

BESTE, E., D.M. CARON, G.Dively, K. Everts, E. Kee, S.D. Walker, J. Whalen, J. Windsor, and T. Wooten.
1998. WATERMELON PRODUCTION GUIDE FOR DELAWARE AND MARYLAND. Cooperative Extension Misc. Pub. 52pg.

BHAMBURE, C.S.
1958. FURTHER STUDIES ON THE IMPORTANCE OF HONEY BEES IN POLLINATION OF CUCURBITACEAE. Indian Bee J. 20: 10-12.

BREECE. J. R.
1962. WATERMELONS. SAMPLE COSTS AND PRODUCTION. Calif. Agr. Ext. Serv. (Imperial County) Cost Data Sheet 10, leaflet.

BREWER. J. W.
1974. POLLINATION REQUIREMENTS FOR WATERMELON SEED PRODUCTION. Jour Apic. Res. 13: 207-212.

CARON. D.M.
1990 WATERMELON POLLINATION. Proc. 21st Mid Atlantic Veg. Workers Conf.: 101-103.

CORDOVA, SAMUAL,
1990. POLLINATION OF WATERMELON BY HONEY BEES. MS Thesis. University of Delaware. 61 pgs.

CUNNINGHAM, C. B.
1939. FRUIT SETTING OF WATERMELONS. Amer. Soc. Hort. SC;. Proc. 37: 811-814.

DUPREE, W. E., J.G. WOODRUFF, and S. SIEWERT.
1953. WATERMELON RINDS IN FOOD PRODUCTS. Ga. Agr. Expt. Stat. Bul. 285. 30 pp.

ELMSTROM, G.W., and D.N. MAYNARD.
1990. ATTRACTIONS OF HONEY BEES TO WATERMELON WITH BEE ATTRACTANT. Proc. Fla. State Hort. Soc. 103:130-133.

FAO
1994. PRODUCTION YEARBOOK 1993. 47th Edition. Food and Agric. Org. United Nations, Rome.

FREE, J.B.
1993. INSECT POLLINATION OF CROPS, 2nd ed. Academic Press, London.

GOFF, C. C.
1937. IMPORTANCE OF BEES IN THE PRODUCTION OF WATERMELONS. Fla. Ent. 20(2): 30-31.

______ 1947. BEES AND WATERMELON GROWERS. Fla. Grower 56(1): 13, 27.

HIBBARD, A. D.
1939. FRUIT THINNING THE WATERMELON. Amer. Soc. Hort. Sci. Proc. 37: 825-826.

HUGHES, G.R., K.A. SORSENSEN, and J.T. AMBROSE.
1982. POLLINATION IN VINE CROPS. North Carolina Agric. Ext. Serv., AG-84.

KENNERLY, A. B.
1960. BEES HELP GROW JUMBO WATERMELON. Glean. Bee Cult. 88: 406-407.

KIHARA, H.
1951. TRIPLOID WATERMELONS. Amer. Soc. Hort. Sci. Proc. 58: 217-230.

MANN, L. K.
1943. FRUIT SHAPE OF WATERMELONS AS AFFECTED BY PLACEMENT OF POLLEN ON THE STIGMA. Bot. Gaz. 105: 257-262.

MCGREOR, S.E.
1976. INSECT POLLINATION OF CULTIVATED CROPS. U.S. Dept. Agr., Agr. Res. Serv. Agr. Hdbk. 496.

MOHR, H. C., H.T. BLACKHURST, and E.R. JENSEN.
1956. F1 HYBRID WATERMELONS FROM OPEN-POLLINATED SEED BY USE OF A GENETIC MARKER. Amer. Soc. Hort. Sci. Proc. 65: 399-404.

MUSSEN, ERIC.
1999 WATERMELON POLLINATION. U.C. Apiaries News Letter, Mar/Apr 99:4-5.

NEWELL, W.
1903. THE RELATION OF BEES TO FRUIT GROWING. Ga. State Hort. Soc. Proc. 27: 58-63.

PARRIS, G. K.
1949. WATERMELON BREEDING. Econ. Bot. 3: 193-212.

PIERCE, L.C.
1987. VEGETABLES. CHARACTERISTICS, PRODUCTION AND MARKETING. Wiley, NY.

PETO, H. B.
1951. POLLINATION OF CUCUMBERS, WATERMELONS AND CANTALOUPES. Iowa State Apiarist Rpt., 1950, pp. 79-87.

POOLE, C. F., and D.R. PORTER.
1933. POLLEN GERMINATION AND DEVELOPMENT IN WATERMELON. Amer. Soc. Hort. Sci. Proc. 30: 526-530.

PORTER D. R.
1931. SOME EFFECTS OF INBREEDING WATERMELONS. Amer. Soc. Hort. Sci. Proc. 27th Ann. Mtg. Dec. 29-30, 1930: 554-559.

______ 1933. WATERMELON BREEDING. Hilgardia 7(15): 585-624.

ROSA, J. T.
1925. POLLINATION AND FRUITING HABIT OF THE WATERMELON. Amer. Soc. Hort. Sci. Proc. 22: 331-333.

SANFORD, M.T.
1990. SEEDLESS WATERMELONS. APIS 8(11):1-2.

SANFORD, M.T.
1998. SEEDLESS WATERMELONS ON THE RISE: THE POLLINATION CONNECTION. APIS 16(5):3.

SEATON, H.L. and J.C. KREMER.
1939. THE INFLUENCE OF CLIMATOLOGICAL FACTORS ON ANTHERS AND ANTHER DEHISCENCE IN THE CULTIVATED CUCUMBERS. Proc. Amer. Soc. Hort. Sci. 36: 327-331.

SEDGLEY, M., and M.S. BUTTROSE.
1978. SOME EFFECTS OF LIGHT INTENSITY, DAYLENGTH AND TEMPERATURE ON FLOWERS AND POLLEN TUBE GROWTH IN THE WATERMELON (CITRULLUS LANATUS) Ann. Bot. 42:609-616

SEDGLEY, M., and P.B. SCHOLEFIELD.
1980. STIGMA SECRETION IN THE WATERMELON BEFORE AND AFTER POLLINATION. Bot. Gaz. 141(4):428-434.

SMITH, B.
1933. HONEYBEES PRODUCE PROFITS FOR MELON GROWERS. Amer. Bee Jour. 73: 349.

SPIVEY, C. D.
1960. GROWING WATERMELONS. Ga. Agr. Ext. Serv. Cir. (leaflet) 466.

SPRANGLER, H.G. and J.O. MOFFETT.
1979. POLLINATION OF MELONS IN GREENHOUSES. Glean. Bee Cult. 107:17-18.

STANGHELLINI, M.S., J.T. AMBROSE, and F.R. SCHULTHEIS.
1997. THE EFFECTS OF HONEY BEE AND BUMBLE BEE POLLINATION ON FRUIT SET AND ABORTION OF CUCUMBER AND WATERMELON. Amer. Bee Jour. 137(5):386-391.

STANGHELLINI, M.S., J.T. AMBROSE, and F.R. SCHULTHEIS.
1998. SEED PRODUCTION IN WATERMELON: A Comparison between Two Commercially Available Pollinators. Hort. Sci. 33(1):28-30.

WADLOW, R. V.
1970. POLLINATION OF CROPS IN FLORIDA. In The Indispensable Pollinators, Ark. Agr. Ext. Serv. Misc. Pub. 127, pp. 61-63.

WATTS, V. M.
1962. A MARKED MALE-STERILE MUTANT IN WATERMELON. Amer. Soc. Hort. Sci. Proc. 81: 498-505.

WEBSTER, J. E., and F.A. ROMSHE.
1951. WATERMELON SIRUP: ITS COMPOSITION AND COMPOSITION OF THE JUICE FROM WHICH IT WAS MADE. Amer. Soc. Hort. Sci. Proc. 57: 302-304.

WEIN, H.C.
1997. THE CUCURBITS: CUCUMBER, MELON, SQUASH AND PUMPKIN. In. Wein H.C., editor. The Physiology of Vegetable Crops. CAB International. Chpt 9.

Text revision by Dewey M. Caron, University of Delaware -- February, 2000.