by Suzanne Stone Department of Horticulture, The University of Georgia, Athens, GA; edited by George Boyhan Department of Horticulture, University of Georgia, Athens, GA.
Pumpkin breeding involves the hybridization of primarily three domesticated Cucurbita species: C. pepo, C. maxima and C. moschata. Cultivars from this diverse genus have been selected for use in culinary, seed oil and ornamental markets. Pumpkins are an easily grown annual crop with an open-pollinated reproductive system. The large monoecious flowers have supported the widespread use of classical breeding techniques for decades. This manual serves as a guide to current general breeding in pumpkin. Future applications of molecular breeding, now only in the early stages of development, will help pumpkin breeders overcome limitations inherent to the crop, namely abundant space and days to maturation requirements.
Pumpkin is a domesticated crop belonging to the Cucurbita genus and may be derived from C. pepo, C. maxima, C. moschata or from a limited number of interspecific crosses with approximately 20 cucurbit species (Whitaker and Davis, 1962). The colloquial name “pumpkin” presents considerable confusion: it can refer to fruits of different characters, bred for pies, sautés and soups, edible seed and seed oil or ornamentals; and shares the same parentage as the diverse groups of fruit referred to as summer squash, winter squash and gourds (Loy, 2012). In 2011, pumpkin was harvested from 19,140 hectares in the United States and the crop was valued at approximately $130 million (USDA, 2012). In Georgia, the 2010 squash harvest including pumpkin was valued at nearly $22 million, however, pumpkins represent less than 10% of this total (Wolfe and Luke-Morgan, 2011).
All Cucurbita species are diploid with 20 pairs of chromosomes (Whitaker and Davis, 1962). The centers of origin for the temperate species C. pepo and C. maxima are located throughout Mesoamerica (Decker, 1988) and the subtropical C. moschata originated in northern Colombia (Nee, 1990). Cucurbits have undergone several independent domestication events (Whitaker and Carter, 1946) beginning as early as 10,000 years ago, which pre-dates maize and bean domestication (Smith, 1997).
Pumpkins are a warm season annual, reaching maturity at around the fifth month. Most cultivars produce fruit on sprawling vines but bush-type hybrids are also available (Loy, 2012). Tendrils on vines make trellising possible, but large-scale vertical production is not feasible since the large fruit require support. Broad leaves are attached to thick stems and, in most cultivars, are covered in trichomes that may induce dermatitis on contact (Whitaker and Davis, 1962).
Pumpkin reproduces sexually via cross-pollination (King et al., 2012). After 8 weeks of growth, pumpkins produce bright yellow-orange monoecious flowers that open to approximately 10 cm from an 8 cm corolla tube (Figure 1). The staminate flowers produce pollen with large granules, which are pollinated by the honey bee (Apis mellifera), bumble bee (Bombus spp.) and, most commonly, squash bee (Peponapis pruinosa) (Shuler et al., 2005). Upon successful fertilization, fruit mature in 35-55 days after pollination depending on the cultivar (Kelley et al., 2001). Summer squash is harvested as immature fruit so that the soft mesocarp and seeds can be eaten entirely; winter squashes and pumpkins are harvested as mature fruit, which has a hard rind, firm mesocarp and seeds encased in a lignified seed coat attached to the fibrous endocarp adjacent to a hollow central cavity (Loy, 2012).
Asexual reproduction is possible with the chemical induction of parthenocarpy in the closely related cucumber and watermelon species (Wien, 1997) but this intervention is not currently utilized in pumpkin production.
Although naturally outcrossing, pumpkins can be bred according to most self-pollinating crop guidelines because inbreeding depression is minimal for most traits (Scott, 1935) and flowers are easy to pollinate. As such, the breeder can self cucurbits for many generations without a significant level of negative traits surfacing. Also, self-incompatibility is not a documented issue for the genus (Whitaker and Robinson, 1986). Hybrid vigor probably varies depending on the trait and cultivar since conflicting reports of heterosis have been documented in various studies (Loy 2012). Wehner (1999) estimated yield heterosis in cucurbit F1 hybrids to be 40-44%.
Pedigree selection is the traditional and most widespread breeding strategy because the pumpkin flower is easily hand-pollinated and even wide interspecific crosses between domesticated and wild species have some success (Loy, 2012). Quite recently, Zhang et al. (2012) established nine interspecific bridge lines for the three domesticated cucurbit species, overcoming F1 male sterility using a number of breeding tactics. These advanced lines may be useful in diverse cucurbit breeding programs.
When the breeder aims to transfer a single trait, such as a specific disease resistance or bush growth habit, into a cultivar, backcrossing using non-elite and wild germplasm donors is an effective strategy (Loy, 2012). Disease resistance was also bred into commercial lines via transgenic biotechnology. Transgenic lines produced higher yields than controls even when exposed to a different virus than that to which they were resistant (Fuchs et al., 1998).
Breeders interested in quantitative traits such as fruit size, percent dry matter or yield face considerable challenges because plants require abundant field space in a nearly half-year maturation period (Loy, 2012) and evaluations must be carried out in different environments and over many years. Likewise, traditional evaluations of the complete germplasm resources are rarely attempted and replicated yield data is limited in the literature (Loy, 2012).
Historically prevalent open-pollinated varieties are now being replaced by F1 hybrids, a trend that is probably a consequence of modern proprietary law (Wehner, 1999). Hybridization of inbred lines has been a successful means for generating cultivars best suited for the edible seed and seed oil markets, where seed uniformity and seed number is selected over mesocarp characteristics (Lelley et al., 2010). Hand pollination makes F1 hybridization less economical for pumpkins, but novel developments in hybridization techniques may change that in the future. At this time, chemical-induction of gynoecious flowers using ethephon is possible in C. pepo (Robinson et al. 1970) but is ineffective for C. maxima or C. moschata (Wehner, 1999). Male and bisex sterile genes have been identified in C. pepo and C. maxima (Carle, 1997; Scott and Riner, 1946), but with the current inability to phenotype this trait in the field, they are not utilized in pumpkin breeding programs (King et al., 2012). Lastly, a method wherein staminate flowers are removed from female parent plants to promote natural, bee-mediated hybridization is possible (Curtis, 1939) but is still too labor intensive for large scale production.
Pumpkins have the genetic capacity to express a wide range of qualities, so there exists the potential of fulfilling many different niche markets. A breeder targeting culinary traits must carry out selections quite differently than one targeting ornamental traits. Palatability usually correlates with smaller fruit size, which is unfortunate because currently profits are generally determined based on weight yields (Loy, 2012). Breeders in the ornamental carving market select for uniform fruit size of 7-10 kg and an open cavity (George Boyhan, personal communication, September 20, 2012). In a similarly unfortunate trade-off for the ornamental breeder, selecting for larger fruit correlates with less fruit per plant and lower yields overall (Loy, 2012). Both breeding scenarios demonstrate the current physiological restraints that the pumpkin presents, as vegetative photosynthetic growth must be balanced with fruit development. Nonetheless, Wien (1997) posits that fruit development is a complex interaction of not just leaf area but number and size of fruits and the position of the fruit on the vine; the potential to further exploit these factors with selective breeding still remains.
Today, growers are interested in cultivars with the bush habit because more individuals can be grown per hectare and the dense leaf cover may better compete with weeds (Loy, 2012). The phenotype is the result of one or two incompletely dominant genes in C. pepo (Shifriss, 1947) and C. maxima (Singh, 1949), respectively, and a single dominant gene in C. moschata (Wu et al., 2007). However, it has been noted that the semi-bush habit results in greater percent dry matter in the fruit (because more photosynthetic energy can be put towards fruit development) and is therefore preferred in pumpkin production (Loy, 2012).
Consumers have begun to show increased interest in the nutritional values of vegetables. In pumpkins, carotenoid content is a quantitative trait with some genes demonstrating modifier effects that complicate breeding (Paris, 1994).
Disease resistance is another complex, quantitative, environmentally variable trait for the breeder to target. In Georgia, pumpkins are threatened by numerous fungal, bacterial and viral diseases (Table 1) and are susceptible to nematode and insect pests (David B. Langston, 2001). Boyhan et al. (2007) successfully bred broad viral and mildew resistance into an elite line well suited for Georgia using C. maxima germplasm in a recurrent phenotypic selection breeding strategy. Transgenic techniques to integrate viral resistance also show promise (Fuchs et al., 1998). Monsanto has developed a virus resistant transgenic pumpkin, but has not brought it to market, presumably because of public resistance (George Boyhan, personal communication, July 18, 2013).
The following procedures are utilized by Univ. of Georgia breeder George Boyhan (personal communication, September 20, 2012) for the development and evaluation of ornamental carving pumpkin cultivars. These procedures can be generalized for any breeding program for pumpkin.
Hand Pollination in the Greenhouse
The greenhouse allows hand pollination and control of natural outcrossing to be effectively managed. Seed is sown in 36-celled flats using generic greenhouse potting mix. Germination in winter months can be augmented using heating mats. Seedlings are ready for transplant after 2-3 weeks into 46 cm pots with no more than two individuals per container. The containers should be placed at approximately 0.5 m apart along a vertical trellis, which is conveniently constructed using ceiling-secured plastic fencing or twine (Figure 2). As vegetative growth develops, the vines are gently secured to the trellis using plastic figure-8 clamps (Figure 3) and identification jewel tags are attached at 30 cm intervals to aid in breeding line identification during daily pollinations.
Pumpkins begin to flower at approximately 8 weeks (Loy, 2012). Adults produce staminate flowers many days prior to pistillate flowers. Fortunately flowering continues in pumpkins for weeks and as the season progresses it is likely that an individual will possess both a pistillate and staminate flower on the same morning. Flowering and anthesis is influenced by temperature (Wien, 1997) so heat can be used to augment flower development in winter months. A flower is viable for just one day, opening at sunrise and remaining turgid until midday (Whitaker and Robinson, 1986). To control pollination, the pistillate flower should be clipped shut or bagged the evening before (Figure 4). Successful pollination is most likely to occur using turgid flowers in early morning hours, although when staminate flowers are limited, using day-old pollen to achieve fruit set is occasionally possible. The large monoecious flowers make hand pollination easy: detach the desired staminate flower from the vine, remove the petals and thoroughly rub the large, loose pollen grains onto the entire stigma surface of the desired pistillate flower (Figure 5). After pollination, the pistillate flower’s turgid petals will likely tear when manipulated so gauze bags secured around the petals and ovary are the most effective means of blocking natural outcrossing once hand pollination has been accomplished (Figure 6). Pollinated flowers are labeled using jewel tags (Figure 7) and fruit set should be verified within a couple of days, at which time gauze bags are removed. Developing fruit is hammocked along the trellis using a breathable material such as a plastic onion bag (Figure 8). Multiple fruit set per vine is possible but fruit and seed development is usually better in proximally located flowers as opposed to distally located flowers (Suzzanne Tate, personal communication, October 1, 2012).
Daily maintenance of sanitation is required in the greenhouse to prevent and control fungal and insect disease (Suzzanne Tate, personal communication, September 10, 2012). Removing the basal leaves from the first couple of nodes will aid in airflow and insecticidal spray coverage, however, excessive removal of leaves will reduce female flower formation. Removing staminate flowers after morning pollinations have been completed will also discourage fungal growth and eliminate the sweet attractants that draw ants and other stray insects. Nonetheless, a weekly application of fungicides and insecticides is generally required to effectively manage greenhouse pumpkins.
For the pumpkin breeder concerned with qualitative fruit traits and yield, phenotyping traditionally takes place after fruit maturity, approximately 60 days after pollination. In evaluation trials, all fruits are harvested and grouped by breeding line or cultivar. Number of fruits and total weight per replication are recorded and then three fruit per replication are randomly selected to undergo further phenotyping. Fruit height and length are measured; rind color, flesh color and seed number may be visually assessed (Figure 9).
Traits related to growth habit can be phenotyped earlier in the pumpkins’ life cycle. Quantifying bush or vine habit can be accomplished by measuring successive internode lengths (Loy, 2012). Disease persistence can be visually assessed on leaves, vines and fruit (Figure 10), but is efficiently quantified in early leaves using an enzyme-linked immunosorbent assay (ELISA) (Boyhan et al., 2007). For this diagnostic test, two young leaves are collected from the developing plant, pulverized and incubated with an antibody to the viral antigen in question. Effective binding of the antibody within the plant sample, which indicates the virus is present, results in luminescence that can be measured by a microplate reader.
Selections during plant development can be accomplished by rogueing plants with non-desirable traits. In a disease resistance breeding program, young plants are inoculated with the virus and those not proving resistant are discarded (Figure 11). Otherwise, selection based on fruit or yield characteristics must be made at the end of the 4-5 month growing cycle. In the field, row selections are made based on average line performance. Seed from desirable plants are extracted and used in the next breeding cycle.
Fruit may be left in the field or in outdoor storage to cure for several days and can be stored for long periods if the skin remains dry and unbruised (Whitaker and Davis, 1962). Seed is extracted from the fruit’s hollow cavity and rinsed to remove attached endocarp fibers. Nonviable seed will float and can therefore be easily skimmed from the rinsing water. After several rinses, the water is drained and the seeds are dried on a mesh screen. Seeds can be used immediately or kept dormant in cool, dry storage. However refrigerated storage usually improves germination.
Evaluation and Seed Increases in the Field
Hybridization can be efficiently carried out in a greenhouse setting, but line or cultivar evaluation and seed increases should be done in the field. This is because pumpkins do not often grow to their true productivity, fruit size or fruit coloration when restricted to indoor potted conditions (George Boyhan, personal communication, September 20, 2012). Transplants are row planted with approximately 6.5 square meter spacing. For evaluations, lines are planted according to a randomized complete block design, which minimizes the effect of environmental variations within the field. Three to four replications are required for statistically relevant evaluations to be made and with high space demands of each individual, there are reasonable limits to the number of lines that can be evaluated in each trial. Plastic mulch may be used, but productivity is generally good without such measures (Suzzanne Tate, personal communication, October 25, 2012). For seed increases or evaluations not involving disease resistance, a regular schedule of fungicides and pesticides are applied.
Challenges in Pumpkin Breeding
Pumpkin are generally easy to grow and especially easy to hand- pollinate. Fruit set is usually predictable, with germination documented at greater than 90% in intraspecific crosses (Loy, 2012). Nonetheless, the space and time requirements of the plants limit breeding efficiency tremendously. Most breeding programs are only able to carry out two cycles per year and field plant 1500 individuals per hectare (George Boyhan, personal communication, September 20, 2012).
Biotechnology can provide the means for enhancing current breeding programs. Marker-assisted selection would be an invaluable way to circumvent the disadvantages that classical pumpkin breeding presents; seedlings in greenhouse flats could be genotyped and selected as opposed to phenotyping 10 m long vining plants with a 150-day growing cycle. To date, this highly efficient complement to phenotyping is not broadly deployed, probably because pumpkin is not a high-priority commodity crop and has achieved sufficient breeding success with traditional techniques (Loy, 2012). Futhermore, molecular marker development is especially difficult for traits that demonstrate environmental variation. Establishing the breeding population and the successful phenotyping required for marker development is a significant undertaking for the pumpkin breeder (King et al., 2012). In the future, as molecular breeding techniques improve, the time and cost required for the development of such markers will likely decrease, making this breeding technology more accessible to pumpkin breeders. In addition, the asexual regeneration techniques that are currently used to develop inbred lines for cucumber and watermelon may be more judiciously applied to pumpkin breeding programs (Loy, 2012).
This manual serves to provide the background, plant development knowledge, breeding objectives and strategies necessary to carry out a pumpkin breeding program. The procedures contained herein, although specific to the Georgia market and environment, can be adapted to any traditional pumpkin breeding program. The future of pumpkin breeding lies in enhancing the established classical techniques with biotechnology. Molecular breeding has the potential to overcome the practical limitations to growing pumpkins.
Procedural training in the breeding procedures described was provided by advisor Dr. George Boyhan and technician Suzzanne Tate, Univ. of Georgia Horticultural Farm, Athens, Georgia, from August through October 2012. S. Stone is supported by an assistantship from Univ. of Georgia, advised by Dr. George Boyhan, Dep. of Horticulture. This manual was created for HORT 6140, Plant Breeding, instructed by Dr. Cecilia McGregor, Univ. of Georgia, on November 2, 2012.
Boyhan G.E., G. Krewer, D.M. Granberry, C.R. Hill, and W.A. Mills. 2007. Orange Bulldog, a virus-resistant pumpkin for fall production in the Southeast. HortScience 42:1484-1485.
Carle, R.B. 1997. Bisex sterility governed by a single recessive gene in Cucurbita pepo L. Report - Cucurbit Genetics Cooperative:46-47.
Curtis, L. 1939. Heterosis in summer squash (Cucurbita pepo) and the possibilities of producing F 1 hybrid seed for commercial planting, Proc. Amer. Soc. Hort. Sci. pp. 827-828.
Decker, D.S. 1988. Origin(s), evolution and systematics of Cucurbita pepo (Cucurbitaceae). Economic Botany 42:4-15.
Fuchs, M., D.M. Tricoli, K.J. Carney, M. Schesser, J.R. McFerson, and D. Gonsalves D. 1998. Comparative virus resistance and fruit yield of transgenic squash with single and multiple coat protein genes. Plant Disease 82:1350-1356.
Kelley W.T., G.E. Boyhan, and D.M. Granberry. 2001 Variety Selection and Culture, in: W. T. Kelley and J. David B. Langston (Eds.), Commercial Production and Management of Pumpkins and Gourds, University of Georgia Cooperative Extension Service, Athens, GA.
King S.R., A.R. Davis, and T.C. Wehner. 2012. Classical Genetics and Traditional Breeding, in: Y.-H. Wang, et al. (Eds.), Genetics, genomics and breeding of cucurbits, Science Publishers; Distributed by CRC Press, St. Helier, Jersey; Enfield, N.H.; Boca Raton, FL. pp. 61-92.
Langston, D.B. 2001. Pumpkin and Gourd Diseases, in: W. T. Kelley and J. David B. Langston (Eds.), Commercial Production and Management of Pumpkins and Gourds, University of Georgia Cooperative Extension Service, Athens, GA.
Lelley T., B. Loy, and M. Murkovic. 2010 Hull-Less Oil Seed Pumpkin Oil Crops, in: J. Vollmann and I. Rajcan (Eds.), Springer New York. pp. 469-492.
Loy, J.B. 2012. Breeding squash and pumpkins, in: Y.-H. Wang, et al. (Eds.), Genetics, genomics and breeding of cucurbits, Science Publishers; Distributed by CRC Press, St. Helier, Jersey ; Enfield, N.H., Boca Raton, FL. pp. 93-139.
Nee, M. 1990 The domestication of Cucurbita (Cucurbitaceae). Economic Botany 44:56-68.
Paris, H.S. 1994. Genetic analysis and breeding of pumpkins and squash for high carotene content, in: H. F. Linskens and J. F. Jackson (Eds.), Vegetables and vegetable products., Springer-Verlag, Berlin; Germany.
Robinson, R.W., T.H. Whitaker, and G.W. Bohn. 1970. Promotion of pistillate flowering in Cucurbita by 2-chloroethylphosphonic acid. Euphytica 19: 180-183.
Scott, D.H., and M.E. Riner. 1946 Inheritance of male sterility in winter squash. Proceedings. American Society for Horticultural Science 47:375-77.
Scott, G.W. 1935. Observations on some inbred lines of bush types of Cucurbita pepo. Proceedings. American Society for Horticultural Science 32:480-480.
Shifriss, O. 1947. Developmental reversal of dominance in Cucurbita pepo, Proc. Amer. Soc. Hort. Sci. pp. 330-346.
Shuler, R.E., T.A.H. Roulston, and G.E. Farris. 2005. Farming practices influence wild pollinator populations on squash and pumpkin. Journal of Economic Entomology 98:790-795.
Singh, D. 1949. Inheritance of certain economic characters in the squash, Cucurbita maxima Duch. Technical Bulletin. Minnesota Agricultural Experiment Station 186:30-30.
Smith, B.D. 1997. The initial domestication of Cucurbita pepo in the Americas 10,000 years ago. Science (Washington) 276:932-934.
USDA. 2012. Vegetables annual summary (2011), National Agricultural Statistics Service, Agricultural Statistics Board, U.S. Department of Agriculture, Washington, DC.
Wehner, T.C. 1999. Heterosis in vegetable crops, in: J. G. Coors and S. Pandey (Eds.), Genetics and exploitation of heterosis in crops, Amer. Soc. Agron., Madison, Wisconsin. pp. 387-397.
Whitaker, T., and R. Robinson. 1986. Squash breeding. Breeding vegetable crops (M. Baset, ed.). AVI Publishing Company, Inc., Westport, Connecticut:209-242.
Whitaker, T.W., and G.F. Carter. 1946. Critical notes on the origin and domestication of the cultivated species of Cucurbita. American Journal of Botany 33:10-15.
Whitaker, T.W., and G.N. Davis. 1962. Cucurbits. Leonard Hill (Books) Ltd., London, and Interscience Publishers Inc., New York.
Wien, H.C. 1997. The cucurbits: cucumber, melon, squash and pumpkin, Physiology of vegetable crops, Oxford ; New York : CAB International, c1997. pp. 345-386.
Wolfe, K., and A. Luke-Morgan. 2011. 2010 Georgia Gate Farm Value Report, in: T. C. f. A. E. Development (Ed.), University of Georgia, Athens, GA.
Wu, T., J. Zhou, Y. Zhang, and J. Cao. 2007. Characterization and inheritance of a bush-type in tropical pumpkin (Cucurbita moschata Duchesne). Scientia Horticulturae 114:1-4.
Zhang, Q., E.D. Yu, and A. Medina. 2012. Development of advanced interspecific-bridge lines among Cucurbita pepo, C. maxima, and C. moschata. HortScience 47:452-458.