rubbed on to the stigma of a previously tied female flower. After pollination, the female flower is bagged, or the corolla re-secured to prevent pollinator contamination, and the developing fruit is then tagged.
Fig. 3.1. Cross-section of female and male flowers from Cucurbita pepo: (left) Spaghetti winter squash and (right) Patty Pan summer squash one day before anthesis, at which time the corollas are tied shut to exclude insects from flowers that are to be hand-pollinated.
Because squash pollen does not store well, the use of fresh pollen is customary in squash breeding. However, pollen can be used from pre-anthesis flowers that are kept for a few days at low temperature and high humidity (Robinson, 2010–2011). Also, pistillate flowers can be pollinated 1 day pre-anthesis. Cucumber can be pollinated 24 h after the flower opens, especially if pollinations are being made in the greenhouse or growth chamber and if the temperature is below 30°C.
Cross-pollination of andromonoecious melons and watermelons by hand is more difficult than for monoecious cultivars, since the anthers need to be manually excised from each perfect flower. The anthers of perfect flowers are usually removed on the afternoon before anthesis. Afterwards, the emasculated flower is enclosed to prevent insect pollination. As with squash, the male flowers of melon are either closed before anthesis or picked and kept indoors overnight at high humidity. Pollen transfer from these flowers takes place the following morning.
Tissue culture
Tissue culture is used in cucurbits for embryo rescue, propagation, embryogenesis and organogenesis. Success in producing interspecific plantlets by embryo culture has been achieved in Cucurbita (C. pepo × C. moschata, C. pepo × C. ecuadorensis) and Cucumis (C. metuliferus × C. zeyheri Sond.). Other crosses, such as African horned melon and melon, have been difficult. So far, there are no reports of the successful use of protoplast fusion in creating interspecific hybrids of cucurbits.
Cultivars within a cucurbit crop often differ in their ability to be regenerated through tissue culture. Embryos have been obtained from various plant tissues (e.g. leaf callus, hypocotyl explants, protoplast-derived callus), but their development into plants has been sporadic. Cotyledons and hypocotyls have proved to be the best explant tissues for organogenesis, but leaves, fruit and embryos have also been used.
In their production of transgenic melon plants, Gonsalves et al. (1994) cultured the developing tissue on Murashige and Skoog medium according to plant developmental stage (i.e. regeneration, shoot elongation and rooting stages). They also performed regeneration experiments in which they found that proximal explants produced a higher percentage of regenerated shoots than distal tissue samples, with most shoots arising from the proximal side of the square-cut piece of cotyledon.
Achieving breeding objectives
Progress in cucurbit breeding has been improved through the use of multivariate statistical analyses (e.g. to increase selection efficiency), research on gene mapping and linkage, the induction of mutations with chemicals and gamma radiation, marker assisted selection, the use of wider crosses within species and gene transfer among species, and the collection and screening of germplasm in genebanks for desirable characteristics.
Although genetic engineering and backcrossing usually focus on single alleles, more attention is being given to multiple gene selections for single or multiple characteristics. For example, in Saudi Arabia, the melon cultivars ‘Najd I’ and ‘Najd II’ were bred for tolerance of both high temperature and salinity. In some cases where multiple insect resistance has been found, it is possible to breed for resistance to several insects from a single cross. ‘Butternut’ squash (Cucurbita moschata), for example, is resistant to squash vine borer, cucumber beetles, pickleworm, melonworm and leafminer. Squash bugs prefer Cucurbita pepo and C. maxima, but if these species are not present, then they will feed on C. moschata. Sources of combined insect and disease resistance are also known for cucumber and melon. Often, quantitative resistance to a disease provides a lower level of resistance, but resistance to multiple isolates of the pathogen. Genetically modified cultivars of squash for virus resistance have been combined with conventionally bred resistance to powdery mildew as well as with fusarium wilt in melon (Quemada and Groff, 1995).
Selection for many morphological characteristics is standard practice in plant breeding. Progeny are visually inspected for the desired character and further breeding proceeds according to objectives. The study of Mendelian genetics in families can help reveal relationships (e.g. dominant, co-dominant, recessive, complementary) among alleles at multiple loci.
Although biochemical qualities can be evaluated with the help of laboratory equipment and analysis (e.g. using a refractometer to select for high sugar content in fruit), linkage or pleiotropy for morphological characters has also been used to breed for these traits. For example, selection for fruit flesh carotenes was initially based on fruit colour, i.e. orange-fleshed fruit were presumed to have higher concentrations of these organic compounds. However, subsequent research in squash (C. pepo) revealed that fruit colour is controlled by the complex interaction of alleles at several genes, and although total carotenoid content is affected by these genes, there are additional genes influencing the percentage of carotenes in the total carotenoid content (Paris, 1994). Also, the major allele (B) for high carotenoid content has various pleiotropic, and not always favourable, effects on fruit and foliage. Direct selection for high carotene content has proved to be more effective. The B gene is not a factor in C. maxima and C. moschata, and selection for high carotenoid content and sometimes specifically for carotenes has been more successful in these species.
Gene linkage, and also pleiotropy, has been a problem in breeding monoecious melons. The use of monoecious melons in F1 hybrid seed production is desirable because emasculation of the female parent is easier with the pistillate flowers of monoecious inbred lines than with the perfect flowers of andromonoecious inbred lines. However, selection for monoecy has been limited by its genetic association with elongated fruit shape. Elongated shape is dominant and F1 hybrids with a monoecious parent generally have undesirably long fruit. Other genes can influence the shape of melons with the monoecious gene, and H.M. Munger of Cornell University was successful in breeding monoecious selections with nearly-round fruit.
Many objectives in cucurbit breeding involve the improvement of traits having complex inheritance. Earliness, yield, adaptation to certain environmental conditions and fruit quality are quantitative traits. Consequently, large populations, efficient experiment designs and multivariate analyses are useful for evaluation of breeding material for future selections. Recurrent selection and pedigree selection have been used to improve cucurbits for quantitative traits. Backcrossing has been used especially for improvement of sex expression and disease resistance (Munger et al., 1993).
Studies with cucumber, melon, watermelon and squash (Whitaker and Davis, 1962; Wehner, 1999) indicate that, in general, there is little or no inbreeding depression, but there is some heterosis for certain morphological traits. Increasingly, wider crosses are being employed to produce F1 populations with more favourable characteristics. The mechanics of F1 seed production are described in Chapter 7. In any case, hybrids are useful for the following reasons.
1. Hybrids permit the protection of the parental inbreds by trade-secret
