20.6 Practical Breeding Procedures for Watermelon
by Cecilia McGregor, Geoffrey Meru and Vickie Waters, Department of Horticulture, University of Georgia
Watermelon (Citrullus lanatus) is a member of the Cucurbitaceae family that includes many economically important domesticated species such as cucumber (Cucumis sativus L.), melon (Cucumis melo L.), squash and pumpkin (including Cucurbita pepo, C. moschata and C. maxima). The genus Citrullus (2n = 2x = 22) includes C. lanatus subspecies (infra-specific) as well as three other diploid species, C. ecirrhosus, C. colocynthis and C. rehmii, which are all cross compatible to some degree (Robinson and Decker-Walters, 1997).
Infra-specific classification within C. lanatus is not clear-cut and various interpretations of the description by Jeffrey (2001) and the designation used by the USDA ARS National Plant Germplasm System (Plant Genetic Resources Conservation Unit, Southern Regional Plant Introduction Station, Griffin, GA; http://www.ars-grin.gov/npgs/index.html) seem to be applied differently by different researchers (Dane and Liu, 2007). From a breeding standpoint, the most practically useful classification is probably: sweet-fleshed cultivated tyes (C. lanatus var. lanatus), citron-types (C. lanatus var. citroides) and egusi-types (C. lanatus subsp. mucosospermus).
Interspecific crosses in Citrullus are possible to varying degrees (Robinson and Decker-Walters, 1997), but not unambiguous. For example, crosses between C. lanatus and C. colocynthis have been successful, but fruit set was low and directionality of the crosses influenced results (Sain et al., 2002). Infra-specific crossed within C. lanatus is relatively easy, however, high levels of marker segregation distortion, low fruit set and diminished pollen viability have been observed (Hawkins et al., 2001; Levi et al., 2004b; Sandlin et al., 2012; Ren et al., 2012).
Cultivated watermelon plants are monoecious, with separate male and female flowers (Fig 1). Therefore, emasculation is not required, but selfing requires hand pollination. C. lanatus var. citroides plants are often androdioecious (Fig. 2) and requires emasculation before cross-pollination. The ratio of male:female flowers vary but for most cultivars it is between 4:1 and 7:1.
Taxonomy and Morphology
Paspalum vaginatum Swartz is commonly known as seashore paspalum, or most simply paspalum. Other common names that are less frequently used include sand knotgrass, siltgrass (Morton, 1973), and saltwater couch (Duncan and Carrow, 1999). Several other scientific names, primarily Paspalum distichum, have also been used in publications since 1788 (Casler and Duncan, 2003). The nomenclature committee for Spermatophyta selected Paspalum vaginatum Swartz as the official scientific name in 1983 (Brummitt, 1983). Seashore paspalum is classified taxonomically as Poaceae (Family), Panicoideae (Subfamily), Panicodae (Supertribe), Paniceae (Tribe), Setariinae (Subtribe), Paspalum (Genus), Disticha (Group), vaginatum (Species), and Swartz (Authority) (Duncan and Carrow, 1999).
Paspalum vaginatum has a chromosome number that is predominately diploid (2n = 2x = 20) (Bashaw et al., 1970; Burson, 1983). However, several reports of tetraploid (2n = 4x = 40) and hexaploid (2n = 6x = 60) accessions have been documented in literature (Casler and Duncan, 2003). Seashore paspalum has the D genome (Bashaw et al., 1970; Burson, 1983), making interspecific crosses with other Paspalum species, and production of viable seed very challenging.
Morphologically, seashore paspalum has both stoloniferous and rhizomatous growth habits (Webster, 1987), with the stolon nodes being noticeably pubescent. It also possesses conduplicate (folded) vernation (Green et al., 2012), and auricles are absent on the plant. Flowering culms (stems) are basally decumbent or erect, and can vary in height from 8 to 60 cm with 5-13 hairless nodes. The leaves of the plant arise on opposite sides of the stem; two vertical rows of leaves are present, and are approximately 50-220 mm long by 1-4 mm wide. The ligule is around 1 mm in diameter, membranous, and hairy. The inflorescence (seed head) commonly consists of two racemes (branches) that are 20-60 mm in length; each of these branches has around 16-32 twin-rowed spikelets. Anthers are 1.2 to 1.6 mm in length, and are yellow and purple striped, while stigmas are feathery and purple (Figures 2 and 3). Paspalum vaginatum seeds (caryopses) are 1.5 mm wide, and 2.5-3.0 mm long, ovate with the narrower end at the base, and concave on one side, while convex on the other (Casler and Duncan, 2003). The weight of one gram of seed is equal to approximately 1760 seeds (Raymer, personal communication, 2012). Seashore paspalum leaf textures range from very coarse (forage types) to fine-leaf types (golf course types).
Germplasm and Cultivars
In 1993, Dr. R. R. Duncan initiated the first seashore paspalum breeding program at the University of Georgia Griffin Campus (Raymer et al., 2008), and continued his work until 2003. During his tenure with the breeding program, he amassed a collection of approximately 250 accessions and began to intensely breed these ecotypes for use in turfgrass. Dr. P. L. Raymer has continued this work since 2003, and maintains what is currently the largest collection of seashore paspalum germplasm (Raymer, personal communication, 2012). Several smaller collections exist in Argentina, Brazil, the University of Florida, and the National Genetic Resources Program (NGRP) of the USDA-ARS in Griffin, Georgia (Casler and Duncan, 2003).
The University of Georgia seashore paspalum breeding program has released several improved cultivars of seashore paspalum since its inception in 1993. Duncan (2003) provides a complete list of seashore paspalum cultivars and their origin prior to 2001. Differentiating these varieties is often difficult, because most commercially grown cultivars are fairly morphologically similar. Amplified fragment length polymorphism (AFLP) is used as a tool to identify cultivars for breeding programs, and to protect intellectual property rights (IPR). The use of simple sequence repeats (SSR) is also increasing being used and provides an additional tool for genotype identification (Raymer et al., 2008).
Reproduction Characteristics and Propagation
Seashore paspalum is a sexually reproducing, cross-pollinating diploid that possesses a degree of self-incompatibility (Carpenter, 1958). Espinoza and Quarin (1997) determined the self-incompatibility is gametophytic, when their studies showed that pollen germinated and penetrated the stigma, but failed to grow in the style. Paspalum vaginatum is heterozygous and rarely cross-pollinates with other paspalum species, because of its D genome. In some instances, crossing closely related species can be accomplished during the early morning hours (0500-0730) at a temperature of around 18-21°C (Burson, 1985). These crosses, with Paspalum vaginatum as the pollen parent, often are not successful, and typically produce sterile hybrids. However, Paspalum vaginatum is capable of successful cross-pollination with other ecotypes and accessions of seashore paspalum, creating genetic recombination that can be utilized in a breeding program (Shin et al., 2006).
Paspalum vaginatum is normally vegetatively propagated from sod or sprigs, but one commercially marketed seeded cultivar is available. Spriging rates can vary from 19-76 m3/ha (5 to 20 bushels per 1000 ft2) (Raymer, personal communication, 2012). In general, seashore paspalum establishes slightly slower than bermudagrass, but is comparable to most other warm-season turfgrasses due to it stoloniferous and rhizomatous growth habit (Duncan and Carrow, 1999). Seaspray® is the only commercially available seeded cultivar at this time and should be seeded at rates between 4.9-9.8 g/m2 (1 to 2 lbs./1,000 ft2) (Raymer, personal communication, 2012).
Breeding Strategies, Techniques, and Hybridization
Until recent years, warm-season turfgrass cultivar development, including seashore paspalum improvement, was rather difficult due to a lack of information on how to make successful crosses. However, laboratory research has helped to determine the proper conditions to make sexually compatible seashore paspalum crosses that produce seeds that lack dormancy and are viable (Shin et al., 2006). Given this information, The University of Georgia seashore paspalum breeding program is able to utilize traditional breeding techniques to hybridize ecotypes, and create genetic diversity that can be used in the development of new varieties.
The creation of F1 hybrids can be done using hand pollination or by using open-pollinated polycrosses. Hybridization usually takes place in a greenhouse rather than in a field setting. Hand pollinations are performed by following these steps: 1) cover racemes (flowers) with small glassine bags supported by wire stakes prior to stigma appearance, 2) transfer pollen from male plant flowers to desired female parent after stigmas have emerged, 3) replace bags over racemes after hand pollination (Figure 4), and 4) wait for viable seed to develop (approximately 1 month) (Raymer, personal communication, 2012). Successful crosses will yield around 2-10 F1 seed that are then harvested by cutting off flower racemes, and threshing seeds after they have dried down. These seeds are then germinated (Figure 5) and grown into mature seashore paspalum plants. The potted plants are then placed in nylon mesh cages (Figure 6), and full-sib mated to produce hundreds of unique individuals per family. Placing flowering mature Paspalum vaginatum plants near one another for approximately 6 to 8 weeks creates F1 hybrids from polycrosses with an unknown male parent. Mass selection is then used to identify desirable phenotypes (Raymer, personal communication, 2012).
In vitro propagation has been used recently to obtain sethoxydim resistance in Paspalum vaginatum (Heckart et. al, 2010). Callus (Figure 7) was plated on an induction medium that contained sethoxydim to select for mutant herbicide resistant cells. This method provides an alterative to genetic modification (GM) to manage bermudagrass and other weedy grasses with herbicide resistance technology. Very little molecular marker technology is currently utilized in seashore paspalum breeding, but the use of simple sequence repeats (SSRs) will likely increase in the near future (Wang et al., 2006), and should help make development of marker assisted selection possible for major traits of interest. At this time, genetic modification is not used in seashore paspalum breeding programs due to pollen escape concerns, and regulatory issues.
Breeding for Abiotic Stress Resistance
Seashore paspalum’s coastal origin exposed the species to several abiotic stresses during its evolution that enabled it to develop numerous stress tolerance and resistance mechanisms. Close approximation to coastal areas, seawater sprays, and some instances of salt waterlogging has made Paspalum vaginatum inherently more salt tolerant than their warm-season turfgrass counterparts in most cases (Duncan and Carrow, 1999). However, Lee et al. (2004) demonstrated that different accessions of seashore paspalum vary greatly in their tolerance to salt with some lines being highly tolerant and others performing no better than a typical hybrid bermudagrass. The trend for turfgrasses to be increasingly irrigated with water that has higher salinity concentrations, and development of golf courses closer to the ocean will drive the need for the next generation of salt tolerant seashore paspalum varieties (Raymer et al., 2008). Screening for salt tolerance (Figure 8) early in the breeding program is a necessary component of breeding for multiple stress resistance and development of improved cultivars. Raymer et al. (2005) describes an effective and efficient procedure to screen for this abiotic stress.
Wear and traffic tolerance is also an important abiotic stress that needs to be selected for in Paspalum vaginatum. Wear and traffic injury tolerance consists of how well a turfgrass stand responds to plant tissue tearing, or physical destruction of an area of turf from events such as divots taken by a golf club on a fairway, cleat marks on an athletic field, and maintenance vehicles driving repeatedly over the same area of seashore paspalum (Duncan and Carrow, 1999). Carrow and Petrovic (1992) showed that turfgrasses with higher shoot and root densities in turn have the best tolerance to traffic and wear stresses, and fill in damaged areas quicker than slow growing, less dense turfgrasses. To test for these stresses, equipment such as the Brinkman traffic simulator can be used to simulate traffic and wear on a turfgrass stand at varying levels of intensity (Cockerham et al., 1990).
Breeding for drought resistance and avoidance in Paspalum vaginatum will be necessary with escalating water quality and quantity concerns (Carrow et al., 1990). According to Huang et al. (1997), seashore paspalum lines have significant genetic diversity in respect to their drought tolerance, and can be assessed using a stress index developed prior to these studies. Other future breeding objectives could focus on cold temperature and shade tolerance of different Paspalum vaginatum ecotypes. Tolerance to mowing at heights that range from greens (3-5 mm) to roughs (25-50 mm) is also a necessary trait that an improved seashore paspalum cultivar must possess (Duncan and Carrow, 1999).
Breeding for Biotic Stress Resistance
Seashore paspalum is planted and maintained perennially using a single commercially available cultivar on golf courses, athletic fields, and for other recreational and aesthetic uses. This uniformity, in conjunction with stresses induced from routine mowing, and the high amounts of clippings and litter present in the turfgrass system from this management practice, can lead to disease problems and epidemics in Paspalum vaginatum (Casler and Duncan, 2003). However, seashore paspalum is not as susceptible to diseases as many other warm-season turfgrasses, likely due to its littoral evolvement that exposed the species to multiple pathogens in a moist environment (Duncan and Carrow, 1999). Seashore paspalum is also no longer exclusively grown on these coastal areas, and when grown inland, may not have the advantage of using salt to manage pathogens and weeds (Duncan and Carrow, 1999). Dollar spot, caused by Sclerotinia homoeocarpa F.T. Bennett, is a major fungal disease that causes turf quality and playability issues. Screening by Raymer et al. (2008) showed substantial differences in dollar spot resistance among eight standard seashore paspalum varieties. Germplasm with resistance can be used as parents in the breeding program to develop improved cultivars with better resistance than what is currently available. Screening for pathogens, insects, and nematodes early in cultivar development, as well as in regional and National Turfgrass Evaluation Program (NTEP) trials, are necessary to produce Paspalum vaginatum lines with resistance to multiple biotic stresses (Raymer, personal communication, 2012).
Experimental Design for Breeding Efforts
Experiments are conducted to select grass genotypes that could later be released as a variety. Most of these trials involve screening for a particular trait of Paspalum vaginatum. Plot size can vary from a pot containing a single clonal propagation early in cultivar development, to larger (6 m2) plots during collaborative, regional evaluation. In the initial field trials, only one replication may be available or used, but several (at least 4 replications) should be planted of the best lines closer to release. Testing improved lines in NTEP also provides replication across multiple locations. Industry standard cultivars of seashore paspalum and other warm-season turfgrass species should be planted with experimental lines as a control (Raymer, personal communication, 2012).
Phenotyping and Data Collection
Phenotyping is often done using visual ratings throughout the course of a breeding program or study according to NTEP guidelines. Before making any ratings, it may be helpful to identify the best and worst performing seashore paspalum lines as a reference. Visual ratings are taken for most aspects of turf on a 1 to 9 scale, with one performing the poorest and nine being the highest rating. Many raters only use whole numbers, as it may be difficult to differentiate similar performing lines with the human eye. Turfgrass quality ratings (6 is acceptable) must take into account the functional and aesthetic aspects of the turfgrass stand, and what the seashore paspalum cultivar will be used for upon commercial release. Turfgrass density reflects the number of plants or Paspalum vaginatum tillers in a given area, with a rating of 9 equaling maximum density. Pests in a seashore paspalum line can also be measured on the 1 to 9 scale, with 1 reflecting 100% injury from the disease or insect, and 9 reflecting 0% injury and a possible resistant line (Morris and Shearman, 2008). Other ratings can be taken for color, percent cover, or traffic and drought tolerance in a similar fashion.
Data from salt screening at various levels of saline water concentration can be obtained by trimming the Paspalum vaginatum lines to a specified height, collecting and drying clippings, recording the dry weight, and then comparing the results to other breeding lines. Greater clipping dry weights is generally a good indication that a possible cultivar performs well in various salt laden conditions over time, but visual ratings, much like those described by NTEP, can be valuable data as well (Raymer, personal communication, 2012).
Richardson et al. (2001) introduced a way to quantify turfgrass cover and color (Karcher and Richardson, 2003) using digital image analysis (DIA). This technique utilizes four incandescent light bulbs inside a light box that are powered by a portable battery or generator (Figure 9), with all components attached to a dolly cart. A small circular hole is on top of the box that allows a camera to take pictures using only the light provided by the light bulbs inside the box. These images are then processed using different macros in SigmaScan Pro (SPSS Inc., 1998) to provide turfgrass breeders with unbiased ratings of their plots for several different facets of turfgrass quality. Numerous other phenotyping equipment options are available for use in a turfgrass breeding program, but are not included in this guidebook for simplicity reasons.
Promising seashore paspalum lines are screened in several phases for multiple abiotic and biotic stress resistances in environmental conditions that emulate where they will be utilized. These screening phases occur after initial F1 populations are created from hand pollination or poly-crosses. After the full-sib mating, approximately 5,000 individual plants are tested for their tolerance to saline water. The visual quality of unreplicated experimental Paspalum vaginatum lines that survive high concentrations of salt are assessed, and then around 2,000 lines are planted into a single plot (space dependent, but no less than 5 m2) at a single location. These plots (Figure 10) are managed rather intensively to find lines that perform well under typical turfgrass growing conditions (routine mowing, as well as regular irrigation and fertilizer applications). Each individual plot is rated monthly for overall turfgrass quality and other factors, depending on the main breeding objectives for creating a new cultivar. Approximately 3-5% (60 to 100) of the best individuals are then selected for further evaluation. These high performing lines are then clonally propagated, tested in replicated greenhouse salt screens, and the best 30 to 50 lines are tested in larger replicated plots (6 m2), with specific management protocols at multiple locations. After two to three years of intensive evaluation, one to three lines may be selected for more extensive evaluation, and under various management protocols. In this stage, prospective cultivars might be entered in NTEP trials, and established on golf courses for evaluation and demonstration. Breeder stock of candidates for release as cultivars begins with a single stolon clonally propagated in the greenhouse, until sufficient material is available to plant a field block of breeder’s material. Plants in this block are closely monitored for off-types and contaminates of other species, and eventually used to develop a larger planting of foundation material, no less than one hectare in size. Upon successful release of the cultivar, foundation plant material is used to establish certified commercial production fields on a sod producer’s farm, which is licensed to then sell the new cultivar to golf courses and other customers. (Raymer, personal communication, 2012).
Dane F., Liu J. (2007) Diversity and origin of cultivated and citron type watermelon (Citrullus lanatus). Genetic Resources and Crop Evolution 54:1255-1265.
Hawkins L.K., Dane F., Kubisiak T.L., Rhodes B.B., Jarret R.L. (2001) Linkage mapping in a watermelon population segregating for Fusarium wilt resistance. Journal of the American Society for Horticultural Science 126:344-350.
Levi A., Thomas C.E., Thies J., Simmons A., Xu Y., Zhang X., Reddy O.U.K., Tadmor Y., Katzir N., Trebitsh T., King S., Davis A., Fauve J., Wehner T. (2004) Developing a genetic linkage map for watermelon: polymorphism, segregation and distribution of markers, Cucurbitaceae 2004, the 8th EUCARPIA Meeting on Cucurbit Genetics and Breeding, Olomouc, Czech Republic. pp. 515-523.
Ren Y., Zhao H., Kou Q., Jiang J., Guo S., Zhang H., Hou W., Zou X., Sun H., Gong G., Levi A., Xu Y. (2012) A high resolution genetic map anchoring scaffolds of the sequenced watermelon genome. PLoS ONE 7:e29453.
Robinson R.W., Decker-Walters D.S. (1997) Cucurbits CAB International Publishing, Wallingford, UK.
Sain R.S., Joshi P., Divakara Sastry E.V. (2002) Cytogenetic analysis of interspecific hybrids in genus Citrullus (Cucurbitaceae). Euphytica 128:205-210.
Sandlin K.C., Prothro J.M., Heesacker A.F., Khalilian N., Okashah R., Xiang W., Bachlava E., Caldwell D., Seymour D., White V., Chan E., Tolla G., White C., Safran D., Graham E., Knapp S.J., McGregor C.E. (2012) Comparative mapping in watermelon [Citrullus lanatus (Thunb.) Matsum. et Nakai]. Theoretical and Applied Genetics 125:1603-1618.