Cucumber (Cucumis sativus L.) originated in India, melon (C. melo L.) and watermelon (Citrullus lanatus) in Africa, and squash, pumpkin, and gourd (Cucurbita spp.) in the Americas. Thus, cucumber, melon, and watermelon (including citron) are relatively recent introductions to the New World. Most domesticated species of Cucurbita were introduced from Mexico, Central America, and South America with the migration of native Americans centuries earlier. Wax gourd (Benincasa hispida (Thunb.) Logr.) is from Southeast Asia. Bottle gourd (Lagenaria siceraria (Molina) Stand.) is of African origin. South Asia is the probable center of origin for cultivated species of Luffa. Bitter melon (Momordica charantia L.) is a tropical Old World species. Chayote (Sechium edule (Jacq.) Swartz) is a New World species from southern Mexico and Central America.

Of the New World taxa, only Cucurbita pepo occurs as a significant weed problem in North America. Cucurbita pepo is a morphologically and ecologically diverse species composed of genetically distinct groups of cultivars and free-living populations (i.e., self-sustaining, including both wild and weedy populations). All of these diverse elements are completely interfertile and are classified as shown in Table 1.

Hybridization among Cucurbita species is also possible, with various of the 15 or so able to hybridize with some difficulty. Diversity in C. pepo is rooted in the ancient widespread distribution of free-living populations. Today, these populations range from northeastern Mexico and Texas, east to Alabama and north through the Mississippi Valley to Illinois. They occupy a diversity of environments and ecological niches—from upland, seasonally dry thornscrub habitat in northeastern Mexico, to primarily riverbanks and moist thickets in Texas, to a variety of riparian and other disturbed lowland habitats (e.g., agricultural fields, railroad tracks, highway embankments, etc.) throughout the Mississippi Valley. Different morphological and physiological adaptations have evolved in these areas, including early fruit abscission from the peduncle in response to riverine dispersal in Texas, as well as relatively quick seed germination in response to a shorter growing season in the more northerly populations (Decker-Walters et al. 1993).

Wild native taxa in the US and Mexico are listed in Table 2.

Table 2. Free-living taxa of Cucurbita native to the US or Mexico. All have 2n=40.

In addition, many Old World cucurbits have been reported as feral species in the US and Mexico (Table 3), particularly in the coastal plain from Florida to Texas and into northern Mexico.

Table 3. Feral Old World cucurbit taxa in the US and Mexico (listed by descending importance).

Production Patterns and Cropping Systems in the US
Cucurbits are grown in several commercial cropping systems and are popular garden crops. Worldwide, there may be more squashes grown in home gardens than are grown commercially for sale in local or distant markets. Although consumption figures are not readily available, production estimates are available for the US and many other countries (FAO 1992).

Cucurbit production in parts of the desert southwest US, e.g., Imperial Valley, California, is done on a large scale in areas of intensive agricultural production of a broad array of warm and cool season vegetables and agronomic crops. Weed control in the immediate vicinity of these production fields is generally very good, but control along river and canal banks is generally not carried out. In some of these areas, it is possible to find one or more cucurbits, usually a Cucurbita sp., grown on a small scale.

In the rest of the US, cucurbits are grown on a smaller scale and are not usually part of an intensive vegetable and/or agronomic crop production area. They are spatially and temporally dispersed. Early season production begins in Florida and moves northward to New York and New England in the East, and Michigan and Wisconsin in the mid-west. Weed control in these systems may be more difficult due to increased rainfall and the resultant native plant populations that may often be found growing immediately adjacent to cucurbit fields.

Pests of Cucurbits
Cucurbits are afflicted with a broad array of insect, pathogen, and nematode pests. With the exception of powdery mildew, which is one of few diseases that may be found in most production areas across the US, each production area requires a different complement of pest resistances. New pests (insect, fungal, viral, and bacterial) continue to be identified. Recently described pests include sweetpotato whitefly and silverleaf whitefly, zucchini yellow mosaic virus, lettuce infectious yellows virus, cucurbit aphid-borne yellowing virus, cucurbit yellow stunting disorder virus, squash leaf curl virus (= watermelon curly mottle = melon leaf curl), bacterial blotch of watermelon and melon, vine decline of melon (causal agent yet to be identified), and Monosporascus cannonballus.

Resistance breeding is the most active area of cucurbit germplasm, breeding, and genetics research in the US and worldwide. Most programs use traditional genetic and plant breeding procedures. Mapping (phenotypic, isozyme, molecular) of cucumber and melon has begun and progressed, but the maps are not yet saturated and few linked markers have been identified (Pitrat 1998). There has been little progress in the development of genetic maps of watermelon and Cucurbita spp.

Most pest resistance genes have been found in US or exotic cultivars or in landraces and cross-compatible relatives from centers of origin or diversity. Unsuccessful attempts have been made to produce fertile F1 progeny from crosses of Cucumis metuliferus with Cucumis melo and Cucumis sativus in order to transfer several pest resistance traits from this distant relative to melon and cucumber. However, Cucurbita okeechobeensis ssp. martinezii was successfully used in crosses with Cucurbita maxima and Cucurbita pepo to transfer powdery mildew resistance to these two species (Contin 1978). Through its Asgrow Seed division, Seminis Vegetable Seeds has introduced transgenic resistance to two potyviruses (ZYMV and WMV) and one cucumovirus (CMV) in summer squash (Cucurbita pepo).

Many sources of pest resistance have been identified in cucurbits, although relatively few have been deployed in commercial cultivars (see McCreight 1998). Few of the identified resistance genes in the other cucurbits have been deployed or stacked in commercially available cultivars.

Weed Complexes
Cucurbits are affected by typical warm season weed species, which are controlled by conventional weed management practices. These include a limited group of nonselective herbicides and standard cultivation (discing, harrowing, hand hoeing, and weeding). Genetic resistance to herbicides has not been identified as a priority for cucurbit production.

There are no known feral species of cucumber in the US or Mexico. Citron, which is cross compatible with watermelon, may be a weed in cucurbit production fields (Robinson and Decker-Walters 1997). In Florida, citron is a weed in citrus groves. Dudaim is one of several cross-compatible groups of cultivated melons (Cucumis melo subsp. melo). It was reportedly feral in parts of Texas (Correll and Johnston 1970) and Florida (Wunderlin 1982). Dudaim was a noxious weed in melon production fields and other crops in the Imperial Valley, California beginning in the mid to late 1960’s (K. Mayberry, pers. comm.). It was declared to have been eradicated from the Imperial Valley as well as from the entire state of California in December, 1998 (C. Valenzuela, pers comm).

A number of wild relatives of the squashes (Cucurbita spp.) occur in parts of the US (Table 2). Of these, only free-living populations of C. pepo occur in agricultural settings. In Arkansas, Louisiana, and Mississippi, C. pepo ssp. ovifera var. ozarkana is an aggressive weed in soybean and cotton fields (Boyette et al. 1984; Oliver et al. 1983). Whereas in wild habitats (i.e., those not directly influenced by human activity) individual plants or small groups of plants are widely dispersed along floodplain corridors, in weedy habitats (i.e., those created by human activities), populations are often very dense and cover large areas in agricultural fields.

Wild-habitat populations from northeastern Mexico, Texas, and many parts of the Mississippi Valley have been accepted as indigenous (e.g., Smith et al. 1992) with long histories of occupation in their general areas. However, morphological and isozymic evidence confirms that some of these populations have experienced hybridization and introgression with cultivated material planted nearby (Kirkpatrick and Wilson 1988; Decker-Walters et al. 1993; Smith et al. 1992). Furthermore, this evidence suggests that some weedy populations in Illinois (Decker and Wilson 1987; Wilson 1990), Kentucky (Cowan and Smith 1993; Decker-Walters et al. 1993), and possibly elsewhere (Asch and Asch 1992) may have evolved purely as ornamental gourd escapes, which may or may not have experienced subsequent introgression with other nearby cultivated, weedy, or wild material of C. pepo. In short, the origins, histories, and genetic compositions of wild and weedy populations of this species are diverse. Consequently, it is difficult to make general conclusions about free-living populations based on observations or research conducted on a limited sampling of these populations.

Because of their similar usage as small, hard-shelled autumn decorations, ornamental gourds are typically thought of as a distinct grouping within C. pepo. Isozymic evidence has clearly shown this not to be true, with cultivars having originated in ssp. pepo, ssp. ovifera, and possibly ssp. fraterna (Decker-Walters et al. 1993). What many of these cultivars do share in common, though, are characteristics often ascribed to free-living populations, e.g., tough pericarps and bitter flesh, which serve to ward off predation in the wild. Among the edible cultivars, human selection pressures have yielded characteristics that hinder the cultivar’s ability to persist in the wild (e.g., large, fleshy, non-bitter fruits). Consequently, most cultivars (e.g., pumpkins, zucchinis, crooknecks, etc.) do not survive as long-lived escaped populations in wild or weedy habitats. Although the supposition has yet to be tested, the occurrence of wild-type characteristics in ornamental gourds has led to the hypothesis that feral populations of C. pepo have been principally derived from ornamental gourd escapes (Asch and Asch 1992).

There is little or no evidence that the introduction of pest resistance genes could increase the ability of any of the Old World cucurbits to become established as a noxious weed species. It is unlikely that New World pests would have affected introduced Old World species and prevented their ability to become established as feral species, as these Old World species were cultivated in highly favorable environments. However, new pests continue to be identified, and pest problems on Old World species in the US have become production-limiting over time as production systems matured. In response to emerging pests, resistance genes have been identified and transferred to acceptable cultivars to maintain production in the face of pressure from pest populations.

Dudaim melon became a feral species before genes from any known dudaim accession were used for resistance in cantaloupe or honeydew. Dudaim can easily intercross with all the other melon groups (Robinson and Decker-Walters 1997). There were 41,000 acres (16,400 ha) of melons grown across the US in 1991 (FAO 1992). Dudaim melon remains a minor weed species and is not a problem in production fields in Arizona, California, or Texas. There were ca. 82,000 acres (32,800 ha) of watermelons grown in the US in 1991 (FAO 1992). Citron was brought to the New World for cultivation and became feral over time in the southern US. Citron can easily intercross with watermelon (Robinson and Decker-Walters 1997).

There is no anecdotal or experimental evidence to suggest that pests have a significant effect on Old World feral cucurbit populations; therefore it is feasible that they are susceptible to the same pests as their cultivated cousins. The consequences of one or more pest resistance traits moving from the crop to their feral cousins are unknown.

The situation for the New World Cucurbita is different. Recent studies have concluded that genes will escape from transgenic crops into cross-compatible wild populations (Hancock et al. 1996). The environmental risk of this gene exchange creating aggressive weeds is believed to be dependent on whether or not the transgene is selectively advantageous in native populations. In evaluating the potential hazards of the transgenic, viral-resistant squash ‘Freedom II’ (Cucurbita pepo ssp. ovifera), researchers concluded that the risk of increased weediness caused by spread of transgenic resistance into wild populations would be minimal because wild-habitat populations were not limited by viral infections (Grumet and Gifford 1998). Not sufficiently tested, however, were weedy-habitat populations of C. pepo, which have been serious pests in the agricultural fields of other crops (e.g., soybean, cotton, and corn) in eastern United States for the last 10 to 50 years. Given their agricultural habitat, which promotes high population density and may be within reach of pests and diseases in nearby cultivated fields of cucurbits, it is more likely that weedy populations of C. pepo are under various pest and disease pressures. Consequently, escape of transgenic resistance into these populations could increase their success as aggressive weeds.

Recent experiments under cultivated field conditions (D. Gonsalves, pers. comm. 1999) have confirmed that transgenes (i.e., genes from transgenic constructs) for viral resistance will pass from transgenic hybrids (i.e., wild x transgenic squash) into wild squash genotypes via natural pollen dispersal, and that viral resistance is advantageous to the wild material when this material is exposed to high viral pressure. Yet to be tested is the fate of introduced viral resistant transgenes in wild or weedy populations themselves. It is particularly important to test the impact of transgene transmission on weedy populations since the habitats that these plants occupy are more likely to be subject to viral pressures.

Although past researchers have not generally been interested in or looked for the occurrence of viruses in free-living populations of C. pepo, there are at least two reports of possible viral infection in weedy-habitat populations. Pathologist Doug Boyette (pers. comm. 1999) saw unconfirmed signs of viral infection in a weedy population near Hope, Arkansas in the 1980s. An herbarium label of a plant (T. C. Andres et al. #293, Cornell University Herbarium, 1994) collected from a weedy population in Issaquena County, Mississippi on November 7, 1994 noted, " . . . in a harvested cotton field. A serious weed problem. One young vine was still green with some slight virus symptoms . . . " This putative viral symptom was not confirmed by biological or laboratory assay.

Also not sufficiently examined in earlier experiments with transgenic squash was the risk posed by spread of transgenic viral resistance to some wild-type cultivars (e.g., ornamental gourds) in which the resistance could increase the ability that these cultivars already have to become successful escapes. Whereas most of the edible cultivars do not survive as long-lived escaped populations in the wild, some persistent weedy populations in Illinois and Kentucky exhibited isozymic and morphological evidence of having originated as ornamental gourd escapes. The cultivation of wild-type ornamental gourds throughout northeastern United States threatens to produce future weedy populations. Those weedy populations that find homes in the agricultural fields of other crops may become more aggressive if they possess resistance to agricultural diseases and pests.

Little specific information exists about the two major Old World weed taxa (citron and dudaim melon) or the wild and weedy Cucurbita populations. Table 4 lists specific types of data desired in order to develop a complete assessment of the consequences of gene flow from cultivated to weedy cucurbits. The taxa of interest for such studies include: Cucurbita pepo ssp. ovifera var. ozarkana, Cucurbita pepo ssp. ovifera var. texana, Cucurbita pepo ssp. fraterna, Citrullus lanatus var. citroides, and Cucumis melo subsp. melo Group Dudaim.

Table 4. Information needed to assess feral and wild native species populations.

Transgenic plants of Cucurbita pepo ssp. ovifera var. ovifera (summer squash) have been released for commercial production in the US. Certain site-specific, eco-geographic data and samples (Table 5) for free-living populations of this species would assist in the risk assessment of a crop becoming a weed or serving as a source of genes for its weedy relatives.