92 – II.2. SQUASHES, PUMPKINS, ZUCCHINIS, GOURDS (CURCURBITA SPECIES)
endemic to Argentina and Uruguay (Martínez-Crovetto, 1974; 1965; 1954), and
C. ecuadorensis, known only from the coast of Ecuador (Cutler and Whitaker, 1969; Nee,
1990). C. maxima ssp. andreana is the most probable wild ancestor of this crop (Millán,
1945; Fursa and Filov, 1982; Nee, 1990; Sanjur et al., 2002).
From the 16th century, several types of C. maxima, such as the turban type, were
transported directly from South America to Europe. Many other cultivars reached
Australia, Africa and Asia, where local landraces evolved. In the 19th century, several
cultivars were introduced into the United States from South America (Decker-Walters
and Walters, 2000). Secondary centres of diversity include Bangladesh, Burma, India and
the southern Appalachians of the United States, e.g. the landrace “Candy Roaster” was
originally developed by the Cherokee people in the southern Appalachians.
These findings suggest that for C. maxima, in addition to the regions of South America
mentioned above, multiple centres of diversity, primarily composed of landraces, exist
around the world.
Cucurbita moschata
The natural distribution of Cucurbita moschata is from the lowlands of Mexico into
Central America (Table 2.1). Cucurbita moschata was domesticated in Latin America
(Whitaker, 1947) but there is no consensus as to the precise area where domestication
likely occurred. It has been proposed that C. moschata was domesticated in Mesoamerica
(Whitaker and Davis, 1962) or alternatively in South America, more specifically in what
is now Colombia. However, available evidence has been difficult to interpret and the
centre of origin/domestication question is still open to debate. The oldest archaeological
remains of C. moschata (4900-3500 BCE) have been recovered from the Ocampo caves,
in the state of Tamaulipas, in northeast Mexico; however, very early dates have been
registered for several localities in Central America (2000 BCE-850 CE) and South
America (2700-300 BCE; Table 2.2). One of the main arguments against the
South American origin hypothesis is that C. moschata is capable of producing highly
fertile hybrids with the wild taxa of the C. argyrosperma group (Merrick, 1990), which
has an identified centre of origin from the southwestern United States to the centre-south
of Mexico. Morphological and ecological studies as well as comparative mitochondrial,
ribosomal and chloroplastic analyses (Wilson, Doebley and Duvall, 1992; Jobst, King and
Hemleben, 1998; Sanjur et al., 2002) suggest that the ancestor of C. moschata might have
derived from a wild taxon of C. argyrosperma ssp. sororia. However, C. moschata and
C. argyrosperma have different isoenzymatic patterns (Sanjur et al., 2002). Although
studies by Merrick (1990, 1991) and Sanjur et al. (2002) support a high level of
relatedness between C. moschata and members of the Argyrosperma group, these authors
do not support the possibility of the ancestor being the subspecies sororia. C. lundelliana
has also been proposed as the C. moschata ancestor, and that wild taxon is indigenous to
the Yucatan Peninsula in Central America (Whitaker, 1974). However, there are several
morphological differences between C. lundelliana and C. moschata (i.e. the greenish-
grayish-blue seed colour in C. lundelliana has not been seen in C. moschata). Results
from molecular biology studies (Puchalski and Robinson, 1990; Wilson, Doebley and
Duvall, 1992) have also lent evidence to exclude Cucurbita lundelliana as a possible
ancestor. Some characteristics associated with C. moschata have been identified in
landraces from Bolivia, Columbia and Panama (e.g. dark-coloured seeds, small fruits, a
lignified and warty rind; Wessel-Beaver, 2000b), suggesting hybridisation between
C. moschata and wild local species in Columbia (Nee, 1990). Based on these sets of
information, some authors have suggested the existence of two independent
