Cucurbit Genetics Cooperative Report 10:69-71 (article 36) 1987
Cucurbita fraterna, the Closest Wild
Relative and Progenitor of C. pepo
Thomas C. Andres
Department of Horticultural Science, New York State Agricultural Experiment
Station, Geneva, NY 14456
Cucurbita fraterna Bailey is one of the least known species
of Cucurbita. Our knowledge has been based entirely on the holotype,
a single herbarium specimen collected by C.L. Lundell in 1937 that contains
no pistillate flowers, fruits, seeds or roots. Recently, T.C. Andres, M.
Nee, and J.J. Wyland collected the first complete specimens of C.
fraterna, including seeds for germplasm, from two populations in
northeastern Mexico. One of these populations represents the topotype (collected
at the type locality near Llera, Tamaulipas). The other population was found
50 km northeast of this site. Some previously unidentified Cucurbita
specimens collected in 1971 by J.V.A. Dieterle (nos. 3800 and 3804, deposited
at the University of Michigan Herbarium), now appear to represent two additional
populations of C. fraterna in Nuevo Leon, ca. 325 km northnorthwest
of the type locality.
When Bailey (3) described C. fraterna, he believed it was
closely related to C. texana (Scheele) Gray, hence the specific
epithet from the Latin 'frater' (brother) of C. texana. (Another
less closely related species, C. sororia Bailey, had already
been named the 'soror' (sister) of C. texana). C. fraterna
is similar to C. texana but differs in leaf shape and lobing,
trichome morphology, and natural distribution.
C. texana has stimulated a considerable number of biosystematic
investigations (5), since it has been considered the progenitor of C.
pepo L. (1,2,3,4,6,11). This hypothesis is flawed because C.
texana has a fairly restricted range and does not occur in Mexico,
where the oldest archaeological records of C. pepo have been
found (7,10) and where primitive landraces of C. pepo are
still being grown today (12). Furthermore, C. texana contains
only a small, atypical subset of the isozymes found in various C.
pepo cultivars (9; Andres, unpub.). Ancestral species typically exhibit
greater genetic diversity than recently derived species, although this may
not necessarily be true in outcrossing species under artificial selection
(8). Thus, no alternative hypothesis existed regarding the origin of C.
C. fraterna has fibrous roots and the plants are annual
like the related taxa. The round fruits are similar in coloration and shape
to C. texana, although most populations of C. texana
also contain pyriform-shaped fruits. The round and pyriform fruit types
resemble 'Miniature Ball Gourd' and 'Striped Pear Gourd', respectively,
both being types of ornamental gourds, C. pepo var. ovifera
The fruits of all wild Cucurbita species and most ornamental gourds
are extremely bitter, but a non-bitter form of C. fraterna
was described on a herbarium label (Dieterle, no. 3804). Similarly, non-bitter
C. sororia fruits have been found (R.A. Bye, personal communication).
The fruits of C. sororia differ from C. fraterna
and C. texana by lacking very definite carpellary stripes
(3). C. sororia appears to represent the wild progenitor of
C. mixta Pang. (L.C. Merrick, personal communication). A population
of C. sororia (Andres and Nee, collection no.177) was found
growing near the type locality of C. fraterna. This species
has not been previously reported in the state of Tamaulipas.
C. fraterna generally occurs in upland, seasonably dry
thornscrub habitat whereas C. texana is restricted to river
bottomland. C. texana fruits readily abscise from the peduncles
earlier in development than those of C. fraterna and C.
sororia. This may represent an adaptation for dispersal by floating
during spring floods.
Accessions of C. fraterna, C. texana, C.
pepo, C. sororia, and other Cucurbita species
were grown in 1986 at Geneva, NY. Crosses were made and isozymes compared
using horizontal starch gel electrophoresis. A total of approximately 25
loci representing ten enzymes was examined. The isozyme data indicate that
considerably more allelic variation exists among the two sampled populations
of C. fraterna than among 20 sampled populations of C.
texana. Unlike C. texana, the alleles of C.
fraterna are all commonly found in cultivars of C. pepo.
A landrace of C. pepo with elongated, medium size Jack-O-Lantern
like fruit, found near the populations of C. fraterna, was
indistinguishable from C. fraterna in the enzymes sampled,
and in its vegetative morphology. This suggests that C. pepo
pumpkins may represent an early domesticated form of C. fraterna.
The sympatric C. sororia population did share many alleles
with C. fraterna, but also contained several novel alleles
not found in C. pepo.
Based on this information, pollinations were made primarily between C.
fraterna and various accessions of C. pepo, C.
texana, C. sororia and, to a lesser extent, other less
closely related species. C. fraterna was readily crossable
with all types of C. pepo, including C. texana,
but was much less compatible with other species of Cucurbita. C.
sororia can be crossed with C. fraterna, but with poor
seed set. The crossing data are therefore congruent with the isozyme data.
C. fraterna appears to be the original wild progenitor
of C. pepo, based on the genetic evidence, occasional occurence
of non-bitter fruits, proximity to archaeological sites of C. pepo,
and similarity to traditional landraces of C. pepo. C.
fraterna and C. texana may have originally been incipient
species or ecotypes. But secondary contact with C. texana may
have occurred when C. pepo (i.e., domesticated C. fraterna)
spread northward by humans into eastern U.S., introducing new sources of
genetic variability. If future evidence proves this to be the case, C.
pepo would represent a compilospecies composed of the originally
separate taxa, C. fraterna and C. texana. Furthermore,
introgression may have occurred with populations of C. sororia.
The astonishing range of cultivars present today in C. pepo
may have arisen over a period of 10,000 years by a complex pattern of (A)
multiple incipient domestications among the small, semi-isolated populations
of C. fraterna, (B) dispersal to new areas, resulting in various
genetic-environmental interactions, and (C) a generally reticulated pattern
of occasional hybridization both among landraces and between these taxa
and related wild taxa. Additional germplasm collecting and molecular analysis
will help clarify these presumptive patterns of domestication.
Based on the evidence given, C. pepo should be classified
as containing three subspecies: C. pepo ssp. pepo,
C. pepo ssp. fraterna, and C. pepo ssp.
- Bailey, L.H. 1929. The domesticated cucurbitas. Gentes Herbarum 2:61-115.
- Bailey, L.H. 1930. Three discussions in Cucurbitaceae. Gentes Herbarum 2:175-186.
- Bailey, L.H. 1943. Species of Cucurbita. Gentes Herbarum 6:265-322.
- Carter, G.F. 1945. Plant Geography and Culture History in the American
Southwest. no.5. New York: Viking Fund Publ. in Anthropol.
- Decker, D.S. 1986. A biosystematic study of Cucurbita pepo.
Doctoral Thesis. Texas A&M Univ., College Station.
- Erwin, A.T. 1938. An interesting Texas cucurbit. Iowa State Coll.
J. Sci. 12:253-261.
- Heiser, C.B. 1985. Of Plants and People. Univ. Oklahoma Press,
- Hutchinson, J.B. (ed.) 1965. Essays on Crop Plant Evolution. Cambridge Univ. Press, London.
- Kirkpatrick, K.J., D.S. Decker, and H.D. Wilson. 1985. Allozyme differentiation
in the Cucurbita pepo complex: C. pepo var. medullosa vs. C. texana. Econ. Bot. 39:289-299.
- Whitaker, T.W. 1981. Archeological cucurbits. Econ. Bot. 35:460-466.
- Whitaker. T.W. and W.P. Bemis. 1975. Origin and evolution of the
cultivated Cucurbita. Bull. Torrey Bot. Club 102:362-368.
- Whitaker, T.W. and R.J. Knight. 1980. Collecting cultivated and wild
cucurbits in Mexico. Econ. Bot. 34:312-319.