a slight thickening of the outer layer of the testa suggests 
the former presence of wings. It seems reasonable to as- 
sign this eroded seed, as well as similar specimens from 
the other sites described, to the broad, winged type de- 
scribed by Whitaker. The site of Castillo de Tomaval 
is essentially of the Gallinazo period which follows the 
Cupisnique and the earlier Preceramic horizons. 
In addition to the Lagenaria seeds described above, 
there are references in the ethnobotanical literature of 
prehistoric Peru to the occurrence of seeds of this plant 
in still other archaeological sites. Wittmack (1880-1887) 
identifies a gourd seed among the plant remains from 
Ancon on the central coast, and he gives an illustration 
(Taf. 107, fig. 17). Also, Costantin et Bois (1910, fig. 12) 
picture a seed recovered from another site on the central 
coast of Peru. Both of these seeds belong to the broad, 
winged type recovered at Huaca Prieta. Harms (1920) 
found fragments of gourd shells and seeds in the mummy 
wrappings from Ancon, but he neither illustrates nor de- 
scribes this material. Carter (1945), in his study of certain 
archaeological cucurbit seeds from Peru, mentions the 
presence of 25 seeds of Lagenaria from Chincha on the 
south coast and attributes them to the late Inca period, 
ec. 1300-1500 A.D. He gives their size as ‘‘17x7x8; 
15x7x3”’ and further states that: ‘*These are dark brown 
seeds with light longitudinal stripes and are of the char- 
acteristic Lagenaria size and shape.’’ In the absence of 
illustrations or any mention of winged protuberances in 
the descriptions of the seeds recovered by Harms or those 
studied by Carter, one can not say definitely to which 
type of Lagenaria seed they belong, although on the 
basis of the few descriptive remarks that are given, one 
leans toward classifying both groups as of the broad, 
winged type of Lagenaria seed. 
As aresult of the present study of Lagenaria seeds in 
[ 179 | 
