WP/L at the 99.9% level of confidence, as determined by a 
Student’s f-test. Statistical analyses were not made on the other 
sets, due partly to absence of garden material of C. ruderalis in 
12. 
Garden plants maintained their relative growth patterns: /.e., 
height and branching; but leaves were noticeably larger under 
cultivation (compare 1976 and 1972 versus “wild” in Table 1). 
Although leaves were larger, dimensional ratios of the central 
leaflet (W/L and WP/L) did not vary significantly between 
garden-grown and wild samples. 
Massed data for all three collection groupings are given for the 
four categories in Table 2. None of these categories is significantly 
distinct for leaflet length, although C. ruderalis consistantly has 
small leaves. The small-seeded, narrow leaved C. sativa (SS) from 
India and Pakistan is not so significantly distinct from C. sativa 
(S) as is C. indica. No formal infraspecific status is proposed for 
those small-seeded plants, but further study is warranted. 
The three species, C. sativa, C. indica, and C. ruderalis, are well 
defined in leaflet width/length ratios, and the latter two are also 
distinguished from C. sativa in their oblanceolate leaflets 
(WP/L). Leaf morphology groupings reported here are 
compatible for those of the holotypes of C. sativa and C. indica; 
the type specimen of C. ruderalis was not available for 
measurement. 
Modal leaflet number for C. sativa was 7 with a mean of 6.35. 
Leaves of C. indica had a mode of 9 (mean, 8.20), and C. ruderalis 
had a modal leaflet number of 5 (mean, 4.59). No significant 
differences were found in leaf morphology between sexes of a 
given species, but in a few populations the female plants had 
wider leaflets than did the males. 
Leaf character sets were generally reinforcing: i.e., leaves of C. 
indica that were unusually narrow and thereby somewhat like 
those of C. sativa were also very oblanceolate. The most nearly 
intermediate leaf morphology was found in a stout Japanese 
cultivar of C. sativa with W/L = .143 and WP/L = .524. It was 
unique in having smooth stalks (few if any trichomes) and no 
THC content. Multivariate analysis would have dramatized the 
distinctness of the foliar characteristics of the species more 
completely, but f-tests were considered ample. 
63 
