182 BOTANICAL GAZETTE [MARCH 
in the general population. Perfect coupling of certain factors 
together with independent combination of others may be possible; 
partial coupling arising from a peculiar gametic distribution may 
be equally possible. 
TABLE Ill 
CorRELATION BETWEEN LENGTH AND SPREAD OF COROLLA IN Fz GENERATION OF 
CROSS BETWEEN J. forgetiana AND N.. alata grandiflora 
Breadth of corolla in millimeters 
Length of corolla in millimeters 
25 30 35 40 45 50 55 60 65 7° det BOBS 
25 I 4 gee 
Boe 5 fa | toll 3 |. 3 ee 
35 I 5 17 26 a1 6 3 Bik 
40 6 2 25 1. 41 31 7 I 136 
45 8 20°} 26 {> 31 26 8 2 I ibe 
5° 3 a2} 730 |. aA | 225 7 I 132 
55 I A ei Na es a 9 4 Fag 
60 I 3 II 32 20 19 9 I 105 
65 Ot eF 1b. at | 82 7 are is 
70 [ I 5b as 4 4 30 
75 7 3 4 I 15 
80 I I 3 2 6 
85 I I an 
4: | 16 50 | 122 | 164 208°} 167 | 67 |. 28 4 ae cot! 
Coef. cor. 0.6100.015 
These are the principal facts collected regarding this cross, 
if the small leaf differences and other minor variations are left out 
of consideration. How unimportant the latter are, can be seen 
by a reference to pl. VIII. There are several suggestions that may 
be made regarding the simple facts obtained, however, that may 
be helpful in further Mendelian interpretations of size complexes. 
Elsewhere‘ it has been shown that the behavior of such char- 
acters in crosses is adequately represented by the segregation and 
4 Amer. Nat. 44:65-82. 1910. 
