CHARACTERS IN AX UPLAND-EGYPTIAN COTTON HYBRID. 
47 
A few of the disherences listed in Table 13 may be explained as 
due to physical or physiological correlation. This is certainly true of 
the correlations leaf index with leaf extension index and leaf vein 
angle with leaf extension index, both of which gave positive and very 
significant coefficients in F t , as well as in F 2 . It is probably true also 
in the following cases, for the F, coefficients, although apparently not 
significant, were of the same sign as the F, coefficients: 
Bract length with boll diameter. 
Corolla length with stamen length. 
Boll length with boll diameter. 
It is evident that the list of disherences may be reduced slightly 
by allowing for inter correlation. This applies to the correlations with 
bract dentation of axis length, internode length, internode number, 
and fruiting branch length, these stem-and-branch characters being 
intercorrelated positively. There remains, however, nearly a score 
of disherences which can not be explained away thus easily. A strik- 
ing example is the correlation of callus color with petiole hairiness. 
The Holdon parent being characterized by a highly colored callus and 
by hairy petioles, the correlation in F 2 , if of a genetic nature, should 
have been positive, but it proved to be both negative and highly sig- 
nificant, the coefficient having been 6.5 times its probable error. It 
does not seem probable that physiological correlation as ordinarily 
understood is involved in this case. 
Xo recognized principle in genetics seems adequate to explain the 
occurrence of disherence, of which so many apparently significant 
examples were encountered in this cotton hybrid. It may be noted 
that Collins and Kempton (.9, pp. 33, 34) were confronted with the 
same dilemma in studying the correlations of characters in a hybrid 
of teosinte and maize. 
Table 14. — Significant and coherent correlations in F 2 of the Holdon-Pima hybrid which 
are not apparently physical or physiological. 
[The value r -s- E is that obtained by dividing the coefficient of correlation by its probable error.] 
Correlation. 
r 
E 
Correlation. 
r 
E 
Axis length and fiber color 
3.7 
4.2 
4.0 
4.1 
4.3 
5.2 
3.9 
4.6 
4,8 
5.1 
4.5 
Corolla length and petal spot 
Fruiting branch lenglh and calyx dentation. 
Fruiting branch internode and rnidloek 
furrow. 
Vein angle and petal color 
Corolla length and boll apex 
Petal color and fiber length 
3.8 
6.0 
Ftamen length and anther color 
5. l 
Callus color and boll diameter 
Petiole hairiness and bract connation 
Petiole hairiness and bract tooth spread 
Calyx dentation and petal spot 
4.6 
, Pistil length and Iock number 
3.7 
Pistil length and boll apex 
Boll length and fiber length 
Calyx dentation and midlock lurruw 
4.4 
Calyx glands and stigma index 
3.6 
Corolla length and petal color 
The cases of significant coherence become reduced to 22 if correla- 
tions of a more or less obviously physical or physiological nature be 
omitted. These 22 correlations are listed in Table 14. To what 
extent they are of a genetic rather than a physiological nature it is 
impossible to judge in the absence of direct evidence. A positive 
correlation between such characters as corolla Length and petal spot 
would seem to belong to the genetic category, but it happens that in 
connection with another investigation the relation between this pair 
