564 
Journal of Agricultural Research voi. xxvii, No. 8 
there are two or more measures of the same character. For instance, 
days to pollen and days to silk both obviously measure the same thing— 
season, while alicole index, number of double female alicoles and number 
of single female alicoles form another intercorrelated group, measuring 
essentially the same thing. With all such cases considered and reduced 
to a single coefficient, there remain 13 characters which are correlated 
directly with type of plant. These are given in italics in the last column 
of Table IV. 
Of these 13 characters 12 are coherences showing a tendency for 
maize characters to be associated with the crinkly type of plant, while 
one is a disherence. This disherence is with leaves above the best in¬ 
florescence. Thus the crinkly plants have fewer nodes above the best 
lateral inflorescence than do the noncrinklies; a condition the reverse of 
that of the parental combination. In view of this fact it may be well 
to examine all the correlations with the character “leaves above the 
best inflorescence” shown in Table II. 
In the second generation there are 36 possibilities for correlations with 
the character “leaves above best inflorescence,” and 26 of these indicate 
disherences. Of the 36 coefficients, 17 are in excess of three times the 
error, but when the Fj correlations are considered and the physiological 
cprrelations eliminated, these 17 are reduced to 12. Of the 12 correla¬ 
tions the one dealing with leaves above the upper branch may be elimin¬ 
ated on morphological grounds, leaving 11, of which 10 are disherences. 
These disherent correlations are concerned with the following characters: 
1. Height. 
2. Rows in central spike. 
3. Length of leaf. 
4. Width of leaf. 
5. Width index. 
6. Position of longest leaf. 
7. Leaves on best inflorescence. 
8. Rows in terminal spike of best inflorescence. 
9. Degree of crinkly. 
10. Type of plant. 
It may be of some significance that, of these 10, 6 are characters in* 
volved in the crinkly variation. It will be recalled that there is a tend¬ 
ency among crinkly plants to develop ears at the base of the tassel or in 
the axil of the upper leaf, and it seems not improbable that this tend¬ 
ency is manifested in the present hybrid. If this be true, it follows that 
the coefficients observed are coherences rather than disherences, since 
the character of few leaves above the upper branch becomes another 
measure of the crinkly variation. In the cross between Tom Thumb 
maize and Florida teosinte there were no significant disherences with 
the number of nodes above the uppermost inflorescence; a fact indicating 
that this character was associated with other characteristics of maize. 
It may be concluded that “few leaves above the best inflorescence^ 
in this case is a characteristic of the crinkly variation; it follows that all 
eight of the characters of maize not directly involved in the classification 
of the plants are correlated with the crinkly type. 
It may be urged that while a character such as total leaves would not 
influence directly the classification of plants into two types, it might 
become correlated with type simply as a secondary relationship through 
its association with height. If this were true, the partial correlation of 
type with total leaves for constant height should be zero, when it actually 
is r = - 0.433 ±0.059. 
