562 
Journal of Agricultural Research voi. xxvii, no. s 
in dominance are constitutional, affecting many or all of the characters. 
A striking instance is afforded by hybrids between maize and Euchlaena 
perennis. In these hybrids there is a much higher degree of dominance 
of all the teosinte characters than has been found in hybrids with E. 
mexicana. 
Regarding the three characters graded for the degree of resemblance to 
the crinkly parent and their relations to the measured characters and to 
themselves, since in these cases there was no variation in the F* plants, 
all being normal, there can be no F x correlations. 
Taking up the degree of crinkly which represents an attempt to measure 
the wrinkling of the leaf blade, this character is found in the F 2 to be sig¬ 
nificantly correlated with eight of the eleven characters and with the type 
of plant as arbitrarily classed. These correlations, some of them large, 
are. all in the direction of coherences. 
The degree of lobing which evaluates the amount of lobing at the base 
of the leaf is found to be associated with ten other characters and with 
the type of plant, while the degree of color is correlated with eight 
characters and with plant type. 
Although only a few of these correlations are large, they indicate that 
the crinkly complex of characters forms a correlated group suggesting 
loose linkage. 
Since the plants were classed into crinkly and noncrinkly from their 
general appearance, it becomes of interest to discover to what extent the 
three characters, leaf lobing, crinkling, and color, determined the class 
in which the plants were placed. From an examination of Table III 
it is found that the degree of lobing, the least conspicuous of the three 
characters, has the closest correlation with the type of plant, but that 
this correlation is not as close as that of width index with type of plant, 
this index being a measured character. It would appear, therefore, 
that the relative width of the leaves was an important factor in the 
classification of the plants. 
Since general characteristics were considered in classifying the plants 
into two groups, it may be urged with reason that it was a combination 
of the degree of crinkly, lobing and color that influenced judgment in 
classifying the plants. To test this, the grades of the three characters 
were combined in several ways and the correlations of these various 
combinations with plant class calculated. The coefficients are given 
with the frequency polygons in figures 40 to 43. It is found that the 
combination of degree of crinkly with degree of lobing is very closely 
correlated with plant class, the coefficient being r~ 0.99 ±0.03, the high¬ 
est obtained with type of plant. While it is apparent that any plant 
having a high degree of crinkly and lobing would almost certainly be 
classed as crinkly, it is clear also that these two characters are not closely 
correlated; r = 0.348. The regressions are 0.181 ±0.029 and 0.669 
±0.109 grades. Thus, for each change of one degree in the amount of 
crinkly there is a change of 0.18 degree in the amount of lobing, while 
for each change of degree of lobing there is a corresponding change of 
0.67 degree in the amount of crinkly. 
CORRELATION OP THE CRINKLY VARIATION WITH CHARACTERS OP MAIZE 
Since all of the characters which are thought to be involved in the 
crinkly variations show significant correlations with type of plant, it 
seems reasonable to consider this arbitrary grouping as representing a 
definite segregation. The very high degree of correlation with leaf 
