12 BULLETIN 1321, U. S. DEPARTMENT OF AGRICULTURE 
TaBLE 7.—Coefficients of partial correlation between some of the ear characters 
studied 
Designation of coefficient C.I. No. 77 C.I. No. 119 C22. No. 12024). (C2 E.Noos 
MB Dyes sea as ey od Se ee 0. 0398-0. 0192 0. 0630-+-0. 0270 | —0. 0017-40. 0239 | —0. 025240. 0271 
T BY Can ee ee =, Sees . 0665+ . 0191 . 1037+ . 0268 . 0054+ . 0239 . 0614+. . 0270 
THE CD Se See eee es — .0832+ .0191 | — .0188+ .0271 | — .0963+ .0288 | — .0877+ .0269 
TRS CDae ae ee ee oe een . 3486 . 0169 . 2946+ . 0247 . 3291+ . 0213 . 3027-— .0237 Gl 
TCHAD ao eee eens - 1688+ . 0187— - 1081 . 0268 *, 0912+ . 0237 . 1644+ .0264 
TOTS GC eee oe eS eNO eet oa . 1963 . 0185 . 1744+ . 0263 . 1809+ . 0231 . 1575+ .9264 | 
Considering the data in Tables 6 and 7 together, the correlations | 
indicate three natural groups of characters: (1) Weight of ear with 
its component elements, length of ear, circumference of ear at butt — 
and tip, and weight of cob; (2) percentage of grain; and (8) the ~ 
number of rows of kernels and the number of kernels per row. 
Number of rows and number of kernels per row are correlated 
negligibly with weight of ear when circumference and length of ear 
are made constant. Weight of cob, on the other hand, has a direct 
relation to weight of ear apart from the length and circumference 
of the ear. Finally, the number of rows and number of kernels per 
unit of circumference or length do not seem to have as important — 
a relation to the percentage of grain as might be expected from the © 
emphasis that formerly was placed on tight-fitting kernels. 
CORRELATIONS BETWEEN THE EAR CHARACTERS AND YIELD 
COEFFICIENTS OF THE ZERO AND FIRST ORDERS 
The coefficients of correlation of the zero and first orders for yield 
with each of the eight characters of seed ears are shown in Table 8. 
Coefficients more than three times their respective probable errors 
are printed in boldfaced type and will be referred to as significant. 
Coefficients less than three times their errors will be referred to as 
nonsignificant. The smallest of the significant coefficients, that of 
—0.0583 between yield and butt circumference for constant weight 
of cob in C. I. No. 77, is equivalent to one of 0.18+ on the basis of 
interpretation already discussed. On the same basis the largest 
coefficient, that of 0.2088 between yield and weight of cob for constant 
percentage of grain in C. I. No. 77, is equivalent to one of 0.66+. : 
Yield is correlated positively with weight of ear regardless of what 
other character may be constant, except weight of cob in C. I. No. 77. 
Making weight of ear constant decreases the positive and increases 
the negative significant correlations between yield and the other 
characters except weight of cob in C. I. No. 77. 
The relations between yield and length of ear are much lke those 
between yield and weight of ear, but somewhat less important. 
This is true particularly in C. I. No. 133 when butt circumference is 
constant. The similarity in the relations of yield with weight and 
length of ear extends to the effect that making each constant has 
upon the correlations with other characters. That length is less 
important than weight of ear is shown by the smaller coefficients be- 
tween yield and length and particularly by the smaller correlations 
of yield with length for constant weight of ear than of yield with 
weight for constant length of ear. 
