SEED-EAR CHARACTERS AND PRODUCTIVENESS IN CORN 15 
strongly correlated with length of ear, a character apparently favor- 
able to yield. When length of ear is made constant, the correlations 
between yield and number of kernels per row are negative in every 
case, though significant only in C. I. No. 77. The fact that the total 
correlation in this variety is negative and significant probably is only 
a reflection of the negligible correlation between yield and length of 
ear. 
The correlations between yield and number of rows for constant 
butt circumference and for yield and number of kernels per row for 
constant length of ear indicate in a general way a positive relation of 
breadth and thickness of kernel to yield. This is not absolute, 
because other factors also are involved in determining the number 
of kernels per unit of circumference or length. The negative coeffi- 
cients for these relations, however, indicate that the ears with wider, 
thicker kernels were more productive. 
COEFFICIENTS OF MULTIPLE CORRELATION 
The coefficients of multiple correlation between yield and various 
combinations of seed-ear characters are shown in Table 9. Perhaps 
the most striking difference is in the extent to which any characters 
were correlated with yield in the different varieties. The squares of 
these coefficients measure the degree to which variation in yield is a 
function of variation in the seed-ear characters considered. Thus, 
2.5 per cent of the total variation in yield of C. I. No. 133 was a 
function of variation in seven characters of the seed ears and could 
be predicted from the multiple regression equation involving these 
characters. Similarly 6.7 per cent of the total variation in yield 
in C. I. No. 119 could be predicted from the equation for the same 
characters. 
TABLE 9.—Coefficients of multiple correlation between yield and various combina- 
tions of ear characters, ranked according to their relative importance 
Meee C.I. No. 77 | C. I. No. 119 | C. I. No. 120 | C. I. No. 133 
age Designation of coefficient (a [ana SS 
rank R |Rank| R |Rank| R |Rank| R |Rank 
De eACROD RR HD 2 see eee eee LS et) 0. 2114 1 | 0.2590 1 | 0.1771 1 | 0.1582 1 
aeRO CRED OTD) sta a ae ce REL . 1913 PAA lem a PA VE! 2, .1684 25 staan 2 
ae Pen CBED IDE) eet ee tee ee ee Te . 1905 3 . 2389 5 . 1657 4 . 1565 3 
AON AC MCBGRILT a= eee ee See eee ot ee . 1903 4 . 2466 4 . 1681 3 . 1328 7 
FEN ik! ARTE) BI TE Ges a . 1888 7 . 2487 3 . 1640 if . 1462 4 
PEAT AT EDD | (Cre 1) ea ae a a . 1900 . 2388 6 . 1657 5 . 1306 8 
7 Vay e027 ha ee ee ee ee ee 1215 9 . 2324 8 . 1636 8 . 1446 D 
Rie RRA NCEP ELT ie atest aes See ee 1894 6 . 2339 7/ Jw 10 . 1254 9 
Ul Riera yeyenh | eee 5 Vea es SIP eee See ee 1345 8 . 1901 10 . 1653 6 1151 il 
Oa eer G NC Pucd FAVS creek ee oe ee ek 1172 10 . 1545 11 . 0435 11 - 1346 6 
MU AGED ER) esse ee eee SoS ee ete 0458 11 - 1917 9 . 1633 9 . 1206 10 
se 
The coefficients of multiple correlation do not show whether the 
relations between yield and the independent variables are positive 
or negative. This may be determined, however, from the regression 
equations. Im the regression equations involving all seven ear 
characters, yield is related to the different characters as shown in 
Table 10. 
Yield is related positively to weight of ear and length of ear and 
negatively to number of rows and number of kernels per row in all of 
