Feb. 23, 1924 
Teosinte Maize Hybrids 
555 
The proterogynous habit is one which may be influenced greatly by 
photoperiodism or other environmental factors. Many varieties of 
maize from South America, indigenous near the equator, are so decidedly 
proterandrous when grown in the United States as to interfere seriously 
with the production of seed. 
Euchlaena seems to be proterogynous, regardless of the flowering 
date, and even when hastened into flower by artificially shortening the 
day. 
Even in maize, though the plant as a whole usually is proterandrous, the 
individual inflorescences seem to be inherently proterogynous. If con¬ 
ditions force pistillate flowers to develop in an inflorescence normally 
staminate, the resulting flowers always are proterogynous. When gyan- 
drous inflorescences are normal to the strain, as in tassel seed, they also* 
are always proterogynous, while in the andromonoecious dwarf variation 
the full development of the anthers occurs after the silks have appeared.. 
None of the characters of the crinkly variations showed discontinuous 
inheritance, though most of the crinkly characteristics were recovered 
in the P 2 in an even more extreme form than was found in the parental 
progeny. Crinkly and teosintelike plants from the F 2 , together with 
their inflorescences, are shown in Plates 2, 3, 4, and 5. 
Notwithstanding the failure to obtain bimodal distributions of these 
characters, the frequency polygons in figures 1, 7, 8, 9, 11, 13, 14, 15, i6 r 
34, 35, 36, 37, and 39, all clearly show the high degree of development, 
in the plants empirically classed as crinkly, of the characters commonly 
associated in this variation. 
These relationships are measured by the coefficients of biserial corre¬ 
lations presented in Tables IV and V. 
CORRELATIONS BETWEEN CHARACTERS 
This brings us to the consideration of the whole problem of character 
relationships. Two questions may be answered by the data at*hand: 
First, to what extent the character complex comprising the crinkly 
variation tends to remain associated; second, to what extent all or any 
of these characters are associated with other maize characters not involved 
in the crinkly variation. 
Difficulties are encountered at the start in the proper interpretation of 
the correlation coefficients, since by direct means it is not possible to 
distinguish purely physiological relationships from those which are 
genetic. 
Recourse may be had to the method suggested by Collins of comparing 
the degree of correlation in the F t with that found in the F 2 (2). Cor¬ 
relations in the Fj are held to be due to physiological relationships, 
while any differences between Fj and F 2 correlations may be ascribed to 
genetic causes. The coefficients of correlation between all the characters 
studied in this hybrid are given in Table II. 
