98 BULLETIN 754, U. S. DEPARTMENT OF AGRICULTURE. 
The results of this analysis show that the number of seeds with 
the waxy endosperm reappearing in the perjugate generation of 
waxy X horny crosses is less than the expected for a simple Men- 
delian character, but this deviation, though significant, is too small 
to warrant the predication of additional factors. In making this 
determination, more than 100,000 seeds were classified, and it has 
been possible therefore to establish*the actual percentage within 0.3 
of 1 per cent. 
The same material was used to test the inheritance of aleurone 
color, which also was found to depart from the theoretical ratios. 
Unlike the waxy endosperm texture, no definite trend above or below 
the expected ratios was observed, but many abnormal ratios were 
obtained which necessitated further refinements in the factorial 
analysis of this character. 
From the results of the Mendelian analysis of aleurone color and 
endosperm texture it must be concluded that in many cases uniform 
Mendelian reactions are obtained which allow certain predictions to 
be made with respect to the behavior of these characters in subsequent 
generations. But that these predictions based upon the gametic an- 
alysis will be uniformly fulfilled must not be supposed. As we have 
seen, aberrant behavior is far from uncommon, increasing with the 
progress of investigation and the refinement of analysis. The fact is 
coming to be appreciated that instead of a few simple unassociated 
factors most characters are composed of many complex units which 
may no longer be considered singly but that their interrelations or 
correlations must be taken into account. 
_ The present investigations show that certain of the more definitely 
alternative characters of maize are subject to variation or fluctuating 
behavior that renders these supposed Mendelian factors too irregular 
to justify a belief that the very definite relationship predicted in the- 
ories of gametic coupling could exist between such characters. There 
can be little doubt, however, that at least with several combinations 
of characters the gametic ratios are to a certain extent regular, but 
that these ratios fit any arbitrary series is not so well demonstrated. 
For the breeder of crop plants where most of the desired characters 
are almost infinitely complex, seldom alternative, and often intangi- 
ble, Mendelism seems to have little of practical value to offer, whether 
the attempt of some investigators to so extend the theory to embrace 
such characters be approved or not. While Mendelism may assist in 
making desired combinations, there is nothing to show that it can 
serve as a substitute for selection either in finding the best stocks or 
in preserving them from subsequent deterioration. 
