438 Journal of Agricultural Research voi. vi, no. 12 



any attraction between colored aleurone and horny endosperm, but 

 rather that one of the parents had colored aleurone and horny endos- 

 perm, while the other parent had white aleurone and waxy endosperm. 

 This tendency fcr parental combinations to reappear has been called 

 "coherence," and, so far as known, all genetic correlations thus far re- 

 corded are of this nature. 



Many investigations have been devoted to correlations in agricultural 

 plants, but unless the special class of correlations covered by coherence 

 is kept in mind the results are likely to be disappointing to the breeder. 

 Cylindrical ears of maize may be correlated with high yield in one popu- 

 lation and the opposite result be reached in another case, depending on 

 whether these characters were introduced into the population under 

 investigation from the same parent or from different parents. 



There are doubtless many physiological correlations that may be de- 

 tected by elaborate measurements, but unless the observations are con- 

 fined to asexually propagated groups or to those of which the ancestry 

 has been carefully studied, there will always remain the uncertainty 

 whether there is an inherent physiological relation between the develop- 

 ment of the two characters or whether the correlation is the result of 

 ancestral combinations. The distinction is not without practical impor- 

 tance, for a physiological correlation can not be reversed by any direct 

 means at the disposal of the breeder — that is, without evoking mutation 

 or some form of evolutionary change — while, if the correlation is genetic, 

 the relation between the characters may usually be reversed by a few 

 generations of selection in the desired direction. 



Two principal theories have been advanced to explain genetic correla- 

 tions. These are the theory of reduplication (Bateson, Saunders, and 

 Punnett, 1906) and the theory of linkage developed by Morgan and his 

 students (1915) from studies of the fruit fly Drosophila ampelophila. 

 Both of these theories deal with characters which are alternative, both 

 having been derived from the study of Mendelian inheritance. 



With the idea that continuous inheritance is to be looked upon as a com- 

 plicated form of alternative inheritance, it should be interesting to 

 learn what light the study of genetic correlations between characters 

 that are blended in inheritance may throw on the theories of reduplica- 

 tion and linkage. The experiments described below constitute a prelim- 

 inary attempt to extend the study of genetic correlations to characters 

 that are continuously inherited. 



METHODS OF DISTHNTGUISHING BETWEEN PHYSIOLOGICAL AND 

 GENERIC CORRELATIONS 



To determine with certainty that a given correlation is physiological 

 and not genetic, it would be necessary to demonstrate the existence of 

 the correlation in material where all the individuals possessed the same 



