34O RAYMOND PEARL. 



The general results which follow the crossing \ a yellow dent 

 (9 ) with a white sweet (cf) maize are well known, Yellowness 

 of endosperm is dominant over "whiteness" of endosperm, and 

 "starchiness" over "sweetness." Consequently the F t kernrk 

 are externally indistinguishable (in fact as well as in theory) 

 from those of the pure yellow dent parent. These FI kernel- 

 planted give rise to plants bearing ears of which each should 

 have four distinct kinds of F* kernels which ought, by theory, 

 to occur in the simple dihybrid ratio, 9 yellow dent , 3 white dent, 

 3 yellow sweet, I white sweet. 



The present experiments 1 entirely confirm in all essential re- 

 spects this general Mendelian result. Certain novel points 

 arose, however, in the course of the work, which led to the present 

 investigation. These points may now be considered. 



A large quantity of ears bearing F2 kernels was raised. These 

 ears were well matured. This was indicated both by their 

 appearance and by the way the seed from them germinated. 

 One of the assistants in the laboratory, Miss Maynie R. Curtis, 

 undertook the sorting and counting of these p2 kernels on an 

 extensive scale. In this work the following situation immediately 

 developed and was called to the writer's attention. While in 

 general the F2 kernels fell without any doubt or difficulty into 

 the four classes or categories, yellow starchy, white starchy, 

 yellow sweet and white sweet, yet there were a number of kernels 

 on each ear that were extremely difficult of classification. These 

 kernels were, in short, intermediate in respect to their external 

 visible somatic characters, and might, in the individual case, 

 be put with equal propriety into either of two classes. Into 

 which class such an intermediate kernel would actually be put 

 plainly depended upon the personal bias of the observer, rather 

 than upon any peculiarity of the kernel itself. This result ap- 

 peared to be of enough interest and potential significance to 

 warrant a more extended and thorough investigation of the 

 matter. The present paper deals with tin- results of such a 

 study. 



1 A detailed description of the conditions and manner of these experiments has 

 been given elsewhere (Pearl, he. cit.) and need not be repeated. 



