PRODUCTION OF MONSTERS BY HYBRIDIZATION. 307 



During the summers of 1914 and 1916 a series of experiments 

 was performed in order to test the theory that retardation is 

 always correlated with abnormal development. In hybrid 

 experiments in which Fundulus heteroclitus eggs are fertilized 

 by the sperm of the mackerel there is noticeable in comparatively 

 early cleavage stages a marked difference in number and uni- 

 formity of blastomeres in different eggs of the same batch. The 

 different types of eggs were divided into three lots as follows: 



' A. Those that were most advanced and exhibited the largest 

 number and greatest regularity of cells. 



B. Those in which the cells were fewer and less uniform. 



C. Those in which the cells were relatively fewest and of 

 various sizes. 



The development of these three lots was watched from day 

 to day with the result that all of those in A reached a fairly 

 advanced condition before abnormalities appeared and ab- 

 normalities concerned chiefly the eyes and heart; those in B 

 showed all sorts of monsters including cyclopeans, eyeless types, 

 and those with reduced bodies ; those in C produced solely apical 

 parts without bodies, amorphous masses of tissue or at best 

 isolated organs such as eyes and hearts. 



The converse of this experiment was tried using the eggs of 

 F. heteroclitus and the sperm of the closely related F. diaphanus. 

 In this experiment there was very little acceleration noticeable 

 during the cleavage stages, but during gastrulation it was a 

 simple matter to separate out three classes A, B, and C as follows : 



A. Those in which the germ ring was distinctly more than 

 half way round the yolk. 



B. Those in which the germ ring was approximately half way 

 round the yolk. These were about like the control. 



C. Those in which the germ ring was distinctly less than half 

 way round the yolk. 



The eggs in lot A nearly all hatched on the average distinctly 

 earlier than the control (pure F. heteroclitus'), and were distinctly 

 more active, grew faster and lived longer than the control. 

 Those in lot B showed a wide range of variability, some hatching 

 as early as the control, others hatching later, and still others 

 failing to hatch. Various' common types of abnormality oc- 



