ON THE EXPERIMENTAL HYBRIDIZATION OF ECHINOIDS. 327 



believes that they can be influenced to a certain degree by external conditions even 

 after fertilization. He found that injury to one of the sexual elements usually 

 resulted in unhealthy hybrids, but the character of the inheritance remained 

 unchanged. Herbst made further experiments which led him to conclude that it 

 was the relative sizes of the female and male pronuclei at the time of fertilization 

 that influenced the character of the inheritance. 



When we turn from Echinoderms toother groups of the Animal Kingdom, we have 

 a number of instances where temperature has been shown to play an important part 

 in modifying the character of the inheritance. The great difiicultv, however, in much 

 of this work seems to be that we know so little about the effect of somatic modifica- 

 tion on the germ-plasm. 



It will have been noticed that in the above-mentioned papers of Vernon, 

 DoNCASTER, and Herbst, the effects of external conditions actine: on the eggs and 

 sperm at the time or after fertilization are considered. The most striking instance, 

 however, of the effect of temperature is shown where the temperature acts only 

 during the growth period of the germ cells. In this instance there is little or no 

 effect beyond the production of some minor somatic changes, if the temperature acts 

 as in Vernon's and Doncaster's experiments, at the time or after fertilization. 



Tower (93), in his hybridization work on Leptinotarsa, was able to modify the 

 character of his hybrids by keeping the parent forms under unusual conditions of 

 temperature and humidity while they were maturing their germ cells. If the parent 

 forms were kept under these conditions after their germ cells were ripe, then these 

 changes did not appear. This result agrees more with our own experience than do 

 those of Vernon, Duncaster, and Heubst. 



Tower (98) made the cross Le/ptinotarsa signaticollis ^ X ^. divei'sa $ under one 

 set of conditions, and F, gave offspring all intermediate. These, in Fj, split into 

 three types in a Mendelian proportion. When, however, identical material was 

 crossed under another set of conditions of moisture and temperature, Fj were half 

 intermediates and half X. siguuticoUis. The intermediates split into the three types 

 in Fn in the Mendelian proportion, while the L. signaiicollu bred true for five 

 generations. These experiments were repeated eleven times with the same results. 



Crosses between L. undccimlineata and L. signaticollis under five different 

 combinations of temperature and moisture gave : — 



(1) F, heterozygous intermediate, mendelising in F., in the ordinaiy way. 



(2) Fj purely maternal, breeding true for five generations. 



(3) Fj half heterozygous intermediates and half maternal, breeding true for four 

 generations. 



(4) Fi partly maternal which bred true, partly paternal which bred true, and 

 partly heterozygous intermediates. 



(5) Complete series from maternal to paternal types. 



