198 CERTAIN LAWS OF VARIATION. 



more important; for nearly similar variations some- 

 times arise under, as far as we can judge, dissimilar 

 conditions; and on the other hand, dissimilar variations 

 arise under conditions which appear to be nearly uni- 

 form." The reasons of these differences of reaction 

 are, as a rule, quite unknown to us, but in the present 

 instance a satisfactory explanation was arrived at. 

 Thus, by heating up portions of the ova to various tem- 

 peratures at various periods of development, it was 

 found that the temperature at which they were killed 

 was by no means constant. It increased regularly with 

 progress in development, so that it was about 12 C. 

 higher for full-grown plutei than for ova at the time of 

 impregnation. 



STAGE OP TIME AFTER DEATH 



DEVELOPMENT. IMPREGNATION. TEMPERATURE. 



Strongylocentrotus ova, 28.5 



Semi-blastulae, 4 hours 33.5 



Blastulae and semi-gastmlse, 12 hours 36.5 



Plutei and semi-plutei, 28 hours 39.5 



Plutei, 6 days 40.3 



Supposing, therefore, a temperature of 33.5 is fatal to 

 four hours' blastulge, then all temperatures a few de- 

 grees lower than this must be unfavourable to their 

 growth. Lower temperatures still we know to be highly 

 favourable. Now we saw that a temperature of 26 

 i. e.j one 3 or 4 below the death temperature was 

 very unfavourable to the growth of ova shortly after 

 impregnation; so to produce an equally unfavourable 

 effect on four hours' blastulse, we should probably need 

 a temperature of 30 or so, and on 12 hours' blastulse, 

 one of 33. Or conversely, the temperature of 26, 



* " Origin of Species," Ed. 6, p. 6. 



