EVOLUTION 361 



not unreasonable to conclude that in the large number of 

 insect embryos in which the primitive germ-cells seem to 

 arise from ordinary mesoderm, they are really descended 

 from certain cells which were already defined in the early 

 segmentation stage, though showing no distinctive aspect 

 then recognisable by the investigator. J. A. Nelson, though 

 unable to trace early differentiation in embryos of the hive- 

 bee (1915), believes it "probable that in all insects the 

 germ-cells are segregated at an early period." 



We have already noted that the various stages and the 

 final result of a creature's development depend upon its 

 surroundings as well as upon its inherited constitution. 

 This is clearly shown by many Lepidoptera which, passing 

 through two Ufe-cycles in the year, appear in alternating 

 warm or cool, w^et or dry, seasonal forms. Members of the 

 spring brood of our native common white butterflies, which 

 emerge from wintering pupae, have their wings more darkly 

 marked than those of the summer brood which pass through 

 the pupal stage in June or July. These differences, brought 

 about partly at least by the great difference in the tempera- 

 ture to which the pupae are subject during development, 

 can to some extent be reproduced by artificially induced 

 conditions. A. Weismann (1882) showed that by cooling 

 summer pupae of the Green-veined White Butterfly {Pieris 

 napi) the resulting imagos were rendered so dark that they 

 approached in appearance the dusky arctic and alpine 

 variety hryoniae. F. Merrifield (1891), by refrigerating 

 summer pupae of the geometrid moth Selenia, induced such 

 dark scaling in the resulting imagos that they resembled the 

 strongly banded spring form of the insect. Attempts to 

 rear the summer form of such seasonally dimorphic insects 

 by subjecting the winter pupae to artificial heat have been 

 only occasionally successful. In nearly all these cases it is 

 found that the abnormal temperature-conditions must be 

 repeated in every generation if the colour-modification is 

 to be maintained ; none of the acquired changes are as a 

 rule inherited. What is always inherited is a power of 

 response to the conditions that may be imposed naturally 



