238 SCIENCE PROGRESS. 



may be looked upon as so many local races. Thus, for 

 instance, whilst shrimps obtained near Plymouth had a 

 mean carapace length of 250*05, those from Southport had 

 one of 248*50, and those from Sheerness one of 247*51. 

 The degree of variability also differed in each case, the 

 "probable errors" being respectively 4*53, 3*17 and 3*05. 

 Also, different organs varied to different extents. For 

 instance, the post-spinous portion of carapace reached the 

 maximum length in Sheerness shrimps and the minimum 

 in those from Plymouth, whilst as we have just seen the 

 carapace lengths varied in the opposite direction. Still 

 again, the correlation between the various organs also 

 varied in the different local races, though this variation was 

 small and might perhaps have been due to experimental 

 error. 



A few remarks may be made as to the bearing of the 

 results obtained in these experiments on the theory of 

 Natural Selection. It has been shown that a change in the 

 salinity of the water in which the larvai are developing may 

 produce a variation of 15 per cent, in their size. Now these 

 individual larvae are found to vary on an average by about 

 6 per cent, from the mean size of the whole number. If, 

 therefore, half of a number of larvae developing in any 

 region were exposed to water of one degree of salinity and 

 the other half to water of the other degree, the variability 

 of the whole group of larvae would be more than trebled. 

 Other changes in the environment would produce similar, 

 though perhaps smaller, changes. Now, other conditions 

 equal, the more variable an organism, the greater oppor- 

 tunity has Natural Selection of stepping in and, by pre- 

 serving the most favourable individuals, modifying the 

 species. But, it may be objected, this is not actually the 

 case, because the variations thus produced are, in Weis- 

 mann's terminology, merely somatogenic and not blasto- 

 genic. They are merely variations in the body tissues, and, 

 the germ plasm not being affected, are not transmissible by 

 heredity. A little reflection will show that this objection, 

 though in some respects valid, yet is by no means fatal to 

 our proposition. Thus, suppose that for instance in sea- 



