1S99] EVIDENCE OF NATURAL SELECTION 41 



as much as G5 units. If we took into account crabs less than 10 mm. 

 long and adults, the difference in their relative frontal breadths would 

 be much more than 6 b units. As a proof that the cause of death in the 

 experiments with fine clay or mud was the entrance of the latter into 

 the gill chamber, Professor Weldon states that the gills of the crabs 

 which died were generally covered with the sediment, while those of the 

 survivors were not. But the dead crabs were taken out at the end of the 

 experiment, so that they had been in the muddy water some time after 

 death, while the survivors were taken out alive. It is possible, there- 

 fore, that the sediment found its way in after death, and the evidence is 

 no proof that filtration was more efficient in one case than in the other. 



From Professor Weldon's own point of view the result of the 

 moulting experiment does not bear the interpretation he puts upon it. 

 He states that the crabs measured before the moult were on the 

 average narrower than those on the shore. Therefore, immediately 

 after the moult these same crabs would still have been narrower than 

 their fellows on the shore which were of the same size before the 

 moult, and moulted at the same time. Consider 500 crabs, say all 

 10 mm. in length, in the bottles, and 500 of the same size on the 

 shore. The former have narrower frontal regions. Both undergo 

 a moult at the same time; if there is no difference in the growth 

 of the two groups the crabs in the bottles must be still narrower, 

 for they are the same crabs in another stage. But those on 

 the shore are exposed to selective destruction after the moult. This 

 cannot well reduce the mean frontal breadth below that of those in the 

 bottles, for it now has a narrower breadth to act upon, and in the 

 previous stage it was not able to reduce the frontal breadth so low as 

 that of those in the bottles. Of course I understand that after the 

 moult some of those on the shore are supposed to be killed off, and 

 none of those in the bottles. But the 500 on the shore were broader 

 to start with by actual observation, and if they could survive with this 

 greater breadth at the size before the moult, I fail to see why many 

 should be killed off when they reach a larger size and a narrower relative 

 frontal breadth. Moreover, there are, in fact, multitudes of crabs on 

 the shore a little smaller than those in the bottles after the moult, and 

 of the same mean frontal breadth. 



It seems to me, as I suggested in a letter to Nature, that all 

 Professor Weldon's observations may be completely explained by varia- 

 tions in the amount or rate of growth. The difference in different years 

 would be at once explained if the amount of change in frontal breadth 

 was constant for each moult, while the amount of growth was variable. 

 The fact is, that in 1893 crabs of a given frontal breadth were larger 

 than in 189 5 and 1898 ; and I have shown that the summer of 1 893 

 was exceptionally fine and warm. Either the warmth alone, or warmth 

 and food together, very probably made the crabs grow more in that year 

 for the same number of moults. On this view the broad-fronted crabs 



