INHERITANCE OF ACQUIRED CHARACTERS. 449 



conditions are maintained, a further increase occurs in all 

 the adaptive characters, and the larvse can live even longer 

 on land. 



By prematurely opening the eggs laid in the water, and keep- 

 ing the embryos in a poor light in a large quantity of cold water 

 with a small food-supply, an Alytes larva was produced which 

 remained for four years and eight months in the larval state, 

 and produced eighteen eggs, which were artificially fertilized 

 with the sperm of a normal male. The larvae reared from these 

 eggs under normal conditions were distinguished by the long 

 persistence of the external gills, and after two and three-quarter 

 years showed no signs of preparation for the metamorphosis. 



Toads kept at a temperature below 17° C. take two years to 

 become sexually mature. If kept above 25° C, this occurs in 

 one year. Offspring of the latter in normal conditions became 

 mature in a year and a half. 



Now these results cannot be explained by a direct action of 

 the environment on the germ -cells ; for in most cases very little 

 if any effect seems to be produced in the offspring so long as it 

 is necessary to keep the parents in abnormal conditions to get 

 them to change their spawning habits. As soon, however, as 

 the changed habits have been acquired by the parents — that is 

 to say, as soon as the parents can be replaced in normal con- 

 ditions — then inheritance of these changed habits appears to 

 occur. 



We know that in such animals as change colour to adapt 

 themselves to their environment the change is brought about, 

 not directly by the action of the altered environment on the 

 chromatophores of the skin, but indirectly by way of the nervous 

 system. It is often supposed that the action may once have 

 been a direct one, but that the perfection of the nervous system 

 has led to its interposition in the process. Of course, the 

 nervous system is not so directly connected with the germ- 

 plasm as it is with the pigment-cells in the skin, but is it not 

 possible that changes in the nervous system may affect the 

 germ-cells ? And there can be no doubt that a change in the 

 characteristic habits of an animal is in some way correlated 

 with a change in its nervous system. 



In this connection it is interesting to notice that the chief 



