PROTOZOA AND HEREDITY 197 



heredity (if there were any) , instead of the probably much Expert 

 larger ones caused by the environment. To avoid this 

 difficulty Jennings began new experiments with a quite 

 different protozoan, the Difflugia corona (Fig. 40, B). 

 This is a shelled rhizopod, the shell being made of grains 

 of sand embedded in a chitinous secretion, and presenting 

 a variable number of projecting spines. When division 

 takes place, to form a new individual, the shell is formed, 

 and it cannot be altered subsequently. The shells are 

 readily preserved, so that many successive generations 

 may be directly compared. It was found that the ani- 

 mals differed in the size and shape of the shell, the length 

 and number of the spines, etc. After many generations, 

 the descendants of a single ancestor, selected for various 

 characters, were found to have actually diverged from 

 one another, the difference being inherited. This ap- 

 pears to contradict flatly the evidence derived from 

 Paramecium, but it may well be that species differ in 

 the mutability of their germ plasm, or are mutable at 

 certain times and not at others. It must be remem- 

 bered that the species of Difflugia are extremely widely 

 spread over the world, and are essentially constant in 

 their characters. This proves that they are of great 

 antiquity, and suggests that however the hereditary 

 qualities may have varied, they have very rarely done 

 more than oscillate about a mean. It may be that the 

 complex molecules forming the determiners are never 

 rigidly constant for great lengths of time, but change 

 within small limits, which usually elude our powers of 

 observation. This might be true, and yet the chemical 

 oscillation, if we may so term it, might be strictly 

 limited under ordinary circumstances, so that the 

 termination of a series of generations would find the 

 organism practically as it was at the beginning. Only 



