20 THE VARIATION OF ANIMALS IN NATURE 



inspection and write-off many forms as ' mere fluctuations ' 

 or ' due to the environment.' It may be claimed that this 

 procedure is justified by analogy with effects known to be 

 produced by experiment. But actually a number of experi- 

 ments has been claimed to show that certain effects are due 

 to the environment, though no examination was made of the 

 behaviour of the affected characters in heredity. Further, the 

 amount of variation that is treated as non-heritable is far in 

 excess of the number of cases that have been experimentally 

 verified. 



It is not easy in fact to obtain more than relatively few 

 instances of characters which have been shown experimentally 

 to be non-heritable. Among the Mollusca, the form albo- 

 lateralis of Arion empiricorum (ater) (Collinge, 1909), the carinate 

 and ecarinate forms of Paludestrina jenkinsi (Robson, 1929), and 

 various forms of Limnea peregra (Boycott, Oldham and Waters- 

 ton, 1932) seem to be definitely fluctuations. Pelseneer (1920, 

 p. 641) catalogues a list of 'variations non hereditaires ' in 

 the Mollusca ; but in all his cases, except that of Arion ater, 

 there is no evidence that the character in question was not 

 acting as a simple recessive, since the breeding test was not 

 extended to more than one generation. In the insects, which 

 have been so much used for genetical research, rather more 

 cases are available. Some of the naturally occurring colour- 

 variations of the bug Perillus bioculatus (Knight, 1924) and of 

 the parasitic wasp Microbracon brevicomis (Genieys, 1922) are 

 certainly not inherited. As for variations known only under 

 artificial conditions, we may mention a white variant of the 

 moth Ephestia kiihniella (Kiihn and Henke, 1929) and a number 

 of variants in Drosophila, especially reduplications of various 

 organs (Morgan, Bridges and Sturtevant, 1925, p. 71 et seq.). 

 Amongst birds, Beebe's (1907) experiments on the effect of 

 a humid atmosphere on doves of the genus Scardqfella are 

 well known. In the rotifers, Kikuchi (1931) shows that in 

 Brachionus pala lateral spines are developed when the animal 

 is fed on the alga Scenedesmus ; the spines are lost when it is 

 fed on Polytoma, and the action is completely reversible. 



A point worth remembering in discussing this question is 

 that a given character may be heritable in one form and not 

 in another. This is especially evident in the matter of the total 

 size of an organism which is determined not only by the 



