24 THE VARIATION OF ANIMALS IN NATURE 



abnormality. As to 6 the position is uncertain. Polyploidy is 

 not completely absent from animals, but according to Gates 

 (1924, p. 177) there is nothing comparable to the condition 

 found in plants. Varieties univalens and bivalens with 2X and 

 4X chromosomes have been recorded in Ascaris megalocephala, 

 Artemia salina, etc. In three out of the four cases noted 

 by Gates ' no particular significance seems attached to 

 the bivalent or tetraploid conditions' {I.e. p. 177). In the 

 Phyllopod Artemia salina it appears to be associated with 

 differences in reproduction, a tetraploid form of that species 

 being parthenogenetic. Tetraploids have been found in 

 Drosophila (Morgan, Bridges and Sturtevant, I.e. p. 21), but 

 ' as yet their chromosomes have not been studied.' As regards 

 the appearance of entirely new characters from any of the 

 various modifications of chromosomes (either those treated 

 here as abnormalities or those figuring in 4 to 6) in Haldane's 

 list, it seems clear that new characters or at least new com- 

 plexes of characters have arisen, e.g. as seen in the appearance 

 of the ' Diminished ' mutant due to the loss of a ' fourth 

 chromosome' (Morgan, Bridges and Sturtevant, I.e. p. 136). 

 But, owing to low viability (I.e. p. 137), it certainly seems that 

 this type and probably other similar ones are of small 

 evolutionary importance. 



Up to the present we have had little opportunity outside 

 the study of Drosophila to distinguish between the various causes 

 of mutation (in the broad sense, p. 4), so that it is not 

 possible to distinguish between gene-mutation and chromosomal 

 abnormality, etc., from the evolutionary point of view. On 

 the other hand, in the many experiments on induction, etc., 

 that have been carried out, we do not know what kind of 

 mutation is involved. From Mavor's experiments (1922) it 

 seems clear that X-ray treatment causes non-disjunction of the 

 X-chromosome. 



3. Recombination 



It is sometimes not realised what an enormous scope for 

 variation lies in the permutation of a relatively small number 

 of gene-differences. Fisher (1930, p. 96) points out that in 

 a species with a hundred segregating factors the number of 

 different true-breeding genotypes would be so large as to 

 require thirty-one figures to express it, or forty-eight if the 



