THE GENETIC NATURE OF TAXONOMIC DIFFERENCES 27I 



segregation takes place in the hybrids. A well-known example is in the 

 guinea-pig;^ the wild Cavia rufescens differs from the domesticated 

 C. porcellus in an allelomorph of the agouti locus, which causes the 

 agouti colouration to occur on the belly as well as on the back. Many 

 such cases are also known in plants, e.g. in Antirrhinum^ and Crepis. 



It is perfectly clear, therefore, that specific differences may involve 

 gene differences. There is no reason to suppose, as some authors sug- 

 gest, that the gene-controlled characters are all trivial. We know genes 

 which, within a species, can cause such important changes as an altera- 

 tion in the number of segments in the tarsus of a fly, or the number of 



Fig. 122. A Gene with Taxonomically Important Effects. — A shows the 

 normal two-carpel seed capsule of Datura stramonium. B is the three-carpeiled 

 capsule produced by the gene Tricarpel. The three-carpel form suggests a relation 

 with the family Polemoniaceae, which belongs to a different sub-order to the 

 Solanaceae in which Datura is classified. The capsules are represented as cut 

 across to show the internal structure. 



(After Blakeslee, Morrison, and Avery.) 



carpels in the seed-capsule of Datura.^ These are certainly more im- 

 portant differences than many on which species have been separated. 

 A beginning has been made with the genetic analysis of the important 

 serological differences between related species.^ 



In the past, groups which differ only in a single gene have been 

 described as different species, particularly in plants, but as soon as 

 this was discovered, the species have been amalgamated. In the differ- 

 ence between true species, very many genes are involved. The fullest 

 knowledge we possess on this relates not to fully-fledged species but 

 to local races within a species. One might expect such races to be more 

 alike than two species would be, but even so the differences, in terms 

 of genes, are fairly complicated. We will take two of the best analysed 

 examples. 



(a) Lymantria. — In a very long series of crosses, Goldschmidt^ has 



^ Detlefsen 1914. For a reviev/ of animal hybnds, see Hertwig 1936. 

 - Baur 1932, Lotsy 191 1. ^ Blakeslee, Mornson and Avery 1927. 



* Irwin and Cole 1936. ^ Goldschnaidt 1934a, c, 19356. 



