698 HEREDITY AND CELL DIVISION 



per cent, between h and v, of 6.4 per cent, between h and p, 

 and of 10.8 per cent, between p and v. Since 6.4+10.8 is 

 approximately 16.3, we have support for the theory that the 

 factors are Hnearly arranged, and that that for purple is 

 situated about one-third of the way along from black to 

 vestigial. The proof is almost perfect when it is found that 

 hundreds of factors can be arranged in a consistent order on 

 the chromosome. Such a plan is called a chromosome map. 

 Many of the slight apparent anomalies in the recombination 

 values can be satisfactorily explained in detailed ways which 

 cannot be considered in an elementary book. 



There are three other classes of evidence. There is sometimes 

 an obvious connection between the number of chromosomes 

 present and a variety of an animal — a characteristic form of 

 the fruit-fly, for instance, has lost one of its fourth pair of chromo- 

 somes. In the salivary glands of insects the chromosomes are 

 abnormal^ large, and it is sometimes possible to see structural 

 changes in them w^hich regularly correspond to known genetical 

 characters or to variations in the linkage maps. Finally, in a 

 few instances there is ' non-disjunction ' of a pair of chromosomes 

 in the formation of gametes, so that a germ cell contains two 

 factors for each allelomorphic pair carried by the chromosome. 

 This can be seen in the cells, and the heredity corresponds. 

 These are the cases where the fundamental principle of Mendelism 

 is untrue (p. 683). It should be noted that in this argument the 

 genetical results have usually preceded the cytological observa- 

 tions, which have often been predicted long before they were 

 confirmed. Mendel's work long antedated any knowledge of 

 meiosis (or even, in any full sense, of fertilisation) ; the separate 

 identity of chromosomes was postulated while cytologists still 

 believed that they first appeared as a single thread or spireme ; 

 and crossing-over was predicted two decades before it was 

 observed. 



GENES 



The factor situated on a chromosome — or perhaps making 

 part of a chromosome — is now caUed a gene, and we must 

 consider shortly what may be its nature. The simple Mendelian 

 story which we have given so far is incomplete and in general 

 incorrect in two respects. In the first place, the presence of a 

 pair of genes does not necessarily lead to the production of the 



