CHARACTERS AND CHROMOSOMES 697 



in gamete production and mating is random. If each gamete 

 contains one factor the zygote must contain two ; there must be 

 reproduction of these factors at cell division, and at some stage 

 before the formation of the next generation of gametes there 

 must be an orderly reduction of the factors in the cell to one. 



It was pointed out by Sutton as early as 1903 that an obvious 

 physical basis for all this was to be found in the chromosomes, 

 but there were at first some difficulties. Now, however, the 

 evidence for the connection is overwhelming, and its develop- 

 ment is a pretty exercise in biological argument. Since, with 

 rare exceptions, the results of reciprocal crosses are the same, 

 we must look for the factors in the nucleus, for what little 

 cytoplasm the male gamete contains often takes no part in 

 fertilisation ; in the nucleus, the chromosomes are the only 

 things with any individuality, and their splitting before mitosis 

 and reduction at meiosis provide a means of reproduction and 

 segregation of the factors. But similarity is not proof of identity, 

 and there is the objection that the continued separate existence 

 of the chromosomes in the resting cell is an assumption, though 

 a probable one ; more evidence is needed, and is provided by the 

 phenomena of linkage.^ In all the organisms which have been 

 deeply studied (it must be admitted that these are very few : half 

 a dozen animals and rather more plants) it has been found that 

 all the allelomorphic characters fall into groups, such that all the 

 members of one group are linked together, and that the number 

 of linkage groups is equal to the haploid number of chromosomes. 

 The linkage is, as we have seen, not perfect, and departs from 

 perfection by a constant amount for any two characters. This 

 could be explained if the factors were carried in a linear arrange- 

 ment on the chromosomes ; when crossing-over occurred at 

 meiosis previously linked factors would be separated, and the 

 chance of their being so should depend on the distance between 

 them, for if the crossing-over can take place with equal ease 

 at any point it is more likely to happen between two given points 

 the farther they are apart. We can thus use the recombination 

 values of different sets of factors to express their relative position 

 on the chromosomes. Thus in another experiment with black 

 vestigial fruit flies, in which another recessive character, purple 

 eye, w^as also studied. Bridges found a recombination of 16.3 



^ The special case of the determination of sex and sex linkage, which is also 

 important evidence, is referred to below (p. 701). 

 M.Z. — 23 



