236 SCIENCE PROGRESS 



period before the formation of the gametes or sexual cells, so 

 that each gamete contains one or other member of each pair. 



Later work has extended, or, as some will prefer to think, 

 restricted, this generalisation, by showing that the unit-factors 

 are not necessarily all independent of each other. As a matter 

 of fact, when organisms are thoroughly studied, and a great 

 number of their Mendelian variations genetically investigated, 

 it is found that the factors concerned fall into a series of associa- 

 tions, the linkage-groups. A factor belonging to one linkage- 

 group is completely independent in heredity of factors 

 belonging to any other linkage-group. But the members of 

 one and the same linkage-group are not independent of each 

 other. On the contrary, they tend to stick together in the 

 course of transmission from one generation to the next. When 

 a cross is made involving two or more linked characters in each 

 parent, the combinations carried by the parents tend to recur 

 more frequently in the offspring, instead of all possible recom- 

 binations being produced in equal numbers, as will happen 

 when the factors involved are not linked. 



Further, linkage may be of differing intensity. Factor A 

 may be strongly linked to B, but weakly to C ; but, in the same 

 conditions, the intensity of linkage between any two particular 

 factors remains the same. The number of linkage-groups 

 appears to be constant for any given species. 



So far, then, the facts elicited by experimental breeding 

 permit us to make the following generalisations. In the higher 

 animals and plants, many characters are represented in the 

 hereditary constitution of the species by discrete units, or genes, 

 whose segregation from each other and whose arrangement in a 

 series of linkage-groups gives us the various ratios found in 

 genetic experiments. 



The next question is whether all hereditary characters of 

 organisms are in some way or other represented by these genes. 

 On this point, as on many others throughout this paper, I shall 

 be unable to present the available evidence. I must be content 

 to state my position dogmatically, and let time and the critical 

 judgment of the reader pronounce upon it. Personally, then, it 

 appears to me, in common with a number of others, that 

 certainly the majority, and possibly all, characters are due to 

 Mendelian genes. There is no evidence which at present 

 definitely contradicts this view, and, until there is, " William 

 of Occam's razor " makes it preferable for us to adopt, as 

 covering all cases, the one hypothesis for which there is tangible 

 evidence. One fact should be pointed out — that a great deal 

 of early Mendelian work was done on striking characters which 

 are to be considered abnormal, or even pathological. As the 

 analysis is pushed home, however, it has been found that a great 



