CYTOLOGY AND MENDELIAN HEREDITY 179 



It will be noted that the recombination percentage for sui-gh is somewhat 

 less than the sum of the percentages for sui-Tu and Tu-gU. This is 

 because genes, when rather widely s(^parated in the chromosome, may 

 ha\'(! two crossovers between them, leaving them still in the same chro- 

 matid. This makes the observed recombination percentage lower than 

 it would have been if no such double crossing over had occurred. Linkage 

 maps, lik(! those for Zea and Drosophila (Figs. 127, 128), are built up 

 l)y plotting ]:)Ositions only on the basis of closely linked genes, thus 

 without taking double crossing over into consideration. This means 

 that map distance (number of units) between two genes represents 

 recombination percentage only when these genes are rather closely 

 linked — within about 10 units in Drosophila and maize. Thus the map 

 may exceed 100 units in length, whereas the upper limit of observed 



sui Tu gl. 



Fig. 126. — Diagram illustrating the method of determining the serial order of linked 

 genes by comparing recombination percentages. The four parallel lines represent the 

 chromatids in the tetrad at pachytene in the plant to be testcrossed. Further explanation 

 in text. 



recoml)i nation is 50 per cent owing to the fact that each crossover alters 

 only two of the four chromatids. At the present time the most fully 

 developed linkage maps for insects, vertebrates, and plants are those of 

 Drosophila melanogaster, the common fowl, and maize, respectively. 

 The map as just described is a convenient record of the linkage 

 relationships and serial order of the genes in the chromosome, but there 

 are limits to its reliability as a picture of the chromosome itself. This is 

 because such a map, especially when only a few genes have been placed 

 upon it, does not necessarily show the actual position of the genes in the 

 chromosome. If, after numerous linkage studies, no gene has to be given 

 a place beyond the one already at the end of the map being constructed, 

 it is a reasonable inference that this gene is actually at the end of the 

 chromosome, but it might be that the chromosome beyond this point 

 is inert or occupied only by unmutated genes. A gene is detectable 

 in normal material only when it is present in the mutated as well as the 

 unmutated form, thus giving a character contrasting with the normal 

 one and revealing the fact that a corresponding normal gene exists. 

 Another method, therefore, must be used to ascertain actual gene position. 



