24 Linkage Group I. 



overs in 48 flies), and the other two agree in giving it a position approximately 6 

 units above that of singed. In both of the latter a difficulty in classification arises 

 through the presence of oblique and sepia together. This necessitates omitting the 

 oblique flies from the calculation or else apportioning them between the oblique and 

 the oblique sepia classes. Both methods are used, and the results are in such close 

 agreement as to cause no difficulty in estimating the approximate values. 



The map location of oblique has been calculated from the average of the three 

 experiments just considered. When only the not-oblique flies are used (in the last 

 two cases) the result is 11 cross-overs in 186 flies, or a value of 5.9 per cent. When 

 all flies arc used the result is 15 cross-overs in 263 flies, or a value of 5.7 per cent; 6 

 per cent is therefore used as the approximate value in constructing the map, and 

 oblique is placed 6 units above singed. It should be noted, however, that in both 

 of the experiments involving sepia, the sepia-singed value is unusually low, ranging 

 around 19 instead of about 40, as it should. The singed-short value, on the other 

 hand, agrees approximately with expectation. This suggests that in these experi- 

 ments a factor was present which reduced crossing-over in the "upper end" of the 

 chromosome. If so, the locus of oblique may be placed too close to that of singed. 



Hairy. 



Hairy has been used very little in Unkage tests because of its unsatisfactory nature. 

 In a test with forked (Metz, 1918) it gave about 3 per cent crossing-over, Weinstein 

 (1920) found it to be at the "left" of (above) magenta, and obtained a cross-over 

 value of 5.4 per cent with magenta. Our experiments 26, 28, 29, 37, and 42 indicate 

 the same sequence of genes and give very nearly the same cross-over value \vith 

 magenta (5 per cent). The combined data give a value of 5.3 per cent, placing it 

 at about 62 of the map. 



Magenta and Forked. 



Since magenta and forked have been used in combination almost exclusively, 

 they may be considered together. They were among the first sex-linked characters 

 obtained and have been used extensively ever since. The order of their genes with 

 reference to the other principal loci is indicated by previously published data and 

 by experiments 27, 29, 30, 31, 37, and 40 to 48. The cross-over value of the magenta- 

 forked region is the best known one in D. virilis. The data of Metz (1918) give a 

 value of 3.7 units, based on 2,529 flies and those of Weinstein (1920) give a value of 

 4 per cent based on 1,642 flies. Data in the present paper (table 3) give 3.4 units, 

 based on 6,208 flies. Combined these give a value of 3.6 units, which is used in 

 constructing the present map. 



Magenta and forked are located with reference to singed by using the singed- 

 magenta value of 25.1 units given by experiments 41, 45, and 48 (table 3). This 

 places magenta at about 67 and forked at about 71 units from the zero end of the 

 map. These loci should be placed more accurately by obtaining large numbers in 

 an experiment involving vermilion (or crossveinlcss), singed, magenta, and forked 

 simultaneously. Our experiments 41 and 48 are of this nature, but the counts only 

 include 718 flies. 



Triangle and Short. 



Triangle is shown by experiments 13 and 32 to be closely linked to short. The 

 latter character appeared later, chronologically, than triangle, but is much better for 

 linkage studies and consequently is used more than triangle. The average cross-over 

 value between the two (table 3) is 5.1 per cent. Experiment 32 shows that the 

 position of short is between that of triangle and droop, and experiment 33 shows 

 that it is above rugose. 



In locating triangle and short on the map, four different lines of evidence have 

 been considered, as summarized on the following page. 



