BIOLOGY. 



327 



will be observed that these curves are not all alike in shape, that they are 

 specific for the specific regions involved. From this it follows that no general 

 formula can be devised that will convert recombination percentage into map- 

 distance. 



The curves of figure 2 show the results of the application of the corrections, 

 shown by the first set of curves, to the map. For each curve of the first 

 series there is a corresponding curve in the second series which gives the 

 recombination percentage corresponding to each pair of loci. 



This map enables one to read off at once the expected percentage of recombi- 

 bination for any two loci in chromosome III, while to use an ordinary map 

 one has to apply a correction on account of multiple crossing-over. Here 

 this correction has been included in the curves. The map-distance can also 

 be read off, if desired, from the straight line that accompanies each curve. 



Fig. 2. — Each slanting straight line shows in the height at which it cuts the ordinate from a 

 given locus the amount of crossing-over between that locus and the locus from which the line 

 originates. Similarly, each curve shows the amount of recombination for the locus of origin 

 and any given locus. 



The number of mutants has been increasing so fast from year to year that it 

 has become confusing to adhere to the original simple system, or even lack of 

 system, of naming them. We have been forced, therefore, to the uncongenial 

 task of systematizing our nomenclature, but have followed the old lines as 

 closely as possible. These changes are not recorded as final, but only transi- 

 tional. They serve to simplify the body of facts that has so far accumulated. 

 A full account of the revision is given in our forthcoming monograph on the 

 characters of the third chromosome of Drosophila. 



In order to construct and perfect the maps of that species most closely 

 similar to Drosophila melanogaster, namely Drosophila simulans, it has been 

 necessary to find new mutants and to get additional data concerning the loci 

 already known. Since these two species can be crossed, it has been possible 

 to demonstrate that 12 of the simulans mutant types are allelomorphic to 



