12 Introductory. 



METHODS EMPLOYED. 



Detailed methods of treatment of the data are given under the 

 respective headings below, but the following general features may 

 be noted at this point. 



For the sake of convenience, a description of each mutant type 

 is given in the present paper, whether it has been published previously 

 or not. Some of these include features that have previously been 

 overlooked or omitted. In connection with the descriptions brief 

 notes are added (under "Comparison"), indicating resemblances 

 to characters in the other species, if such exist. The absence of 

 such comparison indicates that no similar characters are known to us. 

 These comparisons are given very briefly, for a full consideration 

 of such relationships is included in later sections. 



The terminology used is that commonly employed at the present 

 time in animal genetics. Symbols have been avoided as far as is 

 practicable in the tables and text, although they have necessarily 

 been used extensively. Recessive mutant characters, or genes, are 

 designated by small letters and dominants by capitals. Super- 

 scripts are employed only in the case of multiple allelomorphs. 



In order to simplify treatment of the experimental data, the 

 name of a mutant type is frequently used to designate either the 

 character itself or the gene responsible for the character or the locus 

 of the gene, without distinction. Although this usage is possibly 

 open to some criticism, its great convenience justifies its adoption, 

 we believe, at least when care is taken to qualify the expression where 

 clarity demands. 



Since the purpose of this study is comparative, the linkage experi- 

 ments have been carried only far enough to indicate the relative 

 order of the genes and their approximate "locations" on the chro- 

 mosome map. No corrections have been made for differential 

 viability or double crossing-over in constructing the maps. Likewise, 

 map values have in nearly all cases been given in whole numbers. 

 It is believed that any attempt to make a more refined calculation or 

 to express values in decimals would in most cases give a false impres- 

 sion of accuracy. It is for this reason, for instance, that the method 

 suggested by Fisher (1922) has not been used. For certain types 

 of work the latter method promises to be useful, but, as indicated 

 by Fisher himself, when applied to cases like the present it gives 

 values only slightly different from those obtained by the more simple 

 method of calculation used here. 



Several characters have been included which overlap normal to 

 such an extent as to make them unsatisfactory for detailed linkage 

 studies, but which are valuable for comparative studies. On the 

 other hand, sex-linked lethals have been omitted, since they are of 

 little value for purposes of comparison. 



