262 



THE AMERICAN NATURALIST [Vol. XLVII 



method, of course, made the fifth and sixth pairs of much 

 lower frequency than the other five, since so many ani- 

 mals had but ten mammae. In figuring percentage fre- 

 quencies for each position the total number of mammae 

 possible in all of the animals, if each had a pair in the 

 position under consideration, was used as a basis for 

 computation. The columns 1^, 2£, etc., show the frequen- 

 cies of mammae forming the apex of a triangle, in the 

 position between pairs 1 and 2, 2 and 3, etc. 



The total number of mammae for each position which 

 could occur if there were no variations is 1,138 for the 

 first four and inguinal pairs. The constancy of the last 

 named pair is particularly marked, only nine variations 

 obtaining. 



Arranging the pairs in order of frequency of variation, 

 the following sequence holds : 6th, 2d, 5th, 3d, 4th, 1st and 

 inguinal. It is probable that the rank of the sixth is not 

 significant, due to the lack of numbers in proportion to 

 the others, but there is no question about the second pair 

 being a leading seat of variation. In a former article on 

 this subject the writer called attention to this fact with- 

 out presenting figures. It is worth while noting that this 

 variance in the second pair is perhaps not characteristic 

 of swine as a whole, but simply due to the fact that both 

 boars showed in this pair their variation from normal 

 patterns. 



The point of most frequent appearance of the triangu- 



