652 Wisconsin Academy of Sciences, Arts, and Letters. 
distinctions when the parts compared are of only a few milli¬ 
meters difference in length; for the probabilities seem to be that 
an examination of a sufficiently large number of specimens will 
show some individuals that reverse the proportions. 
Of the measurements taken a number are omitted from some 
of the species. Selecting the sixteen comparisons summarized 
in the above table that are common to all of the species, and ob¬ 
taining the average of the extremes of all of these for each 
species, they arrange themselves as follows (to which, for con¬ 
venience, I have added the number of specimens of each species 
used in the computation) : 
Molossus rufus ......... 
Desmodus rotundus .... 
Qlossophaga soricina ... 
Chilonycteris ruMginosa 
Myotis velifer . 
ISJyctinomus braziliensis 
Natalus stramineus 
Mormoops megalophylla 
No. Av. per cent. 
Specimens, of variation. 
8 
5.5 
9 
10.1 
25 
11.1 
41 
12.4 
167 
13.1 
59 
13.4 
73 
13.4 
33 
14. 
The above table shows that with the exception of Myotis veli¬ 
fer and Mormoops megalophylla, the species have been arranged 
in the order of the number of specimens examined at the same 
time that they were arranged in the order of their average of ex¬ 
tremes of variation in proportions. 
If the amount of variation per species was equal we would ex¬ 
pect that the larger the number of specimens examined of any 
one species, the greater would be found to be the extremes of 
variation of proportions. 
The fact that this table so nearly parallels in these two com¬ 
parisons suggests the possibility that the variation in the eight 
species under consideration is approximately equal, and that 
were a sufficiently large and equal number of specimens of each 
of them examined the results would have been less diverse. 
It is doubtful if measurements of any other group of mammals 
would show, much less individual variation in proportions. It 
is my belief that no working zoologist really takes seriously the 
fine differences in measurements so frequently oven as diagnos¬ 
tic of species that are known by only a few specimens ; but 
