638 Wisconsin Academy of Sciences, Arts, and Letters. 
pears to be no possibility of sucb mistakes having been made as 
all the measurements of each individual were made and re¬ 
corded one after the other before the next specimen was taken 
in hand. These were written in figures from left to right across 
the ruled pages and are now in my possession. Where such 
anomalies appear in the graphs I have reproved them from the 
original sheets to assure myself that errors have not been made 
in the plotting. 
The third metacarpal may be as long as the forearm or it may 
be 7 mm. shorter. The second phalanx of the fourth digit may 
equal the first or be only two-thirds as long. The thumb may 
equal the first phalanx of the third digit or be only seventy 
per cent, of its length. In the usual points of comparison there 
is so much difference in the lengths of the various members that 
their measurements do not overlap but merely approach or recede 
in varying degrees. 
It may be considered that hTo. 809 has an abnormally short 
second phalanx to its fourth digit as the next shortest one is 
3 mm. longer. However, as this joint showed no signs of injury 
and the other phalanges indicate no tendency towards abnor¬ 
mality, we are not warranted in excluding it or considering it 
other than a normal variation whose isolation a larger series 
probably would bridge. 
Natalus stramineus Gray. 
Seventy-three specimens collected near Jalapa, Veracruz, Feb¬ 
ruary 23, 1891, by H. L. Ward and OL M. T'eran. 
With the exception of the graphs of length and tail, which 
are rather remarkably dissimilar, there is a considerable corre¬ 
spondence between those of this species. Those of the head and 
ear show rather less of similarity than any of the others, yet as 
the range of variation in these measurements is so slight, three 
and four millimeters respectively, the real amount of propor¬ 
tional variation between these two members is considerably less 
than in others in which the trends of the graphs more closely 
coincide. The first and second phalanges of the fourth digit 
as plotted are rather exceptionally parallel; yet the extremes' of 
proportional differences is five per cent, greater than between 
