62 Wisconsin Academy of Sciences , Arts and Letters . 
from forty-five to fifty millimeters. The legs' are steel-blue, 
with the exception of the tarsi, which are pale yellow. The 
length of the body measures from twenty to twenty-two milli¬ 
meters. 
The male differs so much from the female that it might be 
taken for a different species. The body is longer and nar¬ 
rower than that of the female, measuring twenty-five milli¬ 
meters or slightly more in length. Either the abdomen is of 
a steel-blue color, or the anterior half is bluish-black with the 
posterior portion reddish-brown; the whitish or yellowish 
markings which are so conspicuous in the female are wanting. 
The wings are either smoky brown or semi-transparent and ex¬ 
pand from fifty to fifty-five millimeters. One noticeable 
peculiarity of the male is a transverse oval hole, covered with 
a white membrane, between the alitrunk and the abdomen; 
this is hardly perceptible in the other sex. 
Individual variation exists in the number of whitish or yel¬ 
lowish markings on each side of the dorso-lateral sides of the 
abdomen. Harris (3), in describing the female, writes as fol¬ 
lows :—'“Her hind body is oval and of a steel-blue or deep vio¬ 
let color, with three or four oval, yellowish spots on each side.” 
Comstock (1), in his general description of the American saw- 
fly, writes:—“The female is about three-fourths of an inch 
long and has a black head and thorax, a steel-blue or purplish 
abdomen, with four yellowish spots on each side.” In order 
to ascertain the most frequent number of whitish or yellowish 
markings on each side of the abdomen, we examined seventy- 
five specimens with the following results:— 
8 specimens with three spots on each side of dorso-lateral 
abdomen. 
23 specimens with four spots on each side of dorso-lateral 
abdomen. 
44 specimens with five spots on each side of dorso-lateral 
abdomen. 
Although the number of specimens examined is rather small 
to draw any definite conclusions from, it appears that the 
most frequent number of markings on each side of the abdo¬ 
men is five. Undoubtedly Harris and Comstock failed to ex- 
