450 
[March 
men is sometimes piceous below. In two 9 spongifica the thorax is 
almost rust-red, (as observed in a single 9 G. q. coccinese by Osten Sacken, 
Proc. Enf. Soc. Phila. I, p. 244,) and the abdomen piceous red; in 
another 9 the thorax is tinged with rust-red and the abdomen is piceous; 
and in the fourth 9 the thorax is black and the abdomen is piceous red, 
the remaining 9 specimen being uniformly black as are also both S % . 
In the closely allied or identical species q. inanis^ however, one of my 
two % % has a piceous red abdomen and all my 9 9 9 ^ black 
thorax and a piceous red abdomen. 
7th. Viewed laterally, the upper edge of the second abdominal joint 
(counting the peduncle as the first joint) describes a circular arc of 
about 60° in both forms. Taking the chord of this arc as a definite and 
permanent basis of measurement, in spongifica 9 the lower or ventral 
edge proceeds straight downwards exactly at right angles with this 
chord for a distance equal to J or i the length of the chord, before it 
curves gradually backwards to form the ventral arch. In aciculafa 9 on 
the contrary, instead of being at right angles (90°) with the chord, it 
forms with it an angle of about 110°, so as to exhibit a most extraor¬ 
dinary bulge in front, and it curves much further downwards from the 
peduncle and in a more compressed and knife-edged form, so that the 
abdomen is vertically at least as wide as long and almost always much 
wider, whereas in spongifica 9 it is always longer than wide and gene¬ 
rally much longer. The above variation in each form is caused by the 
terminal abdominal joints being more or less telescopically drawn out 
in different specimens, so that in each form the second abdominal joint 
sometimes occupies dorsally I the entire length of the abdomen exclusive 
of the peduncle, and sometimes almost |. St. Fargeau has observed the 
same thing of the genus Megachile, {^Hymenopt. II, p. 338,) and I only 
notice it here because Osten Sacken, having only a few specimens of 
each form on hand, supposes the relative length of the 2nd abdominal 
joint with regard to the terminal joints to be a permanent character of 
each. i^Proc. Ent. Soc. Phil. I, p. 246.) 
8th. In consequence of the above bulge on the anterior abdomen in 
aHculata, (See Aqypcndix., Fig. I,) the distance from the ‘‘ ventral 
valve” (Fig. I, z;,)* to the “dorsal valve” (Fig. I, 7)* is proportion- 
These terras are explained in the Appendix. 
I 
