402 Trans. Acad. Sci. of St. Louis 
record the details at hand, in the hope that additional 
data later will reveal the system in this complicated piece 
of work. 
One tunnel containing five cocoons (‘‘A’’ in fig. 52) 
was 12 inches in length and approximately 1% inch in 
diameter. The bottom half inch was filled with finely bit- 
ten pith; the cell above it was one inch in length, the 
lower half of which was filled snug with a cocoon made 
of earth, and the upper half was empty, but the walls 
were dirt-smeared. It seemed that in all probability the 
mother had made this room thus, filling the lower half 
with egg and provisions, and the upper half with dirt, so 
that the full-grown larva, when ready to pupate, could 
pull the loose earth down around itself, moisten it some- 
how and with body contortions shape a cocoon against 
the sides of the cell. The outside of this cocoon was 
rough and granular, while the inner surface conformed 
to the shape of the occupant’s body, and was glazed with 
some substance which made it smooth and brittle, though 
more tenaceous than mud. It is reasonable to surmise 
that in this species, as in many others, the excrement 
that accumulates in the body during larval growth is 
at the time of pupation ejected in toto and is worked to 
a smooth, glazed surface. While the sides and bottom 
of the cylindrical cocoon were outwardly rough with 
natural granules of dirt, the top (headward end) of the 
cocoon was a transparent disc, made either by spinning 
or by skillful manipulation of the varnishing material. 
Exactly these same conditions were present in all five 
of the cocoons which utilized the bottom 8 inches of the 
tunnel. Above the open space of the top one was a plug 
of mud, loosely thrown in, an inch deep, and then an 
empty channel to the top. 
Other nests were similar in the nature of their con- 
