1318 Rocer C. SMITH 
parent. If a larva has been observed as having just started to spin, 
it may be removed to a glass slide or placed in a cell slide, and after 
some restlessness it will usually begin to spin a new cocoon. In this 
way the actual start can be seen. A larva spinning on the bottom of a 
vial, in the angle, first makes a framework by attaching viscid silk thread 
from the side of the vial to the bottom. The spinning is done entirely 
by the tail, which, as previously noted, can be extended and retracted in a 
remarkable manner. The colorless, gelatinous, silky secretion issues 
from the anal opening in the center of the tenth segment. The larva 
hes on its back or on one side, and reaches with its tail in all directions. 
When the secretion touches the glass, the soft silk readily adheres, form- 
ing interesting attachment disks. Then the abdomen is moved over to 
another place and the silken thread issues as it moves from one place to 
another. The thread is fastened again, and so the process goes on if 
supports are located. 
But if supports for the thread are not found, the tail searches aim- 
lessly about. Sometimes it seeks in vain and may attach a thread to 
one of its own setae. The tail can be brought forward over the head 
of the larva and attach a thread, or it can be twisted to either side to a 
relatively considerable distance. If no supports are available, the 
threads are fastened on one side of the larva and then carried up and. 
fastened to a seta, or carried over to the other side and fastened. During 
the first few hours the spinning proceeds rather slowly, and the more so 
if much time has been wasted by the larva in seeking places to attach 
the thread. The larva spins for a few minutes in one position and then 
shifts and continues in the new position. Threads are attached to other 
threads and by constantly shifting to a new position the larva keeps the 
cocoon spherical. 
The spinning from thread to thread is carried on in a fairly regular 
triangular design. The shifting continues at intervals, but instead of 
shifting in a true cirele the larva moves to one side a little, and in this 
way the wall of the cocoon is made of the same thickness. As the larva 
shifts, the prominent dorsal and lateral setae are broken off and go into 
the construction of the cocoon. Perhaps these add a degree of strength 
and rigidity, like the ribs of a basket. Long before the larva is hidden 
from view it is without setae. 
The spinning continues without cessation day or night unless the 
larva is disturbed. The triangular design continues until the cocoon is 
at least half completed. Observation from this stage on is difficult, as 
the larva is partly hidden. There appears to be a fairly gradual change 
to a figure-8 pattern. The end of the spinning appears to be a general 
plastering over the inside of the cocoon, which completely hides the 
