COMPARATIVE COCOONING INDUSTRIES. 



163 



I was, however, made painfully conscious of the fact that she very 

 rarely presented the same attitude consecutively. As she made her rounds 

 she would almost invariably appear at a different point each 

 Wind- time, now above, now below, now at the middle, and anon 

 J^j^ , emerging from beneath and coming back upon her course. (Figs. 

 198, 199.) It thus became a matter of much greater difficulty 

 to secure good drawings than I had calculated upon, although I eventually 

 satisfied myself. But in the meantime it appeared that by this peculiar 

 mode of progress the spider equalized the distribution of spinning material 

 upon her cocoon, and prevented any part of it from growing dispropor- 

 tionatelj' to otlier parts. The method was not unlike that of a person 

 winding a ball of silk or wool from a skein of thread. 



One cannot but recognize in this action a manifest purpose, however 

 directed or originated, to build up her cocoon mass symmetrically, and 

 cover all parts thereof equally. While thus engaged in spinning, the feet 



Fig. 198. Fk;. 199. 



Equalizing the output of thread. 



Fig. 198. Position below the cocoon. Fig. 199. Position above on next round. 



were extended upwards, grasping tlie shoulder of the flossy mass or the 

 edges of the supporting top piece. As the mass increased, the legs, of 

 course, were stretched out further, but at no time was there any difficult}' 

 in enclosing the entire structure within the long legs of the animal. 



Brief rests were taken at long intervals of time, but the periods were 

 very short, three or four minutes as a rule, rarely more. The spinning 

 continued without intermission, sometimes more slowly and again 

 ^ . more rapidly. At seven o'clock and eight minutes the cocoon 



appeared to be completed, as far as its general shape and size 

 were concerned, but the spider continued working on it until ten o'clock 

 and fifteen minutes, when I was compelled to cease my observations. 



During the last three hours the spinnerets were more frequently squeezed 

 against the cocoon, as though to pack the mass and fasten the threads 

 more closely. The filaments now, instead of being bent upon the surface 



