Pupal suspension of Thais. 213 



the rule that rotation is made once only for a fraction of 

 a second, about twenty-four hours after the moult. 



The rotation is on an axis, i. e. it does not sweep the 

 head round in a circle, maintaining the venter towards the 

 stem of attachment all the time, but presents towards the 

 stem successively one side, the dorsum, the other side, and 

 so on. The rotation is effected by the abdominal seg- 

 ments being bent forwards, i.e. virtually making the pupa 

 present a deep curve towards the stem. Then by sweep- 

 ing round the abdominal segments in a way that is common 

 to many pupai the hollow of the curve remains towards 

 the stem, but affects successively each aspect of the pupa 

 as it rotates. 



The mature pupa a few days later is very stiff and 

 straight, and its possessing such motility twenty-four hours 

 after moulting is somewhat unexpected. 



In order to observe the way in which the girth is made 

 by the larva of Papilin, I obtained a few larva of P. 

 machaon, and kept rather a close watch on them, but four 

 succeeded in suspending themselves, girth and all, without 

 my having surprised them at work. They do it apparently 

 somewhat quickly, and give no clear indication beforehand 

 of when they are likely to do so. With the fifth I was 

 more fortunate, happening to look at it when the making 

 of the girth was under way, indeed more than half-finished; 

 he was working at one end, passing the spinneret too and 

 fro about the attachment of the loop. The loop itself, 

 consisting of a number of quite detached threads, passed 

 between the fore-legs of the larva, forwards, the larva being 

 so bent round that its head and prothorax were directed 

 backwards. Then well above the meso-thoracic leg and 

 rather higher on the meta-thorax, then well above the 

 spiracle of the 1st abdominal, not quite so far above that 

 of the second, and then passed down to its other attach- 

 ment. The most anterior portion of the larva in this 

 position was the side of the meta-thorax and 1st abdominal 

 segment just below the loop. Then the larva began to 

 add another thread to the loop. The head and front seg- 

 ments are well raised and the loop falls into the incision 

 between the 1st and 2nd pair of legs, the head is pressed 

 down and the spinneret lies between the first two legs 

 and reaches very close to the loop ; the larva sweeps the 

 head across from the one end of the loop to the other 

 slowly and by short jerks and stops, as though the silk 



