PRESKRVATION OF LARV/E BY INFLATION. 231 



is Id be inflated ; and in a few minutes the operation is com- 

 pleted. The exact amount of time and heat required are 

 points on which we can only say, "Experientia docet." The 

 head of the larva always dries last, so it is easy to tell if it is 



Fifir. 3. 



quite done by touching the head or first pair of legs with a 

 needle; if they move in the slightest the larva must go in 

 again. Sometimes on raivsing the spring the larva will fall off 

 without trouble ; more often it sticks to the tube, and can 

 only be removed by softening the last segment with a little 

 water drop])ed from the end of a match : the larva can then 

 soon be got off with the help of a needle ; the superfluous 

 water must be taken up at once with a little blotting-paper, 

 and the flap closed down; put the larva on the cork, tail 

 foremost, and let it remain in the oven a few moments to dry 

 off the moisture. Attention to these details is specially 

 urged, as we have seen otherwise good specimens spoiled by 

 the distended aperture in tlie tail, which, while it gives a 

 good view of the thorough cleansing the skin has undergone, 

 can hardly be called desirable. Fortunately larvae have a 

 tendency to take the position natural to them : a Noctiia 

 could hardly be made to loop, while a Geometer can readily 

 be fixed in its characteristic position with the help of a few 

 short pins. Geometers are best laid sideways on the cork, if 

 it is wished to arch them ; indeed, any larva that is to be 

 specially posed is best laid sideways. The larvae must each 

 be dried separately if they are to lake a particular position, 

 otherwise there is nothing to prevent inflating two together, 

 the air-pressure being more than sufficient. To connect them 

 with the india-rubber tube, have a piece of metal made at a 

 tinman's of the shape of the Y used tor butlerfly-nels, only 

 nuich smaller: it must be made hollow, so that l)y blowing 

 in at the stem the air will pass out at each of the l)rancl)es; 

 the thickness of llie metal should be such that it will lit 



