98 HYDROCAMPA NYMPH2EATA. 



end, and I made sure that it was full of water ; and 

 once I could notice a small silvery air-bubble clinging 

 to the side of the case. Probably when the larva 

 fastens up its case for pupating, it joins the two 

 pieces so closely together as to render it watertight, 

 but before that time the openings at the ends are 

 generally so wide that the water cannot be kept out. 

 I am, however, disposed to agree with Reaumur's 

 statement that the larva can control the admission of 

 water to its case, and probably this is regulated by 

 its requirements in the matter of respiration, and 

 must be managed by enlarging or contracting the 

 orifice at either end. 



When a larva makes use of Myosotis, after selecting 

 the end of a leaf, and detaching a piece of proper 

 length by a semicircular cut, it floats oft' upon it to- 

 wards the end of another leaf. There it either mounts 

 on the upper surface, and turns over the piece it is 

 carrying, so as to form its new roof, or else carries it 

 underneath and fastens it there as its new floor ; 

 either way, after fastening the edges of the detached 

 piece in place, the larva seems to be at leisure in 

 cutting out what is needed from the leaf to complete 

 its case, eating away the surroundings, and not merely 

 making a cut ; the cavity between the roof and the 

 floor seems formed by having one of them broader in 

 the first instance than the other, so that when the 

 edges are joined there is necessarily a bulging out of 

 the broader piece. 



When Sparganium is the material, the case is of a 

 narrower and more elongated form, the floor being the 

 flatter side, and thus shorter than the roof, which is 

 arranged so that the keeled surface of the leaf is out- 

 side, and the thin side-edges drawn inwards, and thus 

 made to help in the formation of the cavity. Not 

 unfrequently the case is formed of two sorts of leaf, 

 for a larva is not particular to have its case all of the 

 same material. Apparently in its growth between two 

 moults it changes only one half of its case at a time 



