1893.] 271 



Three forms, then, of frass arrangement may be recognised : — 

 (1) collected into a central thread, with a wide free margin on either 

 side ; (2) scattered irregularly over the mine, with comparatively 

 little or no margin ; (3) arranged in regular rows from side to side, 

 or, as I propose to term it, " coiled." The larva, however, does not 

 by any means limit itself to one form, for it will deposit its frass 

 after one fashion in one part of its mine, and after another in another 

 part, the change generally indicating that a moult has taken place. 

 And here naturally comes the question, how often does the Nepticula 

 larva change its skin ? Looking at the size of the objects, the enquiry 

 is one of much delicacy, so much so, indeed, that for a long time I 

 quite overlooked the first moult. Three ways of proceeding seemed 

 open to one. One was to follow an individual larva through its 

 whole life, and to count the number of times it lay up. In most 

 species an opportunity of doing this would be purely accidental, for I 

 can scarcely imagine the patience requisite to search for a Nepticula 

 larva just hatched, but there are a species or two that stain the leaf 

 directly they begin to burrow, of which angulifasciella is one, and by 

 looking for its little purple spots in the rose leaves it is comparatively 

 easy to catch it immediately after it has begun to mine. The second 

 method was to compare together different individuals of a species at 

 all ages, and to reckon the number of moults from the number of «izes 

 in the head. Here again angulifasciella is a convenient species, for 

 however small may be the mine of most other species we are likely to 

 drop on, the first moult will probably have been passed, and in that 

 case the only plan left to ascertain the existence of this moult is to 

 open the mine and search for the cast head among the debris — a 

 troublesome proceeding, which I have been contented to undertake in 

 suhhimaculella ^n^fulgens only. The other possible method was, it 

 seemed to me, to note the characters of different parts of the mine, 

 and so to cut it up into segments answering to the stages or skins of 

 the larva. An excellent plan in some respects, but misleading, because 

 it took no account of the first moult, and consequently broke up the 

 mine into three portions only instead of into four. By these methods, 

 separate or combined, I have examined many species so far as to 

 ascertain, and at the same time allocate, the last two moults, but in 

 only the three species, angulifasciella, suhhimaculella, siudfulgens, have 

 I carried the examination to the point of demonstrating the existence 

 of the other or primary moult. When examining the heads, I made use 

 of a compound microscope with ^ inch objective, and I found that the 

 four sizes might be roughly represented thus — the length of the first 



