OBSERVATIONS ON THE FAMILY COLEOPHORIDES, 87 



when taken from the mine, crawls in the ordinary Avay without 

 arching the body, and I think it probable that some of the seed-feeding 

 larvae which only make cases late in life may be found to crawl in the 

 same manner during the earliest stages. But even in the case 

 mentioned the true legs afforded the means of progress. The 

 Coleophorid larva has three or four pairs of ventral proleijs and a pair 

 of fairly well developed anal daspers. These four or five pairs together 

 are employed during progression to maintain the case in its proper 

 position, and also when the larva is more or less exposed while feeding 

 to withdraw the body into the case when necessary. Besides these 

 functions the prolegs and claspers must perform an important office 

 during the process of skin-casting. It is probably by their means that 

 the old skin is firmly fixed, and while the larva creeps out of it the 

 old skin will be retained in position by the old crotchets belonging to 

 it. These ventral prolegs of the Coleophorids are small, short, and 

 weak, and more difficult to see than those of many smaller larvit 

 belonging to other families. When taken from its case the larva has 

 a habit of contracting the prolegs to such a degree tbat Avhen viewed 

 in the lateral aspect it appears to have none at all, but when viewed in 

 the ventral aspect the dark crotchets will usually show where the pro- 

 legs are situated. When examining the prolegs it is always well to 

 use the microscope. The larva may be placed on its back on a piece 

 of moistened stamp paper where it will remain long enough to allow 

 of an examination being made. Normally the crotchets on each pro- 

 leg lie in an anterior and a posterior row not connected, each row 

 consisting of about half-a-dozen points. They vary in number and 

 are occasionally absent. In one larva of (fenistae there were none on 

 one leg and the other legs each carried a different number, while in a 

 larva of siccifolia all these legs were without crotchets. 



In the genus Coleophnra there are a greater number of characters 

 present which may be used to distinguish the different species than is 

 perhaps found in any other genus. Besides these characters which 

 may exist in the imago, pupa, larva and ovum, we also have the larval 

 case in its essentially different forms, to consider when we attempt to 

 divide into groups the very numerous species. With so many factors 

 to work on this division would seem to be an easy task but as certain 

 species stand out alone and others show gradations of these characters 

 or apparently strange combinations of them, no hard and fast lines 

 can be drawn, at least with our present knowledge. Nevertheless 

 certain natural groups do exist though we cannot yet clearly define 

 their limits. From certain evidence to be considered later we 

 may conclude that there are two main divisions, the one consisting of 

 those species whose larvfe feed on the seeds of plants, the other con- 

 taining the species whose larv?e mine in the leaves of plants and 

 further that the seed-feeders are the more ancient of the two. Tbat 

 the loss of the prolegs belonging to the sixth abdominal segment 

 which occurs in some of the larviB has been brought about l)y 

 redaction, lies beyond dispute. The cause of this redaction I propose 

 to consider later. There is a group of rather small species whose 

 larvae feed on the seeds of Juncaceae, and in their earliest stages some, 

 probably all of them, live without a case in the seed heads of these 

 plants. This group prohal^ly contains the most primitive species now 

 known. All those examined have four pairs of ventral prolegs. They 



