i DASYGASTRES ANTIIIDIUM 47 



larval life it commences the expulsion of " frass " in the shape of 

 small pellets, which it fastens together with silk, as they are 

 voided, and suspends round the walls of the chamber. This 

 curious arrangement not only results in keeping the embarrassing 

 material from contact with the food and with the larva itself, 

 but serves, when the growth of the latter is accomplished, as the 

 outline or foundations of the cocoon in which the metamorphosis 

 is completed. This cocoon is of a very elaborate character ; it 

 has, so says Fabre, a beautiful appearance, and is provided with 

 a very peculiar structure in the form of a small conical pro- 

 tuberance at one extremity pierced by a canal. This canal is 

 formed with great care by the larva, which from time to time 

 places its head in the orifice in process of construction, and 

 stretches the calibre by opening the mandibles. The object of 

 this peculiarity in the fabrication of the elaborate cocoon is not 

 dear, but Fabre inclines to the opinion that it is for respiratory 

 purposes. 



Other species of this genus use resin in place of cotton as their 

 . working material. Among these are Anthidium septemdentatitmand 

 A. bcHicosmtt. The former species chooses an old snail-shell as its 

 nidus, and constructs in it near the top a barrier of resin, so as to 

 shut off the part where the whorl is too small ; then beneath the 

 shelter of this barrier it accumulates a store of honey-pollen, de- 

 posits an egg, and completes the cell by another transverse barrier of 

 resin ; two such cells are usually constructed in one snail-shell, and 

 below them is placed a barricade of small miscellaneous articles, 

 similar to what we have described in speaking of the cotton- 

 working species of the genus. This bee completes its metamor- 

 phosis, and is ready to leave the cell in early spring. Its con- 

 gener, A. lellicosum, has the same habits, with the exception 

 that it works later in the year, and is thus exposed to a great 

 danger, that very frequently proves fatal to it. This bee does 

 not completely occupy the snail- shell with its cells, but leaves 

 the lower and larger portion of the shell vacant. Now, there is 

 another bee, a species of Osmia, that is also fond of snail-shells 

 as a nesting-place, and that affects the same localities as the 

 A. septemdentatum ; very often the Osmia selects for its nest 

 the vacant part of a shell, the other part of which is occupied by 

 the Anthidium ; the result of this is that when the metamorphoses 

 are completed, the latter bee is unable to effect its escape, and 



