38 Bulletin American Museiivi of Natural History {Vol. XXIII, 



the workers, and were of the same color, except that a number of them had 

 traces of the black antero-median and parapsidal blotches of the female on 

 the mesonotum. They had a perfectly normal appearance, as the enlarge- 

 ment of the mesonotum led to less distortion and inelegance of profile than 

 in pseudogynic F. sanguinea and ruja, since in iyicerta, which belongs to the 

 pallide-fulva group, the thorax, even of the normal worker, is imusually long. 

 Traces of the scutellum and metanotum were rarely present in any of the 

 specimens, so that the only difference between these and normal workers 

 was the greater prominence of the mesonotum. The extent of this modi- 

 fication may be clearly seen by comparing Plate III, Fig. 42, which represents 

 a large typical worker, and Figs. 43 and 44, which represent pseudogynes 

 from the same colony and drawn under the same magnification. 



On finding these pseudogynes I was convinced that an examination of 

 the incerta nests earlier in the season would reveal the presence of Xenodusa. 

 Accordingly, on revisiting Colebrook during July of the following year (1905) 

 I set to work to look for the beetle and its larva. July 1, I found under a 

 stone on the middle hill a small colony of incerta containing about 75 normal 

 workers, a few worker pupte and larvse, and 6 larvae which, from their resem- 

 blance to Wasmann's figure of the Loviechusa larva^ could at once be 

 recognized as those of A'^. cava. The legs, however, seemed to be much 

 longer and the body more slender and more concave above. They were 

 clinging to the lower surface of the stone covering the nest. I transferred 

 them to an artificial nest together with as many of the ants as I could capture. 

 The larvae associated themselves with the brood which the ants had collected 

 in the cavities of the damp sponge in the dark chamber of the nest. They 

 walked about but little and very clumsily as their legs seemed to be incapable 

 of much movement at the strongly flexed articulation between the femora 

 and tibiae. They were frequently seen in the act of begging the ants and 

 one another for food. At such times they raised their fore feet and stroked 

 the head of the ant or fellow larva. Although the ants usually responded 

 very willingly to this solicitation, the liquid food thus received seemed to be 

 insufficient, for one morning I saw one of the Xenodusa larvae seize and 

 devour an ant larva about 3 mm. in length. On July 7 two of the Xenodusa 

 larvae had disappeared (eaten by the ants?) and the remaining four had 

 become somewhat inactive after having grown appreciably during their 

 week's confinement in the artificial nest. Fearing that the ants might devour 

 the remaining parasites, and concluding from their size that they must be 

 nearly ready to pupate, I removed them from the nest and embedded them 

 in some earth. This proved to be disastrous as I had not taken the precau- 



1 Die Moderne Biolosie unci die Entwicklungstheorie. 2 Aufl., Freilmrg i. Br., 1904, p. 

 223, fig. 30. 



