REVISION OF STEEPSIPTERA PIERCE. 47 



egg at first and later draws up to one pole to give rise to the rudiment of the germ- 

 band by a rearrangement and multiplication of ils cells. Older embryos are of the 

 usual generalized type, although on account of their length they are curled up in the 

 egg in a peculiar manner. 



The foregoing is the oogenesis of AcroscMsmus irlieeleri Pierce 

 (pechii Braes). Whether this process is the same throughout the 

 order will probably remain for some time a matter for conjecture. 



LARVIPAROUS REPRODUCTION. 



There is no possibility of oviposition in this order. The period 

 of parturition is passed in the body of the parent, and the young 

 hexapods reach the open air by passing from the body cavity of the 

 parent into a canal formed between the parent's skin and its uncast 

 pupal skin, and thence through a slit opening on the ventral or 

 uppermost side of the cephalothorax, between the head and thorax. 

 The young swarm over the parent's body in immense numbers. 

 Newport (1851) computes that more than 7,000 triungulinids were 

 produced by one female Stylops aterrima Newport in Andrena trim- 

 merana Kirby. The writer (1904) counted 2,252 young of Stylops 

 swenJci Pierce {Xenos? species) from Andrena solidaginis Robertson, 

 omitting all consideration of multitudes of undeveloped eggs. These 

 young reach a new host in one way or another, already described, 

 and immediately commence the attack. Saunders (1852) found 

 that Xenos larvae became completely buried in the body of Polistes 

 in three hours, and the next day were completely out of sight. 



METAMORPHOSIS. 



The discussion of this subject must be confined almost exclusively 

 to Xenos vesparum. After the triungulinid has entered its host it 

 commences a development which involves some of the most inter- 

 esting phenomena known among insects. The campodeoid hexapod 

 grows so rapidly that it soon loses all semblance to its former self 

 and the body grows out of all proportions to the legs. From this 

 point in the biology of these larvae Nassonow (1892 e) must receive 

 all of the credit, and reference to his plates will be of much value. 

 The second instar is devoid of legs and scarabaeidoid in form. In the 

 first part of this instar the head, three segments of the thorax, and 

 ten segments of the abdomen are distinct and ventrally flattened, 

 the abdominal segments being shorter than the thoracic (Nassonow, 

 1892 e, pi. 1, fig. 5). During the latter portion of the instar the 

 body becomes cylindrical, and the abdominal segments become as 

 long as the thoracic (Nassonow, 1892 e, pi. 1, fig. 6). From this 

 point the development of the male and female is radically different, 

 and it will be best therefore to compare them in parallel columns. 



