1907.] Wheeler, The Polymorphism of A^its. 9 



suppose that it had reached its full growth within the ant-larva and was just 

 breaking through the integument on the nuchal or sternal side. I saw the 

 workers, which evinced the greatest interest in this phenomenon, pull the 

 larva away from the semipupa and throw aside the latter, now reduced to a 

 sickly mass, though still retaining enough of the contents of the abdomen 

 and of its original form to be recognizable as a pupa, notwithstanding the 

 fact that the head, as shown in Fig. 16, was much smaller than in the cor- 

 responding stages of the nonparasitized soldiers and females. 



Shutting my eyes to the correct interpretation of the above stages as 

 indicating that the larva was ecto- instead of entoparasitic, I stained and 

 mounted m toto whole series of soldier and female larvae and young pupa? 

 in the hope of finding the Orasema larvse prior to their eruption. This 

 search proved, of course, to be futile, and I was baffled until I accidentally 

 found the crucial stages represented in Plate II, Figs. 13 to 15. 



The growth of the parasite, after it has plunged its mouthparts into the 

 integument of its host, must be extremely rapid. I doubt whether the 

 stages above described require more than a couple of days for their com- 

 pletion. Such rapid growth, however, is not surprising when we consider 

 the accessibility and high nutritive value of the food on which the larva sub- 

 sists. 



As soon as the full-grown Orasema larva has been separated from its 

 prey, it begins to pupate. Occasionally the ants are either unable or neg- 

 lect to detach the parasite. In such cases, two of which are shown in Plate 

 II, Figs. 18 and 19, the larva begins to pupate in situ. It undergoes an ecdy- 

 sis in which it is undoubtedly assisted by the workers, and then appears as 

 a short, thick-set semipupa, slightly constricted just in front of the middle 

 of its body. Another ecdysis seems to follow almost at once, leaving the 

 semipupa covered with a peculiar envelope studded with large blisters, or 

 pustules. These are arranged segmentally in regular rows along each side 

 of the body but are absent in the middorsal and midventral regions. I am 

 unable to assign any function to these singular organs, which in O. viridis 

 disappear with the semipupa stage. On focussing through the pustulate 

 envelope the semipupa is seen to present the appearance of Fig. 20 (PI. II). 

 The imaginal head, with its large eyes and antennae, is embedded in a hood- 

 like prothoracic mass; the legs and Avings are clearly indicated. A little 

 later the pustulate envelope is shed and the complete, pure white pupa of 

 the Orasema is seen enclosed in a thick membrane (PI. II, Fig. 22) which, 

 in the intersegmental regions of the abdomen, is thrown into prominent 

 transverse welts. The color, which now gradually deepens, becoming 

 first blackish and then metallic green, is dimmed by the rather opaque, 

 white pupal envelope (PI. II, Fig. 23). Soon after this stage is reached, the 



