AND ANATOMY OF CERTAIN HYMENOPTERA PARASITICA. 397 



ventral chain. As shown diagrammatically in PI. 26. fig. 24:, the ventral 

 chain ends at NY, just near where the bulb joins the ninth abdominal 

 segment ; tlie chain does not continue into the vesicle. In text-fig. 3, p. 393, 

 a profile view of the last segments o£ the abdomen shows that the last 

 ganglionic swelling is in segments twelve and thirteen, and lies in both ; this 

 was very clear especially in one case. The point to be noticed is that no 

 ganglia are found in the penultimate or ultimate segments o£ the larva 

 (counting the bulb as a segment). 



The Gonads lie in the eighth abdominal segment (PI. 24. fig. 6, PI. 25. 

 fig. 10, Gr) and open by a short duct in this region. The duct is solid and 

 only forms a blind connection between epidermis and gonad, at this period 

 (see PI. 26. fig. 24, G). In the embryonic larva the gonad is a most 

 conspicuous object (PI. 26. fig. 25, Gr). 



7i. Though I have been able to add many new facts with regard to several 

 other parts of the anatomy oE the larval Microgaster, the main purpose of 

 this paper is to describe the Abdominal Vesicle. 



The latter is a most remarkable organ, and I have come to the conclusion, 

 with some other observers, that its function is respiratory. Its gross outer 

 morphology at different stages has been reported upon already and is shown 

 in PI. 24. fig. 3, a-h. 



In Pi. 24. fig. 6, the vesicle is seen to consist of very large polyhedral 

 cells with large, often irregular, nuclei. The cells are largest at the outer 

 and middle regions of the balb, and where the latter tapers to join the ninth 

 abdominal segment, the vesicle cells become smaller till they pass imper- 

 ceptibly into the ordinary hypoderm cells of the larval body. In PI. 24. 

 fig. 6, the ordinary hypoderm cells in the head, thoracic and body region 

 were too small to be seen individually, but in the bulb they were easily 

 drawn in with the camera lucida. 



The largest vesicle cells were 90 /^ in their greatest length and their 

 nuclei attained a diameter of 40 /x. 



The Histologij of the Vesicle Cells is very remarkable. In PI. 25. fig. 12 

 are four cells from the region enclosed by the square in fig. 10. On the 

 outside of the cells of the vesicle is a cuticular chitinous layer (OC) which 

 in larv93 of various ages differs in thickness. The cells in PI. 25. fig. 12 

 were fixed for four days in 3 per cent, bichromate of potash. Just below 

 the outer cuticle the vesicle cells have a granular darkly staining zone 

 (GLX) which cannot be found in material preserved in alcohol-acetic 

 fixatives. The rest of the cytoplasm is clearer, but in some cases seems to 

 contain granules (mitochondria ?), while in other cases I have found a large 

 siderophile granule, insoluble in acetic acid. The nucleus is large, pale, 

 finely granular, often irregular, and does not contain a karyosome or 

 plasmosome. The inner edge of the cytoplasm is drawn out in a remarkable 



