irregularly (figs. 5, 6 and 7), so that one may have developed into an adult flagellate while the remainder are dividing further and remain attached to its anterior end (fig. 7). The adult flagellates divide by simple longitudinal division (figs. 8 and 9), and eventually collect m the hind-gut of the bug, where they attach themselves in rows; they now shorten (fig. 10), the flagella degenerate (fig. 11) and are shed, and they divide (fig. 12) at least twice (fig. 13), resulting in small pear-shaped bodies, the cysts, which are passed out in the fæces of the bug (fig. 14).
I have already described in great detail the development of the parasite of kala-azar in twelve bed-bugs (Ciniex rotundatus), and if you refer to my illustrations you will find its life-cycle, so far as it is known, is very similar to that of the two herpetomonads I have just described. All three parasites are very similar in their preflagellate stages, except that H. muscæ domesticæ is much larger; they develop and multiply in the same way, and in their adult flagellate stage they have the same general morphological build. They all undergo their developmental stages in the intestinal tracts of insects, but the fact that one of them passes its preflagellate stage in man has led most observers to regard it as something quite distinct from a Herpetomonas. The parasite of kala-azar is certainly not a sporozoon; it remains throughout its life-cycle, as far as we know it, a typical flagellate. With a complete knowledge of the life-cycle of such a closely allied form as H. lygæi it is possible to conjecture that the parasite of kala-azar, in order to get back to man, must pass back again to a non-flagellate stage, postflagellate, and that this change is most likely to take place somewhere in the neighbourhood of the biting parts of the bed-bug. Had I known the
