1. Musca domestica and other non-biting flies frequenting similar feeding-grounds, are probably all liable to infection with a common flagellate. The great variability of this form is shown on comparing the flagellates in the larvae with those in the mature flies of Homalomyia, sp. and of Scatophaga lutama.
2. Infection is casual–i.e. by the mouth. In the case of the dung-flies examined, the larvae ingest faecal matter infected with herpetomonad cysts: the cysts develop into flagellates in the mid-gut, where they multiply with great rapidity: towards the close of larval life, when the larva stops feeding, they round up in the hind-gut, and are for the most part passed out as cysts. A few survive the pupal stage in a half-encysted condition, but it is probable that the infection of the adult fly is usually freshly acquired. The cycle in the fly is similar to that in the larva, and is in agreement with Patton's account of Herpetomonas muscae-domesticae. The parasite was never found in the ovaries or ova. Patton's suggestion that the degree of infection depends directly on the number of cysts ingested, is borne out by the much higher rate of infection in the larvae than in the flies: this is not surprising when we remember the complete restriction of the feeding larva to the infected area.
3. It is important to use some reliable cytological stain such as iron-haematoxylin as a control to Romanowsky stains where possible, seeing that very different results are sometimes given by the two methods.
4. The apparent double flagellum is produced in the course of longitudinal division. The new flagellum grows up alongside the old, and is not merely split off from it: this is best seen in the flagellate from the larva, where the body of the organism usually divides at an earlier period than in the fly. Study of the living flagellate is necessary to a clear understanding of the process of division.
5. I have seen no conjugation in the living herpetomonads. Occasionally flagellates were met with in Giemsa preparations devoid of a tropho-nucleus. Such individuals might be regarded as male gametes; it is more likely that they are simply degenerate forms. Sufficient consideration is not given to the possibility of degeneration in richly-nourished, rapidly multiplying protozoa, such as the trypanosomes and their allies.
6. In encystment the flagellum is not cast off bodily, but is drawn down into the cell by the kineto-nucleus, which moves to a position either alongside of, or posterior to the tropho-nucleus. In this way apparent Crithidia or even trypanosome forms are produced, but there is no hint of an undulating membrane. Small blunt “trypanosomes” were also produced occasionally by adhesion of the flagellum to the body-wall in the “stickiness” resulting from confinement under a cover-glass: the efforts of the flagellum to free itself raised up an undulating membrane, and produced a very deceptive appearance.
7. The early stages of encystment could be induced by keeping the flagellates under waxed-down cover-slips, where they would continue to live for 30—40 hours. Similar results were got by transferring flagellates to agar-agar, or by subjecting the larval host to starvation for a day or two.
8. From what I have seen, I do not agree with Chatton and Alilaire's suggestion to divide the genus Herpetomonas into Leptomonas and Herpetomonas proper. I do not think that these authors are justified in regarding the double flagellum and the presence of a long rhizoplast as generic characters.