1. Of 20 mice (previously ascertained to be free from Hymenolepis infection for a period of at least 42 days), artificially infected with Hymenolepis eggs, only one failed to show definite evidence of Hymenolepis infection when examined from three to 24 days after. Of nine control mice none showed infection during a period of at least 53 days and were found to be uninfected when killed. Of five further control mice which were subsequently artificially infected, four were found to be free from infection when killed.

These results prove conclusively that direct infection, i.e. the infection of one mouse by ingestion of the eggs laid by the Hymenolepis of another mouse, can take place without the intervention of any intermediate host. The one-host account of the life-cycle of Hymenolepis fraterna promulgated by Grassi and Rovelli and Joyeux is thus the correct one, though this does not exclude the possibility of the Hymenolepis egg occasionally developing into a cysticercoid inside the body-cavity of a flea and undergoing the rest of its development when the flea is eaten by a mouse. I have not ascertained whether or not a mouse can re-infect itself by swallowing Hymenolepis eggs contained in its own faeces.

2. Hymenolepis fraterna individuals of the same age (i.e. developed from eggs ingested at the same time) often but not always develop at very different rates in different mice and in the same mouse. Grassi and Rovelli's statement that Hymenolepis fraterna takes from 15 to 30 days to become mature is approximately correct.

Hymenolepis fraterna can become mature, i.e. possess ripe eggs in the hind proglottides, when only 6 mm. or 7 mm. long. In cases of heavy infection (e.g. 100–200 individuals in a single mouse) the worms probably always remain of small size (ca. 10 mm.) as compared with the worms in light infections where they attain a length of 20–25 mm.

Eggs can be shed as early as the fifteenth and probably on the fourteenth day of development.

3. It is probable that mature H. fraterna often sheds all its eggs within a period of a week and often less, the eggless and degenerate worm then being voided in the faeces.

4. In artificial infection of mice definite results can only be obtained with heavily-infected faecal material. Occasionally mice appear to be immune, but, judging from my experiments, such immunity is rare in mice of medium size. I cannot confirm Joyeux's statement that very young mice are most susceptible to infection. Out of some 200 mice examined I found that the smallest and largest mice were least liable to be infected, medium-sized mice supplying all my infected faeces. The largest mouse (over 35 gm.) in my experimental series, on the other hand, became infected with a weak concentration of eggs.