1. Evidence from previous literature shows that puppies are commonly infected with larvae ofToxocara canisat birth and that prenatal infection can be produced by feeding embryonated eggs to pregnant bitches.2. Observations on fifty-eight dogs in Brisbane showed that all of twenty-nine puppies 1–6 months old were infected, while only three out of twenty-nine dogs over 6 months old were infected.3. In naturally infected puppies, 1–3 weeks old, it was found that at birth third-stage larvae were present in the lungs. Third-stage larvae continued to appear in the lungs for the first week of life; their length was 0·6–1·3 mm.4. Third-stage larvae were found in the stomach on the day after birth; at 3 days after birth fourth-stage larvae were found in the intestine.5. It appeared likely that the second moult occurs in puppies before birth, and that the third moult takes place at a length of 1·0–1·3 mm. in the lungs and stomach within the first week of life.6. By the beginning of the second week, fourth-stage larvae were fully grown and had commenced the fourth moult at a length of 5–7 mm.7. Throughout the second and third weeks, adults grew rapidly, reaching a length of about 67 mm. by the end of the third week, but no eggs had appeared in the faeces at this time.8. Experimental infection of mice with eggs showed that the larvae were distributed to the somatic tissues, few reaching the alimentary tract. The larvae did not progress beyond the second stage in the tissues of mice.9. Experimental infection of dogs with eggs showed that in dogs over 5 weeks old the larvae were distributed to the somatic tissues and did not reach the alimentary tract. The larvae did not progress beyond the second stage, though some of them showed signs of commencing the second moult. In contrast, 1 to 3-week-old puppies infected in the same way were found to harbour larvae in the alimentary tract. Though some of these larvae were probably derived from a naturally acquired prenatal infection, it was evident from the progress of development that the experimental infection resulted in the presence of second-stage larvae in the liver and lungs and that these larvae underwent the second moult, commenced development as third-stage larvae in the lungs, migrated into the stomach and developed to the adult stage in the intestine.10. Experimental infection of dogs and foxes with mice harbouring second-stage larvae in the tissues showed that, in some instances, development of larvae proceeded in the alimentary tract. No evidence of somatic migration was found in dogs infected in this way, but in foxes second-stage larvae were found in the lungs.11. The structure and development of the second, third and fourth stage is described in detail and found to resemble closely the development ofT. catilarvae (Sprent, 1956). Second-stage larvae from dog tissues had a length of 0·34–0·44 mm. and the second moult occurred at a length of 0·37–0·44 mm. Third-stage larvae varied in length from 0·46 to 1·36 mm., and the third moult occurred at a length of 0·98–1·3 mm. Fourth-stage larvae measured 1·2–6·3 mm., though moulting fourth-stage larvae were observed up to a length of 7·4 mm. Sexual differentiation occurred during the fourth stage and was evident at a length of about 1·5 mm.12. The migratory behaviour of the larvae ofT. canisandT. catiis compared and discussed in relation to their wide range of hosts. It is concluded that differences in migratory behaviour are adaptations to prevailing modes of nutrition and it is suggested that the somatic migration occurring in dogs is an adaptation to the non-predatory habits of this particular host.