To understand whether mammalian circadian time structure measurably affects the host-cancer balance, we studied tumor-take frequency after s.c. tumor cell inoculation and the number of pulmonary tumor nodules after i.v. tumor cell injections at each of 6 equispaced times of day. We employed 2 genetically distinct mouse strains and 2 different tumor model systems, a methylcholanthrene A-induced fibrosarcoma of C 3 HeJ mice and 2 B 16 melanoma cell lines of vastly different metastatic efficiency in C57 Black/6 mice. Fibrosarcoma cells were injected s.c. in 1 of 8 different doses, at 1 of 8 permutated anatomic sites and at 1 of 6 equispaced circadian times, in 96 female C 3 HeJ mice maintained under a synchronizing schedule of 12 hr light alternating with 12 hr dark. Regardless of tumor cell dose and inoculum location, tumortake frequency depended strongly upon the circadian stage of tumor cell inoculation. Injections of between 2,000 and 50,000 live tumor cells inoculated near the daily sleep/wake interface resulted in the lowest incidence of tumor take compared with inoculation at other times of day. In the experimental i.v. B 16 melanoma metastatic model (N โซุโฌ 110), the capacity of both high and low metastatic potential clones to successfully metastasize to lung depended, to a large extent, upon when in the day each of these clones was injected. Similar to the fibrosarcoma data, the daily sleep/ wake boundary was the time of day associated with the greatest resistance to metastatic spread. Int.