Newly hatched fry were acclimated to 7 or 12A?C and either fed daily (controls) or denied food for varying lengths of time and then fed daily until the end of the study (day 91 at 7A?C and day 43 at 12A?C). Growth was reduced by delays in the onset of feeding of 27 or more days at 7A?C and 7 or more days at 12A?C. Mortality of fry unfed for more than 34 days at 7A?C, or more than 21 days at 12A?C, was higher than among controls. Daily mortality increased with the length of the food deprivation period and did not cease immediately when food was made available, but reached zero by the end of the study. Mortality among unfed fry reached 50% in about 59 days at 7A?C and 32 days at 12A?C. Study results permitted calculation of the 'point-of-no-return' (PNR) mortality, which included the mortality that occurred during the period of food deprivation, and also the delayed component of mortality that was directly attributable to starvation and that occurred after food was made available. The PNR for 50% mortality for food-deprived fry occurred after 52 days at 7A?C and 24 days at 12A?C. Thus, both measures of mortality indicate that lake trout fry would be highly resistant to death by starvation in the thermal habitat they would be expected to occupy in the Great Lakes. We conclude that a more likely adverse effect of reduced food availability would result from a reduction in growth rate that extends the length of time fry remain small and vulnerable to predation by adult alewives and other non-native fishes with which they associate.