The influence of environmental and human factors on group sizes of large ungulates is still poorly understood, especially when considered at multiple timescales. We analyzed long-term data of red deer (Cervus elaphus) group sizes collected at 3 timescales in Białowieża Primeval Forest, Poland, from 1986 to 2003. On the daily scale, group size differed among periods of the day for females in summer (P = 0.001) and autumn (P = 0.04) but not winter (P = 0.18). Group size of males did not differ (P > 0.35) among the periods of day during any season. On the seasonal scale, both sexes had larger group sizes in winter than summer, and males were always more solitary than females. At the annual scale, multiple-regression analysis explained 96% (r² = 0.957) of deer grouping patterns. Human hunting had the greatest influence (sr²_i = 0.414), followed by acorn crop (sr²_i = 0.012) and snow cover (sr²_i = 0.008). Intense hunting by humans during a 5-year period appeared to cause increases in group sizes, even while red deer densities decreased. The disruption of the normally positive relationship of deer density to group size apparently affected predator–prey relationships, consequently red deer numbers continued to decline even after hunting pressure was relaxed. Our results indicate caution should be used when reducing red deer numbers by dramatic increases in hunting because the resulting behavioral and ecological changes in deer have implications that extend well beyond the season and year, with important consequences for the long-term stability of the population, especially in areas where large carnivores are present.