Temporal changes in key factors and key age groups influencing the population dynamics of female red deer

Summary 1. We use structured demographic accounting to decompose the contribution to the variance in relative population growth between 1971 and 1997 in female red deer on Rum, Scotland, in terms of different fitness components and age categories. 2. During the first 10 years (1971–80) the variance in relative population growth was small (0·00246) as the population grew from 58 to 148 having been culled previously. After this the relative variance in population growth increased to 0·00575 between 1981 and 1989, and increased again to 0·00848 between 1990 and 1997. 3. In the first phase of population growth (1971–80) changes in the birth rate contributed most to the variance in relative population growth (30%), then adult winter survival (23%), with calf summer and calf winter survival contributing less than 10%. Birth rate tended to covary strongly with adult winter survival, and to a lesser extent with calf winter survival. 4. The contribution of birth rate to the variance in relative population growth diminished to less than 5% in both the periods 1981–89 and 1990–97. In contrast, the contribution of adult winter survival increased to 36% in 1980s and 40% in 1990s. At the same time, the covariation of adult winter survival with calf winter survival increased successively to reach 40% in 1990–97. 5. Changes in population age structure were small and contributed little to the variance in relative population growth (≈ 1% over the entire study). In the 1970s the 5–7‐year‐old group contributed most (23%) to the variance in relative population growth, though this covaried with 8–12‐year‐olds. In the 1980s the 8–12‐year‐olds contributed most (27%) and now the covariation was with older animals (> 12‐year‐olds). This pattern persisted in the 1990s, though there was covariation with 5–7‐year‐olds too. 6. Birth rate, calf winter survival and adult winter survival were all density‐dependent. The strength of density dependence within each vital rate did not differ between the 1970s, 1980s and 1990s; however, there were differences in the means of these vital rates relative to population size between the three periods. 7. The implications of temporal changes in the most important vital rates, the ‘key factors’, contributing to the variance in relative population growth rate are discussed.