Fawn: female ratio during December was the dependent variable 

 used to represent relative fawn survival. 



In addition to estimates of forage production, other 

 independent variables evaluated included: mean temperature 

 for the winter prior to birth, the number of adults in the 

 population at parturition, the number of adults one year prior 

 to parturition, and the average number of adults for the 2 

 periods. We included mean winter temperature as an index to 

 winter severity because of the possibility that severe winters 

 might reduce condition of the female, thereby affecting 

 subsequent fawn survival. Numbers of adult deer were included 

 to cover possible density effects and lag effects of density. 



We did not expect a regression including all years to 

 result in a significant relationship, and it did not (R 2 = 

 0.195, F = 1.33, P = 0.29). Although fawn survival generally 

 followed forage conditions in most years, some notable 

 exceptions (1974-1975) were apparent. Low survival of fawns 

 during the early to mid-1970s, despite apparently good forage 

 conditions, was one factor leading to initiation of the 

 intensive phase of this study. Because of this, we used a 

 "step-up" multiple regression technique. The regression was 

 initiated with those years which most appeared to fit a 

 forage-fawn survival relationship. Additional years were 

 added to the regression, retained if they significantly fit 

 the regression and dropped if they did not. About 80% of the 

 variation in fawn: 100 female ratios in December was explained 

 by the independent variables of the regression for 21 of 27 

 years, 1960-1986 (R 2 = 0.805, F = 37.25, P < 0.00001). Forage 

 production variables of July-April precipitation and mean May 

 temperature were the only variables that were significant in 

 predicting fawn survival to December [ Y (fawns: 100 females) 

 = 437.5 - 7.39 (May temp.) + 3.15 (July-Apr. ppt . ) ]. Winter 

 severity and deer density did not add significantly to the 

 overall regression over any combination of years. 



We also compared fawn survival directly with the forb and 

 shrub production observed and estimated by regression for the 

 measured plots. A simple linear regression of fawn survival 

 against forage production for the same 21 years included in 

 the multiple regression indicated a significant relationship 

 (r=0.873, P < 0.01). When an index for available forage per 

 capita (kg forage/ha/number of adult deer 1 June) was 

 regressed against fawn survival, the relationship was 

 significant (r=0.627, P < 0.01), but less than that for 

 non-density adjusted forage production. 



Data on relative fawn survival through spring (fawn: adult 

 ratios) were available for 15 years. Fawn survival through 

 spring was significantly (R 2 =0.659, F=11.57, P=0.002) related 

 to the same variables (July-April precipitation and mean May 



132 



