Survivorship 



Arabis fecunda mortality rates were more variable and 

 generally higher at Birch Creek and Vipond Park (Fig. 3). 

 Temporal patterns of mortality were more similar between Charleys 

 Gulch and Lime Gulch and between Birch Creek and Vipond Park. 



Survivorship of the uneven-age 1987 cohort declined 

 significantly more precipitously at Birch Creek than at Charleys 

 Gulch (log-rank test; LR=19.3, P<0.001), and most of this 

 difference occurred between 1987 and 1988 (Fig. 4a) . Decline in 

 survivorship of the uneven-age 1989 cohort was greatest at Vipond 

 Park and least at Charleys Gulch (Fig. 4b) . Survivorship curves 

 from Charleys Gulch and Lime Gulch were not significantly 

 different (LR=1.4, P=0.23), but the Vipond Park survivorship 

 curve was different from both Lime Gulch (LR=10.2, P=0.001) and 

 Charleys Gulch (LR=9.5, P-0.001). At Charleys Gulch and Lime 

 Gulch the probability of death was nearly constant, while at 

 Birch Creek and Vipond Park younger plants had a higher chance of 

 dying. 



Recruitment 



Recruitment of Arabis fecunda varied among sites and years 

 but was generally higher and more variable at Birch Creek and 

 Vipond Park (Fig. 5) . The best years for recruitment were not 

 the same among the four sites. Recruitment was high at Charleys 

 Gulch and Lime Gulch in 1991 but not at Birch Creek or Vipond 

 Park. 1988 was a strong recruitment year at Birch Creek but not 

 Charleys Gulch, and 1992 was the highest recruitment year at 

 Vipond Park but was lowest for both Charleys Gulch and Lime 

 Gulch. Recruitment was high at all sites in 1990 (Fig_5.). 

 There was a strong negative correlation between mortality rate 

 and recruitment rate among the four sites over the six years of 

 the study (N=15, r=.63, P=0.01). ' 



Fecundity 



The percent of each Arabis fecunda populations that flowered 

 and produced fruit varied among years (Fig. 6) . On years of high 

 recruitment a larger percentage of plants are new recruits and 

 less likely to flower. Thus, the percentage of plants flowering 

 is expected to be lower in years of high recruitment; however, 

 this does not appear to be the case (compare Fig. 5 and 6) . 

 Percent reproductive plants was at least 15% at all sites in all 

 years except there were few reproductives at Birch Creek and 

 Charleys Gulch in 1988 and at Lime Gulch in 1989 (Fig. 6) . 

 Percent reproductives was highest in 1990 at all sites except 

 Birch Creek. 



The number of fruits per reproductive A^ fecunda plant 

 varied greatly among sites and years (Fig. 7) . Site was a 



