Seedling establishment in alpine buttercups under experimental manipulations of growing-season length

We tested effects of growing-season length and site fertility on early seed ling growth and seedling establishment in a snowbed population of Ranunculu s adoneus. On average, growing-season length decreases by 30% from shallow to deeper parts of the snowbed. Past work has shown that plant cover and so il fertility decline with delayed snowmelt date along this steep habitat gr adient. We experimentally uncoupled the relationship between snowmelt sched ule and site fertility by adding snow to plots at naturally shallow snow de pths and removing snow prematurely from plots with naturally deeper snow in the summers of 1991-1994. Seeds of R. adoneus were transplanted into manip ulated plots and adjacent control plots in 1990 and 1991. Seedling shoot si ze and survival were monitored over the course of the experiment. Productio n of true (adult) leaves by newly emerging seedlings was significantly redu ced by the delayed snowmelt schedule. Seedlings failing to produce true lea ves during their first summer were at a strong survival disadvantage over t he winter months. Leaf length for seedlings that produced true leaves was n ot affected by snowmelt or location on the snow depth gradient; nor did lea f length affect overwinter survival. We defined seedling establishment as the successful transition from seed to seedling of age two years. At this age, all seedlings have produced true l eaves. Effects of growing-season length on seedling establishment rate were pronounced in inner portions of the snowbed. There, average establishment success increased from <1% under natural conditions to 5% with experimental extension of the growing season. In contrast, establishment rate in plots near the snowbed edge tended to increase when snowmelt was experimentally d elayed. Possibly, delaying snowmelt in relatively fertile outer plots enhan ced establishment by synchronizing seed germination with warmer mid-season soil temperatures in the snowbed environment. Regrowth of other plant speci es into planting microsites occurred more extensively in early snowmelt plo ts than in late snowmelt plots but had little effect on seedling establishm ent. Pooled across the entire cohort of germinating seedlings, emergence, s urvival over the first winter, and survival during the second growing seaso n contributed significantly to variation in establishment rate (R-2 = 37% 7 %, and 32%, respectively) but were poorly intercorrelated (r = -0.15). Thes e findings show that sites favorable at any one stage in establishment do n ot remain favorable at other stages. We conclude that the relationship between growing-season length and seedlin g establishment success in snow buttercups depends on other aspects of micr osite quality. In nutrient-poor habitats, short growing seasons play a majo r role in restricting establishment. However, in more fertile sites, the ti ming of snowmelt and its associated effects on the quality of the microclim ate for seed germination may be more critical than the period of time from snowmelt until the onset of winter. For snow buttercup seedlings much of es tablishment is stochastic, in that microsites conducive to emergence do not ensure survival over subsequent phases of the life cycle.