We assessed the role of growing-season length in regulating absolute a
nd relative cover of six coexisting dominant plant species in an alpin
e snowbed habitat. To help explain disparity in species-specific respo
nses to growing-season length, we examined the developmental phenology
and distribution of each species in relation to natural snow depth va
riation. Season length varies from approximate to 50 d on early-meltin
g edges of the snowbed to 35 d in the late-melting center, 100 m away.
By experimentally altering snowpack, we uncoupled the relationship be
tween spatial location and snowmelt schedule in three consecutive year
s, imposing the same early dates of snow release in a ''long growing-s
eason'' treatment and the same late dates of snow release in a ''short
growing-season'' treatment near the edge and center of the snowbed. O
ver the course of the experiment, growing-season length had significan
t effects on absolute and relative cover of the species studied (P < 0
.025 and P < 0.005, respectively), and these effects were similar near
both the edge and center of the snowbed. Yet, only for the snowbed sp
ecialist, Sibbaldia procumbens, were changes in absolute and relative
cover under early and late snowmelt schedules predictable from the spe
cies' distribution along the historical snow depth gradient. S. procum
bens increased in cover under a long growing-season and was more commo
n in historically early-melting portions of the snowbed. Other species
(e.g., Ranunculus adoneus, Artemisia scopulorum) were equally common
in historically early- and late-melting locations within the snowbed,
but showed discordant responses to experimentally imposed changes in s
nowmelt schedule. That the cover of many species under long- vs, short
growing seasons was not predictable from their current distributional
affinities in relation to snowmelt pattern likely reflects the dispar
ity between the rates of processes exerting long-term control on speci
es' abundances (colonization, soil development) and more immediate eff
ects of growing-season length on plant growth. Consistent with this vi
ew, differences in developmental phenology better predicted species-sp
ecific responses to snowmelt schedule than distributional affinities.
Species having leaf expansion schedules that are poorly synchronized w
ith snowmelt typically had similar cover under early vs. late schedule
s of snow release (Geum rossii, Trifolium parryi, and Poa alpina). In
contrast, species in which leaf expansion schedules are synchronized w
ith snowmelt responded positively to early snow release (Ranunculus ad
oneus and Sibbaldia procumbens). We hypothesize that maintaining metab
olic ''readiness'' under snowcover provides a mechanism for monopolizi
ng nutrient flushes and competitor-free intervals at snowmelt, and exp
loiting occasional long intervals for growth in years of little snow a
ccumulation, but incurs a respiratory cost that is manifest as reduced
growth and vegetative cover when snowmelt is delayed. Our results sug
gest that interspecific differences in growth phenology of coexisting
species will promote shifts in snowbed plant communities with climate
change within generations.