RESUMEN
Delayed flowering due to later snowmelt and colder temperatures at higher elevations in the alpine are expected to lead to flowering phenological adjustment to prevent decoupling of peak flowering from the warmest time of the year, thereby favoring pollination. However, even if flowering is brought forward in the season at higher elevations, an elevational temperature gap is likely to remain between the high- and low-elevation populations of a species at the time these reach peak flowering on account of the atmospheric reduction in temperature with increasing elevation. The negative effect of this temperature gap on pollination could be compensated by plastically-prolonged flower life spans at higher elevations, increasing the probability of pollination. In a tightly temperature-controlled study, the flowering phenology adjustment and flower longevity compensation hypotheses were investigated in an alpine species in the Andes of central Chile. The snow free period varied from 7 to 8.2 months over 810 m elevation. Temperatures were suitable for growth on 82-98% of the snow free days. Flowering onset was temporally displaced at the rate of 4.6 d per 100 m increase in elevation and flowering was more synchronous at higher elevations. Flowering phenology was adjusted over elevation. The latter was manifest in thermal sums tending to decrease with elevation for population flowering onset, 50% flowering, and peak flowering when the lower thermal limit for growth (TBASE) was held constant over elevation. For TBASE graded over elevation so as to reflect the growing season temperature decline, thermal sums did not vary with elevation, opening the door to a possible elevational decline in the thermal temperature threshold for growth. Potential flower longevity was reduced by passive warming and was more prolonged in natural populations when temperatures were lower, indicating a plastic trait. Pollination rates, as evaluated with the Relative Pollination Rate index (RPR), when weighted for differences in floral abundance over the flowering season, declined with elevation as did fruit set. Contrary to expectation, the life-spans of flowers at higher elevations were not more prolonged and failed to compensate for the elevational decrease in pollination rates. Although strong evidence for phenological adjustment was forthcoming, flower longevity compensation did not occur over Oxalis squamata´s elevational range. Thus, flower longevity compensation is not applicable in all alpine species. Comparison with work conducted several decades ago on the same species in the same area provides valuable clues regarding the effects of climate change on flowering phenology and fitness in the central Chilean alpine where temperatures have been increasing and winter snow accumulation has been declining.