RAMET DEMOGRAPHY OF ALLIUM-TRICOCCUM, A SPRING EPHEMERAL, PERENNIAL FOREST HERB

1. Allium tricoccum is a long-lived perennial, harvested for its edible bulbs. This study provides precise information on vital rates in vulnerable populations at the northern limits of its distribution in southern Quebec. 2. A dense population was studied for 5 years, by mapping 18 1-m x 1-m plots and measuring all ramets to estimate annual mortality, recruitment, growth, sexual and asexual reproduction rates. Ramets were classified both by their size and flowering status. A stage-classified projection matrix model was used to reveal population trends. Elasticity analyses and simulations were performed to identify the demographic variables most relevant to population growth and to estimate the effect of harvests. 3. All vital rates estimated were closely size-dependent, and differed greatly between sterile and flowering ramets. The bimodal population structure was dominated by large sterile ramets showing high survival rates and maintained through vegetative recruitment. Seedlings emerged every year, but largely failed to establish. 4. Most flower scapes died before setting seed. Flowering in consecutive years was rare. Flowering ramets showed high rates of asexual reproduction. At the population level, reproductive patterns were clearly biennial; resource availability (critical size and reproductive cost) and the regulating effect of season length, appear to determine reproductive synchrony. 5. Temporal variability was significant for all demographic variables, but especially for seedling emergence and reproductive behaviour. One exceptionally good growth season was observed during this 5-year study; outstanding flowering, mast seed production and seedling recruitment led to the establishment of new genotypes. 6. Population growth rates (lambda) estimated were close to the equilibrium value of 1.00. Elasticity values associated with sexual reproduction usually contributed to less than 1-5% of lambda. Thus, germination rate had essentially no effect on population growth rate. Clonal development of already-established ramets was the most important pathway to population growth. 7. Simulated harvesting rates of only 5-15% were sufficient to bring the population growth rates below the equilibrium value, predicting a population decline. Bulb harvest represents a threat for population maintenance of these largely clonal northern populations, in which recruitment from seeds is unlikely.