Our previous work demonstrated a clear link between the prior history of exposure to paralytic shellfish toxins (PSTs) and toxin resistance in softshell clam, Mya arenaria, populations. Resistance is caused by selection for a naturally occurring point mutation in the sodium channel gene which greatly reduces the binding affinity of PSTs. The present study attempts to identify the life history stage at which natural selection is most likely to operate and the conditions of toxin exposure necessary to cause a shift in the genetic composition of a population from sensitive to resistant. Progeny from a predominantly sensitive population (Lawrencetown River Estuary, LE) and a predominantly resistant one (Bay of Fundy, BF), were compared in laboratory experiments. Growth and survival of 1 week-old larvae exposed to a high-toxicity Alexandrium tamarense isolate (PR18b) in combination with non-toxic algae were not adversely affected in either population compared to non-toxic controls. Thus selection for resistance is not expected at these stages, as A. tamarense cells are too large (35 Âµm ESD) to be ingested. In contrast, juveniles (4-5 mm) exposed to a unialgal suspension of A. tamarense (100 cells ml-1) showed significantly higher mortalities and lower toxin uptake rates in LE than BF clams. LE and BF juveniles suffered 95% and 49% mortalities respectively after 1 week of toxification. Sensitive clams exhibited paralysis of the foot and showed evidence of anoxic patches in the pallial cavity presumably due to reduced irrigation. BF clams accumulated 2.4x more toxins during the first 48h of exposure than LE clams. A comparison of burrowing capacity as an indicator of toxin sensitivity before and after toxin exposure showed an increase in the proportion of resistant individuals in both populations. These results were confirmed by genotypic analysis that showed an increase in the prevalence of resistant genotypes following 1 week of toxin exposure (from 10% to 47% in LE clams). In nature clams can be exposed to mixed toxic and non-toxic phytoplankton assemblages. Therefore a second experiment tested the effects of varying dietary proportions (10%, 50% and 100%) of isolate PR18b on juveniles (4-5 mm) from the sensitive population over a 2-week period. Highest mortalities (50%) were found in the toxic unialgal treatment. Mortalities attained only 15% in the 50/50 mixed diet and were negligible in both the non-toxic control and 10% toxic treatments. Clam toxicities peaked at 13.3x103, 5.3x103 and 1.5x103 Âµg STXeq 100g-1 in the 100%, 50% and 10% treatments respectively. Growth rates were significantly reduced with increasing proportion of A. tamarense in the diet relative to controls, with an 88% and 65% reduction in tissue dry weight in the 100% and 50/50 toxic diets respectively. Growth inhibition was not significant in the 10% toxic treatment, yet resulted in significant reduction in motility (crawling) of juveniles. Our study thus demonstrates that a highly toxic A. tamarense bloom can cause rapid lethal and sublethal effects in M. arenaria postsettlement stages and result in strong selection for resistance within a few days of exposure to PSTs.