Technical Abstract: Yields and berry weights of two widely grown, commercial, highbush blueberry cultivars, ‘Bluecrop’ and ‘Duke’, were evaluated for 3 or more harvests every season over ten years, and seed set was determined at each harvest over the last four. Across 10 years, yield and berry weight had no significant correlation. Under our study conditions, the limitations of yield for ‘Bluecrop’ did not include pollination. In fact, pollination (expressed as s/gfw) had an inverse relationship to yield. For the years in which replicated seed counts were made, we found, in general, that ‘Bluecrop’ berry weight and seed/fruit decreased linearly between Harvest 1 and Harvest 3, and berries with similar seed numbers varied in weight as much 39% between years. Hence, reduced berry weight was not necessarily due to poor pollination. For ‘Duke’, trends of berry weight and seed/fruit across harvests were inconsistent, but ‘Duke’ appeared to have sufficient environmental plasticity to produce fruit with increased berry weight between successive harvests given favorable inputs. For ‘Duke’, berries with similar seed numbers varied in weight as much as 86% between years. A cyclical yield variation was seen in ‘Duke’ that generally followed a high-medium-low pattern.