Sex expression was examined in two island populations of the monoecious Taxus canadensis at the Apostle Island National Lakeshore, USA. Long-term observations of tagged plants supported earlier published conclusions based on comparisons among populations of the effect of size and age on sex expression, measured as standardized phenotypic gender. As populations of plants increased in size and age, strobilus ratios became more male-biased and gender concordance increased. Male-biased plants remained male-biased and female-biased plants remained female-biased; gender switching was not a feature of either of these populations.
The two Taxus canadensis populations were mapped to examine spatial dispersion of plants and plant attributes, such as size, strobilus production, and gender. Frequency distribution of gender in both populations was significantly bimodal supporting earlier studies; this conclusion held even when small (typically male) plants with unstable gender were excluded from the analysis. Distribution of T. canadensis in both populations was clumped. Plant attributes showed some spatial autocorrelation, although the results were not consistent across populations or at different lag distances. Size and male strobilus production tended toward spatial autocorrelation in both populations, female strobilus production showed significant spatial autocorrelation on Manitou Island only at a few lag distances, and gender showed no spatial autocorrelation at any lag distance in either island population. The implication of these results for genetic versus environmental determination of sex expression in T. canadensis are briefly discussed.