Local and regional species richness in north American lacustrine fish

1. Species-area and local-regional species richness patterns in lacustrine fish in North America were investigated by reduced major axis regression, using a database of 651 lakes. 2. Small lakes (area < 1.5 km(2)) had steeper species-area slopes than large lakes. In one area open water species declined in frequency in small lakes, suggesting that within-lake habitat diversity is partly responsible for the species-area slope. Local lake characteristics (area, mean depth and altitude) explained only about one-third of the variation in species numbers. 3. Species-area slopes were not significantly different across regions but there were considerable differences between regions in local richness for standardized lake sizes. Local richness was linearly related to regional richness when the regional pool consisted only of native species. 4. Pseudosaturation, the detection of curvilinearity when the true relation is linear, was found when introduced species were included in the regional species pool. Introduced species tend to have more localized distributions than native species, and their inclusion over-estimated regional pool sizes but had little effect on local richness estimates. Additionally, undersampling of small species, coupled with an increase in the relative importance of small species in southern regions, led to pseudosaturation by underestimating local richness in species-rich regions. 5. Turnover diversity was not correlated with regional pool size. However, comparison of three eastern regions showed that the percentage of the region occupied per species was less in the more speciose, southern regions, consistent with increased turnover diversity and pseudosaturation. 6. These results support the notion that regional-scale processes play a major role in determining the number of species in North American lacustrine fish communities.