The Investigation of Saturation in the Species Richness of Communities: Some Comments on Methodology

Small‐scale species richness in forest canopy gaps: the role of niche limitation versus the size of the species pool

“…In all cases, we examined the form of the relationship between mean quadrat and gap species richness first, by including both variables alone in a simple regression model and second, after including other factors (such as gap area, age, shape and orientation) in a multiple regression model to statistically remove their potentially confounding effects. Cresswell et al (1995) point out that values for local and regional species richness will not be independent if there are species restricted to only one local area within a region. If uncorrected, this non-independence could bias the outcome of statistical tests and, depending on circumstances, make a significant positive slope more or less likely to be observed.…”

“…Because distinguishing among the alternative hypotheses relies on a test of the slope of the relationship between local and regional species richness, it is important to correct for this source of bias. Unfortunately, we could not apply the method that Cresswell et al (1995) advocate (calculating regional species richness with the local area of interest excluded from the regional data set) because our measure of local species richness is the average of many local quadrats. However, the slope of the relationship between local and regional species richness will be unbiased if the number of species restricted to only one local area within a region is uncorrelated with regional species richness (see Fig.…”

“…However, the slope of the relationship between local and regional species richness will be unbiased if the number of species restricted to only one local area within a region is uncorrelated with regional species richness (see Fig. l a in Cresswell et al 1995). We tested if this was true for our data by summing the number of species found in only one quadrat within each canopy gap and correlating this with the total number of species in each gap (excluding those species found in only one quadrat).…”

“…We would then expect a curvilinear relationship between mean quadrat and gap species richness. We tested for a relationship of this form by testing if the inclusion of a second order polynomial in the regression model resulted in a significant improvement in model fit (Cresswell et al 1995;Cornell & Karlson 1996;Caley & Schluter 1997). In all cases, we examined the form of the relationship between mean quadrat and gap species richness first, by including both variables alone in a simple regression model and second, after including other factors (such as gap area, age, shape and orientation) in a multiple regression model to statistically remove their potentially confounding effects.…”

Are There General Laws in Ecology?

“…or (for parasites and plant-feeding insects) different host-species within one region. Srivastava's second point is that formal statistical tests to distinguish between Type I (linear) and Type I1 (curvilinear, saturating) relationships are far from straightforward: for example local and regional species richness are not statistically independent (Cresswell et al 1995. Zobel 1997, and data on local richness may be pseudoreplicated.…”