We gathered lizard population density data from the literature, expanding from Rodda et al. (2001) (see Appendix D for literature sources). Mean density was used when a single author reported multiple density estimates that were separated temporally for the same location or by less than 0.5 degrees latitude/longitude. The observations were geo-referenced using text description of study locations (www.biogeomancer.org). The database was filtered to remove islands (defined as dispersal limited and with areas less than 9000 km2) and to restrict the number of observations per species to 10 to avoid bias. The majority of studies reported number of individuals per unit area.
While the studies vary in methodology and quality, there is little evidence of systematic biases in density estimates. The estimation method varies and includes mark-recapture studies (35% of studies), continuous observation of marked individuals in life-history studies (25%), censusing of quadrats (20%), and strip transects (9%). Sampling method does significantly, but weakly, account for residuals in the density-body mass relationship (F[4,470] = 16, P < 10-11, r2 = 0.12). Transect sampling, which account for just 9% of studies, has the strongest effect by depressing densities. However, there is no interaction between body size and sampling method in accounting for densities. This suggests that combining different sampling methods in the analysis may increase scatter but does not bias results. Sampling method is not a significant factor in the relationship between energy use (Bx) and population density. The slopes of the relationships between population density and body mass, energy use (Bx), and energy availability (Pmin) within data collected using a single sampling method are similar to those reported in the text, which were derived by combining sampling methods. Sample areas approximately ranged from 100m linear transects to 500,000 m2 (which has the potential to bias abundance estimates, Gaston et al. 1999). Studies conducted in non-representative habitats (e.g., particular rock piles) were excluded. Sampling was worldwide, though geographically biased (North America: 148, Central America and the Caribbean: 67, South America: 25, Europe: 28, Africa: 49, Asia: 36, and Australia and Pacific: 130). Five species were represented by more than 10 populations and 128 species were represented by single populations. Data from the literature was used to assign diet (references including Pough 1973, references including Cooper and Vitt 2002, Perry and Garland 2002, Metzger and Herrel 2005).
To estimate energy availability, DOLY predictions were only available as annual averages. We thus used satellite-based average monthly NDVI (normalized difference vegetation index, a measure of the photosynthetic activity of vegetation) data to seasonally scale the DOLY productivity predictions. This approach assumes within-site comparability of NDVI as a correlate of productivity across seasons. We used monthly NOAA/NASA Pathfinder AVHRR based maximum value composite NDVI images averaged across 18 years between 1981 and 2000 and provided by Clarks Labs in 0.1° resolution (Anonymous 2001). For each 0.1° pixel we averaged monthly NDVI values across four three-month periods (DecFeb, MarMay, JunAug, SepNov) and identified the period with the lowest average NDVI value. We then calculated proportional seasonal NDVI for this minimum three-month period by dividing its NDVI by the summed NDVI across all four periods (i.e., in an aseasonal environment the proportional seasonal NDVI would be 1/4, in a seasonal environment < 1/4). Subsequently, we averaged this value across the 0.5° grid in which the DOLY predictions were available. An estimate of minimum NPP (Pmin) of a 0.5° grid cell was then given by the product of DOLY total annual NPP predictions and the respective proportional minimum three-month NDVI.LITERATURE CITED
Anonymous. 2001. Volume 4: 0.1 Degree Global Monthly Vegetation Index (NDVI) 1981-2000. Clark Labs, Worcester, Maine, USA.
Cooper, W. E., and L. J. Vitt. 2002. Distribution, extent, and evolution of plant consumption by lizards. Journal Zoological Society of London 257:487517.
Gaston, K. J., T. M. Blackburn, and R. D. Gregory. 1999. Does variation in census area confound density comparisons? Journal of Applied Ecology 36:191204.
Metzger, K. A., and A. Herrel. 2005. Correlations between lizard cranial shape and diet: a quantitative, phylogenetically informed analysis. Biological Journal of the Linnean Society 86:433466.
Perry, G., and T. Garland. 2002. Lizard home ranges revisited: Effects of sex, body size, diet, habitat, and phylogeny. Ecology 83:18701885.
Pough, F. H. 1973. Lizard energetics and diet. Ecology 54:837844.
Rodda, G. H., G. Perry, R. J. Rondeau, and J. Lazell. 2001. The densest terrestrial vertebrate. Journal of Tropical Ecology 17:331338.