JONATHAN ROUGHGARDEN (Department of Biology, Stanford University, 
Stanford, California 94305) has utilized Anolis species to test theoretical 
measures of niche width. He distinguishes within phenotype (WPC) and 
between phenotype (BPC) components of niche width, the measure of the 
second being the variance of the frequency distributions of the different 
resource utilization phenotypes and of the first, the average variance of 
individual phenotypes' utilization functions. Total niche width is the 
variance of the population's resource utilization functions and is the sum 
of WPC and BPC. 
Roughgarden, operating with samples collected under GB 37731X, finds 
the within phenotype component of niche width to be strongest in Anolis. 
Anolis ferreus, as a solitary anole unique in its large size and hence 
perhaps in its size variation, was chosen for special study as possibly 
showing the upper bound for the between phenotype component in niche width. 
Calculation was based on total sample of 44 adult male lizards. The 
resource axis was prey size and data taken were jaw size and length and 
weight of insects from the stomachs of the measured lizards. The per cent 
composition of total niche width so determined was 32.5% BPC and 67,5% WPC. 
If A. ferreus represents, indeed, the upper bound for BPC in Anolis (poly- 
morphism), then this component is never the primary one in Anolis niche 
width and Anolis is characteristically monomorphic and generalist in strategy. 
Roughgarden makes also a number of predictions which he tests for 
various Anolis communities: (1) A population has a smaller niche width and 
is less polymorphic (has less BPC) as environmental productivity decreases 
or the number of competing species decreases. (2) If both productivity and 
the number of competing species decrease, there is little change in total 
niche width but BPC decreases. (3) Change in variance (e.g. in jaw length) 
is slower than displacement of the average phenotype (e.g. mean jaw length). 
Hence jaw size variance patterns should have a strong lineage component, 
whereas patterns in average jaw size will not. Ecological events in the 
relatively remote past should be more important in explaining jaw size 
variance than average jaw size, 
Roughgarden uses an indirect approach in testing these predictions. 
The coefficient of variation (CV) in lizard jaw size is taken as an index 
of BPC. Then CV should be explained by a function whose value strongly 
increases with productivity, weakly decreases with species diversity and 
involves a lineage component, 
His data for 24 Anolis populations do show (1) a strong productivity 
effect, (2) a weak species diversity effect, (3) lineage components are 
necessary to explain the pattern of jaw size CV but (4) not necessary to 
explain the pattern of jaw sizes. 
