The "Niche-Variation" Hypothesis: A Test and Alternatives

It is highly anthropocentric to speak of the width of ecological niches. This and other reasons militate against the utility of the idea. 2. There seems to be no correlation in birds between the variety of foods taken and the variation of bill characters. Hence, the "niche width-variation model" is rejected for complex characters. 3. One or more of the following is likely to cause unusually high variability of a canalized character in a morph of a stable population: (a) immigration from genetically different populations, (b) directional selection affecting the exposure of theretofore hidden genetic variation, (c) mixing individuals from different localities (artifactual variation). 4. It is suggested that a permanent high level of variation in a complex character, even if "adaptive," is improbable because other characters overlap its genetic and epigenetic base and stabilizing selection on these related characters will set limits on its flexibility. 5. It follows from (4) that the fewer the genes influencing a character, or the fewer it shares with other characters, the greater its freedom to vary.