STATISTICAL DIFFICULTIES IN THE ANALYSIS OF PREDATOR FUNCTIONAL RESPONSE DATA

Abstract

The utility of the ‘random predator equation’ of Royama (1971) and Rogers (1972) as a summary of the functional response relationship between predatory feeding behavior and prey density is questioned on the grounds that statistical assumptions in the regression analysis are not met by the linearized form of the equation. The absence of an alternative linearization that does not violate these assumptions necessitates the use of Holling's (1959) disc equation for the description of experimentally derived functional response relationships, when the comparison of parameters of different populations of predator or prey is a more important objective than a precise estimation.The statistical validity of the traditional linearization of the disc equation is questioned. An alternative transformation is proposed, which removes the statistical problems associated with the former transformation, and which permits a higher degree of explanation of variance in the independent variable by the regression.