Premium
Allometry and prediction in Hominoids: A solution to the problem of intervening variables
Author(s) -
HartwigScherer Sigrid,
Martin Robert D.
Publication year - 1992
Publication title -
american journal of physical anthropology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.146
H-Index - 119
eISSN - 1096-8644
pISSN - 0002-9483
DOI - 10.1002/ajpa.1330880105
Subject(s) - allometry , interspecific competition , biology , statistics , estimator , body weight , mathematics , multidimensional scaling , ecology , zoology , endocrinology
To avoid misinterpretation of allometric exponents determined from interspecific allometric comparisons, specific conditions must be met with respect to the common reference variable. Body weight is considered to be the best general indication of overall size and is hence widely acknowledged to be the most suitable reference variable. However, because of the paucity of recorded body weights for museum specimens, various comparative studies have used other size indicators as intervening variables, although the allometric relationships to body size/weight were often unknown and possibly differed between species. Because of differences in the scaling properties of alternative intervening variables across the species investigated, conflicting conclusions may be drawn if different variables are chosen as substitutes for overall size. This is illustrated with two examples. In this study, series of skeletons with associated body weights of Gorilla , Pan , Pongo , and Homo were investigated. Both ontogenetic and static adult allometric relationships between several widely used reference variables and body weight were determined. Neither these variables nor additional estimators investigated in this study displayed allometric exponents and coefficients similar enough across species to justify direct interspecific comparison. To generate an alternative size estimator for both ontogenetic and static interspecific investigations, equations for combined sexes were derived to predict body weight from various long bone dimensions for individual hominoid species. From a total of 25 predictors, 12 prediction equations per species (six for nonadults and six for adults) were selected according to their relative suitability for reliable prediction of body weight. It is shown that the derived reference variable “predicted body weight” avoids problems of intervening variables, is valid for any interspecific ontogenetic and static allometric comparison, and displays less fluctuation in comparison to actual body weight. © 1992 Wiley‐Liss, Inc.