ZELDITCH, M.L.*; SHEETS, H.D.; FINK, W.L.: Conservative ontogenetic trajectories: artifacts or data?

Most comparative studies support the hypothesis that ontogenetic trajectoriesare historically conservative. The primary basis for this generalization is the typically high correlation between ontogenetic allometric vectors, a measure of their similarity. More than 76% of 233 comparisons yield vector correlations as high as 0.96. However, these correlations are misleading if taken to mean that ontogenetic allometries are highly similar, according to several lines of evidence. Among these are the high correlations obtained from biologically absurd comparisons, such as between allometric coefficients of a chimpanzee (based on cranial measurements) and a lizard (based on combined cranial and postcranial measurements); this comparison yields a correlation 0.87. Also, correlations between randomly shuffled coefficients and those based on actual data should yield correlations near 0.0 but instead yield correlations averaging 0.98 or higher. And correlations as high as 0.99 can be obtained by comparing allometric growth vectors to vectors of isometric growth, even when growth is far from isometric. We suspect that correlations between ontogenetic allometric vectors may be high merely because most taxa enlarge as they grow. When enlargement is removed from consideration, using geometric morphometrics, biologically absurd comparisons yield correlations near 0.0, as they should. And taxa that seem highly conservative in their allometries, based on traditional measurements, reveal considerable repatterning of relative growth rates.