FLAWED INFERENCE - WHY SIZE-BASED TESTS OF HETEROCHRONIC PROCESSES DONOT WORK

Allometric (or size-based)heterochrony is widely used:in drawing infer ences about heterochronic processes. Whereas problems associated with substituting size for age in heterochronic analysis are acknowledged, it is also generally believed that comparisons of ancestral and descen dant growth allometries often reveal real processual shifts in growth processes. At the very least, they are presumed to discriminate system ic perturbations in ''size'' and ''shape'' (''ontogenetic scaling'') f rom local dissociations of ''size'' and ''shape'' (''neoteny,'' ''acce leration,'' and so on). When ancestor and descendant follow the same g rowth trajectory in size (relative to age), it is further presumed tha t size-based heterochronies will match ''true'' (age-dependent) hetero chronies. In this paper we argue that growth allometries are a poor ve hicle for inferring heterochronic process; the actual processual shift s that produce them can be counter-intuitive. Systemic shifts can resu lt in dissociated allometries, and overlapping growth allometries can conceal perturbations in ''shape'' relative to ''size''. Changes in gr owth allometries are influenced as much by the initial ancestral relat ionship between ''size'' and ''shape'' as they are by any perturbation s that may occur during the evolution of the descendant from the ances tor. Finally, it is emphatically not the case that size-based heteroch rony will reveal true heterochrony as long as the size trajectories of ancestor and descendant remain constant relative to age. We urge care ful consideration of the actual shapes of biological growth trajectori es, recognition of the fact that they arl rarely simple power function s of age, and calculation of possible relationships between complex an d/or lazy-S growth trajectories.