Convergent growth regulation in arthropods: biological fact or statistical artifact?

Abstract

Convergent growth regulation, where individuals adjust their growth trajectories to reach a targeted final body size, has been reported for many arthropod taxa. Divergent growth, where larger individuals grow proportionately more than smaller individuals, is seldom observed. Most studies based their conclusions on growth increment analysis: correlation or regression between body size at a particular molt and the increment grown during the next molt. These studies interpreted a negative relationship as evidence for convergent growth regulation, since smaller individuals appeared to grow more during the subsequent molt than larger individuals. Using random data simulations and an analysis of the statistics, I demonstrate that autocorrelation in these statistics generates false evidence for convergent growth, even when divergent growth actually occurred. I suggest model II geometric mean (GM) regression as an alternative method because it does not suffer from these statistical problems. A GM regression reanalysis of two published studies revealed evidence for divergent growth or no growth regulation in cases where the original studies reported convergent growth regulation, suggesting that the reported prevalence of convergent growth may be a statistical artifact.