AbstractConspicuously absent from plant allometric models are insect herbivores, despite their ubiquity in pantropical forests as well as the importance of gauging their impact on plant growth for understanding tree regeneration dynamics. Conceivably, the scaling relationships of plant populations with and without exposure to insects could be compared to gain insight, but this has yet to be tried. Here, I examined the size allometry of three canopy‐statured species (Fabaceae) by fitting standard major axis regressions of stem height, number of leaves, and maximum area per leaf against stem diameter (root collar) of their juveniles (seedlings/saplings) in 41 canopy gaps after ca. 2 years of herbivore exclusion (using mesh cages with mesh‐rooftop controls), in a central African forest. Herbivores did not change the slope or intercept, nor shift the location of Tetraberlinia bifoliolata and T. korupensis populations, whose seedlings tolerate shade well. In stark contrast, all three size allometries were altered in the faster‐growing Microberlinia bisulcata, whose seedlings cannot tolerate shade indefinitely. More pronounced than either an altered slope (scaling exponent) or intercept was the strong rightward shift in location (to larger sizes) of protected M. bisulcata, whereas its normal (control) population was left stunted, with fewer and smaller leaves. Nevertheless, all species had higher than theoretically predicted scaling exponents, even in the absence of herbivores, suggesting a convergent response early in tree ontogeny across the shade‐tolerance spectrum prioritizing height gain over mechanical stability in forest gaps. But insect herbivores were able to significantly strengthen this deviation in M. bisulcata, and perhaps in T. korupensis. These results demonstrate an allometric approach may prove useful for revealing how insects differentially impact trees' growth and form (slenderness) when they are most vulnerable to herbivory.