Which is the best field method for assessing volume and surface area of arboreal termite nests

Arboreal nests of eusocial insects are typically roughly egg- or pear shaped with superficial irregularities, which makes the correct determination of their volume or surface area in the field difficult. Here, we evaluated the accuracy (benefit) of different geometric proxies with varying number of underlying measurements (cost) when assessing the volume and surface area of arboreal termite nests of Microcerotermes indistinctus and Constrictotermes cyphergaster. We compared size measures based on the 3D models with the most common shape models in literature: spheres (1–3 parameters), ellipsoids (3 parameters), cylinders (3 parameters), and a combination of superimposed cylinders with spherical caps (at least 9 parameters). The most complex model returned the most accurate results (low estimation error and low error variability). However, estimation errors did not differ from simpler shape models, based on only three (ellipsoids, spheres using half of the arithmetic mean of height, width and depth as radius) or even a single parameter (spheres using half of depth as radius). Cylindric proxies and spheres based on nest height or width, on the other hand, returned far less accurate results with estimation errors of up to almost 300%. Although nests of M. indistinctus showed a higher variation in shape and more surface irregularities than those of C. cyphergaster, each geometric proxy tested in the present study determined the nest size of both termite species with equal precision. This suggests that each shape model may generate similarly accurate estimates of volume and surface area over a wider range of termite species that build egg-shaped arboreal nests.