Traditional sampling methods have limited our understanding of the behavioral ecology of leaf litter fauna. Substrate-borne vibrations provide a window that addresses this shortcoming. Under natural conditions, we implemented a passive monitoring system that measures vibrations produced by two miniaturized Puerto Rican geckos: Sphaerodactylus grandisquamis in the mesic karst forest and Sphaerodactylus townsendi in the coastal dry forest. The dichotomy between habitats allowed us to test the hypothesis that microclimate conditions drive divergence in thermoregulatory behavior. For each species in its native habitat, thirty individuals were monitored inside a field enclosure, excluding other vibration sources. We measured vibrations at two strata depths (i.e., surface and bottom) to characterize behavioral responses to diel cycle, microclimate, and predatory bird calls. Vibrational data revealed contrasting activity patterns across strata. Mesic S. grandisquamis showed continuous daytime activity at all strata levels, regardless of microclimate. Dry forest S. townsendi featured a biphasic activity pattern, peaking midday and at dusk, and increased activity in deeper litter strata during hot and dry conditions (i.e., midday). Analysis at fine temporal scales for S. grandisquamis revealed a reduction in surface activity after the onset of bird calls from three species that pose a predation risk. Our results demonstrate interspecific differences in thermoregulatory behavior that likely increase the fitness of each species in their respective habitats. Furthermore, vibrational monitoring strongly suggests Sphaerodactylus exploit their niche in a three-dimensional manner, mirroring habitat use patterns of arboreal fauna and highlighting the importance of incorporating multiple strata levels when studying leaf litter microhabitats.

The vibrational dataset represents field recordings for 1) Sphaerodactylus grandisquamis in the mesic karst forest at Mata de Plátano and 2) Sphaerodactylus townsendi in the coastal dry forest at Cabezas de San Juan, along with concurrent microclimate data from the leaf litter. Processing of vibrational data was conducted in MATLAB, and analyses were performed in R. All data and scripts are provided.

Funding provided by: University of MissouriROR ID: https://ror.org/02ymw8z06Award Number: Funding provided by: Animal Behavior SocietyROR ID: https://ror.org/031nh9x49Award Number: George W. Barlow Award