1. Anuran (frog and toad) jumping power varies greatly across species, yet muscle power does not. Given that the jumping power of some species is up to five times higher than typical muscle power, power amplification by elastic elements is suggested to explain this discrepancy. However, the ecological reasons for this variation in jumping power are unclear. One hypothesis is that small jumpers are limited by the time available to accelerate their body during take-off, leading to small species needing greater power production than larger species to achieve similar jumping performance. Another (non-mutually exclusive) hypothesis is that the microhabitat species inhabit may drive variation through trade-offs with performance in microhabitat-specific, non-jumping behaviors. 

2. We compared jumping power across 68 anuran species that were diverse in evolutionary relationships, microhabitat use, and body mass. We used phylogenetic comparative methods to compare the role of microhabitat and body mass in explaining variation in jumping power across species. 

3. We found the strongest support for a model that included two factors and their interaction. First, as body mass increased, jumping power decreased. Second, species that burrowed showed lower jumping power than species that did not burrow. Third, the interaction between body mass and burrowing behavior showed that jumping power declines more rapidly with body mass in burrowing species than non-burrowing species. 

4. The effect of body mass suggests that interspecific variation in jumping power might be partly explained by scaling relationships. Anurans with small body mass may be able to achieve similar locomotor performance (e.g. takeoff velocity) as those with larger body mass, by more effectively amplifying muscle power. Additionally, the effect of burrowing behavior suggests that species that use hindlimbs to burrow may experience a reduction in their ability to generate jumping power. This may indicate a functional trade-off between jumping and burrowing performance.

Frogs jumps were recorded with a high-speed video camera at 500 frames per second. Each individual's video with their best effort was digitized with ImageJ. We digitized from the beginning of movement until three frames post take-off. We took the derivate of the displacement profile to get velocity and the derivate of the velocity profile to get acceleration. Body-mass-specific power was calculated by multiplying the velocity profile by the acceleration profile. Muscle-mass-specific power was calculated by dividing the body-mass-specific power by the muscle-mass to body-mass ratio.