The relationship between body condition, body composition, and growth in amphibians

Body condition of animals is often assumed to reflect advantages in survival or reproduction, but body condition indices may not reflect body composition, or condition may be unrelated to fitness-associated traits. The relationship between body condition indices and composition has rarely been quantified in amphibians, and body condition has not previously been related to growth in adult amphibians. We used laboratory (quantitative magnetic resonance) and field methods to evaluate the relationship between body composition and the four common body condition indices for wildlife studies (body mass index, Fulton's index, scaled mass index, and residual index) in two frog and one salamander species in Montana, USA. We then assessed the relationship between body condition and summertime somatic growth during a 3-yr mark-recapture study of one of our study species (Columbia spotted frogs, Rana luteiventris). Correlation of body condition indices with fat and lean mass differed across species, sexes, and whether components were represented as percentages or were scaled based on size. Scaled mass index, residual index, and Fulton's index were most often well correlated (r > 0.6) with scaled body components, but Fulton's index was strongly correlated with body length. Scaled mass and residual indices predicted scaled fat relatively well and were uncorrelated with body length. Heavier condition predicted higher growth rates of Columbia spotted frogs, regardless of the index used. Frogs of heavy body condition (90th percentile residual index) grew 0.04 and 0.05 mm/day greater than frogs of light condition (10th percentile) for average length males and females, respectively. Frogs of short body length (10th percentile) grew 0.11 and 0.19 mm/day more than long (90th percentile) males and females, respectively. By examining the relationship between body condition indices and body composition and revealing a link between condition and future growth, our results provide an empirical basis for choosing the most appropriate condition index, as well as a potential link to fitness-related traits.