How diverse animal communication signals have arisen is a question that has fascinated many. Xenopus frogs have been a model system used for three decades to reveal insights into the neuroendocrine mechanisms and evolution of vocal diversity. Due to the ease of studying central nervous system control of the laryngeal muscles in vitro, Xenopus has helped us understand how variation in communication signals between sexes and between species is produced at the molecular, cellular, and systems levels. Yet, it is becoming easier to make similar advances in non-model organisms. Here, we summarize our research on a group of frog species that have evolved a novel hind limb signal known as ‘foot flagging.’ We have shown that the evolution of foot flagging in multiple species is accompanied by the evolution of higher androgen hormone sensitivity in the leg muscles and an increased density of spinal interneurons in the neuromotor system that controls the hind limb. Comparing this work to prior work in Xenopus, we highlight which patterns of hormone sensitivity and neural circuit properties are shared between Xenopus and foot-flagging frogs and which appear to be species-specific. Overall, we aim to illustrate the power of drawing inspiration from experiments in model organisms, in which the mechanistic details have been worked out, and then apply these ideas to a non-traditional model species to reveal new details, further complexities, and fresh hypotheses.

Data on androgen receptor (AR) expression was generated using radioactive in situ hybridization. See manuscript supplemental files for complete methods. Data were collected from photomicrographs. Tissue slides were developed in photographic emulsion and silver grains were developed over cells that contained androgen receptor mRNA. AR expression data were collected from photomicrographs taken on a light microscope. Tissue sections were counterstained with Nissl stain so we could also see the cells. We quantified the number and area of developed silver grains in each image, then subtracted the number of background silver grains outside for the target area, and divided by the total number of identified neurons in the image/field of view. Thus, AR expression data can be expressed as the total area covered by AR grains, percent area covered by AR grains, number of grains above the background, and number of grains above the background per cell. Additionally, we can average the values for each frog and compute a ratio of AR grains for the lumbar vs. brachial spinal cord segments. Data on neuron cell numbers (motoneurons, interneurons) was generated using ImageJ. A trained observer manually classified and counted all cells in the photomicrographs.