July 13, 2016

Fish, robots, and math help scientists understand how early animals crawled onto land

by Megan Cartwright

A robot named MuddyBot and six tiny fish have helped American scientists figure out how the first tetrapod animals moved from oceans onto land about 360 million years ago. According to the study published this month in Science, the first land tetrapods used their powerful tails to push up the slippery, sandy beach slopes that could tumble them back into the water. The study helps scientists puzzle out how early animals moved—literally—from swimming in the ocean to crawling up beaches. The answer involves a lurching, tail-propelled movement that the scientists call “crutching.”

Based on fossils of these early animals, the scientists hypothesized that the tail was a critical body structure for getting animals onto land. But they couldn’t exactly go back in time to test their hypothesis on early tetrapods. Instead of time traveling, “we used three complementary approaches,” says Benjamin McInroe, the study’s first author and a Georgia Tech student at the time of the study.

Gambian Mudskippers
Mudskippers like these helped scientists understand how early tetrapods moved onto land. (©Bjørn Christian Tørrissen, Wikimedia Commons)

For their first approach, scientists at Clemson University observed the movements of an animal they considered very similar to early tetrapods: the mudskipper. This small fish with bulging eyes lives in tidal areas and propels itself across the muddy shore with its land-adapted fins. The Clemson scientists recorded video of how six mudskippers made their way across a flat sandy surface, and then the same surface when it was tilted up 10° and 20°. The videos showed that the fish used only their fins to move across the flat surface. However, to climb the steep slopes, the fish had to lift their bodies up out of the sand with their fins (like a person using crutches), and then “kick” forward with their tails.

With the fish data, the Georgia Tech team used a 3D plastic printer and servomotors to build MuddyBot, a small robot with two front limbs and a powerful tail. By programming the robot to move the same way as the fish, McInroe says that the team was “able to simplify the complexity of the mudskipper” and control individual variables affecting how the robot moved. But when they wanted to test how well MuddyBot climbed the sandy slopes, they realized there was a problem. The tiny sand grains would get inside MuddyBot and destroy its servomotors.

MuddyBot
MuddyBot poses with Dan Goldman, the study’s senior author. (©Rob Felt, Georgia Tech)

The team solved the problem by replacing the sand with poppy seeds—which are too soft to destroy the motors—and plastic particles too big to get inside MuddyBot. Using these seed- and particle-covered slopes, the scientists were able to rigorously test how well MuddyBot climbed with and without using its tail. Like the mudskippers, it turned out that MuddyBot’s tail was essential for climbing the steep slopes.

Using these experimental data, scientists at Carnegie Mellon simulated the movements of early tetrapods crawling onto land. “With the mathematical model,” McInroe says, “we were able to understand the physics behind what was going on.” Physics such as how the crutching movement might have helped early tetrapods reduce friction with the sand, so they could climb up those steep beaches.

Ultimately, the study is part of the team’s ongoing effort to understand how natural selection can modify body structures like tails and fins, says the study’s senior author Dan Goldman of Georgia Tech. “Swimming and walking on land are fundamentally different,” says Goldman, “yet these early animals had to make the transition.”

Megan Cartwright

Written by

Megan Cartwright is a freelance science/medical writer near Seattle. Before becoming a full-time writer, she worked as a scientist studying infectious diseases and vaccines, and earned her Ph.D. in Toxicology from the University of Washington. Megan has written for Slate and Bitesize Bio, and helps write blog posts and learning modules on chemistry for Visionlearning.