Looking at an object and then reaching out and grabbing it is something humans do without thinking. To do this, a brain must encode an object's position in two-dimensional visual space, and then transform the information into activation of motor neurons, which finally move a limb through three dimensions. This sensorimotor transformation has been thought to be too complex for smaller-brained invertebrates to pull off. In a recent issue of Proceedings of the Royal Society B, Jeremy Niven, Swidbert Ott and Stephen Rogers tested this assumption by studying how horse-head grasshoppers walk across gaps.

Horse-head grasshoppers (Orthoptera: Proscopidae) have evolved to look like sticks and spend their lives climbing and hiding among twigs and branches. Niven and colleagues wanted to know whether horse-heads use their visual systems to target forelimb movements as the creatures navigate through environments with the kind of complex substrates that they usually encounter in the wild. The team set up two horizontal rods at right angles to each other with a gap in between. They filmed individuals walking along one rod and reaching across the gap to the other rod. During a reach, animals moved their forelimbs directly to the rod on the other side and made no rhythmic searching movements with their forelimbs. Furthermore, proscopid antennae are too short to be effective as long-range mechanoreceptors. These results suggest that horse-heads do not ‘feel their way’ across gaps in a substrate with either legs or antennae.

However, the team's initial results did not preclude the possibility that gap detection triggers a stereotyped leg movement that simply sweeps through a large space until contact is made with a substrate. To test this hypothesis, Niven and colleagues varied the separation between rods in both horizontal and vertical dimensions, and then measured trajectories of limb movements during reaches. They found that horse-heads do adjust limb trajectories depending on the position of the target rod. The team also found that limb placement was very accurate (less than 10% of all reaches within range missed their mark) and no attempts were made to reach out to targets beyond the range of the forelimbs. These results suggest that proscopids have specific targets in mind when reaching for footholds and can judge reaching distances accurately.

Next, the team next wanted to determine the role that vision plays in judging target distances. They observed that when confronted with large (but not small) gaps, horse-heads would make side-to-side ‘peering’ movements with their heads. When Niven and co-workers deprived individuals of binocular vision (by painting over one of their two eyes), the amount of peering at small gaps increased and individuals almost always initiated reaches with the limb on the same side as the unobstructed eye. They were still able to make accurate reaches with the forelimb on their sighted side; however, the manipulation did prevent accurate reaching by the limb on the sightless side and supressed its use. Taken together, these results suggest that horse-head grasshoppers do rely on visual cues to target limb movements and that they are able to switch flexibly between two different strategies for targeting limbs to points in space (peering versus binocular cues).

The work of Niven and co-workers reminds us not to underestimate the capabilities of invertebrate nervous systems. The work also demonstrates the value of taking a comparative approach to studying the neural basis of animal behavior. If we take the time to study unusual animals, we can up-end long-held assumptions in biology.

J. E.
R. O.
S. M.
Visually targeted reaching in horse-head grasshoppers
Proc. R. Soc. Lond. B