Animals are known to exhibit different walking behaviors in hilly habitats. For instance, cats, rats, squirrels, tree frogs, desert iguana, stick insects and desert ants were observed to lower their body height in traversing slopes, whereas mound-dwelling iguanas and wood ants tend to maintain constant walking kinematics regardless of the slope. This paper aims to understand and classify these distinct behaviors into two different strategies against toppling for climbing animals by looking into two factors, (i) the torque of the center of gravity (CoG) with respect to the critical tipping axis, and (ii) the torques of the legs, which have the potential to counterbalance the CoG-torque. Our comparative locomotion analysis on level locomotion and inclined locomotion exhibited that primarily only one of the proposed two strategies was chosen for each of our sample species, despite the fact that a combined strategy could have reduced the animal's risk to topple over even more. We found that desert ants of Cataglyphis fortis maintained their upright posture primarily through the adjustment of their CoG-torque (geometric strategy), and wood ants of the Formica rufa species group controlled their posture primarily by exerting leg-torques (adhesive strategy). We further provide hints that the geometric strategy employed by Cataglyphis could increase the risk for slipping on slopes since the leg-impulse substrate angle of Cataglyphis' hind legs were lower compared to Formica's. In contrast, the adhesion strategy employed by Formica's front legs not only decreased the risk for toppling. It also explained the steeper leg-impulse substrate angle of Formica's hind legs which should relate to more bending of the tarsal structures and therefore to more microscopic contact points potentially reducing the risk for hind leg slipping.

Please refer to the README_woehrl_et_al_2021_JEXBIO2021242677.txt file. Funding provided by: Deutsche ForschungsgemeinschaftCrossref Funder Registry ID: http://dx.doi.org/10.13039/501100001659Award Number: BL 236/20-1 to R.B.

Please refer to the method section in https://doi.org/10.1242/jeb.242677.