A most aggressive bear: Safari videos document sloth bear defense against tiger predation

Abstract

AbstractSloth bears are non‐carnivorous yet they attack more people than any other bear. They often stand up and charge explosively if a person mistakenly gets too close. Here, we argue that their aggression toward humans is an extension of their behavior toward tigers, which are their only natural predator. Interactions between sloth bears and tigers have not previously been studied because scientists have rarely observed such events. We collected and examined 43 videos or photo documentations of sloth bear–tiger interactions posted on the internet or social media from 2011 to 2023, mainly by tourists visiting tiger parks in India. We observed that sloth bears were most likely to stand up and charge if they first became aware of the tiger at close range (<3 m away). This aggressive–defensive strategy, intended to dissuade the tiger from attacking, appeared to be successful, in that 86% of interactions ended with no contact, whereas four (9%) culminated in the bear's death. We propose that a myrmecophagous diet led to this species' aggressive behavior: (1) their long, blunt front claws, well adapted for digging termites and ants, hamper their ability to climb trees for escape, and (2) they walk with their head down focused on scents underground, and make considerable noise digging and blowing soil, enabling tigers to approach quite closely without being detected. Sloth bears have coexisted with tigers or other (now extinct) large felid predators for their entire evolutionary history. Whereas their aggressive behavior has served them well for millions of years, more recently, people's fear of and retaliation against sloth bears represents a major threat to their survival. Understanding how sloth bears react to tigers provides guidance for reducing attacks on humans, thereby contributing to sloth bear conservation. Our investigation was made possible by passive citizen scientists, who unknowingly collected valuable data.