Tool use, defined as “the external employment of an unattached or manipulable attached environmental object to alter more efficiently the form, position, or condition of another object, another organism, or the user itself” (Shumaker et al. 2011:10), is a rare phenomenon that is reported for <1% of animal genera, yet is found in diverse taxa ranging from urchins to chimpanzees (Biro et al. 2013). Among mammals, tool use is most commonly observed in primates (41 species), while only 23 non-primate species are reported to have used tools (Bentley-Condit and Smith 2010). The behavior can be limited to a single captive individual of a species or include all members of the global population (e.g., Rutz et al. 2016). Tool use is generally observed in the context of foraging, but also during physical maintenance, courtship, predator defense, or antagonistic interactions (Bentley-Condit and Smith 2010). Despite extensive interest in tool use by animals, however, the taxonomic distribution and evolutionary context of the behavior are still poorly understood.

Here, we report an observation of apparent tool use in a striped skunk (Mephitis mephitis), captured on 13 February 2018 by a camera trap installed by one of us (SD) in a suburban front yard of Colorado Springs, Colorado, USA. Nine pictures, taken over a period of 15 s, show an individual using a rock to break through a layer of ice, presumably to drink the water below (Fig. 1). The skunk is shown approaching the water dish while carrying a rock in its left front paw. The rock is then struck onto the ice surface in a dynamic yet controlled manner, as indicated by the left paw's grip combined with the change in posture. Unsuccessful in this attempt, the skunk pulls the rock out of the dish with the same paw. Four seconds later, the skunk is again captured standing on the ice with a rock that resembles the one used previously. The skunk then moves across the dish, likely to drink water. No other animals were recorded on the camera trap that night. Inspection of the water dish the following morning revealed two holes with a diameter of ~2 cm in the ice surface.

Skunks (Order Carnivora; Family Mephitidae) are opportunistic generalists, inquisitive and flexible foragers with great dexterity and a manipulative approach to potential prey items (Azevedo et al. 2006). The small- to medium-sized mammals are ubiquitous in North America and among the most familiar mammals throughout their range, mostly due to their unusual use of scent glands as a principal defense mechanism. They tend to be more insectivorous than other members of the order Carnivora (Dragoo 2009). To date, striped and spotted skunks (Spilogale spp.) were categorized as proto-tool users because they often drop food items during extractive foraging (Shumaker et al. 2011). In captivity, they have been observed opening eggs by throwing them between their legs at hard objects (Stebler 1938, Van Gelder 1953, J. Dragoo personal observations). Similarly, skunks overcome the defenses of Tenebrionid beetles that produce toxic chemical secretions when disturbed by rolling the prey in dirt until a tentative bite confirms that the insect has become palatable (Slobodchikoff 1978). The observation of rock pounding or throwing provides plausible evidence for tool use; it involves a mobile, external object, the rock, which is manipulated to change the condition of the ice surface to access an important resource, a drink of water.

Other carnivores regularly use tools, but the behavior is most frequently observed in captivity and only rarely reported for wild animals (Bentley-Condit and Smith 2010). Polar bears (Ursus maritimus) have been reported to throw chunks of ice or rocks onto resting pinnipeds, often successfully injuring their prey (Harington 1962). Spectacled bears (Tremarctos ornatus) re-arrange large trees in their cages in order to reach desired food (Lang 1974), while Alaskan brown bears (Ursus arctos) and giant pandas (Ailuropoda melanoleuca) use objects from their environment to scratch their skin and maintain their fur (Eisenberg and Kleiman 1977, Deecke 2012). Several mongoose species, including ring-tailed (Galidia elegans), dwarf (Helogale parvula), and crab-eating (Herpestes urva) mongoose, are all proto-tool users that extensively manipulate objects to extract food (Bentley-Condit and Smith 2010). Banded mongoose (Mungos mungo), like skunks, are known for throwing rocks between their hind legs to break ostrich (Struthio camelus) eggs (van Lawick—Goodall 1970).

Tool use and proto-tool use are particularly common in the family Mustelidae (weasels, badgers, and sea otters), closely related to Mephitidae. California sea otters (Enhydra lutris) use rocks to break shells of mussels (Mytilus californianus), sea urchins (Strongylocentrotus spp.), and spiny lobsters (Panulirus interruptus; Hall and Schaller 1964). North American badgers (Taxidea taxus) use rocks, clay bricks, and wood blocks to cut off escape routes for ground squirrels (Urocitellus spp.) before attempting capture (Michener 2004). In contrast to the egg throwing between the hind legs observed in mongoose, yellow-throated martens (Martes flavigula) lift eggs up vertically and drop them directly in front of themselves (Wemmer and Johnson 1976). Except in the case of sea otters, our knowledge of tool use in carnivores has generally been based on single observations or repeated observations of single individuals, often in captivity (Bentley-Condit and Smith 2010).

The skunk's ice-breaking behavior, best described as rock pounding or throwing, therefore joins that of other carnivores using tools in the context of food extraction or capture. It shares the uniting feature of foraging carnivore (proto-) tool users—the need to access inaccessible resources by overcoming physical barriers. Such complex problem-solving tends to be associated with increased cognitive ability found in animals with large relative brain size. Among carnivores, experiments show that species with large residual brain size (i.e., larger brains than expected for their body weight) and extensive behavioral repertoires are more successful in opening artificial puzzle boxes to extract food (Benson-Amram et al. 2016). Furthermore, broad comparative analyses across carnivores show that animals in the families Mephitidae and Mustelidae evolved a steeper allometric relationship between body mass and brain volume, so that larger species also have greater relative brain sizes (Finarelli and Flynn 2009). Interestingly, brain size variation in mammals is often associated with group size, yet this relationship does not hold for carnivores (Finarelli and Flynn 2009). Instead, problem-solving and behavioral flexibility in novel environments appear to be important drivers of brain size variation (Sol et al. 2005, 2008). Striped skunks, who are among the heaviest members of Mephitidae (~5 kg; Wade-Smith and Verts 1982, Dragoo 2009), are therefore strongly encephalized which likely provides the neural substrate for their complex behavioral repertoire. As generalist opportunists with high dexterity, they are primed to develop sophisticated foraging strategies, including tool use, in novel environments of suburban backyards (Sol et al. 2008, Benson-Amram et al. 2016). We hypothesize that several other species of carnivores have the capacity to innovate while foraging, including the potential to learn how to use tools. We also predict that the discovery of tool use in other carnivore species will occur at an accelerated rate because of the broad use of remote sensing technology, particularly trail cameras by community scientists and lay people, and the dissemination of animal behavior videos in social media.

Community Science and Social Media: A Recipe for the Discovery of Rare Phenomena?

This case study highlights the emerging synergy between broad public access to technology developed for specific audiences, in this case scientists and hunters, and social media in their contribution to the understanding of the natural history of understudied species or phenomena. While community science (often called citizen science) is usually based on a pre-existing question stated by scientists (Newman et al. 2012), the flow of information was reversed in our study. SD first acquired a commercially available camera trap in 2008 with the aim of documenting the wildlife that resides on a remote small-acreage lot. She recorded a diverse animal community, including black bears (Ursus americanus), coyotes (Canis latrans), elks (Cervus canadensis), and ground squirrels (Spermophilus sp.). Curious to learn more, she installed more units in her yard. To achieve her goal of obtaining a close-up portrait of her nightly visitors, she eventually bought better cameras, lenses, and even engineered her own battery wiring to improve longevity.

But with whom to share her bounty? Family and friends soon tired of the wildlife photographs, even after she gifted many with cameras. Through a friend, SD discovered a Twitter community that centers around showcasing animals captured on trail cameras and also provides feedback on equipment or techniques used to obtain high-quality images. In this context, SD inquired about previous reports of skunk tool use, which was shared and liked widely (https://twitter.com/CameraTrapSue/status/963438772065992704). The post elicited responses from several animal behavior researchers, including the first author, and SD quickly agreed to co-author a publication on her discovery.

Other researchers have used observations shared on social media to obtain relevant data on rare phenomena or species. Rutz and Deans (2018), for example, report the observation of tool use in Eurasian nuthatches (Sitta europaea), which was recorded by a birdwatcher, posted online, discovered, and published by a scientist. Schachner et al. (2009) studied vocal mimicry and entrainment (dancing) in parrots by analyzing contributions posted on the social video platform YouTube. Davies et al. (2012) supplemented the data from a community science platform (www.whaleshark.org) by searching the photo platform Flickr and YouTube for whale shark images and videos captured in their study area.

The synergy between broadly distributed technology and social media provides a powerful discovery tool for rare phenomena and species. The challenge for scientists seeking such data is now to develop information pipelines that connect self-motivated community scientists and incidental observers of animal behavior with their professional counterparts. While there are several community science projects that invite users to post their camera trap images, continued work in science communication to connect lay people with specialists should ultimately catalyze discoveries that expand our understanding of natural history. We predict that the rate of discoveries of rarely witnessed behavior, such as that of tool use in uncommon species, will increase as scientists and enthusiasts become more connected online.