The puzzling pattern of tool use in the animal kingdom


Exploring the ecological and cognitive variables that contribute to the rare and diverse instances of tool use among animals. Image credit: Allanah Booth, The Oxford Scientist.

Tool use was previously considered an exclusively human behaviour. However, in 1960, Jane Goodall observed chimpanzees using twigs to extract termites from termite mounds. Since then, many animals have been found to use tools, from sand flicking by antlions to complex tool modification for foraging in birds and primates. However, whilst tool use is diverse and taxonomically widespread, it is also very rare. This rarity gives rise to a complex and confusing distribution of tool use across the animal kingdom, making it difficult to understand the evolution of this behaviour. To make sense of this puzzling pattern both the ecological factors driving the evolution of tool use and the constraints preventing it need to be explored.

The many definitions of tool use

Tools are not just the conventional objects we tend to think of like hammers or saws but instead can be a huge range of objects such as twigs, stones and even coconut shells. Tool use can be defined many ways, but a popular definition is by St Amant and Horton (2008). They define tool use as ‘the exertion of control over a freely manipulable external object (the tool)’ to either alter ‘the physical properties of…the target’ or to mediate ‘the flow of information between the tool user and the environment or other organisms in the environment’. This first goal has many obvious examples, including using tools to access food resources, as observed by Goodall. Whilst examples of tool use for the second goal in this definition are less obvious, a very clear illustration is the use of sticks to test water depth by wild gorillas.

As well as this broad definition, tool use is often split into two types: stereotyped and flexible. Stereotyped tool use is predominately seen in fish and invertebrates and is typically widespread within a species. It develops with little or no social interaction and occurs in a highly specific context with relatively little variation. Additionally, it’s known to involve minimal working memory. In contrast, flexible tool use, which is mainly found in mammals and birds, is suggested to involve multiple, complex cognitive skills. It often takes a long time to develop in an individual, arising from distinct development stages and involves learning from other individuals.

Costs of tool use

The rarity of tool use in non-human species may best be explained by its costs. There are direct costs of time and effort, as well as opportunity costs of missing out on food that is accessible without tools whilst using tools. There is also likely cognitive expenditure for flexible tool use, although this is difficult to quantify. The development of tool use has also proven costly in New Caledonian crows (NCCs) who, for the first one to two years post fledging (after being capable of flight), spend a lot of time using tools but with little success. So, what are the benefits that enabled tool use to evolve despite these costs?

Ecological factors

There are different hypotheses for how ecology may drive the evolution of tool use. The “necessity hypothesis” suggests tool use evolves to enable access to difficult-to-obtain food sources when accessible resources are scarce. A study on woodpecker finches supports this hypothesis. During the dry season in the arid zone, where food is limited and hard to access, finches use tools to gather half their prey, including the use of twigs or cactus spines to access arthropods in tree holes. In contrast, tools are rarely used in zones where food is abundant. Patterns of tool use by chimpanzees in Bossou, Guinea also fit this hypothesis with tool use to access backup food sources (oil palm nuts and pith) correlating with the scarcity of their accessible food, fruit. However, this hypothesis fails to explain tool use in some species. For example, tool use by a Goualougo chimpanzee population does not increase to compensate for low fruit availability.

Instead, the “opportunity hypothesis” explains this population’s tool use patterns better. This hypothesis suggests tool use evolves due to repeated exposure to appropriate conditions (i.e., the inaccessible resource and tool materials). Contrary to the necessity hypothesis, under the opportunity hypothesis you would expect tool use to increase when the inaccessible resource is abundant rather than when the accessible resource is not abundant. This is the pattern of tool use found in the Goualougo population, with tool use seeming to track fluctuations in army ant and honey availability (inaccessible resources). Further evidence comes from chimpanzee populations in Gombe, where termite fishing occurs during the rainy season when termites are more accessible, despite fruit being more abundant.

The “necessity hypothesis” suggests tool use evolves to enable access to difficult-to-obtain food sources when accessible resources are scarce.

Tool use can also allow access to vacant niches that are inaccessible to non-tool-using species, thus reducing interspecific competition (competition between individuals of two different species for the same resource). For example, despite dolphins typically using echolocation (the location of objects by using the reflection of soundwaves) to locate prey, one population in Australia use sponges to protect their rostra while probing the sea floor for prey. This population is targeting prey in rubble-littered substrate that lack swim bladders, which echolocation is less effective at detecting. Using sponge tools allows the Australian dolphin population to exploit a food resource inaccessible to non-tool-using dolphins.

However, like the necessity hypothesis, the opportunity hypothesis does not explain all cases of tool use (e.g., in woodpecker finches). The inconsistency in support for these two hypotheses suggests that neither theory fully explains the ecological factors driving tool use evolution. Instead, a more appropriate hypothesis, which includes both the necessity and opportunity hypotheses, has been proposed, called the relative profitability hypothesis. In this hypothesis, tool use is expected when it is more profitable than conventional (non-tool-assisted) foraging strategies. The tool use patterns of the woodpecker finch support this hypothesis. In arid zones during the dry season, foraging success per minute is similar for both tool-aided and conventional feeding. Still, the larger prey gathered by tools makes tool use more profitable. However, in the areas with higher food availability, greater foraging success with conventional methods outweighs the advantage of larger prey accessed by tools.

The “opportunity hypothesis” suggests tool use evolves due to repeated exposure to appropriate conditions.

Therefore, certain environmental conditions increase the relative profitability of tool use and facilitate its evolution. For example, the isolated islands NCCs inhabit are proposed to have promoted the evolution of their tool use. On these islands the presence of profitable, inaccessible food and lack of direct competition for these resources resulted in an empty niche accessible via the evolution of tools.  Also, the low risk of predation on these isolated islands likely facilitates tool use by allowing more time and energy to be spent on object interaction and exploration. Distantly related Hawaiian crows are proposed to have convergently evolved tool use in a similar isolated island environment to NCCs.  Convergent evolution is when distinct lineages evolve similar traits independently of one another, often due to experiencing similar environmental conditions. This convergent evolution provides further support for tool use evolution being facilitated by specific environmental conditions. Therefore, a very complex interplay of ecological factors determines the evolution tool use.

Lack of utility or excess of opportunity?

Suppose this mix of ecological factors determines tool use evolution. In that case, the rarity of tool use suggests a lack of necessary environmental contexts in cases where it is advantageous for tool use to evolve. This argument is called the lack of utility hypothesis. However, Hunt et al. (2013) argue there is actually an “excess of opportunity” that contradicts the lack of utility hypothesis. This excess of opportunity problem is that there are many opportunities for tool use to evolve, but these opportunities remain unexploited. For example, termites have incredibly high nutritional value but are difficult and energetically costly to exploit without using tools. However, only certain chimpanzee populations have evolved tool use to exploit termites, despite other populations being present in similar ecological contexts that should facilitate the evolution of tool use. This is just one example of a rich food source exploited much more efficiently with tools that remains rare, contradicting the lack of utility hypothesis. So why is tool use so rare if a lack of ecological opportunity is not constraining it?

Constraints on tool use

Other constraints have been proposed to limit the evolution of tool use, differing between the two types. Stereotyped tool use is proposed to evolve from a pre-existing behaviour pattern that undergoes a phenotypic change (i.e. an observable change in a behaviour). For example, antlions dig pits in the sand by flicking out material, but have also evolved to flick sand at prey on the rim of the pit to capture it from this pre-existing flicking behaviour. Therefore, the evolution of stereotyped tool use is said to be constrained by the frequency of pre-existing behaviour that can evolve into tool use by a phenotypic change. Only a specific, very limited category of pre-existing behaviour has the potential to become tool use via this process, which explains the low frequency of independent occurrences of tool use in invertebrates and fish.

The evolutionary pattern of tool use in animals results from a complex interplay of ecological factors influencing the profitability of tool use relative to alternative non-tool methods.

Flexible tool use, on the other hand, is thought to be constrained by cognitive demands. Whilst there is no simple association between cognitive ability and tool use across all tool-using animals, flexible tool use seems to be associated with more complex cognition, which could constrain its evolution.  For example, tool use in birds correlates with overall brain size as well as neostriatum size. The neostriatum is thought to be the avian equivalent of the mammalian neocortex and is involved in several kinds of learning.  Tool use and brain size are also related in primates.  Additionally, in a study comparing physical cognitive abilities in the closely related tool using NCC and non-tool using carrion crow, NCCs outperformed the carrion crows in the cognition tasks, which suggests complex cognitive processes are required for tool use. Therefore, the evolution of flexible tool use could be constrained by these cognitive demands, explaining its rarity.

Future directions

The evolutionary pattern of tool use in animals results from a complex interplay of ecological factors influencing the profitability of tool use relative to alternative non-tool methods. Additionally, the evolution of tool use has further constraints that can explain its rarity despite the ecological opportunity for it. Understanding how tool use has evolved in other animals is essential for understanding how it has evolved in humans. The more we study flexible tool use in large-brained animals, like crows and primates, the more it seems that certain complex cognitive processes, like physical reasoning and planning, may have originated deeper in the phylogenetic tree than we thought.

However, a lot is still unknown about the cognition underlying tool use in these animals, so further work uncovering the nature of these cognitive processes is important and can help us understand the evolution of human cognition and tool use.