Dogs have demonstrated their excellent socio-cognitive abilities in several other tasks as well. Although no differences were found between dogs of blind versus sighted owners in their way of communicating visually about an inaccessible toy object [8] or their sensitivity to human pointing cues [9], dogs are able to take into account the attentional state of humans in a wide range of situations. For example they prefer to beg for food from a human whose eyes are visible [10], [11], and they are less likely to approach forbidden food when a human’s eyes are open than when they are closed [12], (but see [13] for controversial results about dogs and wolves in a similar experiment). Besides being sensitive to the open eyes of a human, dogs also tend to seek eye-contact with the human partner when facing an unsolvable task, contrary to wolves [7]. Face-to-face communication is of great importance for humans [14] and seems to be a crucial aspect of the dogs’ behaviour as well [15].

Many think that these abilities have been formed by the cognitively challenging complex human social environment [5], [16] and, as a consequence of the shared environment, some rudimentary social-cognitive skills such as interspecific attraction and/or sensitivity to human social cues may have developed in some of the domestic species. Through this evolutionary process, the dog as a species has moved from the niche of its ancestor to the human niche [3]. In this new niche dogs have formed a close social relationship with their human partners (e.g. “attachment") [17], and a flexible system for interspecific communication has also emerged [18]. Alternatively or in parallel to these hypotheses, one might expect the socio-cognitive abilities of dogs resulting from their extensive hand rearing and individual socialization to the human environment from a very early age on. One way to find out the role of domestication in the emergence of these special abilities is to study other domesticated species as well.

Although surprisingly little is known about the socio-cognitive abilities of domesticated species other than dogs, the effects of domestication are probably not limited to canids and therefore the comparative exploration of the phenomenon is important. Recent studies found that domestic cats [19], horses [20] and goats [21] are also able to follow human pointing gestures in order to locate hidden food. Furthermore, experimentally domesticated fox kits (selected for tameness for over 45 years) were also found to be more skilled to follow human pointing gestures than fox kits from a control population [22]. These findings indicate that, in line with previous findings on dogs, domestication as a special evolutionary process leads to increased susceptibility to human communication.

First, subjects’ ability to tolerate eye contact was tested both with a familiar (owner) and an unfamiliar (experimenter) human (Tolerance of eye-contact test). At the beginning of the trial the human lifted the subject so that it was positioned at his/her face level, established eye contact with it and tried to maintain it’s attention by emitting sounds and/or gently moving the animal during a 30 sec period. Half of the subjects in each group were first tested with the owner and then with the unfamiliar experimenter. This was reversed for the other half of the subjects. We found that both domestic species looked more at the owners’ than at the experimenter’s eyes (paired samples t-test, ferrets: t₍₁₅₎=6.088, p<0.001; dogs: t₍₁₇₎=6.093, p<0.001), while no such effect was found for the group of wild Mustela hybrids (t₍₁₅₎=1.092, p=0.292) (Figure 1). In accordance with this result, the preference for the owner (measured by subtracting the duration of looking at the experimenter’s eyes from the duration of looking at the owner’s eyes) was higher in the domestic ferret group than in wild Mustela hybrids (independent samples t-test, t₍₃₀₎=3.488, p=0.001), but no difference was found between domestic ferrets and dogs (t₍₃₂₎=0.006, p=0.995). The above difference between domestic ferrets and wild hybrids resulted from the latter group looking less at the owner’s eyes (t₍₃₀₎=3.572, p=0.001), while no such difference was found between the two groups in case of the unfamiliar experimenter (t₍₃₀₎=0.389, p=0.700). Thus we may conclude that the key difference between domestic ferrets and wild Mustela hybrids is the lack of increased tolerance for eye contact with the owner in the latter group.

Next, subjects had to choose repeatedly (six trials) between a female experimenter and their owner – both of them holding a piece of food – in a two-way choice’ task (Social-preference test). Some of the subjects were not willing to participate or completed only part of the trials (see Materials and Methods for details), but no difference was found between groups in this respect (Fischer exact test, p>0.1). Both domestic ferrets and dogs chose their owners (as opposed to the experimenter) significantly more often than expected by random selection (Wilcoxon Signed Rank Test; ferrets: T+=53.5, p=0.004; dogs: T+=143.0, p<0.001), while the wild Mustela hybrid group displayed a marginally significant preference for the unfamiliar experimenter (T−=38.0, p=0.074) (Figure 2). Domestic ferrets, in comparison with wild hybrids, selected their owners significantly more often (Mann-Whitney U-Test; U=19.0, p=0.001), while no difference was found between the domestic ferrets and dogs (U=132.5, p=0.985).

It should be noted that while domestic dogs and ferrets all had female owners, some of the wild Mustela hybrids had male owners. No difference was found in the Tolerance of eye-contact test between wild Mustela with male and female owners in the duration of eye-contact with the owner (N₁=9, N₂=7, t₍₁₄₎=0.849, p=0.409) and with the experimenter (N₁=9, N₂=7, t₍₁₄₎=0.262, p=0.796). However wild Mustela with male owners showed higher owner preference in the Social preference task than those with female owners (N₁=8, N₂=5, U=5.1, p=0.037). This might possibly be explained by the fact that it is easier to make a distinction between a male versus a female, and suggests that wild Mustela hybrids were involuntarily tested in an easier version of the Social preference task. Interestingly however, despite their ‘advantage’, wild hybrids as a group showed lower preference towards their owners than the two domesticated groups.

Finally we measured the subjects’ responsiveness to two types of human directional gestures (sustained touching and momentary pointing) in two-way object choice tasks (Responsiveness to human gestures test). Dogs and ferrets had to choose between two containers – both of them baited with a piece of food – based on the experimenter’s signals (6 touching and 6 pointing trials). In this test wild Mustela hybrids were less willing to participate than domestic ferrets both in the Sustained touching (Fischer exact test, p=0.04) and the Momentary pointing (Fischer exact test, p=0.03) task. Furthermore those subjects in the wild Mustela hybrid group that did complete all 12 trials had a higher domestic ferret blood ratio (t₍₁₃₎=2.12, p=0.05) than those that did not.

Both domestic ferrets and dogs followed the human directional gestures above chance level in the Sustained touching (Wilcoxon Signed Rank Test; ferrets: T+=120.0, p<0.001; dogs: T+=171.0, p<0.001) and the Momentary pointing (ferrets: T+=66, p=0.001; dogs: T+=66, p=0.001) conditions. Wild Mustela hybrids however, did not succeed in any of these tasks (touching: T+=26.5, p=0.652; pointing: T+=23, p=0.109) (Figure 3). No effect of the owners’ gender could be observed in case of the wild Mustela hybrids (Sustained touching: male owner (N=8): 52.78%, female owner (N=5): 55.56%, U=23.5, p=0.343; Momentary pointing: male owner (N=6): 52.78%, female owner (N=4): 50.00%, U=10, p=0.999). Domestic ferrets outperformed their wild hybrid counterparts in both the Sustained touching (Mann-Whitney U-Test; U=15.5, p<0.001) and the Momentary pointing (U=20.0, p=0.015) tasks. At the same time no difference was found between the domestic ferrets and dogs in any of the two tasks (Sustained touching: U=114.0, p=0.231; Momentary pointing: U=68.5, p=0.584). Furthermore when analyzing only the first trial (it was a sustained touching trial for all subjects) both domestic species succeeded in choosing the indicated cup (binomial tests, test proportion: 0.5; ferrets: p=0.001; dogs: p=0.008) while wild Mustela hybrids did not (p=1.0).

The results obtained from ‘Tolerance of eye contact’ and ‘Social preference’ tests show striking differences between domestic ferrets and wild Mustela hybrids in their specific responsiveness towards their caregivers. Although based on the present results we cannot decide whether the behaviour of ferrets is guided by auditory and/or visual cues, the differences found between domestic and wild individuals allow us to draw a parallel between domestic ferrets and dogs with regard to the studies comparing domestic individuals (ferrets/dogs) with their wild counterparts (Mustela hybrids/wolves). Our findings suggest that the owner-preference showed by domestic ferrets is a consequence of their genetic differences from the wild Mustela hybrids emerged during their domestication history and that behaving distinctively towards the owner may be a basic trait shared by different domestic species.

It is widely accepted that during the process of domestication dogs were selected for preferring the eye contact with humans and for exploiting this form of social interaction as a potential source of information [19]. Propensity to make eye-contact with humans has already been shown to be an important factor in tasks where subjects have to base their choice on human communicative gestures and accounts for the failure to follow human gestural cues in the case of wolves [7] and apes [31]. The present findings are in line with these arguments as domestic ferrets and dogs – both showing increased tolerance of eye-contact in some sense – were equally successful in following human gestural cues while wild Mustela hybrids were not. It is frequently claimed that the utilization of gestural signals presupposes some cognitive skills on the part of the receiver beyond the ability to generalize from everyday communicative interaction with humans to a more controlled experimental situation. However, the fact that domestic ferrets and dogs can rely on human cueing in directing their behaviour in a choice situation is not surprising, since with appropriate human social contact and training, non-domesticated species such as monkeys [32], dolphins [33] and seals [34] are also able to rely on this cue in a two-way object choice test. However, subjects in the present study (contrary to the above mentioned species) did not receive formal training prior to the experiment and were not habituated to the cups containing the food reward. Yet, members of the two domesticated species were successful from their first trial on, whereas members of the wild Mustela group were not. This provides evidence of both domestic ferrets and dogs spontaneously relying on human communicative cues and further confirms the assumption that domestication involves genetic changes that lead to enhanced socio-cognitive abilities toward humans. This is in line with previous claims suggesting that relying on human gestural cues may be a basic ability of the domestic species including domestic cats [19], horses [20] and goats [21] that are also able to follow human pointing gestures in order to locate hidden food.