Maternal care plays a crucial role in the life of the offspring; in most mammals it ensures neonatal survival and represents the main source of stimuli in the early postnatal environment. Its influence on neuro-behavioural development has been well documented in a variety of species [1
] including rodents [4
], non-human primates [8
], and humans [10
In mice and rats there is a connection between maternal grooming and ano-genital licking received by pups in the early postnatal period, and their later behavioural patterns [12
], including those related to stress responses, fear and anxiety [1
]. Highly responsive maternal behaviour is known to promote a stress neurobiology that is less reactive as well as more resilient to challenge. The mechanisms are highly specific and involve relatively permanent modifications of DNA that control the expression of glucocorticoid receptors [14
In monkeys, such as the Geoffroy’s marmoset [9
] and Rhesus macaque [15
], low quality or abusive maternal care during the early postnatal period produces more pronounced hypothalamic-pituitary adrenal (HPA) responses to environmental stressors throughout development and into adulthood, compared with animals reared with higher quality early maternal care.
In human infants, sensitive and responsive maternal caregiving has an equivalent role in buffering the HPA system [16
], and a higher quality of maternal behaviour has been found to predict a better cortisol recovery after a mildly stressful event in three-month old infants [18
With respect to domestic dogs, the study of maternal behaviour has previously received comparatively little attention [3
]. Rheingold’s landmark text on maternal behaviour in dogs [19
] has only recently been followed up by other peer reviewed works in this area [3
]. With regard to the effects of maternal care on the behavioural development of puppies, and on the subsequent behaviour of adult dogs, only few works based on observational methods have been published [26
]. Results have also been inconsistent. Foyer and colleagues [26
] found that, in female military German shepherd dogs, there was individual variation in the expression of maternal behaviour. They also found that mother-offspring interactions were associated with individual differences in physical and social engagement, and aggression, in the offspring. Guardini et al. [27
] studied the effects of morning maternal care on 8-week-old puppies living in standardised rearing conditions with a low level of socialisation towards people. This study showed that, as previously found in rodents (see [6
]), the quantity of maternal care received by puppies reared in standardised conditions during the first three weeks of life mediates a set of responses which allows the individual to cope with stressful situations and to better adapt to the environment. Similarly, Tiira & Lohi [29
] found, through the use of a questionnaire administered to dog-owners, that fearfulness in family dogs was associated with poorer maternal care during puppyhood. However, in contradiction to the findings of Foyer et al. [26
] and Guardini et al. [27
], Bray et al. [28
] found that increased maternal behaviour was positively associated with undesirable anxiety-related behaviours and performance in young adult dogs.
To our knowledge, the effects of maternal care on the response of family reared puppies towards unfamiliar environmental and social stimuli has not previously been investigated. This was the focus of the present study.
The offspring-directed maternal behaviour of 12 mothers, from a range of breeds, was recorded during the first three weeks after the birth of the puppies. The behaviour of these puppies was then assessed at two months of age in two potentially stressful situations (arena and isolation tests).
In a previous study we applied the same methodology to a population of laboratory dogs [27
]. In that study it was found that a higher level of maternal care was associated with more exploratory and fewer anxiety related behaviours during the isolation test. No systematic relationship was found between maternal care and arena test behaviour. However, human socialisation and external stimulation provided to laboratory puppies were significantly lower than for family reared puppies.
We therefore hypothesised that family puppies would behave differently from laboratory puppies in those situations in which social (unfamiliar person) and inanimate stimuli (toys) were present, as was the case in the arena test. For the isolation test, we expected results similar to those in our previous study [27
The PLS of the arena test data produced a model with a single predictive component (R2
Y = 0.613, Q2
= 0.462, CV-ANOVA p = 8.2 × 10−9
). This indicates a strong and highly significant systematic relationship between individual puppy behaviours in the arena test and maternal care (see Figure 1
for a loadings plot for this PLS model). Both loadings charts (Figure 1
for the arena test and Figure 2
for the isolation test) should be read as illustrations of the patterns of associations between the X and Y variables in the two models, with level of contribution being broadly related to loading magnitude. However, although it is useful to mention key variables when attempting to interpret the meaning of the models, neither the magnitude of individual loadings, nor the confidence intervals associated with them, should be taken out of context; the overall pattern of behaviour is the key result. Confidence intervals are included for completeness, but overall model evaluation should be based only on R2
In the arena test, the strongest positive loadings (variables positively associated with maternal care score) were Yawning, Whining/Yelping, Near the enclosure, Attention oriented to the stranger and Approach. The strongest negative loadings (variables negatively associated with maternal care score) were Number of squares crossed and Individual play-rope. However, rather than focus on the individual loadings, the key finding is that there is a systematic relationship between maternal care and the pattern of the puppies’ arena test behaviour, characterised by stress signs, approach and attention oriented to the stranger, time spent near the enclosure, and a lack of exploratory or play behaviour.
The PLS of the isolation test data produced a model with two predictive components (R2
Y = 0.507, Q2
= 0.287, CV-ANOVA p = 0.000562). Since it is hard to interpret PLS models that have more than a single predictive component, orthogonal signal correction was used to reject systematic variance in X that was uncorrelated with Y. The resultant cross-validated OPLS model included a single orthogonal component (representing variance in X that is not correlated with Y), and a single predictive component (representing variance in X that is correlated with Y) and there was minimal effect on model quality compared with PLS (R2
Y = 0.507, Q2
= 0.282, CV-ANOVA p = 0.000626). This indicates a significant systematic association between individual puppy behaviours in the isolation test and maternal care (see Figure 2
for a loadings plot for this OPLS model).
In the isolation test, the strongest positive loadings (variables positively associated with maternal care score) were Standing, Paw lifting and Howling, although the value for Paw lifting was undermined by its very wide confidence interval. The strongest negative loadings (variables negatively associated with maternal care score) were Lying, Yawn, and Nose licking. Again, rather than focus on the individual loadings, the key finding is that there is a systematic relationship between maternal care and the pattern of isolation test behaviour, characterised by stress signs, passive standing, paw lifting and not resting.
Early life experiences are known to shape the behavioural development of animals. Previous studies in different species have demonstrated that maternal care plays a key role in the offspring’s ontogeny (e.g., [1
]), and such results have also been found in domestic dogs [26
]. However, to our knowledge, this is the first study investigating the effect of maternal care on the behaviour of two-month old puppies that have been reared in a family environment.
In the present study, the use of multivariate statistical methods (PLS and OPLS) identified statistically significant systematic relationships between maternal care and behaviour in both the arena and isolation tests. This suggests that the amount of maternal care (in terms of nursing, body licking, ano-genital licking, and physical contact between mother and puppy) received during the 21 days after birth has an effect on the pattern of behaviour of family-reared puppies when they are faced with unfamiliar environmental and social stimuli at two months of age. As opposed to individual correlations between variables, multivariate relationships of this kind indicate the presence of patterns of altered behaviours that point to an underlying latent process.
The puppies’ behaviour in the arena test was systematically associated with the amount of maternal care received during the first 21 days of life (as summarised by the maternal care score). The strongest positive loadings (variables positively associated with maternal care score) within that model were certain stress behaviours (increase of yawning and whining/yelping), the time spent near the enclosure and the puppies’ interest in the unfamiliar human figure (increased orientation and approach). In contrast, the strongest negative loadings (variables negatively associated with maternal care score) were the puppies’ activity, interest in his/her surroundings and the inanimate stimuli (reduced number of crossed squares and time spent playing with the rope). Increased maternal care was associated with increased social interest shown by family reared puppies towards the stranger in the arena. The pattern of positive loadings found in the arena test in the present study are comparable with those behaviours observed during the activation of the attachment system when a young mammal is separated from its attachment figure in an unfamiliar environment, as observed in the Ainsworth Strange Situation Test (ASST, [10
]), which is a paradigm created for the assessment of the attachment bond between children and their caregivers. The stress signs and the vocalisations emitted by the family reared puppies in the arena test are comparable with the protest behaviours children show when separated from their mothers [10
], which have also been documented in infant monkeys [32
], lambs [34
], and domestic dog puppies [35
]. Remaining near the enclosure boundary could be interpreted as an attempt to regain proximity to the attachment figure. This behaviour is similar to “stay by” and “be oriented to the door” in children [10
] and adult dogs [38
] when tested in the ASST, and to the attempts to break the barriers separating infant rhesus monkeys from their mothers [39
]. It seems that the more maternal care a family reared puppy receives during early life, the more distress he/she shows during separation, and the more orientated he/she is to reunification with the attachment figure or obtaining social support from another social partner (in this case the stranger in the arena). This is likely to be due to the higher quality of the attachment bond between the mother and the offspring. The puppies stay close to the unfamiliar person, and attempt to seek attention in order to obtain comfort and support in the absence of the attachment figure. The same tendency to seek contact and support from an unfamiliar individual, in the absence of a familiar individual or attachment figure, has been previously observed in children [11
], young chimpanzees [40
], and adult dogs [38
]. In the present study, the level of maternal care seems to influence the puppy’s motivation and/or strategy when faced with a stressful situation, in this case seeking support and comfort from an unfamiliar human being. The puppies’ attempt to buffer stress through social contact with a human is supported by the work of Pettijhon et al. [43
], in which it was observed that the human figure was very effective in alleviating separation-induced distress vocalisation in young puppies, especially if the person behaved actively rather than passively. In the current study, the stranger was instructed to behave passively, and therefore any relief from separation distress that the puppies might have obtained from contact with that person would only have been partial. Stress behaviours observed in the arena test in the present study might therefore also have been influenced by this failed distress buffering.
The fact that maternal care has an impact upon the puppies’ interest in a social stimulus is an important finding, and is in agreement with studies in other mammalian species (e.g., rats: [44
]; prairie voles: [46
]; rhesus macaques: [47
], including humans [49
] and adult military German shepherds dogs [26
]. With regard to an underlying biological basis for this effect on social behaviour, it has been observed that the quantity and the quality of maternal care influences oxytocin receptor density in the brain in rodents [2
], humans [31
], in rhesus monkeys [51
], as well as various neurochemicals which including vasopressin, prolactin, catecholamines, endogenous opioids, adrenocorticotropic hormone, and gamma-aminobutyric acid [52
]. Although there is a lack of knowledge about the impact of maternal care upon oxytocin levels in Canis Familiaris
puppies, it is likely that a similar mechanism is present in this species.
The tendency to seek security from strangers is potentially very important for domestic dog puppies, who have to forge a close relationship with humans when they are taken away from their mothers and introduced into a new home. A positive attitude towards human strangers at two months of age, which is the usual time of adoption, may in fact make the separation from the mother and the establishment of a bond with the new owners easier. Further research is needed in order to more deeply understand the link between the amount of maternal care, oxytocin levels and puppies’ attitude towards people.
In our previous study [27
], the same protocol was used to investigate the influence of morning maternal care on two-month old puppies reared in standardised conditions with limited exposure to human beings. The findings were quite different: In that study, no significant systematic relationship was found between maternal care score and the puppies’ behaviour in the arena test. This difference may be due to general effects of maternal care on the puppy’s adaptability to the environment and his/her inclination to seek social support, combined with the higher level of interspecific socialisation of the family reared puppies that enabled them to regard an unfamiliar person as a potential source of security. The presence of this source of security may also have enabled the puppies to orientate their coping strategy more consistently, leading to a more systematic organisation of behaviour that was not detectable in laboratory puppies.
Contrary to the original hypothesis, higher scores for maternal care were associated with reduced interest in the environment or the objects that were present in the arena test. These findings also contradict some of the previous literature on the effects of maternal care in other species. In rodents, increased ano-genital licking and grooming by the mother is correlated with an increase in play behaviour by the offspring [31
], as well as in exploratory behaviour [4
]. This could, again, be explained as being a consequence of the developmental environment; puppies reared in a family are likely to be more focussed on people than on physical stimuli during the arena test. This explanation is supported by data from a previous work [27
], which found that higher morning maternal care scores were associated with increased exploratory behaviour in puppies reared in an environment with low interspecific socialisation.
The puppies’ behaviour in the isolation test was also systematically associated with the amount of maternal care received during the first 21 days of life, although this model was weaker than the arena test behaviour model. The strongest positive loadings in this model were for standing posture, paw lifting and howling. Howling is a distress vocalisation, which can be interpreted as et-epimeletic or care soliciting behaviour [35
], emitted by the puppies in order to call the mother and to regain contact with her. Standing posture could be interpreted as the puppy waiting for signs of social reunification (i.e., with the mother). As in the arena test, it seems that the more maternal care a puppy received during his/her early life, the greater his/her need to regain proximity with his/her mother during separation (and hence the greater the distress observed during isolation). The important difference between the arena and the isolation test was the absence, in the latter, of any form of potential social support. When family puppies were completely alone they vocalised, but if a stranger was present, puppies tried (although unsuccessfully) to seek comfort and support from that person (otherwise they showed separation distress). It can be hypothesised that the level of maternal care influences the type of stress responses shown by family puppies in this stressful situation (absence of the mother and littermates and being in an unfamiliar environment). In the isolation test model, the strongest negative loadings were with lying, yawn, and nose licking. The latter two are subtle stress signs [57
The initial hypothesis that we would find similar results in the isolation test to those of our previous study involving laboratory reared puppies [27
] was not confirmed. In the laboratory puppies, the amount of maternal care was most strongly positively associated with engagement with the environment inside the metal fences, less destructive behaviours and reduced non-exploratory locomotion. Those results are more similar to those from the literature in rodents (see [1
]). The difference may relate to the diversity of breeds in the present study, and the different sensory environments to which the puppies were exposed. A family home is a much more complex and stimulating environment than a laboratory kennel, and the family reared puppies had far greater contact with people. It is possible that family reared puppies found the arena environment less engaging because they were already used to more complex and stimulating environments. The opposite could be said for the laboratory puppies.
Another potentially influential factor in this study is litter size. Litter size can have an effect on the amount of maternal behaviour available to individual puppies; when the mother’s time and energy are divided between members of a larger litter, we might expect individual mother-pup interactions to be reduced. Unfortunately, due to the small number of litters and relatively uniform litter size in the present study, it was not possible to study the effect of litter size. Given the confounding effects of environmental variation within and between homes, if the effect of litter size were to be investigated, it would probably be easier to identify an effect in laboratory reared pups with a standardised environment.
The results of the present study are likely to be more applicable to dogs that live in an enriched environment that includes social stimuli, as is the experience of most companion and working dogs. Interestingly, Bray et al. [28
] also found an association between the amount of maternal care received and anxiety behaviours in young guide dogs between 14 and 17 months of age; dogs that received higher levels of maternal care showed behaviours related to stress and anxiety when isolated in an empty room (e.g., high activity and short latency to vocalisations when presented with a novel object). Bray et al. [28
] took into account the nursing style of the mothers and found that mothers whose nursing style (vertical nursing) required greater effort by puppies were more likely to produce offspring that were successful in cognitive and temperament tasks, whereas mothers that used a ventral nursing style that required less effort from the puppies were more likely to produce offspring that failed. The authors proposed that there might be benefits from a moderate amount of stress during early life: vertical nursing style may provide opportunities for puppies to cope with small challenges which are beneficial and adaptive in the long term, enhancing arousal regulation and resilience.
Future research may clarify how maternal factors, the social environment and also genetic effects contribute to the expression of pet dogs’ behaviour in different phases of development, from puppyhood to adulthood.