Exploring breed differences in dogs (Canis familiaris): does exaggeration or inhibition of predatory response predict performance on human-guided tasks? Monique A. R. Udell a, b, * , Margaret Ewald b , Nicole R. Dorey b , Clive D. L. Wynne b, c a Department of Animal and Rangeland Sciences, Oregon State University, Corvallis, OR, U.S.A. b Department of Psychology, University of Florida, Gainesville, FL, U.S.A. c Department of Psychology, Arizona State University, Tempe, AZ, U.S.A. article info Article history: Received 2 September 2013 Initial acceptance 25 September 2013 Final acceptance 14 November 2013 Available online 21 January 2014 MS. number: A13-00723R2 Keywords: breed Canis familiaris cognition dog genes pointing predatory motor patterns social Domestic dogs’, Canis familiaris, responsiveness to human action has been a topic of scientific interest for almost two decades. However, are all breeds of domestic dog equally prepared to succeed on human- guided object-choice tasks? In the current study we compared three breeds of dog with distinct pred- atory motor pattern sequences still under direct selection pressure today based on their traditional working roles. Airedale terriers (hunting dogs) are bred for a fully intact predatory sequence, matching the wild-type form. Border collies (herding dogs) are bred for an exaggeration of the eye-stalk-chase component of the predatory sequence. Anatolian shepherds (livestock-guarding dogs) are bred for the inhibition of the full predatory sequence. Here we asked whether and how these opposing selection pressures correspond with each breed’s tendency to track and follow a human point to a target in an object-choice task. Our results suggest that the presence or exaggeration of key components of the predatory sequence may in fact predict superior initial performance on pointing tasks when compared to a breed selected for its inhibited predatory response. This is the first time relative success on a pointing task has been tied to a known heritable behavioural mechanism (breed-specific motor patterns). How- ever, we also demonstrate that breed-specific differences can sometimes be overcome with additional experience. Thus, an individual’s performance on human-guided tasks is still best predicted by a com- bination of genetic and lifetime factors. Broader implications for the understanding and investigation of canine social cognition are discussed. Ó 2013 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved. The idea that dog breeds should differ in social demeanour and intelligence is heavily disseminated in popular media (Coren, 2006). While it is known that some breeds differ in develop- mental rate (Scott & Fuller, 1965), motor pattern presentation (Coppinger & Coppinger, 2001) and approach avoidance response (Plutchik, 1971), it has been more difficult to empirically demon- strate consistent breed differences in other areas, including social cognition and gesture responsiveness (Dorey, Udell, & Wynne, 2009; Pongrácz, Miklósi, Vida, & Csányi, 2005). Although there have been some exceptions (Buttelmann & Tomasello, 2012; Jakovcevic, Elgier, Mustaca, & Bentosela, 2010; Vas, Topál, Gácsi, Miklósi, & Csányi, 2005), it is possible that breed differences in social cognition are not as pronounced as popularly thought. On the other hand, only a small percentage of studies are specifically designed to analyse breed differences. Many negative reports come from post hoc analyses, based on small numbers of included breeds, which may not provide sufficient power to detect true differences even if they exist (Dorey et al., 2009). The human-guided object-choice task has been extensively used as a measure of canine social cognition (reviewed in Udell, Dorey, & Wynne, 2010a). In this task, an experimenter points to one of two containers where the dog can obtain food upon approach. Several studies have reported breed-group differences on this task. For example, Wobber, Hare, Koler-Matznick, Wrangham, and Toma- sello (2009) and Gácsi, McGreevy, Kara, and Miklósi (2009), found that ‘cooperative’ breeds followed the experimenter’s point to the target significantly more often than ‘independent’ breeds. How- ever, one possible setback to this approach is the subjective nature of breed stereotypes. In other words, how do we know whether a particular breed should be labelled cooperative or independent? In fact, in Gácsi, McGreevy, et al. (2009), Siberian huskies were placed in the independent worker group, and were ultimately deemed less socially sensitive (less accurate on the pointing task), while in Wobber et al. (2009), Siberian huskies were placed in the * Correspondence: M. A. R. Udell, Department of Animal and Rangeland Sciences, Oregon State University, 308 Withycombe Hall, Corvallis, OR 97331, U.S.A. E-mail address: moniqueudell@gmail.com (M. A. R. Udell). Contents lists available at ScienceDirect Animal Behaviour journal homepage: www.elsevier.com/locate/anbehav 0003-3472/$38.00 Ó 2013 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.anbehav.2013.12.012 Animal Behaviour 89 (2014) 99e105