Hartpury Student Research Journal

Home » Articles posted by hartpuryjournal

Author Archives: hartpuryjournal

To what extent are whales and their surrounding environments affected by whaling?

Author Name: Daisy Harrison, BSc (Hons) Animal Behaviour and Welfare



Two centuries of intensive whaling have reduced whale populations worldwide by up to 99%. While most whales are granted protection as endangered species, there are still human activities regularly carried out that negatively affect the recovery of whales. Whale watching has grown in popularity since the fall in commercial whaling. The presence of humans in the natural habitat of species such as the Humpback Whale can evoke unnatural behaviours that interfere with the mating and breeding habits of these animals. This can hinder population regrowth. The connection between the two practices is discussed in this review.

Data on whale numbers are often taken from the records of whaling boats in order to compare whale population numbers pre, during and post whaling. However the accuracy of these records could be increased by acknowledging instances of hit and miss. Through this, the low numbers of Northern Bottlenose and Right Whales could be better understood and the impact of whaling could be discussed and analysed with increased accuracy.

The loss of such vast numbers of whales has impacted the ecosystems they once occupied in abundance. This is partially by way of knock on effect as their role as a top down predator has led to the ‘Krill surplus hypothesis’, resulting in a rise in Krill and Mink Whale populations. Their absence has also led to a decrease in nitrogen distribution to areas of lower productivity and a decrease in carbon levels in the ocean. Whales also benefit the ocean with their carcasses which provide habitats for deep-sea animals  and encourage both biodiversity and adaptive radiation. The impact of their reduced numbers and the quality of the literature within this research area is discussed in this review.


1.0 Introduction

The commercial whaling of the nineteenth century and illegal whaling carried out by the Soviet Union in the twentieth century resulted in dramatic declines in whale numbers from which populations are still struggling to recover  (Jackson et al., 2008). These declines have made the monitoring and recovery of whales a priority in wildlife conservation attempts (McCauley et al., 2012). For whales such as the Humpback, rates of recovery are so low that they are not currently measurable (Schaffar et al., 2013). Data are therefore often calculated from whaling records (Torres et al., 2013). Catch data are used to estimate historical abundance in contrast to present day numbers (Jackson et al., 2008; Torres et al., 2013). In addition, exploring the impact that the decline of these marine mammals has had on the environments they once inhabited in abundance has been of great interest to marine biologists (Kuo, Chen and McAleer, 2011; Schaffar et al., 2013). A consequence of this loss to the oceans is a substantial reduction of mtDNA haplotype richness (Jackson et al., 2008) which enables chemical energy to be absorbed as a substance of nutrition. Other contributions derived from whales include whale fall carcasses (Smith et al., 2015). These provide nourishing and ideal habitats for small ocean floor life (Butman, Carlton and Palumbi, 1995; Roman et al., 2014). This wide scale removal of whale populations is evident throughout the species affected and the many ecosystems they supported. By understanding the impact of whaling better, rehabilitation and rebuilding of reduced species may be encouraged.


2.0 To What Extent are Whales and Their Surrounding Environments affected by Whaling?

2.1 The Effect of Whaling on Whale Populations

Post intensive whaling, some populations are estimated to have been reduced by 99% (Clapham, 2016). After sustainable whaling was derailed for profit, large-scale illegal whaling was carried out by the Soviet Union (Jackson et al., 2008; Clapham, 2016). Data on whaling numbers has been drawn in most cases from catch records in order to assess numbers of Whales pre, during and post whaling (Jackson et al., 2008; Torres et al., 2013). It should be noted however that not all data can be assumed to be accurate given the high abundance of whales in the past and the use of the medium to give approximate figures (Jackson et al., 2008; Torres et al., 2013). Sperm Whales have been a protected species for approximately 25 years but to date there is no clear evidence for their recovery in any exploited area of their habitat (Carroll et al., 2014). The data presented for this particular study were collected over only three months and therefore may not show the full extent of population trends if the behaviour and migration patterns of the whales vary throughout the year.

In some areas, particularly Scandinavian regions, The Humpback Whale does not currently have an increased population trend due to low numbers and slow paced recovery (Schaffar et al., 2013). Those whales that currently reside in the South Pacific Ocean are threatened by the highly popular and economically valuable activity of whale watching (Schaffar et al., 2013; Bertulli et al., 2016). Whale watching has long been engrained into the tourism culture of many Scandinavian countries and offers those visiting the chance to experience whales in their natural environments and at a much closer proximity than they may be viewed on land. This is an appealing and popular activity for many tourists. The decline of whaling may also have made countries previously known for whaling more attractive to tourists  (Kuo, Chen and McAleer, 2011; Schaffar et al., 2013; Bertulli et al., 2016). It is hypothesised that whaling and whale watching cannot coexist and that tourists may have boycotted countries who participated in whaling on moral grounds or else been unable to enjoy the activity in those areas due to practical issues and the conflict of whaling and whale watching boats (Kuo, Chen and McAleer, 2011).

When approached by the whale watching boats, Humpback Whales were found to significantly change their behaviour and path predictability (Kuo, Chen and McAleer, 2011; Schaffar et al., 2013). This is thought to be an avoidance strategy upon being approached by the whale watchers and also a method of energy preservation (Schaffar et al., 2013). It is unknown to what extent whaling and whale watching have influenced one another but this regular and stressful interruption to their normal behavioural patterns may have a negative effect on the growth and survival of the already vulnerable Humpback Whale. Their risk of extinction puts them in a position where they cannot afford further human interference during their attempts to recover. However it should be noted that this same study (Kuo, Chen and McAleer, 2011) presenting a correlation between the decline of whaling and the rise of whale watching also indicated that 31% of the tourists who participated in the study were unaware of the country’s history of whaling (Kuo, Chen and McAleer, 2011). These tourists may have supported the country’s whale watching industry regardless of on-going whaling if they were ignorant to the activities. It is also unclear if the remaining 69% of tourists who participated in the study who were aware of the history of whaling in the country would have refused to visit and ceased to contribute to the tourist economy if whaling were on-going. This calls into question how far whaling has really influenced whale watching and the strength of the correlation between the two in regards to the Humpback Whale population growth.

Along the east coast of Brazil, Humpback Whales are beginning to re-occupy areas which have been identified as conservation priorities as they include important breeding areas that were previously overtaken by human activity (Martins et al., 2013). This study exclusively documents this trend along the Brazilian Coast, which is limiting to understanding how widespread are these changes. In recent years the Humpback Whale has experienced conservation success in Australia, bringing its population number in that region to 50% of what it was pre whaling and an estimated 80,000 worldwide (Bejder et al., 2016). The world wide population trends of the Humpback Whale currently show no signs of diminishing despite a slow recovery and several scientists have suggested it ought to be down-listed on the official endangered species list (Bejder et al., 2016). This report does not fully disclose the methods of data collection, citing only the Australian monitoring of the whales (Bejder et al., 2016). Full discussion of the methods would improve the validity of this study.

Prior to intensive whaling, approximately 1000,000 Northern Bottlenose Whales existed worldwide but by 1920 this number had been heavily reduced (Whitehead and Hooker, 2012). Similarly, Right Whales around New Zealand and Australia also struggle to recover after having been the focus of world wide whaling from the 1500’s to the 1900’s (Carroll et al., 2014). Data gathered from log books show an increase on previous kill estimates as this records incidences of struck and loss rates where the whale was supposedly killed but not successfully retrieved by the Whaling boat (Carroll et al., 2014). The commercial extinction of Southern Right Whales appears to have occurred between 1830 and 1849, indicating an intensive period of whaling in the New Zealand and East Australian areas (Carroll et al., 2014).


2.2 The Environmental Effects of Whaling

Marine mammals are credited with enhancing primary production within ecosystems due to the great abundance in which they occur (Estes, 2009; Roman and McCarthy, 2010). One effect of intensive whaling on ecosystems is thought to be an excess of Krill, often called the ‘Krill surplus hypothesis’ (Surma, Pakhomov and Pitcher, 2014). This puts forward the plausible but unproven theory that intensive whaling has led to an abundance of other smaller ocean life due to a competitive release for smaller Krill eating species caused by the reduction of the top down predator (Ruegg et al., 2010; Wiedenmann et al., 2011). While samples are often drawn from a limited geographic area, they have been estimated to reflect genetic diversity through the ocean due to the variety of  samples from the Japanese meat market where the data were gathered (Ruegg et al., 2010). However there is no evidence put forward from this study that the markets from which the samples were collected were particularly diverse or that the meat they sold reflected various areas of ocean diversity equally in order to provide data that can be applied globally. The study may be improved by collecting samples from various locations worldwide or by providing detail and evidence that the samples from the market are as diverse as suggested.

As the structure of the ocean continues to change with the increase of whale populations, similarly as it did in response to their decline, an effect on PH Biomass has been observed as Humpback and Fin whales slowly grow in numbers. This is thought to be due to the top down effects of increased predation by the recovering whale populations (Heithaus et al., 2008; Surma and Pitcher, 2015).

The alterations of deep sea biodiversity brought on by whaling is thought to have diminished a prime source of organic matter for chemosynthetic based communities that are associated with hydrothermal vents and other organic sources of input (Butman, Carlton and Palumbi, 1995; Gibson and Atkinson, 2003). Marine mammals enhance primary productivity in their feeding areas by distributing recycled nitrogen near the surface through the release of flocculent faecal plumes to facilitate the transference of nutrition (Roman and McCarthy, 2010; Roman et al., 2014). The physical changes to the ocean that occurred during the period of intensive whaling could have brought about a decline in biodiversity contributing to a decline of southern ocean productivity (Surma, Pakhomov and Pitcher, 2014).

Despite being only a small portion of the ocean’s biomass, the loss of whale abundance has impacted the ocean’s ability to store and sequester carbon (Pershing et al., 2010). Carbon management is often considered an environmental responsibility to maintain for the sake of biodiversity. Whale conservation to rebuild their numbers would provide a similar attempt to restore and manage carbon in the ocean (Pershing et al., 2010). Even in reduced numbers, whales sustain productivity in the regions they occur. Their reduced numbers and possible extinction would have an extreme effect on other marine life in these locations (Roman and McCarthy, 2010).

Research has only developed sufficiently to allow the relationship between whales and ocean carbon levels to be explored in approximately the last few decades as seen by the publication dates of the papers surrounding the topic. This discussion is therefore relatively recent and the full extent of the connection may not be apparent yet. These findings gain validity by the fact that they are extensively tested and recently published with such sponsors as the Marine Mammal Commission of the Fulbright scholar program (Pershing et al., 2010). The results and conclusions they have suggested are most likely of the highest quality currently available. It is important for the sake of validity to note that recent scientific research is often subject to change as new findings and methods come to light though.

Whale carcasses produce an organic and sulphide rich habitat on the sea floor due to their high bone lipid content that allows them to support heterotrophic and chemosynthetic microbial assemblages in the low energy environment of the deep sea (Smith et al., 2015). The communities the carcasses support contain many new species, allowing for adaptive radiation and high trophic diversity (Baco and Smith, 2003; Smith et al., 2015). This study presented evidence supported elsewhere (Smith, 2015) but is limited by its sample sizes. Only 1-7 bones of the 356 bones whales can possess were used and the samples were collected from just three whale carcasses, which may not accurately represent the average whale. The paper also acknowledged that fauna may have fallen from the bones during the initial collection, thus hindering how far the samples reflect those found in the ocean (Baco and Smith, 2003). Larger scale research that protects the samples from damage would strengthen and improve knowledge of whale carcass biodiversity.

Two studies have specifically recorded nematode communities in the remains of an implanted whale carcass (Debenham et al., 2004; Pavlyuk, Trebukhova and Tarasov, 2009). The organic enrichment of the carcasses saw an increase in nematode density, demonstrating the abilities of whale falls to facilitate increased life on the ocean floor (Debenham et al., 2004; Pavlyuk, Trebukhova and Tarasov, 2009). It has been speculated that low numbers of nematodes in recent years can be attributed to the reduced whale numbers (Pavlyuk, Trebukhova and Tarasov, 2009). Data between the studies conflict slightly as one study (Pavlyuk, Trebukhova and Tarasov, 2009) showed nematode density in the context of mink whales, which have been in abundance as a result of whaling (Ruegg et al., 2010; Wiedenmann et al., 2011). While neither of the papers suggest that the results would be different with another whale not in abundance since whaling, the impact of those carcasses removed from the ocean due to whaling ought to be further considered.


3.0 Conclusion

Current conditions for the regrowth of whales are often not ideal despite most whale species occupying a place on the endangered species list. Whale watching may have increased in popularity since the decline of whaling but this human interference can cause abnormal behaviour which can conflict with breeding and mating behaviours. However in some areas along the Brazilian and Australian coast, Humpback whale populations thrive to almost 50% of what they were pre-whaling. This has brought about hope for future whale recovery and suggestions that Humpback whales ought to be down-listed as an endangered species.

The decline of whales has also seen a shift in ecosystem structure due to their effect on the PH Biomass. The consequences of this are a surplus of Krill and the Mink Whale. Deep-sea biodiversity has also reduced as a result of whaling due to the whale’s ability to distribute nitrogen and enhance productivity. The reduced role of whales as facilitators of primary productivity has impacted and reduced the ocean’s ability to store carbon. Whale carcasses are another loss to ocean life as they provide ideal resources for a multitude of species and the decline of carcasses due to whaling has resulted in a loss of deep-sea trophic diversity, biodiversity and adaptive radiation. Overall the removal of so many whales from the ecosystems they once inhabited in abundance has resulted in reduced deep-sea biodiversity and a less enriched ocean.


Reference List

Baco, A. and Smith, C. (2003) ‘High species richness in deep-sea chemoautotrophic whale skeleton communities’, Marine Ecology Progress Series, 260, pp. 109–114. doi: 10.3354/meps260109.

Bejder, M. et al. (2016) ‘Embracing conservation success of recovering humpback whale populations: Evaluating the case for downlisting their conservation status in Australia’, Marine Policy, 66, pp. 137–141. doi: 10.1016/j.marpol.2015.05.007.

Bertulli, C. G. et al. (2016) ‘Can whale-watching and whaling co-exist? Tourist perceptions in Iceland’, Journal of the Marine Biological Association of the United Kingdom. Cambridge University Press, 96(04), pp. 969–977. doi: 10.1017/S002531541400006X.

Butman, C. A., Carlton, J. T. and Palumbi, S. R. (1995) ‘Whaling Effects on Deep-Sea Biodiversity’, Conservation Biology. Wiley/Blackwell (10.1111), 9(2), pp. 462–464. doi: 10.1046/j.1523-1739.1995.9020462.x.

Carroll, E. L. et al. (2014) ‘Two Intense Decades of 19th Century Whaling Precipitated Rapid Decline of Right Whales around New Zealand and East Australia’, PLoS ONE. Edited by A. C. Tsikliras. Public Library of Science, 9(4), p. e93789. doi: 10.1371/journal.pone.0093789.

Carroll, G. et al. (2014) ‘No evidence for recovery in the population of sperm whale bulls off Western Australia, 30 years post-whaling’, Endangered Species Research, 24(1), pp. 33–43. doi: 10.3354/esr00584.

Clapham, P. J. (no date) ‘Managing Leviathan: Conservation Challenges for the Great Whales in a Post-Whaling World’, Oceanography. Oceanography Society, pp. 214–225. doi: 10.2307/24862723.

Debenham, N. J. et al. (2004) ‘The impact of whale falls on nematode abundance in the deep sea’, Deep Sea Research Part I: Oceanographic Research Papers. Pergamon, 51(5), pp. 701–706. doi: 10.1016/J.DSR.2004.02.004.

Estes, J. A. (2009) ‘Ecological Effects of Marine Mammals’, Encyclopedia of Marine Mammals. Academic Press, pp. 357–361. doi: 10.1016/B978-0-12-373553-9.00086-9.

Gibson, R. N. and Atkinson, R. J. A. (2003) Oceanography and marine biology : an annual review. Volume 41. Taylor & Francis. Available at: https://books.google.co.uk/books?hl=en&lr=&id=64crGFXWn5gC&oi=fnd&pg=PA311&dq=smith+et+al+1994+deep+sea+biodiversity&ots=LZEJLoNk7W&sig=uT3gY-Qvg-sBU7BTUjjdcN8D-MY#v=onepage&q=smith et al 1994 deep sea biodiversity&f=false (Accessed: 13 October 2018).

Heithaus, M. R. et al. (2008) ‘Predicting ecological consequences of marine top predator declines’, Trends in Ecology & Evolution. Elsevier Current Trends, 23(4), pp. 202–210. doi: 10.1016/J.TREE.2008.01.003.

Jackson, J. A. et al. (2008) ‘How few whales were there after whaling? Inference from contemporary mtDNA diversity’, Molecular Ecology. Wiley/Blackwell (10.1111), 17(1), pp. 236–251. doi: 10.1111/j.1365-294X.2007.03497.x.

Kuo, H.-I., Chen, C.-C. and McAleer, M. (2011) ‘Estimating the Impact of Whaling on Global Whale Watching’, Documentos de Trabajo del ICAE. Universidad Complutense de Madrid, Facultad de Ciencias Económicas y Empresariales, Instituto Complutense de Análisis Económico. Available at: https://ideas.repec.org/p/ucm/doicae/1123.html (Accessed: 27 September 2018).

Martins, C. C. A. et al. (2013) ‘Identifying priority areas for humpback whale conservation at Eastern Brazilian Coast’, Ocean & Coastal Management. Elsevier, 75, pp. 63–71. doi: 10.1016/J.OCECOAMAN.2013.02.006.

McCauley, D. J. et al. (2012) ‘Assessing the effects of large mobile predators on ecosystem connectivity’, Ecological Applications. Wiley-Blackwell, 22(6), pp. 1711–1717. doi: 10.1890/11-1653.1.

Pavlyuk, O. N., Trebukhova, Y. A. and Tarasov, V. G. (2009) ‘The impact of implanted whale carcass on nematode communities in shallow water area of Peter the Great Bay (East Sea)’, Ocean Science Journal. Korean Ocean Research and Development Institute and The Korean society of Oceanography, 44(3), pp. 181–188. doi: 10.1007/s12601-009-0016-1.

Pershing, A. J. et al. (2010) ‘The Impact of Whaling on the Ocean Carbon Cycle: Why Bigger Was Better’, PLoS ONE. Edited by S. Humphries. Public Library of Science, 5(8), p. e12444. doi: 10.1371/journal.pone.0012444.

Roman, J. et al. (2014) ‘Whales as marine ecosystem engineers’, Frontiers in Ecology and the Environment. Wiley-Blackwell, 12(7), pp. 377–385. doi: 10.1890/130220.

Roman, J. and McCarthy, J. J. (2010) ‘The Whale Pump: Marine Mammals Enhance Primary Productivity in a Coastal Basin’, PLoS ONE. Edited by P. Roopnarine. Public Library of Science, 5(10), p. e13255. doi: 10.1371/journal.pone.0013255.

Ruegg, K. C. et al. (2010) ‘Are Antarctic minke whales unusually abundant because of 20th century whaling?’, Molecular Ecology. Wiley/Blackwell (10.1111), 19(2), pp. 281–291. doi: 10.1111/j.1365-294X.2009.04447.x.

Schaffar, A. et al. (2013) ‘Behavioural effects of whale-watching activities on an Endangered population of humpback whales wintering in New Caledonia’, Endangered Species Research, 19(3), pp. 245–254. doi: 10.3354/esr00466.

Smith, C. R. et al. (2015) ‘Whale-Fall Ecosystems: Recent Insights into Ecology, Paleoecology, and Evolution’, Annual Review of Marine Science.  Annual Reviews , 7(1), pp. 571–596. doi: 10.1146/annurev-marine-010213-135144.

Surma, S., Pakhomov, E. A. and Pitcher, T. J. (2014) ‘Effects of Whaling on the Structure of the Southern Ocean Food Web: Insights on the “Krill Surplus” from Ecosystem Modelling’, PLoS ONE. Edited by E. L. Hazen. Public Library of Science, 9(12), p. e114978. doi: 10.1371/journal.pone.0114978.

Surma, S. and Pitcher, T. J. (2015) ‘Predicting the effects of whale population recovery on Northeast Pacific food webs and fisheries: an ecosystem modelling approach’, Fisheries Oceanography. Wiley/Blackwell (10.1111), 24(3), pp. 291–305. doi: 10.1111/fog.12109.

Torres, L. G. et al. (2013) ‘From exploitation to conservation: habitat models using whaling data predict distribution patterns and threat exposure of an endangered whale’, Diversity and Distributions. Edited by J. Franklin. Wiley/Blackwell (10.1111), 19(9), pp. 1138–1152. doi: 10.1111/ddi.12069.

Whitehead, H. and Hooker, S. (2012) ‘Uncertain status of the northern bottlenose whale Hyperoodon ampullatus: population fragmentation, legacy of whaling and current threats’, Endangered Species Research, 19(1), pp. 47–61. doi: 10.3354/esr00458.

Wiedenmann, J. et al. (2011) ‘Exploring the effects of reductions in krill biomass in the Southern Ocean on blue whales using a state-dependent foraging model’, Ecological Modelling. Elsevier, 222(18), pp. 3366–3379. doi: 10.1016/J.ECOLMODEL.2011.07.013.

What are the welfare issues relating to injury and injury rates in racehorses?

Author Name: Katie Salter, BSc (Hons) Equestrian Sport Science



Horseracing has always had a high injury rate. Some of the key factors influencing injury rate are training regimes, surface type, horse age and type of race. Research has shown that different surfaces can induce different loads under racing conditions increasing fracture rates. Exercise can have a positive effect on young horses by increasing tendon cross section area and reducing fracture rates later in life. Exercise at a young age can also have a negative effect though, via causing osteochondrosis and having a damaging effect on the growth plate due to the rapidly growing skeleton being unable to adapt appropriately under vigorous training. Training regimes have seen changes over the years with the popular use of uphill/incline exercise becoming standard practice alongside lack of rest and recovery. This has increased fracture and hindlimb lameness rates. Studies have suggested that this is due to an increase in strain placed on the limb. A larger number of injuries were also seen in national hunt races in comparison to flat races, which was due to either a collision or a fall at a fence. Horserace Betting Levy in the UK are funding research into factors affecting injury rates and what can be improved to decrease fatalities. Using the findings of this research, it is then hoped that initiatives can be set up to help tackle the ongoing issue of equine racing injury.


1.0 Introduction

Injury rates in racehorses are a major welfare concern. In Britain, 90% of injuries involve the horse’s limb (Perkins, Reid and Morris, 2005). This can cause lameness which is a significant welfare issue (Dyson et al. 2008). Furthermore, joint injuries are observed in 24% of UK Thoroughbreds with the carpal, metacarpophalangeal and metatarsophalangeal areas being worst affected due to repetitive loads (Reed et al. 2012). Per month 1.15 out of 100 UK flat racehorses in training are reported to have fractures (Verheyen and Wood, 2004). Fractures can be due to excessive repetitive high speed, high intensity over loading, cumulative distance training or external trauma, which can all result in micro-damage (Verheyen et al. 2006; Dyson et al. 2008; Martig et al. 2014). Out of 616 flat racehorses assessed, 248 injuries were recorded, with fractures of the tibia (20.7%) and proximal Phalanx (14.5%) being the most common (Ramzan and Palmer, 2011). National hunt racing sees more musculoskeletal injuries with 89% being the superficial digital flexor tendon (Ely et al. 2009; Clegg, 2012).

Training regimens, which differ between yards, remain one plausible factor that influences injury rate due to varying surfaces, frequency of high speed exercise and differing rest periods (Woodward et al. 2000; Boston and Nunamaker et al. 2000; Verheyen et al. 2006). It is hard to distinguish if this is due to a trainer’s personal preference or lack of knowledge though. Misunderstanding the bone remodelling process for instance could, lead to fatigue stress fractures (Riggs, 2002; Verheyen et al. 2006; Dyson et al. 2008; Ely et al. 2009; Martig et al. 2014). Further factors related to injury rates are longer distances and increase in speed, race frequency and the experience and age of the racehorse (Pinchback et al. 2004; Martig et al. 2014). In addition, similarly to as is seen in sports horses, training young horses, ground conditions and excessive workload increase injury (Dyson, 2002). Firmer surfaces have also been associated with higher risk of fatal injury (Williams et al. 2001; Verheyen and Wood, 2004; Perkins, Reid and Morris, 2005). Studies have found that differences between turf and all-weather surfaces (inclusive of surface material, gradient, maintenance and season) influenced the way of going, resulting in the perceived injury rates (Williams et al. 2001; Dyson et al. 2008; Martig et al. 2008; Ramzan and Palmer, 2011). Differing injury rates and types, which have a significant impact on the horse’s welfare, will be discussed during this review to try to determine the factors influencing catastrophic injuries (Perkins, Reid and Morris, 2005). One of the most comprehensively researched areas related to injury is track surface (Reiser et al. 2000; Firth and Rogers, 2005; Setterbo et al. 2009).


2.0 Critical Review

2.1 Influence of track surface on racehorse injuries

All-weather, synthetic surfaces have become popular alternatives for training, enabling all year round preparation (Dyson et al. 2008). Synthetic surfaces have low peak accelerations, peak ground reaction forces and reduction in concussion. These are all factors thought to decrease musculoskeletal associated fatalities (Setterbo et al. 2009; Jacob et al. 2009; Symons, 2016). However, a sudden change in loading environment when under racing conditions, such as ploughed dirt and turf, alters loading forces through bones increasing fracture rates (Kohnke, 2007; Carke, 2009). Supporting research found that horses training on sand during their first year had an increase in fatal fractures during racing (Parkin et al. 2010), although previous research states that these injuries could be due to other influencing factors, such as training at lower speeds or being predisposed to injury (Perkins, Reid and Morris, 2005). Data from a computer system designed by Japan Racing Association to enable the exchange of racing information was used to look at 183,465 flat racehorses, racing on 99,803 turf courses and 83,662 dirt courses at ten different racecourses (Maeda et al. 2011). The validity is questionable as Maeda et al. (2011) did not report the amount of data excluded due to disqualification and any human error in updating due to inaccurate or incomplete information. It was found that the track condition and type had an effect on speed. Horses were found to gallop faster on dirt tracks; however, as conditions deteriorated, their speed decreased and they found it harder to run on a sloped track than a muddy track (Maeda et al. 2011). Velocity on grass (9.06, SE 0.292 m/sec), is faster than on sand (7.90, SE 0.132 m/sec) (Rogers and Firth, 2004), which increases the excitement of spectators. Faster going increases injury risk such as epistaxis and distal limb fracture increasing concerns for equine welfare (Rosanowski et al. 2017). The data by Maeda et al. (2011) could be assessed further by analysing data from other countries due to the variation in track conditions caused by differences in weather (Williams et al. 2001). Weather conditions have been seen to play a significant role in the ground conditions (Perkins, Reid and Morris, 2005). Firmer going can cause an increased risk of injury (Mohammed, Hill and Lowe, 1991; Pinchbeck et al. 2004; Reardon et al. 2013). New Zealand in winter is associated with a reduction in the odds of injury due to higher moisture contents (Perkins, Reid and Morris, 2005). This suggests that regions of low rainfall increase fracture associated fatalities due to structural changes within the track surface (Williams et al. 2001). In Symons (2016) study, horses had greater fetlock hyperextension or dorsiflexion and reduction in forces to the tendon and ligaments during gallop on synthetic surfaces compared to dirt tracks. Synthetic surfaces reduce musculoskeletal failure through changes in distal limb motions due to different race surface mechanics altering musculoskeletal tissue loads, and predisposition for injury (Symons, 2016). The validity of this study is debatable however because only 5 horses were used, there was no repetition of the test and the surface was only harrowed prior to testing. The key aim for the British Horseracing Association is to improve racehorse welfare (BHA, 2018). Therefore following changes seen in other disciplines; a shift towards synthetic surfaces would be beneficial in racing (Hobbs et al. 2016).


2.2 Influence of training on racehorse injuries

A high number of injuries occur in training due to insufficient recovery periods after chronic fatigue states from training or racing. This results in inadequate recovery and depletion in energy stores leading to poor performance, fatigue and exhaustion (Evans, 2017). Inadequate rest and excessive overloading results in an increase in matrix and mineral deposition, greater activation of osteoclastic resorption and overlaying of cartilage, increasing brittleness and accumulation of micro-damage in bones (Norrdin, 1998). Repetitively increasing loads exceeding the rate of bone adaption which can take place can result in a rise in the process of ossification leading to dorsal metacarpal disease and stress fractures (Bailey, 1998; Lawrence, 2003b; Davies, 2003). Similar overloading-induced reactions are seen in racing greyhounds when exercised at high speed and high frequency at young ages (Boemo, 1998; Ireland 1998).

Ramzan and Palmer (2011) observed that tibia stress fractures are the most common in flat racing. However previous studies reported third metacarpal, pelvis and carpal fractures to be more common (Bathe, 1994; Verheyen and Wood, 2004; Cogger et al. 2008). Differences may be due to excessive use of uphill training in the Ramzan and Palmer (2011) study. Yards are now using an increase in incline which can increase the risk of hindlimb lameness (Dyson et al. 2008). Incline is thought to produce larger peak forces, greater stride frequency and compression which is usually found to be unilateral (Pieter and Ramzan 2014). Ramzan and Palmer (2011) found that hindlimb stress fractures were the most common injury type. This was also seen in a study looking at high intensity training regimes, which found hindlimb locomotion asymmetry increased during spring when the 2-year-olds training commenced (Ringmark, 2016).  A subjective score to measure lameness was utilised; therefore the results may not be reliable so a more objective measure would increase reliability (Ringmark, 2016). Ramzan and Palmer (2011) also reported that 25% of horses in training each year sustained a significant musculoskeletal injury. This study excluded dorsal metacarpal disease though which is an explanation for a higher percentage reported by Cogger et al. (2008).

Two flat yards in the UK observed 291 horses over one year and found 18 fractures occurred in training (Verheyen and Wood, 2004). Validity is questionable though as the study was only conducted on two flat yards; however other studies have reported similar findings (Verheyen and Wood, 2004). A two year study on fracture rates in training found 245 fractures occurred in training in Newmarket (Bathe, 1994). These rates were not just seen in the UK. Throughout California, 100 horses in training over one year were euthanized (Blackwell, 2011) and in 9 months there were 71 fatal fractures in training, compared with 66 during racing (Estberg, 1995; Verheyen and Wood, 2004).

It is important to simulate more realistic demands of racing to create a more representative adaption response without having a detrimental effect (Verheyen et al. 2006). The average gallop is around 60km/h however it is essential not to exceed this, as extreme loading and increased limb cycles result in accumulation of micro-damage, increasing fracture risk (Verheyen et al. 2006; Martig et al. 2014). Although no gallop work during training within the first year was associated with an increase in injury risk (Parkin et al. 2008), inappropriate preparation for competition resulted in earlier fatigue during racing, increasing injury risk and exhaustion (Evans, 2007). A minimum distance gallop of 805-2012m per week was thought to decrease fracture rates (Parkin et al. 2008). Variation of exercise intensity was found between yards; some include six days canter work and a gallop 2-3 times per week (Verheyen et al. 2006), whereas others are thought to not include gallop work during their normal training regime. The use of bone scintigraphy and ultrasonography in 3 month intervals during training would be a useful method to track the remodelling process and evaluate information as well as increasing early recognition of injury (Verheyen et al. 2006; Ely et al. 2009; Clegg, 2012).


2.3 Effect of horse development in racing

National hunt racehorses begin training at 3-4 years, whereas flat racehorse training starts at 18 months (Ely et al. 2009). The age at which a horse should start training is debatable as age has been found to have both positive and negative impacts (Williams et al. 2001; Dyson et al. 2008). Firth (2006) found that the cross sectional area of the third metacarpal (MC3) was significantly larger and bone mineral density was higher in foals at pasture compared to stabled foals. Proximal sesamoid bones and trabecular bone mineral density were higher in exercised foals. Differences between groups disappeared at 11 months of age though, suggesting normal development continued once out on permanent pasture (Firth, 2006). There were indications that sprinting led to overstimulation of bone resulting in less active mineral deposition long-term. One explanation is the horse’s lack of skeletal maturity to withstand training (Williams et al. 2001; Dyson et al. 2008), although there is evidence that exercise has benefits for the immature skeleton and reduction of fractures later in life (Ely et al. 2009). The foals which did perform sprint exercise appeared to have greater MC3 cortical size and mineral content (Firth, 2006). This study did not investigate horses after the age of 11 months however it suggested exercise is beneficial (Firth, 2006). This suggestion is supported by evidence showing that early training is associated with an increase in the cross section area of tendons and greater thickness of hyaline cartilage (Rogers and Firth, 2005; Firth, 2006). In contrast, high frequencies of high speed exercise were discovered to have detrimental effects on bone adaptation (Judex and Zernicke, 2000). This causes osteochondrosis, trabecular micro-fracture and retained cartilage and affects the growth plate (Firth, 2006).

In 2001, 537 horses entered flat racing training for the first time. Only 61% went on to race as 2 year olds and of those unable to continue onto racing 62% required veterinary interventions (Wilsher et al. 2006; Dyson et al. 2008). Supporting research showed that horses starting intense training at 2-3 years showed a higher risk of joint injury (Reed et al. 2012), shin soreness (Perkins, Reid and Morris, 2005) and an inability for the rapidly growing skeleton to adapt appropriately under vigorous training (Dyson et al. 2008). Horses are 90% of their mature height and 66% of their mature weight when they reach 12 months (Firth, 2006). Maturity in thoroughbreds is around 2 years, however full growth is not complete until 4 years (Firth, 2006). The approach used by national hunt trainers that allows their horses time to adapt skeletally and provides more rest periods than flat trainers is beneficial (Verheyen and Wood, 2004). This was found to reduce micro-damage, limit fatigue accumulation and allow appropriate bone adaptation to take place (Verheyen and Wood, 2004). By reducing the magnitude of the loads generated within the limb and the number of cycles, it enables the body to repair and adapt thus decreasing injury risk (Martig et al. 2014), whereas it was seen that flat racehorses have a shorter racing career (Verheyen and Wood, 2004).

There are suggestions that as horses get older there is a decreased tendon matrix (Smith et al. 2002). Research has found that older horses were more susceptible to injury due to the inability to remove partially degraded collagen from the matrix leading to reduced mechanical competence (Clegg, 2012). Ageing decreases adaptive bone responses and increases the occurrence of injury, in particular increasing the risk of tendon and ligament injury (Verheyen and Wood, 2004; Perkins, Reid and Morris, 2005; Reardon et al. 2013; Rosaowski, 2017).


2.4 Injury rates in race type

The frequency of injury is higher in national hunt racing than flat racing (Williams et al. 2001) with fatality rates of 3.0/1000 starts associated with obstacles (McKee, 1995). Injuries to the shoulder/Humerus region (1.29/1000 starts) were commonly associated with a fall at a fence (41%) or a collision with a jump (78% at a hurdle and 47% at a fence) (Williams et al. 2001; Ely et al. 2009). It was reported that horses being whipped progressively throughout the race were 7 times more at risk of falling (particularly at a fence) compared with those not being whipped (Clegg, 2011). The use of the whip has been controversial as it causes a flight response, increasing speed which is associated with increased fall rates, which can compromise welfare (McLean and Mcgreevy, 2010). Whips are used to reduce drifting and collisions and the British Horseracing Authority have rules in place to monitor usage to protect horse welfare (Aid, 2016). Jockey experience was also linked to falls and injury rates. Inexperienced jockeys approached fences differently (Powers and Kavanag, 2007; Kang, 2010) and it was found that 68% of horses with tendon injuries completed the race before the injury was detected (Ely et al. 2009), although it was suggested that the horses’ adrenaline may mask an injury (Sukumarannair et al. 2002).


3.0 Conclusion

Injuries in the racing industry are a major welfare concern. Investigations into the effect that surface, age, training regimes and race type have on injury rates have been conducted, with fractures, tendon and musculoskeletal injuries being found to be the most common injuries. Research can help tackle misconceptions used in training as several studies have shown major effects of differences amongst trainers. The effect of race type is another key influencing factor with the frequency of injury seen to be higher in National Hunt racing than flat racing. Age is undoubtedly an important factor affecting injury rates although it is an area which needs to be investigated further as evidence from previous studies have been controversial. Shifts towards synthetic surfaces, particularly in training and other disciplines, need further consideration for the use on racing tracks to help minimise injury. Further studies need to be conducted to enhance understanding of racehorse injuries and ways to reduce injury rates to improve equine welfare.


Reference List

Aim, C. (n.d.). Briefing: ‘The use of the whip in horse racing’ campaign aim, 44 (1787309),1-2.URL: https://www.animalaid.org.uk/wpcontent/uploads/2016/07/whipbriefing.pdf

Anil, S. S., Anil, L., & Deen, J. (2002). ‘Challenges of pain assessment in domestic animals’. Journal of the American Veterinary Medical Association, 220(3), 313–319. https://doi.org/https://doi.org/10.2460/javma.2002.220.313

Bathe, A. . (1994). ‘245 Fractures in Thoroughbred racehorses: results of a 2 year prospective study in Newmarket.’ Proc. Am. Ass. Equine Practnrs, 40, 275–276.

BHA. (2018). ‘The role of the BHA in horse welfare.’ Retrieved from https://www.britishhorseracing.com/regulation/role-bha-horse-welfare/

Blackwell. (2011). ‘Equine Welfare.’ Retrieved from https://books.google.co.uk/books?hl=en&lr=&id=bPMJKvJRYZ4C&oi=fnd&pg=PT210&dq=welfare+issues+relating+to+injury+and+injury+rates+in+racehorses&ots=nP2-RDROhw&sig=Oj490SdT7W1hqTkqK8D7t_4CIpQ#v=onepage&q=welfare issues relating to injury and injury rates in

Carke, A.F., 2009. ‘Review of safety of turf versus non-turf (synthetic and dirt) racing surfaces.’ In AmAssoc Equine Pract (Vol.55, pp.183-186).

Clegg, P. D. (2011). ‘HBLB’s advances in equine veterinary science and practice: Musculoskeletal disease and injury, now and in the future. Part 1: Fractures and fatalities.’ Equine Veterinary Journal, 43(6), 643–649. https://doi.org/10.1111/j.2042-3306.2011.00457.x

Clegg, P. D. (2012). ‘Musculoskeletal disease and injury, now and in the future. Part 2: Tendon and ligament injuries.’ Equine Veterinary Journal, 44(3), 371–375. https://doi.org/10.1111/j.2042-3306.2012.00563.x

Cogger, N., Evans, D., Hodgson, D., Reid, S., & Perkins, N. (2008). ‘Incidence rate of musculoskeletal injuries and determinants of time to recovery in young Australian Thoroughbred racehorses.’ Australian Veterinary Journal, 86(12), 473–480. Retrieved from https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1751-0813.2008.00359.x

D.L, E. (2006). ‘Welfare of the Racehorse During Exercise Training and Racing.’ Faculty of Veterinary Science, 1(1), 181–201. Retrieved from https://link.springer.com/chapter/10.1007/978-0-306-48215-1_8

Dyson, P. K., Jackson, B. F., Pfeiffer, D. U., & Price, J. S. (2008). ‘Days lost from training by two- and three-year-old Thoroughbred horses: A survey of seven UK training yards.’ Equine Veterinary Journal, 40(7), 650–657. https://doi.org/10.2746/042516408X363242

Ely, E. R., Avella, C. S., Price, J. S., Smith, R. K. W., Wood, J. L. N., & Verheyen, K. L. P. (2009). ‘Descriptive epidemiology of fracture, tendon and suspensory ligament injuries in National Hunt racehorses in training.’ Equine Veterinary Journal, 41(4), 372–378. https://doi.org/10.2746/042516409X371224

Estberg, L., Gardnerb, I. A., Stoverc, S. M., Johnsona, B. J., Case, J. T., & Ardansa, A. (1995). ‘Cumulative racing-speed exercise distance cluster as a risk factor for fatal musculoskeletal injury :in Thoroughbred racehorses in California.’ Preventive Veterinary Medicine, 24, 253–263. Retrieved from file:///E:/Year 3/Semester 1/Undergraduate Independent study/Assignment/Articles/2.pdf

Firth, E. C. (2006). ‘The response of bone, articular cartilage and tendon to exercise in the horse.’ Journal of Anatomy, 208(4), 513–526. Retrieved from https://onlinelibrary.wiley.com/doi/full/10.1111/j.1469-7580.2006.00547.x

Firth, E.C. and Rogers, C.W., (2005). ‘Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. 7. Bone and articular cartilage response in the carpus.’ New Zealand veterinary journal, 53(2), pp.113-122.

Hobbs, S. J., Northrop, A., Mahaffey, C., Martin, J., Clayton, H., Murray, R., … Peterson, M. (2014). ‘White Paper on Equine Arena Surface Assessment.’ Retrieved from https://inside.fei.org/system/files/Equine Surfaces White Paper.pdf

Judex, S., & Zernicke, R. F. (2000). ‘Does the mechanical milieu associated with high-speed running lead to adaptive changes in diaphyseal growing bone?’ Bone, 26(2), 153–159. https://doi.org/10.1016/S8756-3282(99)00256-2

Kang, O. D., Ryu, Y. C., Ryew, C. C., Oh, W. Y., Lee, C. E., & Kang, M. S. (2010). ‘Comparative analyses of rider position according to skill levels during walk and trot in Jeju horse.’ Human Movement Science, 29(6), 956–963. https://doi.org/10.1016/j.humov.2010.05.010

Kohnke, J. R. (2007). ‘Bone biomechanics: a review of the influences of exercise and nutritional management on bone modeling in the growing and exercising horse.’ Equine. Retrieved from https://en.engormix.com/equines/articles/bone-biomechanics-in-horse-t33552.htm

Martig, S., Chen, W., Lee, P. V. S., & Whitton, R. C. (2014). ‘Bone fatigue and its implications for injuries in racehorses.’ Equine Veterinary Journal, 46(4), 408–415. https://doi.org/10.1111/evj.12241

Mckee, S. L. (1995). ‘An update on racing fatalities in the UK.’ Equine Veterinary Education, 7(4), 202–204. Retrieved from https://onlinelibrary.wiley.com/doi/abs/10.1111/j.2042-3292.1995.tb01225.x

McLean, A. N., & McGreevy, P. D. (2010). ‘Ethical equitation: Capping the price horses pay for human glory.’ Journal of Veterinary Behavior: Clinical Applications and Research, 5(4), 203–209. https://doi.org/10.1016/j.jveb.2010.04.003

Norrdin, R. W., Kawcak, C. E., & Wayne McIlwraith, C. (1998). ‘Subchondral bone failiure in an equine model of OA.pdf.’ Bone, 22(2), 133–139.

Parkin, T. D. H., Clegg, P. D., French, N. P., Proudman, C. J., Riggs, C. M., Singer, E. R., … Morgan, K. L. (2010). ‘Horse‐level risk factors for fatal distal limb fracture in racing Thoroughbreds in the UK.’ Equine Veterinary Journal, 36(6), 513–519. Retrieved from https://onlinelibrary.wiley.com/doi/abs/10.2746/0425164044877387

Perkins, N. R., Reid, S. W. J., & Morris, R. S. (2005). ‘Risk factors for musculoskeletal injuries of the lower limbs in thoroughbred racehorses in New Zealand.’ New Zealand Veterinary Journal, 53(3), 171–183. https://doi.org/10.1080/00480169.2005.36502

Pinchbeck, G. L., Clegg, P. D., Proudman, C. J., Stirk, A., Morgan, K. L., & French, N. P. (2004). ‘Horse injuries and racing practices in National Hunt racehorses in the UK: The results of a prospective cohort study.’ Veterinary Journal, 167(1), 45–52. https://doi.org/10.1016/S1090-0233(03)00141-2

Powers, P., & Kavanagh, A. M. (2005). ‘Effect of rider experience on the jumping kinematics of riding horses.’ Equine and Comparative Exercise Physiology, 2(4), 263–267. https://doi.org/https://doi.org/10.1079/ECP200568

Ramzan, P. H. L., & Palmer, L. (2011). ‘Musculoskeletal injuries in Thoroughbred racehorses: A study of three large training yards in Newmarket, UK’ (2005-2007). Veterinary Journal, 187(3), 325–329. https://doi.org/10.1016/j.tvjl.2009.12.019

Ray C. Boston, D. M. N. (2000). ‘Gait and speed as exercise components of risk factors associated with onset of fatigue injury of the third metacarpal bone in 2-year-old Thoroughbred racehorses.’ American Journal of Veterinary Research, 61(6), 602–608. https://doi.org/https://doi.org/10.2460/ajvr.2000.61.602

Reardon, R. J. M., Boden, L. A., Mellor, D. J., Love, S., Newton, J. R., Stirk, A. J., & Parkin, T. D. H. (2013). ‘Risk factors for superficial digital flexor tendinopathy in Thoroughbred racehorses in steeplechase starts in the United Kingdom’ (2001-2009). Veterinary Journal, 195(3), 325–330. https://doi.org/10.1016/j.tvjl.2012.06.033

Reed, S. R., Jackson, B. F., Mc Ilwraith, C. W., Wright, I. M., Pilsworth, R., Knapp, S., … Verheyen, K. L. P. (2012). ‘Descriptive epidemiology of joint injuries in Thoroughbred racehorses in training.’ Equine Veterinary Journal, 44(1), 13–19. https://doi.org/10.1111/j.2042-3306.2010.00352.x

Reiser II, R.F., Peterson, M.L., McIlwraith, C.W. and Woodward, B., (2000). ‘Simulated effects of racetrack material properties on the vertical loading of the equine forelimb.’ Sports Engineering, 3(1), pp.1-11.

Riggs, C. M. (2002). Fractures – ‘A preventable hazard of racing thoroughbreds?’ Veterinary Journal, 163(1), 19–29. https://doi.org/10.1053/tvjl.2001.0610

Rogers, C.W. and Firth, E.C., (2004). ‘Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. 2. Measurement error and effect of training stage on the relationship between objective and subjective criteria of training workload.’ New Zealand Veterinary Journal, 52(5), pp.272-279.

Rosanowski, S. M., Chang, Y. M., Stirk, A. J., & Verheyen, K. L. P. (2017). ‘Risk factors for race-day fatality, distal limb fracture and epistaxis in Thoroughbreds racing on all-weather surfaces in Great Britain’ (2000 to 2013). Preventive Veterinary Medicine, 148(October), 58–65. https://doi.org/10.1016/j.prevetmed.2017.10.003

S, W., Allen, W. R., & Wood, J. L. N. (2010). ‘Factors associated with failure of Thoroughbred horses to train and race.’ Equine Veterinary Journal, 38(2), 113–118. Retrieved from https://onlinelibrary.wiley.com/doi/pdf/10.2746/042516406776563305

Setterbo, J. J., Garcia, T. C., Campbell, I. P., Reese, J. L., Morgan, J. M., Kim, S. Y., … Stover, S. M. (2009). ‘Hoof accelerations and ground reaction forces of Thoroughbred racehorses measured on dirt, synthetic, and turf track surfaces.’ American Journal of Veterinary Research, 70(10), 1220–1229.

Setterbo, J.J., Garcia, T.C., Campbell, I.P., Reese, J.L., Morgan, J.M., Kim, S.Y., Hubbard, M. and Stover, S.M., (2009). ‘Hoof accelerations and ground reaction forces of Thoroughbred racehorses measured on dirt, synthetic, and turf track surfaces.’ American journal of veterinary research, 70(10), pp.1220-1229.

Smith. R. K Birch, H. L., Goodman, S., Heinegard, D., & Goodship, A. E. (2002). ‘The influence of ageing and exercise on tendon growth and degeneration—hypotheses for the initiation and prevention of strain induced tendinopathies.’ pdf, 133, 1039–1050.

Verheyen, K. L. P. (2013). ‘Reducing injuries in racehorses: Mission impossible?’ Equine Veterinary Journal, 45(1), 6–7. https://doi.org/10.1111/evj.12009

Verheyen, K. L. R., & Wood, J. L. N. (2004). ‘Descriptive epidemiology of fractures occurring in British Thoroughbred racehorses in training.’ Equine Veterinary Journal, 36(2), 167–173. https://doi.org/10.2746/0425164044868684

Verheyen, K., Price, J., Lanyon, L., & Wood, J. (2006). ‘Exercise distance and speed affect the risk of fracture in racehorses.’ Bone, 39(6), 1322–1330. https://doi.org/10.1016/j.bone.2006.05.025

Williams, R. B., Harkins, L. S., Hammond, C. J., & Wood, J. L. N. (2001). ‘Racehorse injuries, clinical problems and fatalities recorded on British racecourses from flat racing and National Hunt racing during 1996, 1997 and 1998.’ Equine Veterinary Journal, 33(5), 478–486. https://doi.org/10.2746/042516401776254808

Woodward, R. F. R. I. M. L. P. C. W. M. B. (2001). ‘Simulated effects of racetrack material properties on the vertical loading of the equine forelimb.’ Wiley Online Library, 3(1), 1–11. https://doi.org/https://doi.org/10.1046/j.1460-2687.2000.00049.x

Dolphin-human Interactions

Author Name: Georgia Oaten, BSc (Hons) Animal Behaviour and Welfare



Dolphins (Delphinus) have long inspired human fascination and desire for interaction. In the wild and captivity, humans attempt to develop relationships with dolphins, and an industry focused on facilitating the observation, swimming and meeting of dolphins has rapidly grown over the last decade. The dolphin interaction industry involves participants paying a fee in exchange for the opportunity to interact with dolphins. Human-animal interaction (HAI) does not constitute a unified field and this is seen in the literature pertaining to dolphin-human interaction. Across studies, there are themes of low statistical power, anthropomorphic projection, varying research designs, and observer bias, all of which hinder comparability of results. There is a clear need for the standardisation of terminology, studies focusing on the impact of HAI on the animal, and metrics to quantify the human interaction behaviours. There are 36 species of dolphin, but their use in organised HAI invariably only includes a few species. There may be species differences in HAI responses, suggesting variation of species would be beneficial. Understanding the effect of these programmes on the animal is important to ensure that they do not negatively impact welfare. This paper highlights the importance of longitudinal research, as behavioural responses detected in long term studies may not be present in short term studies. It additionally presents the argument that methodology should combine behavioural observations and physiological measures of dolphins, as use of behaviour alone can be misleading due to individual variation and a lack of understanding of the mechanisms driving behaviour.


1.0 Introduction 

Human-animal interaction (HAI) refers to the mutual and dynamic interaction between people and animals and how these interactions may affect physical and physiological health and wellbeing (Griffin et al., 2012). The majority of HAI research examines agricultural and companion animals, and there is a distinct lack of HAI research into wild and zoo animals. A literature review by Hosey and Melfi (2014) identified approximately 161 papers on companion animals, 76 on agricultural animals, 21 on wild animals and just 22 on zoo animals. The majority of HAI studies encompass primates, cetaceans, felids and ungulates, with a distinct absence of dolphin research (Baker and Wolen, 2008).  Additionally, the variation between papers’ scope and subject matter make comparison across environments and species challenging (Hosey and Melfi, 2014).

The literature highlights the need for further investigation of the effect on HAI on the animals participating in these programmes (Chandler, 2012).  There are few studies that focus on dolphin-human interaction, which is concerning due to the growing prevalence of dolphin-interaction programmes (Mayes, Dyer and Richins, 2004; Brakes and Simmonds, 2013). It is widely believed that people have an emotional need to connect with animals, which manifests as people visiting zoos and partaking in interaction programmes (Vinning, 2003; Myers, Saunders and Birjulin, 2004). Dolphin-interaction programmes consist of participants paying a fee in exchange for interaction with dolphins (Barnett et al., 2009). Due to the lack of data on the effect of dolphin-interaction programmes on participating dolphins, questions remains as to whether they are negative or positive experiences for the animals involved (López, 2012).


2.0 Critical Review

2.1 Captive Dolphin

Trone, Kuczaj and Solangi (2005) investigated changes in the behavioural repertoire of three dolphins, before, during and after trainer-controlled dolphin-interaction programmes. An increase in interaction initiated by the dolphin to humans outside of the programmes was observed. Trone (2005) concluded that the increase in dolphin-initiated interaction, alongside an increase in “play” behaviour after interaction suggested robust psychological health. However, the description of “play” was vague and included behaviours which may have been aggressive. Janik (2015) argued that the same behaviour reported as play are  indicative of a warning signal in the wild. Trone et al., (2005) interpreted the lack of behavioural changes and high prevalence of dolphin initiated interaction as suggestive that programme participation had no detrimental effect on the dolphin. In contrast, Frohoff and Packard (2015) identify a willingness by dolphin to initiate interaction with humans as assimilation, which is common in captivity.

Connor (2007) supports assimilation theory, evidencing that it can cause behavioural disturbances, as dolphin misread human behaviour and become frustrated. Assimilation is a more likely interpretation of the “play” seen, though this research was not available at the time. However, Schmitt et al., (2010) argue that humans can satisfy the fission fussion roles that are characteristic in dolphin social lives. Trone et al., (2005) findings imply interaction may be negative for welfare, however it is only an indication of situations where dolphins are involved in one trainer-controlled session a day and does not have wider application. Brensing et al., (2005) supported the former study by observing the behaviour of untrained individuals during dolphin-interaction sessions. Using non-trained dolphins indicated that the behaviour observed was self-motivated rather than the result of habituation. Data were collected over four years, enabling high statistical reliability, although environmental differences between the observation sites could account for differential reactions to the programmes (Chandler, 2012).  Findings could be the result of disturbance to the dolphin swim area.  This view is supported by Candelieri, Chiandetti and Cattaruzza (2013), who evidenced swim-with-dolphin programmes can cause suffering as result of environmental disturbance. Sew and Todd (2013) investigated the effect of HAI on dolphin in a robust study of behaviour, enclosure utilisation and social interaction, quantifying them to identify changes between interactions. Significant differences between individuals were seen and anticipatory behaviour was present. The study used a small observation set and observation quantities varied between individuals, which could reduce reliability as results may represent certain individuals’ reactions rather than a collective response.  The category of “play” was removed from analysis due to low incidences, however the lack of “play” could indicate distress, as evidenced by Paulos et al., (2010). Sew and Todd (2013) concluded that the participants’ welfare was not compromised by these programmes and this was supported by an increase in swimming and chasing behaviour that Henderson et al., (2012) and Jenson, Delfour and Carter (2013) describe as common play behaviour in the wild. Between observations there was a 10-minute gap that may have led to the loss of important information as the immediate pre and post effect of the swim with dolphins and meet with dolphins programmes were not recorded. Bejder et al., (2009) argued that if habituation was present in dolphins a change in behaviour may not be sensitive enough to identify a response to the stimulus and so may not have been observed.

Clegg et al., (2018) used sound cues to condition dolphin to positive HAI with a trainer and then measured anticipatory behaviour. Significantly higher frequencies of anticipatory behaviour before HAI resulted in the interaction being perceived as a reward event. The study was pseudo random, which decreased the likelihood of bias and multiple observations enabled a large data set and strong statistical analysis, however, there was variance in the history of the participants. The history is important as captive born or wild caught dolphin responses may differ, and individuals may seek human interaction as a result. The increase in anticipatory behaviour was concluded to have positive intrinsic value, however findings by Moe et al., (2006) suggest anticipatory behaviour is higher for negative experiences, though Yeates and Main (2008) and Clegg et al., (2017) argue that if anticipation is higher in negative states, then low participation would be seen. The findings have potential application to interaction programmes however Hill et al. (2016), Yeater et al., (2014) and Fabienne and Helen (2012) evidence that dolphins discriminate between familiar and unfamiliar humans and show preference for the former, which may indicate the anticipation was toward specific individuals.


2.2 Wild Dolphin 

Filby et al., (2014) investigated long-term responses of dolphins to swim with dolphin vessels from 1998-2013. Responses to swim with dolphin vessels changed across time, with avoidance and approach behaviours increasing significantly with cumulative experience. Greater levels of time and energy were spent avoiding, and approaching boats, potentially decreasing biological fitness by detracting from core biological activity.   A further study by Filby et al., (2017) looked at the effect of swim with dolphin vessels. Their findings showed that there was an effect on the behavioural activity budget. An increase in foraging, milling and socialising and reduction in foraging was seen. This study was carried out over three years, ensuring behavioural responses were steadily documented, increasing both reliability and statistical power. However socialising dolphins are more attractive to swim with dolphin vessels and operators target these groups, meaning the observed increase in social behaviour may be a result of observer bias. The paper was limited by excluding behaviours that occurred rarely. For example, “rest” was excluded from the results however work by Charlton-Robb et al., (2015) suggests that dolphins spend such limited time resting that any disturbance could be detrimental. By including this in the results it may have provided a more accurate representation of the effect of this programme. Though the short- term behavioural budget was significantly altered, the cumulative yearly budget was not, suggesting the programme may not have a detrimental effect on the targeted populations.

Behavioural changes were also observed by Perrtree et al., (2014) who investigated the percentage of sightings of dolphins involved in human interaction in Savannah over two years. The quantity of survey days changed by the year (44 surveys =2014, 25 surveys = 2015). This change may have impacted results as there were more data for one year than another. This means behaviour may have been missed and behaviour over the study years may not have been directly comparable. Begging was observed in 83 sightings and patrolling on 8 sightings, however begging on the ethogram was intermixed with patrolling, therefore levels of patrolling may have been lower or higher but may have been recorded as begging. This study varies from other literature in this area as it used a large sample size, producing stronger data and analysis. Increased aggression amongst conspecifics and towards humans were seen, with incidences of human injury and death; similar findings were documented by Cunningham-Smith et al., (2006).


2.3 Overview of dolphin-interaction literature  

Across the literature there is a theme of low participant quantity, resulting in low statistical power. Thus, caution should be taken when accepting the conclusions of the literature (Ellis, 2010).

Skewed sex ratios contribute to a lack of validity in studies; Samuels and Spradlin (1995) evidence that male dolphin are more likely to increase aggressive and sexual behaviour during non -controlled swim with dolphins sessions. Research highlights that there are differences in responses from the two sexes, however it is not considered in the discussion of the papers. Sex specific responses highlight the need for studies that target the appropriate species, sex and ages of the dolphin for accurate industry application (Marino, 2013).

The history of the participants is rarely explored in the literature but could be important to consider when evaluating findings as many captive dolphins were wild caught and there may be distinct differences in responses between captive born and wild caught dolphin (Marino, 2009). Research into dolphin-interaction programmes predominantly focuses on captive dolphin which is useful for management and welfare considerations, however captive results may be inappropriate for generalisation of the species (Trone et al., 2005).

Care should be taken not to misinterpret results, as when little or no effect is seen, it may be concluded that there is a lack of distress present from the programmes. However, this may be due to habituation, which may lead to inappropriate management decisions that could compromise wellbeing (Lussea, 2007).

There is a lack of diversity in the human participants that the dolphins interact with.  Most papers use the same age range and ethnicities of human participants. Thus, the results may be more indicative of responses to certain group types (Siniscalchi et al., 2012).

Methodologies vary between studies, although all of the studies use behavioural observations, the behaviours deemed important and the terminology to describe them differ (Marino and Frohoff, 2011).  Methodologies would produce more robust and valid results if the effect of participants, visitor presence and the effect of the programmes themselves were investigated, however this would be logistically difficult (Filby et al., 2014).

Due to the nature of the dolphin-interaction programmes, research varies between the type of interaction, e.g. whether they are controlled and involve familiar or unfamiliar humans. Due to the lack of uniformity, cross study comparison is unachievable (Scheer, 2010). The importance of longitudinal research in this area is highlighted throughout the literature, given behavioural responses detected in the long – term would not be detected in the short term. The need for comparability in this area is evident as HAI and the effect on behaviour in different species would inform collection managers and conservation organisations to help them know where to focus their efforts (Filby et al., 2017).


2.4 Comparison to Zoo Literature 

HAI in zoo literature appears in peer reviewed journals relatively few times and dolphin interactions even more infrequently. Zoo research often focuses on the effect of exhibit design and visitor presence on the behaviour of animals (Fernandez et al., 2009).  Methodology in this area is relatively standardised across different species, as most studies focus on the effect of interaction on the animals using behavioural observations. However, Miller et al., (2013) argues that the use of behavioural indicators does not provide a comprehensive and accurate understanding as no single biochemical or behavioural measure can be used to assess behaviour in response to interaction. Behaviour is a useful measure but can be misleading due to individual variation in stress reactivity, ability to cope and a lack of understanding of the mechanisms behind behaviour. This makes behaviour inappropriate to be used as the only indicator of negative or positive impact (Miller, 2011). A combination of physiological and behavioural responses would increase validity across studies. For example, Shepherdson et al., (2013) collected data from 55 bears housed across 20 zoos and used behavioural observations and faecal glucocorticoid levels for a year, which produced a reliable data set.

HAI studies are limited as there is a lack of consistent terminology across the field (Griffin et al., 2012). As the different disciplines use their own terminology there is a lack of agreed terms which means that findings may be interpreted differently depending on the nomenclature used in the paper. Furthermore, search engines do not always pick up relevant literature due to terminology variance (Griffin et al., 2012).   Throughout zoo and dolphin interaction literature it could be argued that there may be anthropomorphic interpretations of the results, for example behaviour associated with play or social attitude in dolphin may be disputed by ethological observations (Donaldson et al., 2012).

Zoo and dolphin interaction literature are similar as they focus on the humans’ perception of the experience rather than the effect on the animal and anecdotal results are often used to evidence the animal’s behaviour. However, this does not make for a reliable account of the behaviour (Carlstead, 2009).

Perceptual and cognitive misinterpretations need to be challenged to understand the effect of HAI on the animal’s wellbeing. Overall, the literature on zoo animals is similar to dolphin interactions but there is variance in the availability and quantity of research and due to the diversity of species, environments and terminology across studies it makes comparison and conclusions difficult. A standardised method of metrics should be produced to quantify HAI and allow application to the field, to enable research to be informative and meaningful (Christiansen et al., 2010).


3.0 Conclusion 

It is important to examine the short-term studies available into dolphin interaction and HAI and concurrently gain evidence into the behavioural changes that have occurred in response to HAI. The literature available highlights the importance of increasing long – term data sets in this area in order to increase the robustness and validity of the results gathered. Across the literature there are themes of low participant quantity, skewed sex ratio, one dimensional human participants and variable life histories which identifies a need for industry continuity. The terminology used within the subject matter requires quantification to ensure HAI and dolphin interactions are investigating unified responses. This will enable standardisation and uniformity across studies in terms of methodology, rationale and recording behaviour. In turn, this would enable comparability, analysis and review of the findings to gather well evidenced and accurate conclusions. It is necessary to create metrics to compare behavioural data against. This would allow for direct comparison, which may provide valuable data that can inform management and conservation efforts. The papers do not need a consensus on whether the behavioural responses should be deemed negative or positive to the individuals as results vary depending on the structure of the programmes, the appropriateness and consistency of methodologies and the collection and analysis of data. Further investigation is required to fully understand the impact of dolphin interaction programmes on the behaviour and welfare of dolphin and the effect of HAI on a broader range of zoo housed species whose daily routine involves interaction with familiar and unfamiliar humans.


Reference List

Allen, S.J., (2014) From exploitation to adoration. Whale-watching: Sustainable tourism and ecological management, p.31.

Bejder, L., Samuels, A., Whitehead, H., Finn, H. and Allen, S., (2009) Impact assessment research: use and misuse of habituation, sensitisation and tolerance in describing wildlife responses to anthropogenic stimuli. Marine Ecology Progress Series, 395, pp.177-185.

Bejder, L., Samuels, A.M.Y., Whitehead, H.A.L., Gales, N., Mann, J., Connor, R., Heithaus, M., WATSON‐CAPPS, J.A.N.A., Flaherty, C. and Kruetzen, M., (2006) Decline in relative abundance of bottlenose dolphins exposed to long‐term disturbance. Conservation Biology, 20(6), pp.1791-1798.

Brensing, K., Linke, K., Busch, M., Matthes, I. and van der Woude, S.E., (2005) Impact of different groups of swimmers on dolphins in swim-with-the-dolphin programs in two settings. Anthrozoös, 18(4), pp.409-429.

Candelieri, I., Chiandetti, C. and Cattaruzza, S., (2015) Healing dolphins? Cognitive and perceptual criticisms in Dolphin-Assisted Therapy.

Carlstead, K., (2009) A comparative approach to the study of keeper–animal relationships in the zoo. Zoo Biology: Published in affiliation with the American Zoo and Aquarium Association, 28(6), pp.589-608.

Chandler, C.K., (2012) Animal assisted therapy in counseling. Routledge.

Charlton-Robb, K., Taylor, A.C. and McKechnie, S.W., (2015) Population genetic structure of the Burrunan dolphin (Tursiops australis) in coastal waters of southeastern Australia: conservation implications. Conservation genetics, 16(1), pp.195207.

Chelluri, G.I., Ross, S.R. and Wagner, K.E., (2013) Behavioral correlates and welfare implications of informal interactions between caretakers and zoo-housed chimpanzees and gorillas. Applied Animal Behaviour Science, 147(3-4), pp.306-315.

Christiansen, F., Lusseau, D., Stensland, E. and Berggren, P., (2010) Effects of tourist boats on the behaviour of Indo-Pacific bottlenose dolphins off the south coast of Zanzibar. Endangered Species Research, 11(1), pp.91-99.

Clark, F.E., Davies, S.L., Madigan, A.W., Warner, A.J. and Kuczaj, S.A., (2013) Cognitive enrichment for bottlenose dolphins (Tursiops truncatus): Evaluation of a novel underwater maze device. Zoo Biology, 32(6), pp.608-619.

Claxton, A.M., (2011) The potential of the human–animal relationship as an environmental enrichment for the welfare of zoo-housed animals. Applied Animal Behaviour Science, 133(1-2), pp.1-10.

Clayton, S., Fraser, J. and Saunders, C.D., (2009) Zoo experiences: Conversations, connections, and concern for animals. Zoo Biology: Published in affiliation with the American Zoo and Aquarium Association, 28(5), pp.377-397.

Clegg, I.L., Rödel, H.G. and Delfour, F., (2017) Bottlenose dolphins engaging in more social affiliative behaviour judge ambiguous cues more optimistically. Behavioural brain research, 322, pp.115-122.

Clegg, I.L.K., Van Elk, C.E. and Delfour, F., (2017) Applying welfare science to bottlenose dolphins (Tursiops truncatus). Animal Welfare, 26(2), pp.165-176.

Cunningham-Smith, P., Colbert, D.E., Wells, R.S. and Speakman, T., (2006) Evaluation of human interactions with a provisioned wild bottlenose dolphin (Tursiops truncatus) near Sarasota Bay, Florida, and efforts to curtail the interactions. Aquatic Mammals, 32(3), p.346.

Donaldson, R., Finn, H., Bejder, L., Lusseau, D. and Calver, M., (2012) The social side of human–wildlife interaction: wildlife can learn harmful behaviours from each other. Animal Conservation, 15(5), pp.427-435.

Ellis, Paul D. The essential guide to effect sizes: Statistical power, meta-analysis, and the interpretation of research results. Cambridge University Press, 2010.

Fabienne, D. and Helen, B., (2012) Assessing the effectiveness of environmental enrichment in bottlenose dolphins (Tursiops truncatus). Zoo biology, 31(2), pp.137150.

Fernandez, E.J., Tamborski, M.A., Pickens, S.R. and Timberlake, W., (2009) Animal– visitor interactions in the modern zoo: Conflicts and interventions. Applied Animal Behaviour Science, 120(1-2), pp.1-8.

Filby, N.E., Christiansen, F., Scarpaci, C. and Stockin, K.A., (2017) Effects of swimwith-dolphin tourism on the behaviour of a threatened species, the Burrunan dolphin Tursiops australis. Endangered Species Research, 32, pp.479-490.

Filby, N.E., Stockin, K.A. and Scarpaci, C., (2014) Long-term responses of Burrunan dolphins (Tursiops australis) to swim-with dolphin tourism in Port Phillip Bay, Victoria, Australia: A population at risk. Global Ecology and Conservation, 2, pp.62-71.

Frohoff, T.G. and Packard, J.M., (1995) Human interactions with free-ranging and captive bottlenose dolphins. Anthrozoös, 8(1), pp.44-53.

Hediger, H., (2013) Wild animals in captivity. Butterworth-Heinemann.

Henderson, E.E., Hildebrand, J.A., Smith, M.H. and Falcone, E.A., (2012) The behavioral context of common dolphin (Delphinus sp.) vocalizations. Marine Mammal Science, 28(3), pp.439-460.

Hill, H.M., Yeater, D., Gallup, S., Guarino, S., Lacy, S., Dees, T. and Kuczaj, S., (2016) Responses to familiar and unfamiliar humans by belugas (Delphinapterus leucas), bottlenose dolphins (Tursiops truncatus), & Pacific white-sided dolphins (Lagenorhynchus obliquidens): a replication and extension. International Journal of Comparative Psychology, 29(1).

Hosey, G. and Melfi, V., (2014) Human-animal interactions, relationships and bonds: a review and analysis of the literature. International Journal of Comparative Psychology, 27(1).

Janik, V.M., (2015) Play in dolphins. Current Biology, 25(1), pp. R7-R8

Jensen, A.L.M., Delfour, F. and Carter, T., (2013) Anticipatory behavior in captive bottlenose dolphins (Tursiops truncatus): a preliminary study. Zoo Biology, 32(4), pp.436-444.

Lusseau, D. and Bejder, L., (2007) The long-term consequences of short-term responses to disturbance experiences from whale watching impact assessment. International Journal of Comparative Psychology, 20(2).

Lusseau, D., (2007) Evidence for social role in a dolphin social network. Evolutionary ecology, 21(3), pp.357-366.

Marino, L. and Frohoff, T., (2011) Towards a new paradigm of non-captive research on cetacean cognition. PloS one, 6(9), p. e24121.

Mason, G.J. and Latham, N., (2004) Can’t stop, won’t stop: is stereotypy a reliable animal welfare indicator?

Miller, L.J., Mellen, J., Greer, T. and Kuczaj, S.A., (2011) The effects of education programmes on Atlantic bottlenose dolphin (Tursiops truncatus) behaviour. Animal Welfare, 20(2), pp.159-172.

Miller, L.J., Zeigler‐Hill, V., Mellen, J., Koeppel, J., Greer, T. and Kuczaj, S., (2013) Dolphin shows and interaction programs: benefits for conservation education? Zoo biology, 32(1), pp.45-53.

Mirimin, L., Miller, R., Dillane, E., Berrow, S.D., Ingram, S., Cross, T.F. and Rogan, E., (2011) Fine‐scale population genetic structuring of bottlenose dolphins in Irish coastal waters. Animal Conservation, 14(4), pp.342-353.

Moe, R.O., Bakken, M., Kittilsen, S., Kingsley-Smith, H. and Spruijt, B.M., (2006) A note on reward-related behaviour and emotional expressions in farmed silver foxes (Vulpes vulpes)—basis for a novel tool to study animal welfare. Applied Animal Behaviour Science, 101(3-4), pp.362-368.

Montgomery, S.H., Geisler, J.H., McGowen, M.R., Fox, C., Marino, L. and Gatesy, J., (2013) The evolutionary history of cetacean brain and body size. Evolution, 67(11), pp.3339-3353.

Myers Jr, O.E., Saunders, C.D. and Birjulin, A.A., (2004) Emotional dimensions of watching zoo animals: An experience sampling study building on insights from psychology. Curator: The Museum Journal, 47(3), pp.299-321.

Paulos, R.D., Trone, M., Kuczaj, I.I. and Stan, A., (201). Play in wild and captive cetaceans. International Journal of Comparative Psychology, 23(4).

Perrtree, R.M., Kovacs, C.J. and Cox, T.M., (2014) Standardization and application of metrics to quantify human‐interaction behaviors by the bottlenose dolphin (Tursiops spp.). Marine Mammal Science, 30(4), pp.1320-1334.

Samuels, A. and Spradlin, T.R., (1995) Quantitative behavioral study of bottlenose dolphins in swim‐with‐dolphin programs in the United States. Marine Mammal Science, 11(4), pp.520-544.

Scheer, M., (2010) Review of self-initiated behaviors of free-ranging cetaceans directed towards human swimmers and waders during open water encounters. Interaction Studies, 11(3), pp.442-466.

Schmitt, T.L., St. Aubin, D.J., Schaefer, A.M. and Dunn, J.L., (2010) Baseline, diurnal variations, and stress‐induced changes of stress hormones in three captive beluga whales, Delphinapterus leucas. Marine Mammal Science, 26(3), pp.635-647.

Sew, G. and Todd, P.A., (2013) The effects of human-dolphin interaction programmes on the behaviour of three captive Indopacific humpback dolphins (Sousa chinensis). Raffles Bulletin of Zoology, 61(1).

Shepherdson, D., Lewis, K.D., Carlstead, K., Bauman, J. and Perrin, N., (2013) Individual and environmental factors associated with stereotypic behavior and faecal glucocorticoid metabolite levels in zoo housed polar bears. Applied Animal Behaviour Science, 147(3-4), pp.268-277.

Siniscalchi, M., Dimatteo, S., Pepe, A.M., Sasso, R. and Quaranta, A., (2012) Visual lateralization in wild striped dolphins (Stenella coeruleoalba) in response to stimuli with different degrees of familiarity. PLoS One, 7(1), p. e30001

Smith, B.D., Tun, M.T., Chit, A.M., Win, H. and Moe, T., (2009) Catch composition and conservation management of a human–dolphin cooperative cast-net fishery in the Ayeyarwady River, Myanmar. Biological Conservation, 142(5), pp.1042-1049.

Trone, M., Kuczaj, S. and Solangi, M., (2005) Does participation in Dolphin–Human Interaction Programs affect bottlenose dolphin behaviour? Applied Animal Behaviour Science, 93(3-4), pp.363-374.

Vining, J., (2003) The connection to other animals and caring for nature. Human Ecology Review, pp.87-99.

Yeater, D.B., Hill, H.M., Baus, N., Farnell, H. and Kuczaj, S.A., (2014) Visual laterality in belugas (Delphinapterus leucas) and Pacific white-sided dolphins (Lagenorhynchus obliquidens) when viewing familiar and unfamiliar humans. Animal cognition, 17(6), pp.1245-1259.

Yeates, J.W. and Main, D.C., (2008) Assessment of positive welfare: a review. The Veterinary Journal, 175(3), pp.293-300

The ability of dogs to follow human communicative cues

Author Name: Emma Walters, BSc (Hons) Animal Behaviour and Welfare



Humans and dogs share a unique relationship, where the dogs’ ability to interpret human communicative skills outperform both humans’ and dogs’ closest living ancestors. Dogs are particularly skilled at following human pointing gestures. Researchers question whether this skill developed due to domestication or learning during ontogeny. Object-choice tasks are used to assess the dogs’ ability to follow human pointing. However, differences between methods used in the studies mean results cannot be compared and some are biased, questioning the validity of the hypothesis they support. The procedure used to carry out object-choice tasks should be consistent, as differences may affect motivation and performance and different species may respond differently. The type of point may also influence results through varying the levels of difficulty, and easier points may cause bias through local enhancement. Momentary distal points may be appropriate as a baseline measure, as this controls for local enhancement and may be more representative of a communicative interaction. When comparing different species, rearing backgrounds should be similar to ensure comparable results and control for performance bias caused by previous training, or fear. Stress should be considered during testing, as this may affect the subjects’ performance and measuring stress levels during testing may be beneficial. When analysing data, individual and group results should be included for subjects’ performance and ability to follow specific gestures, so reliable results are obtained. Ensuring object-choice tasks use the same procedure between studies and provide clear, detailed methods can ensure comparable results, minimise bias and ensure tests are repeatable.


1.0 Introduction

Dogs (Canis lupus familiaris) have developed remarkable abilities to use human social behaviour and communication (Miklósi, 2009; Phillips Buttner, 2016). Dogs have been found able to interpret human emotional (Buttelman and Tomasello, 2013; Nagasawa et al. 2011) and attentional states (Call et al. 2003; Virányi et al. 2004), comprehend human visual perception (Kaminski et al. 2009; MacLean, Krupenye and Hare, 2014) and learn socially from copying human behaviour (Tópal et al. 2006). Dogs are particularly skilled at following human ostensive signals (e.g. pointing) (Riedel et al. 2008; Gácsi et al. 2009a), showing similar performance to human infants (aged 2 years old) (Lakatos et al. 2009). Dogs outperform humans’ closest living ancestors, chimpanzees (Pan troglodytes) (MacLean et al. 2017) and wolves (Canis lupus) (Viryáni et al. 2008). This questions how a distantly related species developed social skills comparable to humans (Elgier et al. 2009; Hare, 2017). Researchers have suggested domestication enabled this ability (Hare et al. 2010; Wobbler et al. 2009), though what part of the domestication process, or if domestication at all is involved, is debated (Wynne, Udell and Lord, 2008; Helton and Helton, 2010; Range and Viryáni, 2015; Udell et al. 2014).


2.0 Critical Review

2.1 Communication

The ethological approach to communication defines it as a sender altering its behaviour to elicit a response from the receiver, benefiting sender and receiver, or just sender (Krebs and Davies, 1993; Miklósi, 2009). Visual communication (e.g. pointing, gaze) has been used to test dogs’ ability to follow human communicative cues (Reid, 2009). This review focuses on the ability to follow human pointing. Pointing is a behaviour unique to humans (Miklósi and Soproni, 2006; Tauzin et al. 2015). For a different species to be able to interpret human pointing suggests a form of interspecific communication (Miklósi and Soproni, 2006).


2.1.1 Hypotheses for dogs’ socio-cognitive abilities

Conflicting hypotheses question how dogs developed their socio-cognitive abilities to interpret human communicative cues. The domestication hypothesis argues that it developed through direct and/or indirect selection during domestication (Kaminski and Nitzschner, 2013); while the two-stage hypothesis claims domestication is not necessary for these skills to develop and they are learned from sharing a human environment during ontogeny (Udell, Dorey and Wynne, 2010a); and the synergistic hypothesis maintains learning during ontogeny and domestication enable dogs’ abilities (Gácsi et al. 2009c). Evidence supporting these hypotheses is based on the reliability and validity of the methodology used to assess them (Pongrácz et al. 2013), yet methodologies vary between studies, meaning results may be biased (Reid, 2009; Miklósi and Topál, 2011).


2.2 Object-choice tasks

Object-choice tasks (OCT) assess the ability of a subject to follow human pointing (Reid, 2009). The general design of an OCT is to have 2 opaque containers where one is baited (usually with food), and the other is sham-baited. Subjects are aware that a reward is hidden but not its location. The experimenter gestures to the baited container and the subject chooses between the containers (see figure 1) (Oliva et al. 2015). The subject is only rewarded if they choose correctly. For a subject to be considered as interpreting the signal, they must perform at above chance levels (e.g. 15 out of 20 correct trials) (Udell, Dorey and Wynne 2010).


Figure 1 EW

Figure 1: An example of an OCT set-up (Pet Behaviour Science, 2016).


2.2.1 Procedure

For most OCTs the reward is hidden, and subjects are aware of this (Hare et al. 2010). One method uses a different procedure where a container is pointed at, but no bait is hidden. If the subject chooses correctly the experimenter drops the reward (Udell, Dorey and Wynne, 2008; 2010a; Zaine, Domeniconi and Wynne, 2015). This approach is used to account for the subjects using olfactory cues (Udell, Dorey and Wynne, 2008). Although this procedure follows the same principle of following a human point to obtain a reward, the motivation for doing so may differ, e.g. following the point to a correct location then being rewarded, compared to locating the hidden reward. In addition, as the subject does not have to find food indicated by pointing there is no link between the reward and the pointing gesture (Pongrácz et al. 2013). Additionally, control trials in previous studies have found olfactory cues do not influence choice (Riedel et al. 2008; Viryáni et al. 2008). A pre-trial is required for this design, whereby the experimenter gains the subject’s attention, then places the reward on a container, repeating both sides until the subject approaches. This builds association with the reward and container and helps control for neophobia (Dorey, Udell and Wynne, 2010; Udell, Dorey and Wynne, 2008). However, this may reinforce an association between the hand and container, potentially influencing performance (Pongrácz et al. 2013).

Furthermore, some studies use clickers to mark correct choices (Udell, Dorey and Wynne, 2008). Clickers are secondary reinforcers (Wilson et al. 2017) used to mark desired behaviour (Feng, Howell and Bennett, 2017). Clicking may reinforce approach behaviour instead of interpreting the pointing gesture, potentially affecting performance (Gácsi et al. 2009c; Pongrácz et al. 2013). These methods have been used to argue for the two-stage hypothesis, but this may be biased for the reasons stated above. Pongrácz et al. (2013) tested these confounding factors, finding no significant improvement for using a clicker alone, even if dogs were clicker trained. Instead, they suggest the clicker, dropping food, and methods used in pre-trials combined may influence performance. Furthermore, testing these variables on wolves’ performance may be interesting, as Udell, Dorey and Wynne (2008) found wolves performed better using this method, although factors such as previous training experience may have influenced this (Hare et al. 2010). This highlights the need for consistency amongst OCTs, to control for potential influences and ensure comparable results.


2.2.2 Type of point

Different points have been used for different studies, potentially influencing results due to varying degrees of difficulty as they can be easier or more difficult to interpret (Miklósi and Soproni, 2006; Viryáni et al. 2008). This means studies using different points may be incomparable and results from studies using easier points may be biased (Gacsi et al. 2009c). Proximal (fingertip is 10-40cm away from the container), dynamic (subject views point occurring, remaining in place until the subject choses), and static points (the point is in place when the subject approaches, remaining until the subject chooses) may cause bias through local enhancement, i.e. the physical association between the point and food location (Miklósi and Soproni, 2006; Viryáni et al. 2008), or the subject approaching the closest part of the experimenter (Reid, 2009). Momentary (point lasts no more than 2 seconds and ends before the subject can choose) distal (container is more than 50cm away from fingertip) points (MDPs) may be most suitable as baseline points for OCTs when assessing the communicative ability to follow pointing. Although the subject may be influenced by the movement of the point, local enhancement is controlled for due to the distance of the point to the container (Viryáni et al. 2008). Furthermore, as the subject must remember the signal before choosing, it is more in line with communicative interaction, where a sender emits a short discrete signal (Miklósi and Soproni, 2006; Viryáni et al. 2008). By setting baseline measures, it may be easier to compare studies assessing different factors more reliably (Viryáni et al. 2008). Gaze and head turning should also be considered during pointing tasks (e.g. looking at subject, not looking at the subject) (Udell, Dorey and Wynne, 2010b; Cunningham and Ramos, 2014) as this can also influence performance, as dogs appear sensitive to human attentional states and whether the cue is intended for them (Kaminski et al. 2012). This highlights the importance of providing clear methods, to ensure that studies are repeatable and comparable (Miklósi and Soproni, 2006; Hare et al. 2010; Udell, Dorey and Wynne, 2010b).


2.2.3 Subjects’ history

To assess whether domestication impacted dogs’ ability to follow human communicative gestures, it might be useful to consider whether their wild ancestors, wolves, can perform this task (Viryáni et al. 2008; Miklósi and Topál, 2011). However, dogs may have an advantage over wolves in their ability to interpret human communicative gestures as living in an anthropogenic environment gives them the opportunity to learn about human social behaviour (Gacsi et al. 2013; Viryáni and Range, 2014, p.35). In addition, if wolves and dogs share genes which emerge indirectly from environmental factors (e.g. exposure to human gestures), different rearing conditions may mean that these are expressed differently, potentially affecting performance (Miklósi and Topál, 2011). Similarly, performance bias may occur if subjects are fearful due to not being socialised to humans, or human environments (Viryáni et al. 2008), or if subjects have already learned similar tasks (Hare et al, 2010). To compare the communicative skills of wolves and dogs, they should be reared similarly, with exposure to humans to control for fear responses.

Viryáni et al (2008) raised 9 wolves and 8 puppies under an intensive socialisation programme, raised by humans without their mothers and littermates shortly after birth (4-7 days wolves, 4-10 days puppies). Both had the same upbringing until 2 months, when the wolves moved into a wolf park, receiving 2 days a week of contact with their carer and daily contact with other people at the park responsible for their husbandry, and the puppies remained with their human. These groups were compared to 9 pet dogs in an OCT using MDP. All subjects were 4 months at the time of testing. None of the wolves performed above chance in the OCT, but 4 pet dogs and 3 hand-raised puppies were successful. This suggests that the ability to follow human pointing cannot be attributed to learning during ontogeny, supporting the domestication hypothesis. Although, at 2 months old the wolves were raised differently, potentially giving the hand-raised puppies an advantage due to spending extra time in human social environments (Udell and Wynne, 2010). However, Gácsi et al. (2009c) compared pet dogs to hand-raised wolves, kept until 3-4 months of age, using MDP. Both were 4 months when tested. They found wolves did not perform above chance but dogs did, supporting Viryáni et al.‘s (2008) findings.

Breed, age and background should also be considered when comparing human-dog communication, to compare how experience with humans, breed traits and development affect dogs’ performance in OCTs (Gácsi et al. 2009b; Riedel et al. 2008; Lazarowski and Dorman, 2015). Studies have compared shelter and pet dogs to assess how experience with humans affects performance (Udell, Dorey and Wynne, 2010a; Barrera et al. 2015). Poor performance of shelter dogs is thought to indicate that learning from experience around humans is a factor in dogs’ social-cognitive abilities (Udell, Dorey and Wynne, 2008; Zaine, Domeniconi and Wynne, 2015). However, caution should be taken when interpreting these results. Shelters are stressful environments (Barnard et al., 2015; Protopopova, 2016) and stress may affect dogs’ cognitive abilities during OCTs (Phillips Buttner, 2016). Furthermore, dogs may be relinquished to shelters due to behavioural problems (Diesel, Brodbert and Pfeiffer, 2009; Kwan and Bain, 2013), or cognitive impairments (Udell et al. 2010; Miklósi and Topal, 2011).

Zaine, Domeniconi and Wynne (2015) compared puppies at a shelter aged 8-12 weeks to those aged 20-24 weeks, finding that younger puppies outperformed older. This did not support their argument for the two-stage hypothesis. Upon further investigation they found that 18 out of 20 younger puppies were given up by their owner, compared to 5 out of 20 older puppies, suggesting that prior experience in a human environment affected younger dogs’ ability. However, this conclusion alone is an assumption as the dogs’ exact rearing history cannot be known (Cunningham and Ramos, 2014; Lazarowski and Dorman, 2015). Another explanation for this result could be due to the socialisation period for puppies which is between 3-12 weeks (Case, 2005, pp. 149-150; Landsberg, Hunthausen and Ackerman, 2013, p.15). During this period puppies are less fearful of new experiences and situations (Landsberg, Hunthausen and Ackerman, 2013, p.15). Due to this reduction in fear, the younger puppies would likely be less stressed than older puppies, who had finished their socialisation period (Phillips Buttner, 2016), supporting the suggestion of stress affecting performance. To investigate the impact stress may have upon performance, experiments could include an independent observer assessing stress-related behaviours.


2.2.4 Statistical analysis

When analysing the results of OCTs for several subjects, individual and group performance should be assessed. Results from group performance may indicate overall success even if only a few individuals perform above chance, or most individuals score just above 50% (random choice) as suggested by Gácsi et al. (2009a). Grouping different gestures together should also be considered. Wynne, Udell and Lord (2008) re-analysed data from Riedel et al. (2008). By grouping together 3 different gestures they found a significant effect on performance, despite the group only performing above chance for one of the cues (Riedel et al. 2008; Hare et al. 2010). Differences have also been found between studies when including no-choices (subject does not choose a container) as an incorrect choice during analysis. Hare et al. (2010) re-analysed data from Udell, Dorey and Wynne (2008) who originally included no-choices as incorrect. When these tests were not included in the statistical analysis, results changed from not significant to significant. No-choices may be related to the subject experiencing fatigue or lack of motivation instead of an inability to interpret the cue (Hare et al. 2010). Interestingly most of the subjects scoring no-choice were shelter dogs, which may indicate stress affecting performance. However, it is debatable as to what constitutes an incorrect choice over a no-choice. Udell et al. (2014) found that one no-choice response was to approach the experimenter. However, in this experiment no food was hidden, but the experimenter dropped the food for the correct choice. Although an incorrect choice was made, the subjects may have approached the experimenter for food, potentially being a previously reinforced behaviour (e.g. begging for food from someone known to have food) (Udell and Wynne, 2010). This highlights the importance of detailed methods and the impact previous experience may have on performance.


3.0 Conclusion

Interspecific communication between humans and dogs has been identified through their ability to follow human pointing. OCTs offer insight into how dogs developed this skill. Differences between methodologies and limitations of OCTs may cause biased results questioning the validity of the hypothesis they support and making studies incomparable. Procedures used by studies should remain the same to control for differences in motivation, performance and between species. MDPs were suggested as a baseline measure when testing point following as this controls for local enhancement. Comparing dogs and wolves may offer insight into the role domestication or living in an anthropogenic environment has on the ability to follow human pointing. However, caution should be taken to ensure that subjects have similar rearing backgrounds to obtain comparable results. Furthermore, subjects’ history, age and breed should be considered. Stress may influence performance, so monitoring stress-related behaviours during trials may be beneficial. Performance and cues should be analysed on an individual level, as grouping subjects together may influence results. Authors should clearly state how they assess correct and incorrect choices. No-choices should be clearly defined as they may not indicate misinterpretation of the pointing cue. Future OCTs should consider the methods used and include detailed methodologies to minimise bias and ensure tests are repeatable and comparable.


 Reference List

Barnard, S. et al. (2015) ‘Development of a new welfare assessment protocol for practical application in long-term dog shelters’, Veterinary Record, 178(1), pp. 18-18. doi: 10.1136/vr.103336

Barrera, G. et al. (2015) ‘Effects of learning on social and nonsocial behaviors during a problem-solving task in shelter and pet dogs’, Journal of Veterinary Behavior, 10(4), pp. 307–314. doi: 10.1016/J.JVEB.2015.03.005.

Buttelmann, D. and Tomasello, M. (2013) ‘Can domestic dogs (Canis familiaris) use referential emotional expressions to locate hidden food?’, Animal Cognition, 16(1), pp. 137–145. doi: 10.1007/s10071-012-0560-4.

Call, J. et al. (2003) ‘Domestic Dogs (Canis familiaris) Are Sensitive to the Attentional State of Humans’. Journal of comparative psychology, 117(3), p.257. doi: 10.1037/0735-7036.117.3.257.

Case. L. (2005) The Dog: Its Behaviour, Nutrition and Health. Iowa: Blackwell Publishing Company.

Cunningham, C. L. and Ramos, M. F. (2014) ‘Effect of training and familiarity on responsiveness to human cues in domestic dogs (Canis familiaris)’, Animal Cognition, 17(3), pp. 805–814. doi: 10.1007/s10071-013-0714-z.

Diesel, G., Brodbelt, D. and Pfeiffer, D. U. (2009) ‘Characteristics of relinquished dogs and their owners at 14 rehoming centers in the United Kingdom’, Journal of Applied Animal Welfare Science, 13(1), pp. 15–30. doi: 10.1080/10888700903369255.

Dorey, N. R., Udell, M. A. R. and Wynne, C. D. L. (2010) ‘When do domestic dogs, Canis familiaris, start to understand human pointing? The role of ontogeny in the development of interspecies communication’, Animal Behaviour, 79(1), pp. 37–41. doi: 10.1016/J.ANBEHAV.2009.09.032.

Elgier, A. M. et al. (2009) ‘Communication between domestic dogs (Canis familiaris) and humans: Dogs are good learners’, Behavioural Processes, 81, pp. 402–408. doi: 10.1016/j.beproc.2009.03.017.

Feng, L. C., Howell, T. J. and Bennett, P. C. (2017) ‘Comparing trainers’ reports of clicker use to the use of clickers in applied research studies: methodological differences may explain conflicting results’, Pet Behaviour Science, (3), p. 1. doi: 10.21071/pbs.v0i3.5786.

Gácsi, M. et al. (2009a) ‘The effect of development and individual differences in pointing comprehension of dogs’. Animal cognition, 12(3), pp.471-479. doi: 10.1007/s10071-008-0208-6.

Gácsi, M. et al. (2009b) ‘Effects of selection for cooperation and attention in dogs’, Behavioural and Brain Functions. 5(1), p.31. doi: 10.1186/1744-9081-5-31.

Gácsi, M. et al. (2009c) ‘Explaining Dog Wolf Differences in Utilizing Human Pointing Gestures: Selection for Synergistic Shifts in the Development of Some Social Skills’, PLoS ONE.  4(8), p. e6584. doi: 10.1371/journal.pone.0006584.

Gácsi, M. et al. (2013) ‘Wolves do not join the dance: Sophisticated aggression control by adjusting to human social signals in dogs’, Applied Animal Behaviour Science. doi: 10.1016/j.applanim.2013.02.007.

Hare, B. (2017) ‘Survival of the Friendliest: Homo sapiens Evolved via Selection for Prosociality’. Annual review of psychology, 68, pp. 155-186. doi: 10.1146/annurev-psych-010416-044201.

Hare, B. et al. (2010) ‘The domestication hypothesis for dogs skills with human communication: a response to Udell et al. (2008) and Wynne et al. (2008)’, Animal Behaviour, 79, pp. e1–e6. doi: 10.1016/j.anbehav.2009.06.031.

Helton, W. S. and Helton, N. D. (2010) ‘Physical size matters in the domestic dog’s (Canis lupus familiaris) ability to use human pointing cues’, Behavioural Processes, 85(1), pp. 77–79. doi: 10.1016/J.BEPROC.2010.05.008.

Kaminski, J. and Nitzschner, M. (2013) ‘Do dogs get the point? A review of dog–human communication ability’. doi: 10.1016/j.lmot.2013.05.001.

Kaminski, J. et al. (2009) ‘Domestic dogs are sensitive to a human’s perspective’, Behaviour. 146(7), pp. 979-998. doi: 10.1163/156853908X395530.

Kaminski, J., Schulz, L. and Tomasello, M. (2012) ‘How dogs know when communication is intended for them’, Developmental Science, 15(2), pp. 222–232. doi: 10.1111/j.1467-7687.2011.01120.x.

Krebs, J. R., Davies, N.B. (1993) Parental care and mating systems. In: Krebs, J.R., Davies, N.B. (1993) An Introduction to Behavioural Ecology. London: Blackwell, pp.  208-243.

Kwan, J. Y. and Bain, M. J. (2013) ‘Owner Attachment and Problem Behaviors Related to Relinquishment and Training Techniques of Dogs’, Journal of Applied Animal Welfare Science, 16(2), pp. 168–183. doi: 10.1080/10888705.2013.768923.

Lakatos, G. et al. (2009) ‘A comparative approach to dogs’ (Canis familiaris) and human infants’ comprehension of various forms of pointing gestures’, Animal Cognition, 12(4), pp. 621–631. doi: 10.1007/s10071-009-0221-4.

Landsberg, G., Hunthausen, W. and Ackerman, L. (2013) Behaviour Problems of the Dog and Cat. Edinburgh: Elsevier.

Lazarowski, L. and Dorman, D. C. (2015) ‘A comparison of pet and purpose-bred research dog (Canis familiaris) performance on human-guided object-choice tasks’, Behavioural Processes, 110, pp. 60–67. doi: 10.1016/J.BEPROC.2014.09.021.

Maclean, E. L. et al. (2017) ‘Individual differences in cooperative communicative skills are more similar between dogs and humans than chimpanzees’, Animal Behaviour, 126, pp. 41-51. doi: 10.1016/j.anbehav.2017.01.005.

Maclean, E. L., Krupenye, C. and Hare, B. (2014) ‘Dogs (Canis familiaris) Account for Body Orientation but Not Visual Barriers When Responding to Pointing Gestures’, Journal of Comparative Psychology, 128(3), p.285. doi: 10.1037/a0035742.

Miklósi, Á. (2009) ‘Evolutionary approach to communication between humans and dogs’, Veterinary Research Communications, 33(S1), pp. 53–59. doi: 10.1007/s11259-009-9248-x.

Miklósi, Á. and Soproni, K. (2006) ‘A comparative analysis of animals’ understanding of the human pointing gesture’, Animal Cognition, 9(2), pp. 81–93. doi: 10.1007/s10071-005-0008-1.

Miklósi, Á. and Topál, J. (2011) ‘On the hunt for the gene of perspective taking: pitfalls in methodology’, Learning & Behavior, 39(4), pp. 310–313. doi: 10.3758/s13420-011-0038-2.

Nagasawa, M. et al. (2011) ‘Dogs can discriminate human smiling faces from blank expressions’, Animal Cognition, 14(4), pp. 525–533. doi: 10.1007/s10071-011-0386-5.

Oliva, J. L. et al. (2015) ‘Oxytocin enhances the appropriate use of human social cues by the domestic dog (Canis familiaris) in an object choice task’, Animal Cognition, 18(3), pp. 767–775. doi: 10.1007/s10071-015-0843-7.

Pet Behaviour Science (2016) Popular Science Notes. Available at: http://petbehaviourscience.org/notes/?p=41 (Accessed: 23 November 2018)

Phillips Buttner, A. (2016) ‘Neurobiological underpinnings of dogs’ human-like social competence: How interactions between stress response systems and oxytocin mediate dogs’ social skills’, Neuroscience and Biobehavioral Reviews, 71, pp. 198–214. doi: 10.1016/j.neubiorev.2016.08.029.

Pongrácz, P. et al. (2013) ‘Test sensitivity is important for detecting variability in pointing comprehension in canines’, Animal Cognition, 16(5), pp. 721–735. doi: 10.1007/s10071-013-0607-1.

Protopopova, A. (2016) ‘Effects of sheltering on physiology, immune function, behavior, and the welfare of dogs’, Physiology & Behavior, 159, pp. 95–103. doi: 10.1016/J.PHYSBEH.2016.03.020.

Range, F. and Virányi, Z. (2014) ‘Tracking the evolutionary origins of dog-human cooperation: the “Canine Cooperation Hypothesis”’, Frontiers in psychology, 5, p. 1582. doi: 10.3389/fpsyg.2014.01582.

Reid, P. J. (2009) ‘Adapting to the human world: Dogs’ responsiveness to our social cues’, Behavioural Processes, 80(3), pp. 325–333. doi: 10.1016/J.BEPROC.2008.11.002.

Riedel, J. et al. (2008) ‘The early ontogeny of human-dog communication’, Animal Behaviour, 75(3), pp. 1003-1014. doi: 10.1016/j.anbehav.2007.08.010.

Tauzin, T. et al. (2015) ‘What or Where? The Meaning of Referential Human Pointing for Dogs (Canis familiaris)’. Journal of Comparative Psychology, 129(4), pp.334. doi: 10.1037/a0039462.

Topál, J. et al. (2006) ‘Reproducing human actions and action sequences: “Do as I Do!” in a dog’, Animal Cognition, 9(4), pp. 355–367. doi: 10.1007/s10071-006-0051-6.

Udell, M. A. R. and Wynne, C. D. L. (2010) ‘Ontogeny and phylogeny: both are essential to human-sensitive behaviour in the genus Canis’, Animal Behaviour, 79, pp. e9–e14. doi: 10.1016/j.anbehav.2009.11.033.

Udell, M. A. R. et al. (2014) ‘Exploring breed differences in dogs (Canis familiaris): does exaggeration or inhibition of predatory response predict performance on human-guided tasks?’, Animal Behaviour, 89, pp. 99–105. doi: 10.1016/J.ANBEHAV.2013.12.012.

Udell, M. A. R., Dorey, N. R. and Wynne, C. D. L. (2010) ‘The performance of stray dogs (Canis familiaris) living in a shelter on human-guided object-choice tasks’, Animal Behaviour, 79(3), pp. 717–725. doi: 10.1016/J.ANBEHAV.2009.12.027.

Udell, M. A. R., Dorey, N. R. and Wynne, C. D. L. (2010a) ‘The performance of stray dogs (Canis familiaris) living in a shelter on human-guided object-choice tasks’, Animal Behaviour, 79, pp. 717–725. doi: 10.1016/j.anbehav.2009.12.027.

Udell, M. A. R., Dorey, N. R. and Wynne, C. D. L. (2010b) ‘What did domestication do to dogs? A new account of dogs’ sensitivity to human actions’, Biological Reviews, 85(2), pp. 327–345. doi: 10.1111/j.1469-185X.2009.00104.x.

Udell, M., Dorey, N. R. and Wynne, C. (2008) ‘Wolves outperform dogs in following human social cues’, Animal Behaviour, 76, pp. 1767–1773. doi: 10.1016/j.anbehav.2008.07.028.

Virányi, Z. et al. (2004) ‘Dogs respond appropriately to cues of humans’ attentional focus’, Behavioural Processes, 66, pp. 161–172. doi: 10.1016/j.beproc.2004.01.012.

Virányi, Z. et al. (2008) ‘Comprehension of human pointing gestures in young human-reared wolves (Canis lupus) and dogs (Canis familiaris)’, Animal Cognition, 11(3), pp. 373–387. doi: 10.1007/s10071-007-0127-y.

Viryani, Z. and Range, F. (2014) ‘On the Way to a Better Understanding of Dog Domestication: Aggression and Cooperativeness in Dogs and Wolves’, ‘in’ Kaminski, J. and Marshall-Pescini, S. The social Dog. Behaviour and Cognition. San Diego: Elsevier, p. 35.

Willson, E. K. et al. (2017) ‘Comparison of positive reinforcement training in cats: A pilot study’, Journal of Veterinary Behavior: Clinical Applications and Research, 21, pp. 64–70. doi: 10.1016/j.jveb.2017.07.007.

Wobbler, V. et al. (2009) ‘Breed Differences in Domestic Dogs’ (Canis familiaris) Comprehension of Human Communicative Signals’, Interaction Studies, 10(2), pp. 206–224. doi: 10.1075/is.10.2.06wob.

Wynne, C. D. L., Udell, M. A. R. and Lord, K. A. (2008) ‘Ontogeny’s impacts on humanedog communication’, Age(weeks), 6(8), p.16. doi: 10.1016/j.anbehav.2008.03.010.

Zaine, I., Domeniconi, C. and Wynne, C. D. L. (2015) ‘The ontogeny of human point following in dogs: When younger dogs outperform older’, Behavioural Processes, 119, pp. 76–85. doi: 10.1016/J.BEPROC.2015.07.004.

Issue 5 (Summer 2019)

Independent Study Articles:

1) To what extent are whales and their surrounding environments affected by whaling? Daisy Harrison, BSc (Hons) Animal Behaviour and Welfare

2) Dolphin-human Interactions: Georgia Oaten, BSc (Hons) Animal Behaviour and Welfare

3) What are the welfare issues relating to injury and injury rates in racehorses? Katie Salter, BSc (Hons) Equestrian Sport Science

4) The ability of dogs to follow human communicative cues: Emma Walters, BSc (Hons) Animal Behaviour and Welfare


Research Updates:

1) Animal and Agriculture 18-19

2) Equine 18-19

3) Sport 18-19

The effect of a horse’s racing career on the success in eventing post retraining

Author Names: Sarah Price (BSc (Hons) Equine Science) and Rachel Collins



Retraining racehorses is proven to be popular within the equine industry, especially due to known good performance in eventing due to factors such as conformation and training methods enabling strengthening of the hors’es psychological and physiological state from a young age. This research aimed to determine whether the effects of the racing career have an impact on the success in eventing post retraining. Retrospective data was analysed from online competition records to determine whether length of racing career, the number and type of races completed or the time between racing and eventing careers affected the highest level completed, Foundation Points (FP) or Grading Points (GP) achieved during eventing careers. Of the racehorses foaled in 2000 that had competed in racing (n=3869), 2.8% (n=107) had completed a minimum one event after their racing career, suggesting high wastage rates. Negative correlations were found between the number of flat races and total FP achieved (p=0.018), FP achieved at BE100 level (p=0.036) alongside time between racing and eventing careers and highest eventing level completed (p<0.001) or FP achieved at BE100 level (p=0.029). Significant positive correlations were found between the number of jump races and total FP (p=0.001), or FP achieved at BE100 level (p=0.036). Horses with longer careers achieved more FP at BE100 level (p=0.040), those completing more jump races achieved more FP (p=0.045) and FP at BE100 level (p=0.034) and those with shorter time between racing and eventing careers (p=0.024) reached higher eventing levels. Overall, the findings suggest that horses who compete for a longer time, completing more races, especially jump racing, with shorter transitions into eventing, often achieve greater success during the eventing career. This research has created a baseline for more in-depth future research to expand upon this knowledge. Understanding which horses are successful when retraining from racing to eventing could improve welfare and health of retired racehorses, alongside enhancing performance which could lead to improved Team Great Britain (GB) success.         


1.0 Introduction

1.1 Industry focus on retraining racehorses for eventing

Retraining Thoroughbreds following their racing career, especially to eventing, is anecdotally reported as common and successful within the equine industry (BHA, 2016; Crossman, 2010). The popular Retraining of Racehorses (RoR) eventing series was launched in 2000 to support the retraining of racehorses to other disciplines from grassroots to elite level (RoR, 2017; Crossman, 2010); registering more than 12,500 horses as active in disciplines including eventing (BHA, 2016). There is limited research into the wider application of retraining and competing at different levels, specifically regarding success and potential barriers to success for horses within the sport. To improve the chance of ex-racehorses having successful eventing careers alongside increasing progression to elite level to enhance Team Great Britain (GB) performance on national and international stages, future research would be beneficial.

Morscher (2010) found that Thoroughbred horses were preferred for eventing during long format as other horses did not have the required stamina and speed, especially for elite level. However, in the new short format, more than half of interviewed riders thought Irish Sport Horse crossed with Thoroughbred to be the most suitable breed, with pure Thoroughbreds the second favourite (Morscher, 2010). The continued desire for Thoroughbred horses shows that research to identify and expand upon best practice for eventing an ex-racehorse successfully as well as maintaining their health and welfare would be beneficial to industry.

Ex-racehorses have recently featured as a prominent part of Team GB, highlighted by ex-racehorse, Arctic Soul, being longlisted for the Rio de Janeiro Olympic Games 2016 after coming third at Badminton International CCI**** Horse Trials in the preceding months (BE, 2016a, BE, 2016b). The industry requires a baseline understanding of how ex-racehorses progress into eventing including the selection of elite horses, to expand knowledge of the links between the racing and eventing industries. Improved understanding of what makes a successful transition onwards from racing can enable riders, owners, and trainers to make the best possible decision when choosing competition horses, especially when selecting horses for specific reasons such as elite competition.


1.2 Thoroughbred wastage

Increased knowledge of how racing career longevity and associating factors impact the suitability of horses to transition to an eventing career, alongside reasons behind horse choice in industry could help to reduce wastage figures within the equine industry (Ducro et al., 2009; Burns et al., 2006). Limited progress has been made during the past twenty years in addressing the racehorse wastage issue, primarily due to the lack of purpose designed studies focusing on wastage and the relevant preventative measures, such as entering other appropriate disciplines (Verheyen, 2012). High wastage figures from the racing industry link to the competitive and profit-oriented nature of the sport, alongside large numbers of horses not achieving the desired results on the track (Stowe and Kibler, 2016; Mayberger, 2011). Introducing discipline specific research on suitability of retraining racehorses for eventing and patterns of success could begin to reduce wastage, especially considering improvement of suitable retirement periods for racehorses to promote long and successful future careers.

Wastage within the equine industry has many financial implications; as a largely underfunded industry it is important to increase understanding of the issue and reduce the occurrence (Ducro et al., 2009; Burns et al., 2006). The racing sector acquires more funding than any other part of the industry, meaning the effect of media and positive perception of the sport is of high priority (Collins et al., 2008). Controlling wastage to reduce the negative implications and perceptions regarding welfare of racehorses and the bad press surrounding the sport could in turn improve wider perceptions of the sport, and encourage further sponsorship and support (Collins et al., 2008). Seeing racehorses transition for a successful second career could benefit viewing figures, encouraging wider involvement in both racing and eventing.

Mayberger (2011) explains a potential cause for racehorse wastage during retraining to be associated with the levels of experience, commitment, and determination to become successful in their retrained career, compared to that of other horses of the same age and respective training experience. Research to develop understanding of which horses have potential for successful eventing careers could help to reduce some difficulties in training due to their natural suitability. Within eventing, dressage, show jumping and endurance horses, Sloet van Oldruitenborgh‐Oosterbaan et al., (2010) found 55% of careers ended due to being sold, often linked to unsatisfactory performance, which is known to make up a large portion of wastage within equestrian sport. It is important to understand both the horse’s suitability for eventing and the current equine market with regards to what are considered appealing traits according to owners, riders, and trainers within the sport to improve understanding of the reasons behind wastage versus success (Hennessy et al., 2008).


1.2.1 Strain during training the young racehorse

It has been suggested that competitive careers could end early due to strain during training the young skeleton causing injuries, or predisposition to injury further into ridden careers (Burns et al., 2006). However, it has also been evidenced that horses racing from younger ages had longer careers, consistent with human research demonstrating that exercising with a growing skeleton has beneficial effects on long-term fracture resistance (Verheyen, 2012). The differing evidence highlights the need for research specifically on longevity of racing career relating to the impact on future health and career potential, for example eventing retraining. This has not been investigated thoroughly leaving many trainers, owners, and riders without definitive guidance on the best practice for enhancing health, fitness, and performance of the ex-racehorse.

Thoroughbreds can start training from two years old and it has been found that within racing, those starting training younger had longer and more successful careers, highlighting the need for research to be continued, enabling further conclusions to be drawn in this area (Martig et al., 2014; Tanner et al., 2013; Velie et al., 2013). Training the young skeleton must be done gradually, allowing remodelling to take place, to ensure the bones are strengthened and not damaged to provide long term benefits to skeletal health (Martig et al., 2014). Exercising the young skeleton is considered to be protective to joints and enhance bone strength, which is potentially beneficial to horses undergoing two physically demanding careers where large forces will be taken through the limbs (Firth et al., 2011; Kawcak et al., 2010). Understanding whether longevity of career impacts the horse’s long term career is vital to prevent suffering and ensure suitable horses are selected for competitive careers.

Positive effects of exercising the young horse have been noted in all equestrian sports, although more heavily so in racing due to young horses being frequently exercised (Rogers et al., 2012). Proven benefits on the homogeneity, strength and quality of articular cartilage were found in Thoroughbred horses exercising at younger ages (Dykgraaf et al., 2008) and horses exercising at pasture showed signs of musculoskeletal disorders earlier than horses in structured training from 18 months old (Rogers et al., 2008). The benefits from early race training strengthening the musculoskeletal system could enable horses to train and compete for longer, potentially enabling higher levels of success to be reached in eventing.

Firth et al., (2011) also found significantly better bone strength and tolerance to increased loads in Thoroughbreds exposed to increased exercise prior to race training, which could reduce the risk of fractures, early breakdown or other musculoskeletal injuries during racing and eventing careers. Injuries involving bone were found to cause large portions of breakdown within racehorses (Martig et al., 2014), therefore training the young horse appropriately could benefit health, career longevity and potential success, although being mindful of the risks associated is also vital.


1.3 Retraining Thoroughbred horses

Thoroughbreds are physiologically capable of retraining for a successful eventing career due to natural suitability to the discipline, such as power and speed (Stowe and Kibler, 2016). International success is highlighted by horses such as Director General, winner of the RoR Elite Eventing Performance Award (RoR, 2016). Training throughout the racing and eventing career will emphasise speed and stamina; focusing training on low intensity canter exercise rather than high-speed exercise decreases the risk of wastage through skeletal injury and benefits joint health (Reed et al., 2013). Most racehorses are thought to retire at an age that allows them to pursue future careers successfully, due to the existing training and musculature providing a good basis for future training in the form of reducing injury risk via superior musculoskeletal strength and cardiorespiratory performance (Stowe and Kibler, 2016). Analysis of 17,962 competition results in dressage, show jumping, and eventing determined that horses competing in more than one discipline, especially those starting younger, had longer competitive careers (Braam et al., 2011), highlighting the possibility of increasing ridden and competitive careers through encouraging retraining after retirement from one area within equestrian sport, such as racing to promote good health and welfare.

Conditioning the racehorse for an eventing career should include an appropriate level of physiological and psychological stress to prepare them for the demands of competition, increasing the potential for success throughout the competing career (McKenzie, 2017). Thoroughbreds are likely to need longer in the habituation phase of retraining due to their high sensitivity, reactivity, and anxiety, which should be less present than in racing where the flight response is important (Lloyd et al., 2008). Assessing appropriate training prior to eventing is vital to ensure longevity of the career, enabling horses to be successful. Graf et al., (2013) explain that when sold, horses are often chosen for good temperament over physiological traits. This highlights the importance of investigating how racing career longevity could impact the duration, type and success of retraining required before achieving competition success in eventing, where limited research currently exists.


1.3.1 Physiological adaptations

All horses should be trained with a combination of schooling to develop neuromuscular coordination and conditioning for physiological and structural adaptation of the body systems, which can be transferred from racing to eventing (Castejon-Riber et al., 2017). However, it is important to consider the length of time required for the musculoskeletal system to adapt to the requirements of the new career to reduce the risk of injury and increase performance (RoR, 2018). It is vital to specifically train systems such as muscular, skeletal and cardiorespiratory to ensure the horse is physiologically strong and capable of carrying out the required work, to avoid strain that could lead to injury, poor performance or poor welfare (Castejon-Riber et al., 2017). Understanding the patterns between type of racing career and the period of retraining would aid production of training plans and further research into enhancing the Thoroughbred’s performance in both racing and eventing.

Skeletal muscle mass of Thoroughbreds is high at 53-57%, compared to 42% in ‘non-athletic’ breeds (Rivero and Hill, 2016). In response to training, protein constituents per muscle fibre increase creating an increased muscle mass, thus generating increased potential power output, creating higher performance potential (Rivero and Hill, 2016). Thoroughbreds have shown increases in muscle fibre type 2A in the forelimbs through training thus enabling enhanced locomotion and speed, however the hind limb is considered the key mechanism of propulsion, with the increase of type 2X fibres enhancing this it will also benefit the horse during the cross-country phase of eventing (Rivero and Hill, 2016; Kawai et al., 2009). Aerobic and anaerobic capacities can also be influenced by training, which should be enhanced through race training due to the goal of training the racehorse, giving horses an improved start to retraining for eventing (Jang et al., 2017).


1.4 Thoroughbred musculoskeletal health

Veterinary problems prevented 35.1% of horses from re-registering with British Eventing (BE) in a study conducted by O’Brien et al., (2005), highlighting the importance of understanding which horses are also not suitable for an eventing career post racing, to hopefully reduce this figure in the future. Horses that are older at their first race or have raced for over five years have demonstrated higher risk of injury (Martig et al., 2014; Ducro et al., 2009; Henley et al., 2006), showing the importance of understanding the interaction between factors in the racing career that could impact eventing performance. Increased injury risk is specified for horses competing in jump racing compared to flat racing, alongside an increased risk when running a different type of race for the first time (Henley et al., 2006).

It is also important to understand which horses are most likely to have experienced injuries during their careers to prevent horses being put under excessive strain during retraining. For example, horses in jump racing careers are older and due to the fatigue damage build-up throughout the longer ridden career they are at higher risk of injury, such as stress fractures due to high pressure through the distal limbs for longer periods (Martig et al., 2014).  Superficial digital flexor tendonitis is common in racehorses due to pressures on the distal limb, however due to strain caused by the cross-country phase it is important to consider horses at higher risk, for example jump racing, and ensure training and future careers are considered appropriately (Witte et al., 2016). Musculoskeletal injury is considered the primary cause of wastage within sport horses, with conformational predisposition to lameness thought to impact heavily (Ducro et al., 2009).


1.4.1 Conformation

It is known that Thoroughbreds make good event horses due to their athleticism and speed, however, considering conformation is important to reduce the risk of injury caused by strain to the limbs and back in particular during eventing (Dyson, 2002). Performing at elite level competition is thought to require stronger conformation, especially in the limbs and more specifically distal limbs (Ducro et al., 2009). Larger angles of the hock and stifle joints have been linked to improved jumping ability (Holmström and Philipsson, 1990), which could link horses with successful jump racing careers to possible eventing success due to natural potential. Ideally racehorses should be physically lightweight to reduce the strain on the body, especially the distal limb; the third metacarpal having smaller width has also been linked to sport horse ability, showing similarities in desired conformation (Holmström and Philipsson, 1990). Knowing that the horses are physiologically well matched could encourage those in industry to choose ex-racehorses for eventing, further evidence of potential for success could also increase appeal, therefore lessening wastage.

Conformational research in racehorses is often focused on development of young horses or injury risk, however significant conformational differences have been found between National Hunt (NH) and flat racehorses (Weller et al., 2006). NH horses presented significantly larger on average in neck, forelimb and pelvis length, tibia circumference and most notably so in front and hind hoof circumference compared to flat racehorses (Weller et al., 2006). It is important to ensure the larger hind hoof circumference does not negatively impact the horse if farriery is not suitable, as uneven hooves are considered undesirable due to the potential of limiting competition career longevity (Ducro et al., 2009). NH horses have larger frames, however, it is considered a possibility for this to be because of widespread anecdotal evidence that horses participating in jump racing should be bigger than flat racehorses, influencing the sales market (Weller et al., 2006). However, Weller et al., (2006) study used a sample of horses from only one yard, meaning the type of horses on that yard may be chosen specifically due to conformational preferences of the owners or trainers, potentially based on previous success rates, potentially impacting the results.


1.5 Research aim

The aim of this study is to determine whether the effects of the racing career have an impact on the success in eventing post retraining.


1.5.1 Objectives

(1) To investigate whether the longevity of a racehorse’s career influences the level of competition reached through British Eventing post retraining.

  • H0: The longevity of a racehorse’s career will have no association with the level of competition reached through British Eventing post retraining.
  • H1: The longevity of a racehorse’s career will have an association with the level of competition reached through British Eventing post retraining.


(2) To investigate whether the longevity of a racehorse’s career influences the number of foundation and grading points accumulated through British Eventing post retraining.

  • H0: The longevity of a racehorse’s career will have no association with the number of foundation and grading points accumulated through British Eventing post retraining.
  • H1: The longevity of a racehorse’s career will have an association with the number of foundation and grading points accumulated through British Eventing post retraining.


(3) To investigate whether the type of races run influences success in eventing post retraining.

  • H0: The longevity of a racehorse’s career will have no association with the level of competition reached through British Eventing post retraining.
  • H1: The longevity of a racehorse’s career will have an association with the level of competition reached through British Eventing post retraining.


(4) To investigate whether the time between racing and eventing careers influences success in eventing post retraining.

  • H0: The time between racing and eventing careers will have no association with the number of grading points accumulated through British Eventing post retraining.
  • H1: The time between racing and eventing careers will have an association with the number of grading points accumulated through British Eventing post retraining.


2.0 Methods

2.1 Study design

To address the aim of the study, quantitative retrospective data from 2000 to 2017 were collected from Weatherbys, Weatherbys Point to Point, following approval to use the data, and open data from British Horseracing Authority (BHA), Racing Post and British Eventing (BE) online databases to gain full data on all aspects of the horses’ competition records, which has been proven as a performance parameter from previous research (BE, 2017; BHA, 2017; Kearsley et al., 2008; Olsson et al., 2008; Racing Post, 2017; Weatherbys, 2017; Weatherbys Point to Point, 2017). Data were recorded and coded in Microsoft Office Excel 2016, before data analysis was undertaken on IBM SPSS Statistics 24; all data were kept confidential and securely in accordance with The Data Protection Act (1998).

Data obtained from Weatherbys included a list of every horse foaled in 2000 in Great Britain that had successfully completed one or more race (n=3869), the names were checked against the BE database to compare full name, dam and sire of the horse to determine whether they had competed in affiliated eventing. Horses that had competition records for racing and eventing then had data collected for the date of first and last race completed, number of runs and the type of races run collected from BHA, Racing Post and Weatherbys Point to Point.  Data collection from British Eventing included the date of first event completed, highest competition level finished, number of BE Foundation Points (FP) and the level at which they were achieved (BE90 or BE100) alongside the number of BE Grading Points (GP) and the level at which they were achieved (Novice, CIC*, CCI*, Intermediate, CIC**, CCI**, Advanced, CIC***, CCI*** or CCI****) (BE, 2018).


2.2 Sampling

The population of racehorses foaled in 2000 in Great Britain (GB) that completed minimum one race (n=3869) was gained from Weatherbys by homogeneous sampling (Etikan, 2016). Inclusion and exclusion criteria were enforced to ensure data was relevant and accurate for analysis to limit the impact of extraneous variables on the results (Smith, 2010). For example, large variations of age could affect the data available regarding career length, retraining and success due to incomplete databases, or horses having unfinished careers. The inclusion criteria consisted of horses foaled in 2000 that had completed a minimum of one competition in racing and affiliated eventing, to enable full competition career data to be gathered. The exclusion criteria specified horses that had not competed in both disciplines or competed in eventing prior to competing in racing.

The sample of horses foaled in 2000 was chosen as BE and The Fédération Equestre Internationale (FEI) changed the competition format from 2004 onwards by removing the roads and tracks phase, significantly changing the athletic demand of the sport (BE, 2009a). Some horses may have been more successful in one version of the sport than the other due to individual variation, therefore standardising age removed the impact of this when determining eventing success. Horses are not able to compete with BE until four years old for four-year-old classes comprising of a different format without a cross country phase, or five years old for all other standard competition, which was recorded during this study (BE, 2009b). All horses in the sample were competing in the same format regardless of the start and end dates of their eventing career. Horses that were foaled in earlier years could have competed in the previous or current format if they evented at a younger age. Horses foaled in more recent years could be at the start of their eventing career and not yet reached peak success, both of which would have influenced the results.


2.3 Data analysis

To assess the quantitative data collected, descriptive statistics including mean, standard deviation (SD), median and range were used to create an overview of the data, before inferential statistics were used to assess the research objectives relating to correlations and differences found (Foster et al., 2015) through the following statistical tests. Kolmogorov Smirnov Test (KS Test) for normality was carried out on all data to determine whether it was parametric (p>0.05) or non-parametric (p<0.05), and therefore which test must follow (Dytham, 2011).

Spearman’s Rho test for correlation was carried out on all variables as they were non-parametric following the KS test (Field, 2013), to determine the significance, strength, and direction of the correlation and whether to accept or reject the null hypotheses (Dytham, 2011). A significance output (p-value) of less than 0.05 indicated a relationship between the two variables; the correlation coefficient, between -1 (negative correlation) and +1 (positive correlation), determined the direction and strength of the relationship between the two variables (Elliott and Woodward, 2016; Dytham, 2011). Effect size is determined by the correlation strength; strong linear relationships were indicated by a correlation coefficient of +/- ≥0.51, moderate relationships at +/-0.31-0.50, weak relationships at +/-0.10-0.30 and no correlation +/-0.10 (Elliott and Woodward, 2016).

To test for difference following the KS test, the Kruskall Wallis Test was carried out on the non-parametric variables (Field, 2013). The output determined the significance, whether a difference was found, if so where the difference was found and therefore whether to accept or reject the null hypotheses (Field, 2013). The significance values of each test were noted, with p<0.05 showing significance; however, the adjusted significance values were taken as the true significance as the data had undergone the Bonferroni correction for multiple tests via pairwise comparisons (Field, 2018). The post-hoc tests in the form of pairwise comparisons compared all different combinations of the groups to provide the adjusted significance values and determine if there is significance between any groups, where the significance lies and to what extent (Field, 2018). Adjusted significance values of p<0.05 demonstrated a significant difference in the distribution of data across the specified categories (Dytham, 2011).


3.0 Results

3.1 Descriptive statistics

Of the 3869 racehorses that made it to the track, 3.6% (142 of 3869) registered with BE and 81.6% of these (116 of 142) completed a minimum of one eventing competition. Horses that competed in eventing before their racing career were excluded from the study (n=8), alongside those only competing in Jumping and Style (JAS) competitions (n=1), leaving 107 horses valid for analysis. The racing career length median was 12 months (IQR=24) with a range of <1-71; see figure 1. The median number of racing runs was 6 (IQR=7.5) with a range of 1-46; throughout the horses’ careers, 17.8% (19 of 107) ran both flat and jump races, 29.9% (32 of 107) only ran flat races and 52.3% (56 of 107) only ran jump races; see figure 2.


Figure 1: Length of horses’ racing careers (months)

Figure 1
Figure 2: Type of races completed during racing careers.

Figure 2


The time between racing and eventing careers had a range of 2-124 months, with a median of 24 (IQR=24). The highest BE level completed was predominately BE100 at 40.2% (43 of 107), followed by 32.7% at BE90 (35 of 107), 18.7% at Novice (20 of 107), 3.7% at BE80(T) (4 of 107), 1.9% at CIC* (2 of 107), 1.9% at Intermediate (2 of 107) and finally 0.9% at CCI** (1 of 107); see figure 3


Figure 3: Highest British Eventing (BE) level completed during eventing careers.

Figure 3

30.8% (33 of 107) of horses achieved FP and 11.2% (12 of 107) of horses achieved GP throughout their eventing careers; 24.2% (8 of 33) of horses with FP achieved points at both BE90 and BE100 levels, whilst 45.5% (15 of 33) of achieved points only at BE90 level and 30.3% (10 of 33) achieved points only at BE100 level; see figure 4. 8.3% (1 of 12) of horses with GP achieved points at both Novice and CCI* levels, whereas 91.7% (11 of 12) achieved points only at Novice level; see figure 5.


Figure 4: Total Foundation Points (FP) achieved during eventing careers, across all horses.

Figure 4

Figure 5: Total Grading Points (GP) achieved during eventing careers, across all horses.

Figure 5


3.2 Testing for correlation with Spearman’s Rho

Testing for normality with KS test showed that all variables were significantly non-parametric (p<0.05). There was no significant correlation (p>0.05) between the highest eventing level completed and the length of racing career (months) or number of racing runs. No significant correlations were found (p>0.05) between racing career length (months) or number of racing runs and the total number of FP, FP at BE90 level, or FP at BE100 level, total GP, GP at Novice level or GP at CCI* achieved.

Weak negative correlations were found between the number of flat races completed and the total FP achieved (n=107, rs=-0.228, p=0.018) and FP achieved at BE100 level (n=107, rs=-0.203, p=0.036). A moderate positive correlation was found between the number of jump runs completed and the total FP achieved in eventing (n=107, rs=0.310, p=0.001) alongside a weak positive correlation between number of jump races completed and FP achieved at BE100 level (n=107, rs=0.203, p=0.036).

No significant correlations were found (p>0.05) between the number of flat or jump races completed and total GP achieved, GP at Novice level or GP at CCI* level. No significant correlations were found (p>0.05) between time between racing and eventing careers (months) and Total FP achieved, FP achieved at BE90 level, total GP achieved, GP achieved at Novice level or GP achieved at CCI* level.


3.3 Testing for difference with Kruskall Wallis

No significant difference was found (p=0.115) between the distribution of racing career length (categorised 6 monthly) across the highest BE level completed however the distribution of the number of racing runs (categorised per 6 runs) across the highest BE level completed was significantly different (X22=16.382 n=107 p=0.022), although there was no significant difference between groups (p>0.05) after post-hoc Bonferroni pairwise comparison.

No significant differences were found (p>0.05) between the distribution of racing career length (categorised 6 monthly) across total FP or FP at BE90 level achieved or between the distribution of number of runs (categorised per 6 runs) across total FP achieved. There was a significant difference between the distribution of number of racing runs (categorised per 6 runs) across FP achieved at BE90 level (X22=15.578 n=107 p=0.029) and across FP achieved at BE100 level (X22=14.167 n=107 p=0.048), although there was no significant difference between any of the groups (p>0.05).

The distribution of racing career length (categorised 6 monthly) across FP achieved at BE100 level was significantly different (X22=19.822 n=107 p=0.48), with horses racing for 49-54 months achieving significantly more FP at BE100 level than horses racing for 7-12 months (p=0.040), as shown in table 1.


Table 1: Significant findings after testing for difference with Kruskall Wallis and post-hoc Bonferroni correction between racing career length and eventing career success.

Dependent Variable Independent Variable Group specific p value Adjusted p value at post-hoc
Number of FP at BE100 level achieved Racing career length (months) p=0.048 p=0.040


There was no significant difference (p>0.05) between the distribution of racing career length (categorised 6 monthly) across the total number of GP achieved or across the GP achieved at Novice level, or across GP achieved at CCI* level. There was also no significant difference (p>0.05) between the distribution of number of racing runs (categorised per 6 runs) across the total number of GP achieved, or across GP achieved at Novice level.


The distribution of number of racing runs (categorised per 6 runs) across GP achieved at CCI* level was significantly different (X22=20.400 n=107 p=0.005). Horses racing for 25-30 months achieved significantly more GP at CCI* level than horses racing for £6 months (a) (p<0.001), horses racing for 7-12 months (b) (p=0.001), and horses racing for 13-15 months (c) (p=0.011), as shown in table 2.


Table 2: Significant findings after testing for difference with Kruskall Wallis and post-hoc Bonferroni correction between number of racing runs and eventing career success.

Dependent Variable Independent Variable Group specific p value Adjusted p value at post-hoc
Number of GP at CCI* level achieved (a) Number of racing runs p<0.001 p<0.001
Number of GP at CCI* level achieved (b) Number of racing runs p<0.001 p=0.001
Number of GP at CCI* level achieved  (c) Number of racing runs p<0.001 p=0.011


There was no significant difference (p>0.05) between the distribution of types of races completed across highest BE level completed, or FP achieved at BE90 level, total GP achieved, GP achieved at Novice level or GP achieved at CCI* level. The distribution of the types of races completed across total FP achieved was significantly different (X22=8.786 n=107 p=0.012); horses competing in jump racing only achieved significantly more FP than those competing in both flat and jump racing (p=0.045). The distribution of the types of races completed across FP at BE100 level was significantly different (X22=7.198 n=107 p=0.027); horses competing in jump racing only achieved significantly more FP at BE100 level than those competing in both flat and jump racing (p=0.034) as seen in table 3.


Table 3: Significant findings after testing for difference with Kruskall Wallis and post-hoc Bonferroni correction between type of races completed and eventing career success.

Dependent Variable Independent Variable Group specific p value Adjusted p value at post-hoc
Total number of FP achieved Types of races completed (flat only, jump only, flat and jump) p=0.012 p=0.045
Number of FP achieved at BE100 level Types of races completed (flat only, jump only, flat and jump) p=0.027 p=0.034


There was no significant difference (p>0.05) between the distribution of time between racing and eventing careers across total FP achieved, or FP at BE90 level, FP at BE100 level, total GP achieved, GP achieved at Novice level or GP achieved at CCI* level. The distribution of the time between racing and eventing careers across highest BE level completed was significantly different (X22=19.582 n=107 p=0.003); horses with 13-24 months between the racing and eventing careers reached a significantly higher level than those with over 73 months between careers (p=0.024) as seen in table 4.


Table 4: Significant findings after testing for difference with Kruskall Wallis and post-hoc Bonferroni correction between the time between racing and eventing careers and eventing career success.

Dependent Variable Independent Variable Group specific p value Adjusted p value at post-hoc
Highest BE level completed Time between racing and eventing careers (months) p=0.003 p=0.024


4.0 Discussion

5194 Thoroughbreds were foaled in 2000 (IFHA, 2018), however this project found that only 3869 of these horses made it to the track with 107 competing in eventing following their racing career, suggesting a high wastage rate. Lack of ability was found to be a large contributor to wastage (More, 1999), which could explain the loss of horses prior to racing and eventing careers if potential is not shown during training. Awareness of where these horses end up and reasons behind unsuccessfulness is largely unknown, due to issues in record keeping as highlighted by Collins et al., (2008) when examining the Irish equine industry, learning that only Weatherbys births were truly regulated.


4.1 Racing career length

Horses racing for 49-54 months achieved significantly more FP at BE100 level than those racing for 7-12 months (p=0.040) suggesting that horses with longer careers are more inclined to succeed at this level. Benefits to musculoskeletal health have been noted due to early and prolonged exercise in horses (Rogers et al., 2012); creating a robust horse through maintenance of appropriate training may aid the length and potential for success in future careers. Knight and Thomson (2011) found that horses racing as two-year-olds earned significantly more prize money throughout their careers (p<0.001), had significantly longer careers (p<0.001) and completed more races (p<0.001) than horses starting racing older. Knight and Thomson (2011) found that horses beginning training at younger ages, thus enabling longer racing careers to follow, were chosen due to demonstrating high athletic ability and trainability. Such qualities are also sought after in the event horse, highlighting a possible explanation for horses with longer careers achieving significantly more FP at BE100 level.

Vigre et al., (2002) found that horses with the same trainer for more than three months had a reduced risk of lameness, as it takes time for the trainer to create optimal training plans to enhance both performance and musculoskeletal health. Horses that are moved on in less time may not be given opportunity to show their full potential which could be a reason for fewer horses succeeding in eventing, with only 30.8% (33 of 107) achieving FP and 11.2% (12/107) achieving GP and no horses exceeding CCI** level. Sobczynska (2007) found that Thoroughbreds with experienced professional trainers had significantly longer racing careers than those trained by amateurs; professional training may increase positive effects such as musculoskeletal development and therefore speed and stamina, meaning improved knowledge as to why professionals are not commonly training ex-racehorses would be beneficial.

Sobczynska (2007) found horses at smaller yards to more commonly be trained by their owners and not professionals; extrapolating this research to eventing suggests that training could have a large impact upon their success and career longevity, potentially explaining why no significant differences were found other than FP BE100. Caston and Burzette (2018) found 77.5% of horses were reported in eventing training with amateur riders from an online survey, suggesting that ambitions for elite competition and associated training may be reduced, potentially impacting the eventing level reached and GP achieved. Caston and Burzette (2018) respondents were predominantly Thoroughbred horses at 44.6%, echoing thoughts that these horses remain popular within eventing. However, it is important to note that findings could be influenced by response rates in professional riders, as only 3% respondents competed at intermediate or advanced level. Understanding the differences between horse choice in amateur and professionals could improve knowledge on why increased points were only noted in FP BE100


4.2 Number of racing runs

Significant differences were found between the distribution of number of racing runs across the FP achieved at BE90 level (p=0.029) and FP achieved at BE100 level (p=0.048). The increase in significance compared to that found across career length could be due to the increased information of how much training and competing the horse undertook during their racing career. The intensity of runs within the horse’s career show the true physiological demands of the career, therefore number of races may be a more representative measure than career length on its own (Castejon-Riber et al., 2017). By assessing the number of runs horses have carried out it was shown that those with increased runs, presented increased success in eventing.

Significantly more GP at CCI* level were achieved by horses running 25-30 races than those running £6 races (p<0.001), 7-12 races (p=0.001) or 13-15 races (p=0.011). It was found that only one horse achieved points at CCI* level meaning the significant differences found may not represent the wider population due to not having enough horses (Etikan et al., 2016). To draw valid conclusions within this area, investigating cohorts of horse with elite eventing results could provide more detailed and representative investigations into the impact of racing career on eventing success, providing better application to industry.

There was a significant difference between the overall distribution of the number of racing runs and the highest BE level completed (p=0.022), however no differences were found between groups. 95.3% of horses’ highest eventing level completed was novice or below, potentially influencing these results as there could be significant differences between number of racing runs compared to higher eventing levels which cannot be fully appreciated with this study population. Levels above novice consist of increased height (≥1.10-1.30m), number (≥28-45) and technicality of jumping efforts, increased speed from £475m/min to ≥520m/min alongside increased technicality required for dressage and show jumping such as the introduction of lateral work and trebles (BE, 2018). Horses capable of competing in these levels require natural ability for the sport, alongside superior musculoskeletal health and training, which can also be impacted by who is training the horse and the time available for development (Sobczynska, 2007), highlighting the importance to understand more on industry practice within training for specific levels of success.


4.3 Types of races completed

Negative correlations were found between the number of flat races completed and FP achieved (p=0.018) or FP at BE100 level (p=0.036). Positive correlations were found between the number of jump races completed and total FP achieved (p=0.001) or FP at BE100 level (p=0.036). Horses competing in only jump racing compared to flat and jump achieved significantly more FP (p=0.045) and FP at BE100 level (p=0.034). This suggests that horses training specifically for jump racing could find large benefit throughout the retraining process due to physiological capabilities and coordination, thus indicating that horses completing more jump races show proclivity for higher success rates at levels up to BE100. Horse completing jump racing tend to be older, therefore may not have the time to do detailed enough training to achieve success at higher levels, including those involving GP (Castejon-Riber et al., 2017; Stowe and Kibler, 2016).

Ducro et al., (2009) highlight that elite sport horses require strong conformation to ensure musculoskeletal health and performance, with the large hock and stifle angles to aid jumping, jump race horses are well suited to the biomechanical principles of eventing (Holmström and Philipsson, 1990). Horses with desirable jumping conformation should ideally compete in just jump racing due to natural ability, allowing jump training to progress when retraining for eventing (Ducro et al., 2009). However, if horses compete in flat racing first, it will add extra strain to the distal limbs, increasing injury risk prior to two strenuous jumping careers (Henley et al., 2006), potentially why these horses achieved less FP and FP at BE100 (p<0.05).

Hausberger et al., (2011) found that personality was affected by the type of work the horse was doing, expression became different in the horses, which is important to consider when looking at the number of horses undertaking flat racing compared to jump racing. Horses start flat racing at a younger age, and can train for jump racing as well, although horses that do not suit flat racing may be immediately retried, although a jump career could improve behaviour and performance (Hausberger et al., 2011). As increased jump racing shows a trend of increased success, it is important to explore whether the training involved in jump racing or physiological characteristics are giving horses the advantage.


4.4 Time between racing and eventing careers

Negative correlations were found concerning the time between racing and eventing careers and the highest eventing level reached (p<0.001), alongside FP achieved at BE100 level (p=0.029). Alongside this, horses with 13-24months between racing and eventing careers reached a significantly higher eventing level than those with 73months or more (p=0.024), suggesting that horses with shorter time between the racing and eventing careers have increased likelihood of achieving greater levels of eventing success.

Conversely, horses that have small breaks between racing and eventing careers may not have had enough time for conditioning and retraining, meaning they are not physiologically and psychologically prepared for a long and successful eventing career (Lloyd et al., 2008).  Thomson et al., (2014) found that more horses left competitive careers due to unsuitable behaviour than any other reason, highlighting the importance of suitable training and habituation. This could explain why no significant differences were found regarding GP achievement, as horses may require different degrees of training and development to succeed at elite level.

McKenzie, (2017) suggests that short transitions between racing and eventing are common and often successful, as demonstrated with these findings. Training the horse for the high intensity cross country phase should focus more on optimising aerobic capacity and speed which can typically be done over short periods, especially as these horses will already have such qualities from race training (Jang et al., 2017; McKenzie, 2017). McKenzie (2017) explains that lengthier periods of training are required to perfect the discipline and technicality required during phases such as dressage. Horses with shorter periods between racing and eventing careers could be missing this increased precision within their performance, potentially reducing success rates with regards to GP achieved and reaching elite level eventing. It is important to consider that horses within this sample did not exceed levels of CCI**, therefore assessing these correlations and differences with horses at top level such as Advanced/CCI**** would improve application to industry with regards of choosing elite event horses.

Negative correlations between time between racing and eventing careers and highest level achieved or FP at BE100 level could signify that horses with lengthier retraining periods have experienced periods out of work, meaning the horse is not capable of reaching high success levels in eventing due to injury rehabilitation and enduring weakness. Racehorses with superficial digital flexor tendonitis were found to have long periods off work, on average 18 months, however the time to competition fitness could be extensive (Witte et al., 2016), elongating the time between racing and eventing careers. During time out of training, bone mineral density (BMD) has been proven to decrease 0.45% in 12 weeks, highlighting issues with reduction of bone strength and increased injury risk upon return to work, especially in those with lengthy periods out of work before returning to strenuous work. (Yamada et al., 2015; Porr et al., 1998).


4.5 Applications to industry

Understanding that horses with longer careers have shown success in eventing can help to improve preventative measures for common injuries and strains seen in both disciplines, to improve career longevity and success alongside reducing wastage. Knowledge on the links between jump racing and eventing success can also aid in horse selection to promote strong conformation and ability, ensuring horses are in a career they can excel in to promote performance, especially at higher levels. The retraining period has been shown to be of importance to eventing success, with shorter periods enabling greater success, potentially due to the horses maintaining training and physiological prowess. Promoting good practice of retraining to ensure physiological benefits from race training are maintained could improve welfare of horses alongside performance. Learning more about why horses are proving successful at lower levels involving FP, but not at higher levels with GP would be useful to improve access for horses to reach the higher levels, improving team GB performance and appealing to wider variety of horse owners.


4.6 Limitations

Limitations within this research link to the nature of retrospective data usage, through lack of access or insufficient and incomplete datasets, meaning study design is important to get the most out of the data accessible (Saunders et al., 2012). However, using databases from Weatherbys and national governing bodies reduced the impact of this due to higher standardisation and regulations. This research focused on creating an understanding of baseline patterns within the existing population to enable further exploration into specific areas with prospective or qualitative research. It is not possible to say that horses could not have exceeded the highest eventing level or number of FP and GP achieved during their career as there are many reasons for progression to stop, even where natural talent and good performance is present. Quantitative research based on purely competition records does not enable full comprehension into why horses did not progress further, for example due to reaching peak performance or injury, rider change or financial factors.

Horses could be exported prior to, during, or after racing, which cannot be controlled, leaving gaps in the data as the horses quickly become untraceable due to insufficient regulations and record keeping (Collins et al., 2008). These horses could have competed in other countries in racing and eventing, which could impact the industry-wide perception of retraining racehorses for eventing due to data potentially being incorrect regarding the longevity and success of their competitive careers. It is also possible that ex-racehorses eventing with BE have competed in racing abroad prior to retraining, where full data is not acquirable, potentially impacting data analysis. Horses may transition to unaffiliated rather than BE due to costs, perception of less competitive pressure and more enjoyment-based compared to more intense competition. Accurate and consistent information also cannot be compiled from unaffiliated eventing due to no overarching standardisation in recording and publishing results (Collins et al., 2008).


4.7 Future research recommendations

Using prospective research to investigate cohorts from other years could improve knowledge on the careers and success rates post racing retirement. Factors such as the financial crisis, which impacted the UK greatly (Hodson and Mabbett, 2009), could influence the number of horses competing at that time. Therefore, determining why horses are not competing through qualitative research could benefit the industry, alongside investigating a variety of time periods. Thirkell and Hyland (2017) explain that quantitative research prevents greater depth of information being collected, therefore now there is information that factors of the racing career longevity and retraining impact some factors of eventing success, it would be recommended to investigate further with qualitative research, enabling more in-depth conclusions to be drawn.

Horses do not reach the racetrack for varying reasons, for instance poor performance or interruptions to training; some horses are also bred with the intention of other purposes despite the full Thoroughbred bloodlines (Bolwell et al., 2013); it would be interesting to investigate the path of these horses to investigate further where the Thoroughbred event horses start off. There is also little knowledge of what horses do between competition record data, therefore establishing methods of gathering more in-depth data via qualitative research following up with horses throughout the retraining period would be beneficial. Information on the length of time in training for racing following the final race and before the first event would give a better indication on the factors relating to training that impact success in eventing, alongside whether time out of training is truly beneficial or detrimental to the horse’s health and performance.


5.0 Conclusion

The research aim was to determine whether the racing career impacts the success of eventing careers post retraining. Overall, horses who competed for a longer time, completed more races, especially jump racing, and had shorter transitions into eventing achieved significantly greater success regarding competition level reached, FP or GP achieved (p<0.05). High wastage rates were found from birth of Thoroughbred foals (n=5194) to racing (n=3869) and finally eventing (n=107), investigating further why those that did not successfully achieve racing or eventing careers could improve knowledge on how to improve training and management strategies to promote good performance in a larger range of horses.

Horses with longer racing careers achieved significantly more FP at BE100 level (p=0.040) suggesting that the physiological conditioning through prolonged training could have lasting benefits to the horse through the eventing career. Underlying conformation, similar demands and training style may contribute to increased jump racing runs resulting in significantly more FP (p=0.045) and FP at BE100 level (p=0.034) being achieved. Horses with shorter periods between racing and eventing careers achieved significantly higher levels in eventing (p=0.003), highlighting the importance of gaining knowledge of horses’ activities throughout their career, to appreciate whether longer periods out of competition equates to injury, therefore potentially limiting future performance. Research to develop this knowledge regarding current success rates of ex-racehorses in eventing, involving qualitative and prospective studies could improve understanding of the interaction between careers further, to promote good welfare, health and performance of Thoroughbred horses.



BE (2018) British Eventing Members’ Handbook 2018. Available from: http://www.britisheventing.com/documents.asp?section=145&sectionTitle=Download+Area#category0001200000010002 [Accessed 4 March 2018]

BE (2017) British Eventing – Dressage, Showjumping and Cross Country. Available from: http://www.britisheventing.com [Accessed 18 December 2017]

BE (2016a) Memorable Mitsubishi Motors Badminton Horse Trials. Available from: http://www.britisheventing.com/asp-net/news/item.aspx?id=6279 [Accessed 25 January 2018].

BE (2016b) Rio 2016 long list announced. Available from: http://www.britisheventing.com/asp-net/news/item.aspx?id=6279 [Accessed 25 January 2018].

BE (2009a) How the sport started and its future developments. Available from: (http://www.britisheventing.com/asp-net/page.aspx?section=1093&itemtitle=history+and+origins+of+eventing [Accessed 4 February 2018].

BE (2009b) Horse info. Available from: http://www.britisheventing.com/asp-net/page.aspx?section=965&sectionTitle=Event+Horse+Information [Accessed 4 February 2018].

Boden, L. A., Parkin, T. D., Yates, J., Mellor, D., and Kao, R. R. (2012). Summary of current knowledge of the size and spatial distribution of the horse population within Great Britain. BMC veterinary research8(1), pp. 43.

Boden, L. A., Parkin, T. D., Yates, J., Mellor, D., and Kao, R. R. (2013). An online survey of horse-owners in Great Britain. BMC veterinary research9(1), pp. 188.

Bolwell, C. F., Rogers, C. W., French, N. P., and Firth, E. C. (2013). The effect of interruptions during training on the time to the first trial and race start in Thoroughbred racehorses. Preventive veterinary medicine108(2-3), pp. 188-198.

BHA (2017) The British Horseracing Authority. Available from: http://www.britishhorseracing.com/ [Accessed 18 December 2017].

BHA (2016) Annual Report and Accounts [online]. Available from: https://www.britishhorseracing.com/wp-content/uploads/2017/03/BHA-Annual-Report-2016-1.pdf [Accessed 5 February 2018].

Braam, Å., Näsholm, A., Roepstorff, L., and Philipsson, J. (2011). Genetic variation in durability of Swedish Warmblood horses using competition results. Livestock Science142(1), pp. 181-187.

Burns, E. M., Enns, R. M., and Garrick, D. J. (2006). The effect of simulated censored data on estimates of heritability of longevity in the Thoroughbred racing industry. Genetics and Molecular Research5(1), pp. 7-15.

Castejon-Riber, C., Riber, C., Rubio, M. D., Agüera, E., and Muñoz, A. (2017). Objectives, Principles, and Methods of Strength Training for Horses. Journal of Equine Veterinary Science56, pp. 93-103.

Caston, S. S., and Burzette, R. G. (2018). Demographics, Training Practices, and Injuries in Lower Level Event Horses in the United States. Journal of Equine Veterinary Science62, pp. 25-31.

Collins, J., Hanlon, A., More, S. J., and Duggan, V. (2008). The structure and regulation of the Irish equine industries: Links to considerations of equine welfare. Irish veterinary journal61(11), pp. 746.

Crossman, G. K. (2010). The organisational landscape of the English horse industry: a contrast with Sweden and the Netherlands. PhD, University of Exeter

Data Protection Act. Chapter 29 (1998) London: TSO (The Stationary Office).

Ducro, B. J., Gorissen, B., Eldik, P. V., and Back, W. (2009). Influence of foot conformation on duration of competitive life in a Dutch Warmblood horse population. Equine veterinary journal41(2), pp. 144-148.

Dykgraaf, S., Firth, E. C., Rogers, C. W., and Kawcak, C. E. (2008). Effects of exercise on chondrocyte viability and subchondral bone sclerosis in the distal third metacarpal and metatarsal bones of young horses. The Veterinary Journal178(1), pp. 53-61.

Dyson, S. (2002). Lameness and poor performance in the sport horse: dressage, show jumping and horse trials. Journal of Equine Veterinary Science22(4), pp. 145-150.

Dytham, C. (2011) Choosing and using statistics. 3rd edition. Chichester: Wiley-Blackwell.

Elliott, A. C, and Woodward, W. A. (2016) IBM ® SPSS by Example, A Practical Guide to Statistical Data Analysis. 2nd edition. Los Angeles: SAGE.

Etikan, I., Musa, S. A., and Alkassim, R. S. (2016). Comparison of convenience sampling and purposive sampling. American Journal of Theoretical and Applied Statistics5(1), pp. 1-4.

FEI (2018) FEI Eventing Rules: 2018. Available from: http://inside.fei.org/fei/regulations/eventing [Accessed 4 March 2018]

Field, A. (2013) Discovering statistics using IBM SPSS statistics. 4th edition. Los Angeles: SAGE.

Field, A. (2018) Discovering statistics using IBM SPSS statistics. 5th edition. Los Angeles: SAGE

Firth, E. C., Rogers, C. W., van Weeren, P. R., Barneveld, A., McIlwraith, C. W., Kawcak, C. E., Goodship, A. E., and Smith, R. K. (2011). Mild exercise early in life produces changes in bone size and strength but not density in proximal phalangeal, third metacarpal and third carpal bones of foals. The Veterinary Journal190(3), pp.383-389.

Foster, L., Diamond, I. and Jefferies, J. (2015) Beginning statistics. 2nd edition. London: SAGE Publications Ltd.

Graf, P., von Borstel, U. K., and Gauly, M. (2013). Importance of personality traits in horses to breeders and riders. Journal of Veterinary Behavior: Clinical Applications and Research8(5), pp. 316-325.

Hausberger, M., Muller, C., and Lunel, C. (2011). Does Work Affect Personality? A Study in Horses. PLoS ONE6, pp. 14659.

Henley, W. E., Rogers, K., Harkins, L., and Wood, J. L. N. (2006). A comparison of survival models for assessing risk of racehorse fatality. Preventive veterinary medicine74(1), pp. 3-20.

Hennessy, K. D., Quinn, K. M., and Murphy, J. (2008). Producer or purchaser: different expectations may lead to equine wastage and welfare concerns. Journal of Applied Animal Welfare Science11(3), pp. 232-235.

Hodson, D., and Mabbett, D. (2009). UK economic policy and the global financial crisis: paradigm lost? JCMS: journal of common market studies47(5), pp. 1041-1061.

Holmström, M., and Philipsson, J. (1993). Relationships between conformation, performance and health in 4-year-old Swedish Warmblood Riding Horses. Livestock Production Science33(3), pp. 293-312.

IFHA (2018) Great Britain Breeding Statistics [online]. Available from: http://www.ifhaonline.org/default.asp?section=Resources&area=4 [Accessed 28 February 2018].

Jang, H. J., Kim, D. M., Kim, K. B., Park, J. W., Choi, J. Y., Oh, J. H., Song, K. D., Kim, S., and Cho, B. W. (2017). Analysis of metabolomic patterns in Thoroughbreds before and after exercise. Asian-Australasian journal of animal sciences30(11), pp. 1633.

Kawai, M., Minami, Y., Sayama, Y., Kuwano, A., Hiraga, A., and Miyata, H. (2009). Muscle fiber population and biochemical properties of whole body muscles in Thoroughbred horses. The Anatomical Record292(10), pp. 1663-1669.

Kawcak, C. E., McIlwraith, C. W., and Firth, E. C. (2010). Effects of early exercise on metacarpophalangeal joints in horses. American journal of veterinary research71(4), pp. 405-411.

Kearsley, C. G. S., Woolliams, J. A., Coffey, M. P., and Brotherstone, S. (2008). Use of competition data for genetic evaluations of eventing horses in Britain: analysis of the dressage, showjumping and cross-country phases of eventing competition. Livestock Science118 (1), pp. 72-81.

Knight, P. K., and Thomson, P. C. (2011). Age at first start and racing career of a cohort of Australian Standardbred horses. Australian veterinary journal89(9), pp. 325-330.

Lloyd, A. S., Martin, J. E., Bornett-Gauci, H. L. I., and Wilkinson, R. G. (2008). Horse personality: variation between breeds. Applied Animal Behaviour Science112(3), pp. 369-383.

Martig, S., Chen, W., Lee, P. V. S., and Whitton, R. C. (2014). Bone fatigue and its implications for injuries in racehorses. Equine veterinary journal46(4), pp. 408-415.

Mayberger, C. L. (2011). Responsibility in the Sport of Kings: Imposing an Affirmative Duty of Care on the Primary Financial Beneficiaries of the Thoroughbred Horseracing Industry. Stanford Journal of Animal Law and Policy, 4, pp. 64.

McKenzie, E. (2017). Foundations of performance–factors that contribute to excellence in equine exercise. Comparative Exercise Physiology13(3), pp. 127-135.

More, S. J. (1999). A longitudinal study of racing Thoroughbreds: performance during the first years of racing. Australian veterinary journal77(2), pp. 105-112.

Morscher, S. (2010). An analysis of conformation and performance variables in potential three-day event horses in Ireland. MSc Thesis. University of Limerick.

O’Brien, E., Stevens, K. B., Pfeiffer, D. U., Hall, J., and Marr, C. M. (2005). Factors associated with the wastage and achievements in competition of event horses registered in the United Kingdom. The Veterinary record157(1), pp.9-13.

Olsson, E., Näsholm, A., Strandberg, E., Philipsson, J. (2008). Use of field records and competition results in genetic evaluation of station performance tested Swedish Warmblood stallions. Livestock Science, 117(2), pp. 287-297.

Porr, C. A., Kronfeld, D. S., Lawrence, L. A., Pleasant, R. S., and Harris, P. A. (1998). Deconditioning reduces mineral content of the third metacarpal bone in horses. Journal of animal science76(7), pp. 1875-1879.

Racing Post (2017) Racing Post Results. Available from: https://www.racingpost.com/results/ [Accessed 18 December 2017].

Reed, S. R., Jackson, B. F., Wood, J. L., Price, J. S., and Verheyen, K. L. (2013). Exercise affects joint injury risk in young Thoroughbreds in training. The Veterinary Journal196(3), pp. 339-344.

Rivero, J. L. L., and Hill, E. W. (2016). Skeletal muscle adaptations and muscle genomics of performance horses. The Veterinary Journal209, pp. 5-13

Rogers, C. W., Bolwell, C. F., Tanner, J. C., and van Weeren, P. R. (2012). Early exercise in the horse. Journal of Veterinary Behavior: Clinical Applications and Research7(6), pp.375-379.

Rogers, C. W., Firth, E. C., McIlwraith, C. W., Barneveld, A., Goodship, A. E., Kawcak, C. E., Smith, R. K. W., and Weeren, P. V. (2008). Evaluation of a new strategy to modulate skeletal development in racehorses by imposing track‐based exercise during growth: The effects on 2‐and 3‐year‐old racing careers. Equine veterinary journal40(2), pp. 119-127.

RoR (2018) RoR Training: Introduction. Available from: https://www.ror.org.uk/care-training/training/introduction/ [Accessed 25 January 2018].

RoR, (2017) Retraining of Racehorses. Available from: http://www.ror.org.uk/ [Accessed 25 January 2018].

RoR (2016) RoR Series: Elite Eventing Performance Award – GB Raced Only. Available from: https://www.ror.org.uk/competitions-series/eventing/elite-eventing-performance-award/ [Accessed 25 January 2018].

Saunders, M., Lewis, P. and Thornhill, A. (2012) Research methods for business students. 6th edition. Harlow, Essex: Pearson Education Limited.

Sloet van Oldruitenborgh‐Oosterbaan, M. M., Genzel, W., and van Weeren, P. R. (2010). A pilot study on factors influencing the career of Dutch sport horses. Equine Veterinary Journal, 42(s38), pp.28-32.

Smith, M. (2010) Research Methods in Sport. 1st edition. Exeter: Learning Matters.

Sobczynska, M. (2007). The effect of selected factors on length of racing career in Thoroughbred racehorses in Poland. Animal Science Papers and Reports25(3), pp. 131-141.

Stowe, C. J., and Kibler, M. L. (2016). Characteristics of adopted Thoroughbred racehorses in second careers. Journal of applied animal welfare science19(1), pp. 81-89.

Tanner, J. C., Rogers, C. W., and Firth, E. C. (2013). The association of 2‐year‐old training milestones with career length and racing success in a sample of Thoroughbred horses in New Zealand. Equine veterinary journal45(1), pp. 20-24.

Thirkell, J., and Hyland, R. (2017). A Survey Examining Attitudes Towards Equine Complementary Therapies for the Treatment of Musculoskeletal Injuries. Journal of Equine Veterinary Science59, pp. 82-87.

Thomson, P. C., Hayek, A. R., Jones, B., Evans, D. L., and McGreevy, P. D. (2014). Number, causes and destinations of horses leaving the Australian Thoroughbred and Standardbred racing industries. Australian veterinary journal, 92(8), pp. 303-311.

Velie, B. D., Knight, P. K., Thomson, P. C., Wade, C. M., and Hamilton, N. A. (2013). The association of age at first start with career length in the Australian Thoroughbred racehorse population. Equine veterinary journal45(4), pp. 410-413.

Vigre, H., Chriél, M., Hesselholt, M., Falk-Rønne, J., and Ersbøll, A. K. (2002). Risk factors for the hazard of lameness in Danish Standardbred trotters. Preventive Veterinary Medicine56(2), pp. 105-117.

Verheyen, K. L. P. (2013). Reducing injuries in racehorses: Mission impossible? Equine veterinary journal45(1), pp. 6-7.

Weatherbys (2017) Weatherbys – Horses and Racing. Available from: http://www.weatherbys.co.uk/horses-racing [Accessed 18 December 2017].

Weatherbys Point to Point (2017) Point to Point Racing Company: The Register. Available from: https://www.pointtopoint.co.uk/register [Accessed 18 December 2017].

Weller, R., Pfau, T., May, S. and Wilson, A. (2006). Variation in conformation in a cohort of National Hunt racehorses. Equine Veterinary Journal, 38(7), pp.616-621.

Witte, S., Dedman, C., Harriss, F., Kelly, G., Chang, Y. M., and Witte, T. H. (2016). Comparison of treatment outcomes for superficial digital flexor tendonitis in National Hunt racehorses. The Veterinary Journal216, pp. 157-163.

Yamada, K., Sato, F., HiguchI, T., Nishihara, K., Kayano, M., Sasaki, N., and Nambo, Y. (2015). Experimental investigation of bone mineral density in Thoroughbreds using quantitative computed tomography. Journal of equine science26(3), pp. 81-87.

Microbiological assessment of canine drinking water: the impact of construction material on the quantity and species of bacteria present in water bowls

Author Names: Coralie Wright (BSc (Hons) Bioveterinary Science) and Aisling Carroll



The number of pet dogs (Canis lupus familiaris) in the common household is continually rising. The increasingly close contact between humans and cohabitant pets is leading to concerns regarding bacterial transmission of zoonoses. The dog water bowl has been identified as the third most contaminated item within the household, suggesting that it is able to act as a fomite for bacterial transmission, particularly where young or immunocompromised individuals are present. Studies in livestock have identified that water trough construction material influences bacterial count; however no similar research has been conducted for dog water bowls. The objectives of the current study were to identify which dog bowl material – plastic, ceramic or stainless steel – harbours the most bacteria over a 14 day period and whether the species identified varies between bowl materials. The study took place over 6 weeks. A sample of 6 medium sized (10-25kg) dogs aged 2-7 (mean= 3.8 ± 1.95) was used. All dogs were clinically healthy, housed individually and located within a rural environment. All bowls were purchased brand new and sterilised prior to a two week sampling period. On day 0, day 7 and day 14 swabs were taken from each bowl and 10-fold serial dilutions were conducted on blood agar. The cultured bacteria were subjected to biochemical testing and the most prominent bacteria from day 14 were further identified using PCR. A significant difference was identified for all bowl materials when comparing total CFU/ml between day 0 and day 7 and day 0 and day 14 (p<0.05). No significant difference was identified between total CFU/ml and bowl material (P>0.05), however descriptive statistics suggest that the plastic bowl material maintains the highest bacterial count after 14 days. Several medically important bacteria were identified from the bowls, including MRSA and Salmonella, with the majority of species being identified from the ceramic bowl. This could suggest that harmful bacteria may be able to develop biofilms more successfully on ceramic materials. Further research is required to identify the most suitable or alternative materials for dog water bowls.


1.0 Introduction

The domestic dog (Canis lupus familiaris) has lived alongside humans for at least 15,000 years (Frantz et al., 2016; Wang et al., 2016; Shannon et al., 2015). Now one of humankind’s closest companions, the number of dogs kept within households is increasing, with an estimated population of 8.5 million in the UK (PFMA, 2018; Damborg et al., 2016; Wang et al., 2016; da Costa, Loureiro and Matos, 2013; Buma et al., 2006). Increasingly close contact between humans and their dogs allows favourable conditions for bacterial transmission both directly and indirectly (Damborg et al., 2016; Chomel and Sun, 2011; Deplazes et al., 2011; Bowman and Lucio-Forster, 2010; Fèvre et al., 2006; Guardabassi, Schwartz and Lloyd, 2004). Dogs spend a considerable amount of time outdoors, where they come into contact with a plethora of bacterial pathogens through contaminated soil, water, faeces and other animals (Lambertini et al., 2016). Additionally, the canine skin and oral microbiome is naturally colonised by a diverse range of bacteria (Hoffmann et al., 2014; Dewhirst et al., 2012; Elliot et al., 2005). Sixty percent of emerging pathogens are considered zoonotic (Dula and Pal, 2017; Cutler, Fooks and van der Poel, 2010), emphasising the importance of minimising the risk of bacterial transmission between humans, particularly the young and immunocompromised, and their cohabitant pets (Damborg et al., 2016; Moyaert et al., 2006).  Lambertini et al. (2016) state that contact between owners and their dog occurs multiple times a day and can also be mediated by surfaces. The dog water bowl has been identified by Donofrio et al. (2012) as the third most contaminated item in the household, suggesting it is able to act as a fomite in bacterial transmission. There is currently no research exploring the most suitable material for dog water bowls with regards to bacterial load and species.


1.1 Development of Biofilms

Water is essential for life and almost every canine bodily function (Kenssington, 2014; Jéquier and Constant, 2010; Aspinall, 2012); however, the visual purity of water can disguise the proliferation of microorganisms (Folorunso, Kayode and Onibon, 2014). Bacterial biofilms can develop on many abiotic surfaces, particularly those in contact with non-sterile water (Flemming, 2011; Morita et al., 2011; Fuster-Valls et al., 2008). Livestock water troughs have been previously identified as reservoirs for bacterial growth and biofilm formation (Folorunso, Kayode and Onibon, 2014; Watson et al., 2012; Cook, Britt and Bolster, 2010; Avery et al., 2008; Sargeant et al., 2003; McGee et al., 2002; LeJeune et al., 2001a; LeJeune et al., 2001b); however research on dog water quality is incredibly limited. Biofilm formation in dog water bowls could serve as temporary or long term habitats for hygienically important bacteria (Wingender and Flemming, 2011). In addition to continuous re-infection of bacteria to dogs via ingestion of contaminated water, it is suggested that biofilms are responsible for 65% of all bacterial infections in humans (Rowson and Townsend, 2016). When washing or refilling a water bowl, humans may come into contact with bacterial biofilms and subsequently transfer pathogens from their hands to household surfaces (Lambertini et al., 2016). This is cause for concern as cohabitant pets often reside within the kitchen, which may ultimately lead to ingestion of contaminant bacteria (Lambertini et al., 2016; Behravesh et al., 2010).

Bacteria in biofilms exhibit increased resistance and tolerance to antimicrobials and environmental stressors (Ciofu et al., 2017; Wall and Mah, 2017; Rowson and Townsend, 2016; Steenackers et al., 2012), suggesting that they may be able to persist for longer within dog water bowls. This is concerning where zoonotic transfer is possible because of the shared antimicrobials in both human and veterinary medicine (Rendle and Page, 2018; Pomba et al., 2017; da Costa, Loureiro and Matos, 2013; Moyaert et al., 2006; Guardabassi, Loeber and Jacobson, 2004). This highlights the importance of minimising the risk of zoonotic spread from water bowls by identifying the most suitable material.


1.2 Influence of Materials on Bacterial Adhesion

The strength of bacterial adhesion depends on the organism’s surface properties and the material being colonised (Ribeiro, Monteiro and Ferraz, 2012; Van Houdt and Michiels, 2010; Faille et al., 2002). Material properties, including surface roughness and hydrophobicity influence the ability of bacterial cells to adhere to the surface and therefore determine the hygienic status of the material (Gharechahi, Moosavi and Forghani, 2012; Van Houdt and Michiels, 2010). A study by LeJeune et al. (2001a) into the hygiene of cattle water troughs identified a significant difference in bacterial coliform count between troughs constructed of different materials, with metal having lower counts than concrete and plastic. This study collected samples of water from each trough, however it is thought that 95% of the bacteria present in water systems are located at the surfaces and only 5% can be identified in the water phase (Flemming, Percival and Walker, 2002). Therefore, these findings may not accurately represent material influence on water quality, despite the large sample used by LeJeune et al. (2001a). Cook, Britt and Bolster (2010) examined decay rates of biofilms on different materials, artificially inoculated with Mycobacterium avium subsp. Paratuberculosis (Map), within glass water troughs. They identified that biofilm decay rates were fastest on stainless steel materials, followed closely by plastic. No significant difference was identified between stainless steel and plastic, however these findings are in agreement with the results of LeJeune et al. (2001a). Despite this, they only examined decay rates for Map and these results are therefore not representative for all bacterial species and their specific adherence to these materials as described by Van Houdt and Michiels (2010) and Faille et al. (2002).

Donofrio et al. (2012) identified high levels of bacteria within dog water bowls; however no specific links between bowl material and heterotrophic plate counts (HPC) were made. Samples of 26 items were tested, in 22 households, and the data revealed that sponge materials had the highest HPC, followed by porcelain, stainless steel and plastic. Dog bowls are commonly constructed of stainless steel, plastic and ceramic. These materials may permit heavy growth of bacterial pathogens as they lack antimicrobial effects (Fatoba et al., 2014). The results of this study suggest that these materials can harbour bacteria in high volumes and thus research specifically applied to bacterial load in dog water bowls constructed with these materials and the species present in them would be beneficial. Furthermore, identification of the most hygienic dog bowl material could allow for evidence-based pet ownership guidelines to be developed and thus reduce zoonotic transmission (Damborg et al., 2016; Leonard, 2014; Moyaert et al., 2006).


1.3 Dogs as a Reservoir of Zoonotic Transmission

There is substantial research which provides evidence of dogs acting as a reservoir of both zoonotic and resistant bacteria (Schwartz, Loeffler and Kadlec, 2017; Damborg et al., 2016; Chomel and Sun, 2013; Martins et al., 2013; Buma et al., 2006; Guardabassi, Loeber and Jacobson, 2004; Guardabassi, Schwartz and Lloyd, 2004). Several hygienically important bacteria have been isolated from dog mouths including Pasteurella spp., Staphylococcus spp., Bacillus spp. and many more (Zambori et al., 2013; Dewhirst et al., 2012; Elliot et al., 2005); therefore these species may potentially be present in dog water bowls.

A pathogen of major importance, Methicillin-resistant Staphylococcus aureus (MRSA), is of growing concern due to increased reports of zoonotic transmission (Han, Yang and Park, 2016; Loeffler et al., 2011; Abbott et al., 2010; Aklilu et al., 2010; Rutland et al., 2009). MRSA can potentially lead to mild or life-threatening skin infections in both humans and animals and it is highly resistant (Pantosti, 2012; Catry et al., 2010; Weese, 2010). Balen et al. (2013) conducted a year-long MRSA surveillance study in a veterinary hospital and discovered that some water bowls were contaminated for 2 consecutive months with the same pulsotype of MRSA, which was also detected on doors and floor surfaces. This study was conducted within a veterinary hospital with a large number of animals present and thus these findings cannot be extrapolated to household environments (Balen et al., 2013). However, the results demonstrate how bacteria, such as MRSA, can persist within a water bowl and can be spread throughout the environment – acting as a possible reservoir of bacterial transmission. Additionally, many studies have identified MRSA strain sharing amongst humans and their cohabitant pets (Faires, Tater and Weese, 2009; Sing, Tuschak and Hörmansdorfer, 2008; Boost, O’donoghue and Siu, 2007; Weese, 2006). However, cross-sectional studies of MRSA colonisation cannot identify direction of transmission and they do not eliminate common source infection (Weese, 2010).

Another microorganism with high zoonotic potential is Salmonella spp., which causes salmonellosis, a severe, common enteric disease in both humans and animals (Lowdon et al., 2015; Verma, Sinha and Singh, 2011). Dogs are thought to be asymptomatic carriers of salmonella and have been shown to act as a vector for bacterial contamination throughout the household (Lambertini et al., 2016; Hoelzer, Switt and Wiedmann, 2011; Bagcigil et al., 2007; Cherry et al., 2004; Sato et al., 2000). Additionally, recent studies have also identified human salmonella outbreaks caused by feeding dogs both dry and raw diets (CDC, 2012; Behravesh et al., 2010; Lenz et al., 2009; CDC, 2008). Weese and Rousseau (2006) state that food bowls can act as a source of infection for humans and animals if not sufficiently disinfected.  They analysed the ability of salmonella to persist within both plastic and stainless steel food bowls, however no significant difference was determined when comparing overall persistence of salmonella and bowl material. A small sample size of 10 bowls was used, which significantly limits the statistical power of this study (Kylie et al., 2017). The presence of bacteria in food bowls may lead to cross contamination to water bowls, which are typically kept within the same area (Lambertini et al., 2016).

Current research suggests that animal drinking systems have the potential to act as a reservoir for multiple bacteria and that material does influence bacterial quantities (Folorunso, Kayode and Onibon, 2014; Martins et al., 2013; Donofrio et al., 2012; Watson et al., 2012; Cook, Britt and Bolster, 2010; Avery et al., 2008; LeJeune et al., 2001a). Weese, Rousseau and Arroyo (2005) state that pathogen survival in dog water bowls has not been adequately evaluated. Furthermore, to the author’s knowledge, there is no current research evaluating the influence of water bowl material on bacterial count and species present.


1.4 Aims and Objectives

1.4.1 Aims

The aim of the current study was to contribute to the limited existing literature by identifying whether the material of a dog’s water bowl – ceramic, plastic or stainless steel – and the length of use influences the quantity of bacteria present. Additionally, it was investigated whether the species identified from the water bowls varied between the three materials.


1.4.2 Objectives

  1. To determine whether there was a difference in bacterial count with different bowl materials.


Null hypothesis: There will be no significant difference between bacterial count with different bowl materials.


Alternative hypothesis:  There will be a significant difference between bacterial count with different bowl materials.


  1. To identify whether the length of use of the water bowl influences the bacterial count.


Null hypothesis:  There will be no significant effect of the length of use of the water bowl on the bacterial count.


Alternative hypothesis: There will be a significant effect of the length of use of the water bowl on the bacterial count.


  1. To qualitatively determine, by observation, whether there is a difference in the species identified between bowl materials.



2.0 Materials and Methods

2.1 Pilot Study

A two week pilot study was conducted in order to eliminate any problems that may be encountered during the six week data collection period. It was acknowledged that the use of nutrient broth to transport the samples encouraged growth and increased contamination risk, and was thus rendering the bacterial counts inaccurate. To rectify this, phosphate buffer saline (PBS) solution was used as the transport medium, similar to that of Donofrio et al. (2012), as this does not encourage bacterial growth.


2.2 Bowl Sterilisation

The three construction materials tested were plastic, stainless steel and ceramic. All bowls were purchased brand new and sterilised in the same way. Gloves were worn throughout the process to prevent contamination. Following adapted methods of Jensen et al. (2013), a sterile basin was filled with boiling water, Dettol and washing up liquid. Each bowl was submerged, scrubbed gently, rinsed thoroughly and shaken dry. Bowls were immediately placed into a sterile autoclave bag and sealed until required.


2.3 Agar Preparation

Prior to each two week sampling period, Blood Agar Base (Oxoid, UK, CM0055) was prepared following the manufacturer’s instructions. This was autoclaved at 121oC, 103 kPa for 15 minutes in order to eliminate microbial contamination (Choi, Rodrigez and Sobsey, 2014). Once the agar had cooled to approximately 50-60oC, defibrinated horse blood was added aseptically at a 5% concentration and approximately 20ml of the solution was poured into petri dishes – in an aseptic environment – and allowed to set. Blood agar was used to culture the bacteria due to it being non-selective, rich in nutrients and able to grow a variety of bacteria (Dortet et al., 2014).


2.4 Broth Preparation

Prior to each two week sampling period MRVP broth (Merck, UK, V964212) and Nutrient Broth (Oxoid, UK, CM0001) was produced according to the manufacturer’s guidelines. Once the MRVP broth had dissolved, 10ml was syringed into individual test tubes and sealed with a lid. The test tubes and nutrient broth were autoclaved at 121oC, 103 kPa for 15 minutes and kept in the fridge until required.


2.5 Sample Collection

A sample of 6 dogs (5 female and 1 male), classified as medium weight (10-25kg) were used in this study (Vicente and Hammond, 2017). The age of the subjects ranged from 2-7 years (mean= 3.8 ± 1.95), and subjects were of various breeds (3 Cockapoos, 1 Border terrier, 1 Springer spaniel and 1 Golden retriever). All dogs were clinically healthy, fed a dry diet and had similar lifestyles. For example, all subjects were walked twice daily, housed individually within a rural location and no other pets were present in the household.

Bowls were placed in the kitchen by the back door in all households and the same position was used for all three construction materials. All owners were informed to thoroughly wash their hands before and after filling the dog bowl, and asked not to rinse, wash or disinfect the bowl.  Samples were always collected at the same time in the morning. Gloves were worn throughout sample collections, sterile cotton swabs were used and all swabs were conducted by the same person (Donofrio et al., 2012), in the same manner – rubbed three times around the inner circumference of the bowl and once over the inner base, before being placed into 3ml sterile PBS. Samples were kept on ice and transported to the laboratory within two hours.

For the controls, on day 0 the bowl was filled with tap water, emptied and immediately swabbed, before being re-filled and placed in the selected position. Samples were collected again at day 7 and day 14.


2.6 Enumeration of Bacteria

2.6.1 Serial Dilution and Bacterial Count

Using modified methods of Folorunso, Kayode and Onibon (2014), Donofrio et al. (2010), LeJeune et al. (2001a) and LeJeune et al. (2001b), spread plate 10 fold serial dilutions were performed for each sample from 10-1 to 10-8. All serial dilutions were conducted within a biosafety cabinet to minimise contamination risk. The original sample was vortexed for 15 seconds using a Mini Vortex Mixer (Fisher Scientific, UK); 100μl aliquots were aseptically withdrawn using a micropipette and transferred into 900μl of sterile PBS (10-1). The solution was pipetted in and out to ensure thorough mixing. A 100μl aliquot was then aseptically withdrawn from the 10-1 solution and transferred into 10-2. This process was repeated until 10-8 was reached. A 100μl aliquot of each dilution from 10-3 to 10-8 was inoculated onto blood agar and spread using an L-shaped spreader. All plates were incubated at 37oC for 18-24 hours (Folorunso, Kayode and Onibon 2014).

Following incubation, total colony forming units (CFU) were counted and recorded. This was performed manually which is considered the gold standard approach (Davey, 2011; Clark et al., 2010). An average total CFU was calculated from the dilutions, for each sample, for each bowl type. Similar to LeJuene et al. (2001a) Plates with less than 30 colonies, or more than 300 colonies were excluded from the results as they are considered not statistically viable (Sutton, 2011; Rawling et al., 2009).


2.7 Isolation of Pure Bacteria

In a biosafety cabinet, the serial dilution plates were observed for individual, varying colonies. All colonies, from each sample, with a morphological difference were individually collected using a sterile, plastic inoculation loop and then streaked, using the four quadrant streaking method, onto blood agar plates to produce pure colonies. Plates were incubated at 37oC for 24 hours (Folorunso, Kayode and Onibon 2014).


2.8 Bacterial Identification

Pure bacterial samples were identified by their morphological characteristics, gram stain, and biochemical test results, similar to the methods of Folorunso, Kayode and Onibon (2014). The catalase, oxidase, Voges-Proskauer and methyl red test were conducted and interpreted following methods presented in literature (Sawain et al., 2018; Cobos-Trigueros et al., 2017; den Bakker et al., 2014; Hemraj, Diksha and Aveneet, 2013).


2.9 Freezing of Bacterial Isolates

The pure bacterial isolates from day 14, which were in the late log phase, were frozen at -80oC until required for polymerase chain reaction (PCR) (Kataoka et al, 2013).


2.10 Identification by Polymerase Chain Reaction

Following biochemical test results, further identification for the presence of potential medically important bacteria, isolated from day 14, was carried out using PCR.


2.10.1 DNA Extraction

The frozen samples were thawed at room temperature. Following 15 seconds of vortexing, using a Mini Vortex Mixer (Fisher Scientific, UK), a 100μl aliquot of each sample was added to an individual, labelled blood agar plate and spread using an inoculation ‘L-shaped’ spreader. The plates were incubated at 37oC for 48 hours. After incubation, DNA was extracted following the guidelines of Dashti et al. (2009).


2.10.2 PCR and Primers

The specific primers (Thermo Fisher Scientific, UK) selected for PCR and expected amplicon product size (bp) are displayed in table 1.

All primers were centrifuged for 30 seconds, hydrated using sterile distilled water according to manufacturer’s guidelines and stored at -20oC until required.


Table 1: A table to show the forward and reverse primer sequences used and the size of the amplicon product (bp)

Organism/ gene Primer Sequence


Size of amplicon product (bp) Reference
E.coli O157:H7/ eaeA (Forward) AAG CGA CTG AGG TCA CT 450 (Holland et al., 2000; Louie et al., 1994)
E.coli O157:H7/ eaeA


MRSA/ mecA (Forward) AAA ATC GAT GGT AAA GGT TGG C 533 (Bühlmann et al., 2008)
Pasteurella multocida/ kmt


TGC CAC TTG AAA TGG GAA ATG 168 (Król et al., 2011)
Pasteurella multocida/ kmt


Pasteurella canis/ sod A (Forward) GTA AAT AAT GCA AAT GCG G 186
Pasteurella canis/ sod A


Staphylococcus species/ TstaG422


GGC CGT GTT GAA CGT GGT CAA ATC A 370 (Martineau et al., 2001)
Staphylococcus species/ Tstag765


Bacillus species/ p8FPL


AGT TTG GAT CCT GGC TCA G 78 (Fernández-No et al., 2011)
Bacillus species/ p806R


Salmonella species/ ttr


CTC ACC AGG AGA TTA CAA CAT GG 94 (Gwida and Al-Ashmawy, 2014)
Salmonella species/ ttr




A QuantiTect® Probe PCR Kit was used to prepare samples for PCR. A total volume of 20μl was used for amplification. Following the protocol, 10μl of QuantiTect® Probe PCR 2x Master Mix, 0.1μl of specific forward primer, 0.1μl of specific reverse primer and 4.8μl of RNase-free water were added aseptically to each sample tube. A 5μl aliquot of the DNA template was added to each tube. A blank was created using 5μl of RNase-free water instead of template DNA. Aseptic techniques were used throughout to eliminate contamination.   All PCR was run using a Techne TC-3000 thermocycler (VWR, UK).


2.10.3 Gel Electrophoresis

A 50x stock solution of Tris acetate EDTA (TAE) was prepared and diluted to a 1x concentration by adding 980ml of distilled water to 20ml of 50x TAE. Each amplified DNA fragment required a different percentage of agarose gel for analysis. The appropriate weight of Electran® Agarose DNA grade (VWR,Belgium) was added to 150ml of TAE in a conical flask and stirred until dissolved. The solution was heated for 2 minutes and then heated gently until the solution was clear. After slight cooling, the gel was poured into a horizontal gel tank and combs were added. Once the gel had set, the combs were removed gently to produce the wells. TAE x1 was poured over the gel until it was covered.

Using a micropipette, 5μl of 100bp DNA marker, containing loading dye, was added to 2μl of 10x GelRed and 6μl of this solution was added to the first and last well. A 5μl aliquot of DNA sample was mixed with 2μl of 10x GelRed and 1μl of 6x loading dye (bromophenol blue) and then 6μl was added to the remaining wells. All runs of gel electrophoresis were examined by UV light using the UVP gel Doc-it reading system. Observations were made for clear bands at the specific amplicon product size (bp), displayed in table 1.


2.11 Statistical analysis

All statistical analysis was carried out using SPSS software (IBM SPSS, version 24, USA). Tests were conducted on the average total CFU/ml for each sample, for each bowl type. Descriptive statistics including mean and standard deviation were calculated. The Kolmogorov Smirnoff test was conducted to determine normality. A Kruskal Wallis was used to determine whether there is a significant difference between the average total CFU/ml for each bowl construction material. A Mann Whitney or independent T test, depending on normality, was used to determine if there is a significant difference in average total CFU/ml between day 0 and day 7, day 7 and day 14, and day 0 and day 14.


2.12 Ethical considerations

Recent studies have found evidence of dogs having the ability to remember the physical properties of objects (Kundey et al., 2010; Pattison et al., 2010). Additionally, exposure to novel surroundings has been shown to cause dogs stress (Rooney, Gaines and Bradshaw, 2007). Changing the material of a dog’s bowl could induce stress. If any signs of stress or adversity to drinking were displayed, owners were advised to remove their dog from the study.

Owners provided informed consent for each subject via the completion of a participant consent form. All data provided was held in accordance of The Data Protection Act (1998). The researcher and owners had the right to withdraw a subject at any time, for any reason, until the point of data analysis.


3.0 Results

3.1 Bacterial Enumeration

For each bowl material on each day, an average total CFU/ml was calculated using all samples. The average total CFU/ml for each bowl material over the two week period is presented in table 2 and figure 1. No bacterial growth was observed from the initial control samples taken on day 0. Bacteria were cultured from all three bowl materials on day 7 and day 14. However, for the plastic bowl material, two of the six samples showed no growth on both day 7 and day 14. On day 7, the stainless steel bowl had the highest bacterial count of 250.33 ± 236.47 CFU/ml x 106. However, although all counts were lower after 14 days, the plastic bowl had the highest remaining bacterial count with 43.26 ± 62.40 CFU/ml x 106. The ceramic bowl had the lowest bacterial count throughout all test days of the study.

Table 2: A table to show the average total CFU/ml ± standard deviation (SD) (1.0 x 106) for each bowl material over time.

Bowl material Average total ± SD CFU/ml (1.0 x 106)
Day 0 Day 7 Day 14
Ceramic 0.00 ± 0.00 18.56 ± 40.50 2.20 ± 2.65
Plastic 0.00 ± 0.00 92.28 ± 191.21 43.26 ± 62.40
Stainless steel 0.00 ± 0.00 250.33 ± 236.40 10.95 ± 16.96


The Kolmogorov Smirnoff test revealed the data to be mostly non-parametric (P<0.05). The Kruskal Wallis identified no significant difference when comparing bowl material and total CFU/ml (P>0.05); therefore the null hypothesis that there is no significant difference between bacterial count and bowl material must be retained. When testing for difference between day and average total CFU/ml, the Mann Whitney U test revealed a significant difference between day 0 and day 7 and between day 0 and day 14 for all bowl materials (P<0.05); thus the null hypothesis that there is no significant difference between the length of use of the water bowl and the bacterial count can be rejected. However, no significant difference in total CFU/ml was identified for all bowl materials between day 7 and day 14 (P>0.05).

Figure 1

Figure 1: A bar chart presenting the average total CFU/ml (1.0 x 106) over time.


3.2 Bacterial Identification

3.2.1 Biochemical test results

If bacterial characteristics identified on day 7 were identified again on day 14, the suspected genus was used for the selection of specific PCR primers, which are displayed in table 1. The colony morphology and biochemical test results for these isolates are presented in table 3. The results for the Voges-Proskauer tests are considered anomalies as no colour change was observed for this test for all isolates on day 7 and day 14.


Table 3: A table to show the colony morphology, biochemical test results and possible genus for all isolates that appeared in samples at both day 7 and day 14.

Colony morphology Biochemical test Possible genus
Macroscopic Microscopic Catalase Oxidase MR VP Gram







White-grey Circular

Low convex



Short bacilli


+ + E.coli









White-yellow Circular

Low convex





+ + + Staphylococcus









Dull brown





Long bacilli + + + Bacillus









Low convex



Short bacilli + + Pasteurella


3.2.2 Polymerase Chain Reaction Results

Following the biochemical test results, all day 14 samples were subjected to PCR for identification of seven hygienically important bacteria, presented in table 1. The results for PCR are presented in table 4. The E.coli O157:H7 primer identified one isolate from the ceramic bowl material, displayed by the faint amplification at 450 base pairs (bp). The MRSA primer detected seven samples with positive amplification at 533bp. From the seven positive isolates, three originated from the stainless steel bowl and four from the ceramic bowl. The primer for the genus of staphylococcus only displayed 3 positive amplifications at 370bp, two from the stainless steel bowl and one from ceramic. No clear bands were present at 168bp for identification of Pasteurella multocida. Two samples, one from both the stainless steel and plastic bowl were identified as positive for Pasteurella canis, displayed by the bands at 186bp. Three ceramic and one plastic bowl displayed amplification at 78bp for detection of Bacillus species. Five Salmonella species were identified by the presence of faint bands at 94bp. One of the positive isolates originated from the plastic bowl and the remaining four from the ceramic.


Table 4: A table to show the results from PCR. S1 – S6 represents the stainless steel bowl samples, C1 – C6 the ceramic bowl samples and P1-P5 the plastic bowl samples. ‘+’ represents the presence of the specific amplicon product.



E.coli O157:H7 MRSA Pasteurella


Pasteurella canis Staphylococcus Bacillus Salmonella
S2 +
S3 + +
S4 +
S5 +
S6 +
C1 + +
C2 + + +
C3 + + +
C4 + + +
C5 + +
P1 +
P3 +
P5 +
– Control (blank)
+ Control N/A N/A N/A N/A + N/A N/A


4.0 Discussion

At present, dogs are the most popular pet in the UK and concerns of zoonotic disease transmission are rising due to their increasingly close contact with owners (PMFA, 2018; Damborg et al., 2016; Wang et al., 2016; da Costa, Loureiro and Matos, 2013; Buma et al., 2006). Dog water bowls have the potential to act as a reservoir of zoonotic bacteria, similar to the findings in livestock water troughs (Folorunso, Kayode and Onibon, 2014; Watson et al., 2012; Cook, Britt and Bolster, 2010; Avery et al., 2008; Sargeant et al., 2003; McGee et al., 2002; LeJeune et al., 2001a; LeJeune et al., 2001b). The current study aimed to explore the effects of material, and length of use, on the quantity and species of bacteria present in dog water bowls. The results indicate that bowl material may influence bacterial count and the species isolated.


4.1 Bacterial Enumeration

4.1.1 Influence of bowl material on bacterial count

On day 7, there was no significant difference identified between bowl material and average total CFU/ml (P>0.05). Although not significantly different, the descriptive statistics suggest that the stainless steel bowl had the highest bacterial count, followed by plastic and then ceramic. Stainless steel is considered a hydrophilic material with a high surface energy and negative charge (Hamadi et al., 2014; Hočevar et al., 2014; Di Bonaventura et al., 2008; Sinde and Carballo, 2000; An and Friedman, 1998). Bacteria frequently attach in higher numbers to hydrophobic materials, such as plastic (Hočevar et al., 2014; Di Bonaventura et al., 2008; Sinde and Carballo, 2000); however Gebhardt et al. (2012) state that cell growth on hydrophobic surfaces is slower than on hydrophilic surfaces. Additionally, Flint, Brooks and Bremer (2000) examined the bacterial adhesion of two streptococcus species to stainless steel, glass and other metal coupons and concluded that negatively charged surfaces attracted more bacteria than positively charged surfaces. This study examined two bacterial species and therefore is not representative of all bacterial species adhesion mechanisms (Greene et al., 2016). Despite this, these findings may suggest why stainless steel had a higher bacterial count than plastic on day 7.

Similarly, on day 14, no significant difference was identified between bowl material and average total CFU/ml (P>0.05). In contrast to day 7, the plastic bowl had the highest bacterial count. Plastic materials are known to degrade rapidly and surface scratches and wear increase bacterial adhesion (Bohinc et al., 2014; Crawford et al., 2012; Verran et al., 2008). An investigation over a longer time period may be beneficial. Cook, Britt and Bolster (2010) examined bacterial decay rate of Map over one year and identified that plastic disks have slower bacterial decay rates than stainless steel disks, although not significantly different. This study was conducted in controlled laboratory conditions, and may not be applicable in vivo (Sorrentino et al., 2018; Cook, Britt and Bolster, 2010). Despite this, LeJeune et al. (2001a) observed metal water troughs to have significantly lower bacterial counts than plastic water troughs over 8 months. A large sample of 467 troughs was studied, suggesting this data to be reliable with high statistical power (Halsey et al., 2015). The lack of statistical significance found in the current study is likely to be due to the small sample used and larger samples should be considered in future research (Halsey et al., 2015).

The ceramic bowl obtained the lowest bacterial count on both day 7 and day 14, although it was not significantly different from the other materials. There have been several reports that ceramic materials exhibit lower bacterial adhesion than other materials due to its higher hydrophilicity (Hofs et al., 2011; Eick et al., 2004). Sorrentino et al. (2018) studied the bacterial adhesion of S.aureus and S.epidermis to ceramic, metal and plastic disks and identified that ceramic materials had a significantly lower bacterial adhesion, slower biofilm development and thinner biofilm formation than the other materials (P<0.05). Additionally, the plastic material showed the highest bacterial adhesion, in line with the findings of this study. Similarly, Fatoba et al. (2014) placed various materials into individual liquid cultures of Bacillus, E.coli and S.aureus and conducted bacterial counts after various time periods. Their findings suggest that the ceramic material obtained lower bacterial counts, followed by plastic and then stainless steel when immersed in E.coli. However, all three materials had very similar counts for both Bacillus and S.aureus. Although there is a body of evidence that suggests ceramic materials obtain lower bacterial counts – similar to the findings of the current study – some studies demonstrate that adhesion is bacteria dependent (Fatoba et al., 2014; Ribeiro, Monteiro and Ferraz, 2012; Hofs et al., 2011; Van Houdt and Michiels, 2010; Eick et al., 2004; Faille et al., 2002). This highlights the need for further research in this area.

Although there was no significant difference identified between bowl material and average total CFU/ml in the current study, descriptive statistics and the above studies support the findings that the ceramic bowl has lower bacterial adhesion. However, further research is required to identify if there is a more hygienic bowl material. Additionally, Crawford et al. (2012) state that it is imperative to have a comprehensive characterisation of the specific surface properties when studying bacterial adhesion to materials and therefore thorough examination of the materials should be implemented in future research.


4.1.2 Influence of length of use on bacterial count

A significant difference was identified for all bowl materials when comparing average total CFU/ml between both day 0 and day 7, and day 0 and day 14 (P<0.05). Bacterial count increased significantly between days 0 and 7. This demonstrates that when bowls are refilled without cleaning, they produce a favourable environment for bacterial proliferation (Folorunso, Kayode and Obinon, 2014). This is in agreement with the findings of Folorunso, Kayode and Onibon (2014) who studied the bacterial count in poultry water troughs at three farms and found that the bacterial load progressively increases over 7 days.

Interestingly, the bacterial counts reduced for all bowl materials on day 14; however they were not significantly different from day 7.  This suggests that high quantities of bacteria can persist within dog bowl biofilms. Many environmental factors, such as competition and nutrient availability, influence bacterial survival and adhesion – which may suggest why the bacterial counts slightly reduce (Crawford et al., 2012; Di Bonaventura et al., 2008; Katsikogianni and Missirlis, 2004). Many cells aggregated as part of a biofilm inhibit their motility, preventing them from searching for optimal environments when nutrient depletion occurs, ultimately causing cell death (Tuson and Weibel, 2013). Furthermore, the bacteria may have been removed by the dog or upon refilling.  Additionally, the swab taken from the bowl surfaces on day 7 is likely to have removed and disrupted bacteria present in the biofilm, reducing the count on day 14. This represents a major limitation of the current study which may be overcome in future by conducting swabs on day 7 and then beginning a new 14 day sampling period with no day 7 swab. Despite this, the lack of significant difference between day 7 and day 14 highlights the importance of daily cleaning regimes as high levels of bacteria in dog drinking water could significantly impact their health and welfare and increase transmission to human cohabitants – particularly immunocompromised individuals (Damborg et al., 2016; Folorunso, Kayode and Obinon, 2014; Weese et al., 2010). Furthermore, this study demonstrates that the water bowl acts as a fomite; therefore it may be a vector of diseases with high veterinary importance.


4.2 Bacterial Identification

4.2.1 Biochemical tests

Bacteria with the same biochemical characteristics isolated from both day 7 and day 14 were identified to the genus level using biochemical tests and morphological characteristics (Vaseekaran, Balakumar and Arasaratnam, 2010). The isolates were identified as E.coli, Salmonella, Staphylococcus, Pasteurella and Bacillus, according to biochemical results presented in literature (Folorunso, Kayode and Onibon, 2014; Patnaik, Prasad and Ganguly, 2014; Verma et al., 2013; Vaseekaran, Balakumar and Arasaratnam, 2010). The limited number of biochemical tests used in this study meant that accurate identification was difficult (Janda and Abbott, 2002). Additionally, the Voges-Proskauer test results were identified as anomalies and excluded from the results. Biochemical tests have low specificity and interpretation of can be difficult if untrained (vanVeen, Claas and Kuijper, 2010). Furthermore, many diverse species can share the same biochemical characteristics, causing unreliability (Srinivasan et al., 2015; Heikens et al., 2005). Despite this, they allowed for selection of appropriate primers for PCR identification. Accurate identification of microorganisms is fundamental in order to implement suitable infection control and improve general hygiene (Nomura, 2015; Pahlow et al., 2015; Srinivasan et al., 2015).


4.2.2 Polymerase Chain Reaction

PCR provides a highly sensitive and specific method for bacterial species identification and avoids the limitations of biochemical tests (Pahlow et al., 2015, Cherkaoui et al., 2010; Vasoo, Stevens and Singh, 2009). Interestingly, the bacteria tested for by PCR were most frequently isolated from the ceramic bowl, followed by stainless steel and then plastic, suggesting an influence of bowl material on species.

E.coli O157:H7 was identified in one sample for the ceramic bowl material. The genus E.coli was identified in many bowls from the biochemical tests; however many species of E.coli are commensal and are therefore typically harmless (Chandran et al., 2017; Odonkor and Ampofo, 2013). E.coli O157:H7 is a highly pathogenic serotype of E.coli often found on the intestinal tract of many animals; therefore dogs are likely to come into contact with contaminated faecal matter when on walks (Lambertini et al., 2016; Mendonça et al., 2012). The presence of E.coli O157:H7 in one ceramic water bowl is concerning as infection can cause bloody diarrhoea, hemorrhagic colitis and fatal haemolytic uremic syndrome and it is consequently a major threat to public health (Albanese et al., 2018; Gould et al., 2016; Li et al., 2014).

Staphylococcus spp. was confirmed by PCR to be present in two stainless steel bowls and one ceramic bowl. The low presence of Staphylococcus is surprising as S.aureus has been previously isolated from dog water bowls (Donofrio et al., 2012). Furthermore, a study by Bean and Wigmore (2013) examined 117 healthy dogs for the presence of S.aureus and S.pseudointermedius and concluded that the mouth and perineum were the most common sites for detection of these bacteria. This suggests that these species are likely to be present within water bowls. However, this study was conducted in Australia and thus cannot be generalised to the worldwide dog population (Jaeger et al., 2010). Despite this, studies in Denmark have also suggested the mouth and perineum as the area’s most highly colonised with S.pseudointermedius (Paul et al., 2012). Additionally, studies assessing the microbiology of canine bite wounds have identified Staphylococcus spp. as one of the most commonly isolated bacteria (Abrahamian and Goldstein, 2011; Thomas and Brook, 2011; Talan et al., 1999). These studies suggest that Staphylococcus spp. are highly prevalent in dog mouths. The low presence of these species in this study may be as a result of an unsuitable primer, as it was designed for qPCR, which was not performed in this study (Martineau et al., 2001); however the positive control amplified with a very clear band.

Interestingly, seven isolates of MRSA were identified, three in the stainless steel bowl and four in the ceramic. This may suggest that false negative or false positive results were obtained due to the discrepancies between the Staphylococcus and MRSA findings. The clearest amplicon band for MRSA was seen in sample C5, which was negative in the Staphylococcus PCR. This suggests the error is in the Staphylococcus primer where minor modifications were made to the cycling conditions and this may have impacted the efficacy. Despite this, the identification of MRSA in the water bowl is concerning as it can be transferred to humans and cause life threatening illness (Faires, Tater and Weese, 2009; Sing, Tuschak and Hörmansdorfer, 2008; Boost, O’donoghue and Siu, 2007; Weese, 2006). Mafu et al. (2013) state that there is limited knowledge about the adhesion of MRSA to surfaces and the current study suggests it preferentially adheres to hydrophilic materials. Alternatively, the other bowls may not have been contaminated with the bacterium. Balen et al. (2013) showed that MRSA can survive within a dog water bowl for 2 months; suggesting the need for a suitable cleaning regime that is effective against this highly resistant bacteria.

Salmonella species were confirmed in four ceramic bowls and one plastic bowl. Salmonella has hydrophobic cell properties and has been previously reported to preferentially adhere to hydrophobic surfaces, such as plastic (Veluz, Pitchiah and Alvarado, 2012; Chia et al., 2009; Chia et al., 2008). This does not agree with the findings of the current study; however, Oliveira et al. (2006) suggest that adherence is entirely strain dependent. Furthermore, the bacterium may not have been introduced into all bowls. Although only 9% of human infections are attributable to animal contact, Salmonella results in severe gastrointestinal disorders (Lowdon et al., 2015; Verma, Sinha and Singh, 2011; Majowicz et al., 2010). It can also lead to abdominal pain, weakness and fever in dogs (Marks et al., 2011). Identifying water bowls as a vector for Salmonella transmission may lead to a reduction of infections if owners’ education is increased.

PCR identified Bacillus in three ceramic bowls and one plastic bowl. The primer identified B.cereus, B.licheniformis and B.subtilis and although it is not clear which exact species were isolated from these bowls, all of these species can be pathogenic (Fernández-No et al., 2011). Bacillus species are often found within soil, suggesting that dogs are likely to come into contact with it on walks (Lambertini et al., 2016; Earl, Losick and Kolter, 2008). Identification of Bacillus species is concerning as they develop spores; this increases their resistance to cleaning procedures and Faille et al. (2002) found that removal was harder on hydrophobic materials.

Pasteurella canis was identified in one stainless steel bowl and one plastic bowl. Pasteruella multocida was not identified. Pasteurella spp. are responsible for several soft tissue infections, occasional respiratory tract infections and in rare cases, meningitis (Faceira et al., 2017; Zambori et al., 2013).The low presence of these species was not expected as Pasteurella species were identified frequently from the biochemical tests and are also frequently isolated from dog mouths and bites (Faceira et al., 2017; Oehler et al., 2009; Talan et al., 1999). This may suggest PCR errors were encountered, for example, the heat lysis method of DNA extraction may not have been successful and thus there was no DNA present to be amplified.  Additionally, the ladder did not migrate clearly for P.multocida which may suggest that the gel did not run correctly.


4.3 Limitations

The limited sample used within this experiment does not allow for findings to be extrapolated to all dogs. In future, a larger sample, including a larger variety of breeds should be used. Additionally, the findings of this study are only applicable to dogs in a rural location of Devon and geographical studies should be considered to identify the difference between urban and rural housed dogs.

Although great care was taken to ensure all components of the PCR were managed correctly, the high number of samples meant that maintaining the reagents at 4oC was difficult. Prolonged exposure to higher temperatures can denature the Taq polymerase enzyme, reducing the sensitivity of PCR (Peters et al., 2004). In addition, non-specific activation of Taq polymerase can cause non-specific amplification (Ashrafi, Yee and Paul, 2009). Furthermore, the lack of positive controls throughout limited the ability to compare the amplifications to a known positive amplicon product.


4.4 Future Research

Although not tested in this study, Fatabo et al. (2014) demonstrated that copper and brass materials had antimicrobial and bactericidal effects against bacteria over time. Additionally, Mehtar et al. (2008) discovered that copper surfaces decreased bacterial loads from 107 CFU/ ml to below detectable limits within 180 minutes. However, these studies are conducted in laboratory settings and may not be representative of the antimicrobial properties of copper in heterogeneous bacterial environments (Gould et al., 2009). Despite this, the antimicrobial properties in these materials should be studied in vivo for the development of antimicrobial dog water bowls, reducing bacterial count and viability. This would subsequently improve dog welfare and health and reduce the risk of zoonotic transmission.

Additionally, due to the high level of antimicrobial resistance in biofilm bacteria, antimicrobial susceptibility testing should be carried out in order to identify effective disinfectants against the bacteria isolated from dog water bowls (Ciofu et al., 2017; Wall and Mah, 2017; Rowson and Townsend, 2016; Steenackers et al., 2012; Reller et al., 2009). This would allow for evidence-based cleaning routines to be developed (Leonard, 2014), improving both owner and dog health and welfare.


5.0 Conclusion

The current study contributes to the limited research on dog water bowl microbiology and begins to determine the impact of bowl material on the bacteria present. The findings demonstrate how ceramic, plastic and stainless steel water bowls allow for proliferation of bacteria – a health risk for owners and their dogs. No significant difference was identified between the bowl material and the bacterial count over two weeks, although descriptive statistics indicate that differences might be present, suggesting that further research involving a larger sample size is required. However, ceramic bowls appear to reduce bacterial proliferation more than stainless steel and plastic bowls. Additionally, other materials, such as copper, should be studied due to their antimicrobial properties. This would allow for identification of a more hygienic bowl that could reduce the risk of zoonotic transmission.

The length of use influences the bacterial count in all three bowl types, which demonstrates the need for daily cleaning regimes and increased owner awareness on the impacts of bowl hygiene. A study over a longer time period would be beneficial in order to identify the bowl material that minimises bacterial load. This would lead to improvements in both human and canine health.

PCR accurately identified multiple bacterial species, many of which have pathogenic potential. This presents the dog water bowl as a possible vector in disease transmission, both to humans and other animals. The frequent observation of MRSA and Salmonella isolated from the bowls is concerning due to their infectious nature for both humans and animals. However, this identifies an area where good hygiene implementation could reduce the number of infections from these bacteria and ultimately improve dog welfare. The ceramic bowl isolated the highest number of harmful bacteria, despite having the lowest bacterial count. This suggests that these bacteria may be able to adhere more easily to ceramic materials, demonstrating how bowl material may impact the species present in the bowl. Ultimately, more research is required before suitable conclusions can be drawn.



Abbott, Y., Leggett, B., Rossney, A.S., Leonard, F.C. and Markey, B.K. (2010) Isolation rates of meticillin-resistant Staphylococcus aureus in dogs, cats and horses in Ireland. The Veterinary Record [Online]. 166 (15), pp.451-455. [Accessed 10 February 2018].

Abrahamian, F.M. and Goldstein, E.J. (2011) Microbiology of animal bite wound infections. Clinical Microbiology Reviews [Online]. 24 (2), pp.231-246. [Accessed 09 March 2018].

Aklilu, E., Zunita, Z., Hassan, L. and Chen, H.C. (2010) Phenotypic and genotypic characterization of methicillin-resistant Staphylococcus aureus (MRSA) isolated from dogs and cats at University Veterinary Hospital, Universiti Putra Malaysia. Tropical Biomedicine [Online]. 27 (3), pp.483-492. [Accessed 10 February 2018].

Albanese, A., Sacerdoti, F., Seyahian, E.A., Amaral, M.M., Fiorentino, G., Brando, R.F., Vilte, D.A., Mercado, E.C., Palermo, M.S., Cataldi, A. and Zotta, E. (2018) Immunization of pregnant cows with Shiga toxin-2 induces high levels of specific colostral antibodies and lactoferrin able to neutralize E. coli O157: H7 pathogenicity. Vaccine [Online]. 36 (13), pp. 1728-1735. [Accessed 08 March 2018].

An, Y.H. and Friedman, R.J. (1998) Concise review of mechanisms of bacterial adhesion to biomaterial surfaces. Journal of Biomedical Materials Research Part A [Online]. 43 (3), pp.338-348. [Accessed 08 March 2018].

Ashrafi, E.H., Yee, J. and Paul, N. (2009) Selective control of primer usage in multiplex one-step reverse transcription PCR. BMC Molecular Biology [Online]. 10 (1), pp.113. [Accessed 09 March 2018].

Aspinall, V. (2012) The Complete Textbook of Veterinary Nursing. 2nd Edition. China: Elsevier Ltd.

Avery, L.M., Williams, A.P., Killham, K. and Jones, D.L., 2008. Survival of Escherichia coli O157: H7 in waters from lakes, rivers, puddles and animal-drinking troughs. Science of the Total Environment [Online]. 389 (2-3), pp.378-385. [Accessed 21 February 2018].

Bagcigil, A.F., Ikiz, S., Dokuzeylu, B., Basaran, B., Or, E. and Ozgur, N.Y. (2007) Fecal shedding of Salmonella spp. in dogs. Journal of Veterinary Medical Science [Online]. 69 (7), pp.775-777. [Accessed 21 February 2018].

Behravesh, C.B., Ferraro, A., Deasy, M., Dato, V., Moll, M., Sandt, C., Rea, N.K., Rickert, R., Marriott, C., Warren, K. and Urdaneta, V. (2010) Human Salmonella infections linked to contaminated dry dog and cat food, 2006–2008. Pediatrics [Online]. 126 (3), pp.477-483. [Accessed 18 January 2018].

Bohinc, K., Dražić, G., Oder, M., Jevšnik, M., Nipič, D., Godič-Torkar, K. and Raspor, P. (2014) Available surface dictates microbial adhesion capacity. International Journal of Adhesion and Adhesives [Online]. 50 (1), pp.265-272. [Accessed 09 March 2018].

Boost, M.V., O’donoghue, M.M. and Siu, K.H.G. (2007) Characterisation of methicillin-resistant Staphylococcus aureus isolates from dogs and their owners. Clinical Microbiology and Infection [Online]. 13 (7), pp.731-733. [Accessed 18 February 2018].

Bowman, D.D. and Lucio-Forster, A. (2010) Cryptosporidiosis and giardiasis in dogs and cats: veterinary and public health importance. Experimental Parasitology [Online]. 124 (1), pp.121-127. [Accessed 17 January 2018].

Bühlmann, M., Bögli-Stuber, K., Droz, S. and Mühlemann, K. (2008) Rapid screening for carriage of methicillin-resistant Staphylococcus aureus by PCR and associated costs. Journal of Clinical Microbiology [online]. 46 (7), pp.2151-2154. [Accessed 15 January 2018].

Buma, R., Maeda, T., Kamei, M. and Kourai, H. (2006) Pathogenic bacteria carried by companion animals and their susceptibility to antibacterial agents. Biocontrol Science [Online]. 11 (1), pp.1-9. [Accessed 18 January 2018].

Catry, B., Van Duijkeren, E., Pomba, M.C., Greko, C., Moreno, M.A., Pyörälä, S., Ružauskas, M., Sanders, P., Threlfall, E.J., Ungemach, F. and Törneke, K. (2010) Reflection paper on MRSA in food-producing and companion animals: epidemiology and control options for human and animal health. Epidemiology & Infection [Online]. 138 (5), pp.626-644. [Accessed 18 February 2018].

Centers for Disease Control and Prevention (CDC). (2012) Notes from the field: Human Salmonella infantis infections linked to dry dog food–United States and Canada, 2012. MMWR, Morbidity and Mortality Weekly Report [Online]. 61 (23), pp.436. [Accessed 22 February 2018].

Centers for Disease Control and Prevention (CDC). (2008) Multistate outbreak of human Salmonella infections caused by contaminated dry dog food–United States, 2006-2007. MMWR, Morbidity and Mortality Weekly Report [Online]. 57 (19), pp.521. [Accessed 23 February 2018].

Chia, T.W.R., Fegan, N., McMeekin, T.A. and Dykes, G.A. (2008) Salmonella Sofia differs from other poultry-associated Salmonella serovars with respect to cell surface hydrophobicity. Journal of Food Protection [Online]. 71 (12), pp.2421-2428. [Accessed 09 March 2018].

Chandran, S.P., Sarkar, S., Diwan, V., Pathak, A., Shah, H., Tamhankar, A.J., Macaden, R. and Stålsby-Lundborg, C. (2017) Detection of virulence genes in ESBL producing, quinolone resistant commensal Escherichia coli from rural Indian children. The Journal of Infection in Developing Countries [Online]. 11 (05), pp.387-392. [Accessed 08 March 2018].

Cherkaoui, A., Hibbs, J., Emonet, S., Tangomo, M., Girard, M., Francois, P. and Schrenzel, J. (2010) Comparison of two matrix-assisted laser desorption ionization-time of flight mass spectrometry methods with conventional phenotypic identification for routine identification of bacteria to the species level. Journal of Clinical Microbiology [Online]. 48 (4), pp.1169-1175. [Accessed 08 March 2017].

Cherry, B., Burns, A., Johnson, G.S., Pfeiffer, H., Dumas, N., Barrett, D., McDonough, P.L. and Eidson, M. (2004) Salmonella Typhimurium outbreak associated with veterinary clinic. Emerging infectious diseases [Online]. 10 (12), pp.2249. [Accessed 22 February 2018].

Choi, W.S., Rodríguez, R.A. and Sobsey, M.D. (2014) Persistence of viral genomes after autoclaving. Journal of Virological Methods [Online]. 198 (1), pp.37-40. [Accessed 28 September 2017].

Chomel, B.B. and Sun, B. (2011) Zoonoses in the bedroom. Emerging Infectious Diseases [Online]. 17 (2), pp.167-172. [Accessed 15 January 2018].

Ciofu, O., Rojo‐Molinero, E., Macià, M.D. and Oliver, A. (2017) Antibiotic treatment of biofilm infections. APMIS Journal of Pathology, Microbiology and Immunology [Online]. 125 (4), pp.304-319. [Accessed 20 January 2018].

Clarke, M.L., Burton, R.L., Hill, A.N., Litorja, M., Nahm, M.H. and Hwang, J. (2010) Low‐cost, high‐throughput, automated counting of bacterial colonies. Cytometry Part A [Online]. 77 (8), pp.790-797. [Accessed 28 September 2017].

Cobos-Trigueros, N., Zboromyrska, Y., Morata, L., Alejo Cancho, I., Calle, C.D.L., Vergara, A., Cardozo, C., Arcas, M.P., Soriano Viladomiu, A., Marco Reverté, F. and Mensa Pueyo, J. (2017) Time-to-positivity, type of culture media and oxidase test performed on positive blood culture vials to predict Pseudomonas aeruginosa in patients with Gram-negative bacilli bacteraemia. Revista Espanola de Quimioterapia [Online]. 30 (1), pp. 9-13. [Accessed 28 September 2017].

Crawford, R.J., Webb, H.K., Truong, V.K., Hasan, J. and Ivanova, E.P (2012) Surface topographical factors influencing bacterial attachment. Advances in colloid and interface science [Online]. 179 (1), pp.142-149. [Accessed 08 March 2018].

Cutler, S.J., Fooks, A.R. and Van Der Poel, W.H. (2010) Public health threat of new, reemerging, and neglected zoonoses in the industrialized world. Emerging Infectious Diseases [Online]. 16 (1), pp.1-7. [Accessed 18 January 2018].

Dashti, A.A., Jadaon, M.M., Abdulsamad, A.M. and Dashti, H.M. (2009) Heat treatment of bacteria: a simple method of DNA extraction for molecular techniques. Kuwait Medical Journal [Online]. 41 (2), pp.117-122. [Accessed 10 February 2018].

da Costa, P.M., Loureiro, L. and Matos, A.J. (2013) Transfer of multidrug-resistant bacteria between intermingled ecological niches: the interface between humans, animals and the environment. International Journal of Environmental Research and Public Health [Online]. 10 (1), pp.278-294. [Accessed 18 January 2018].

Damborg, P., Broens, E.M., Chomel, B.B., Guenther, S., Pasmans, F., Wagenaar, J.A., Weese, J.S., Wieler, L.H., Windahl, U., Vanrompay, D. and Guardabassi, L. (2016) Bacterial zoonoses transmitted by household pets: state-of-the-art and future perspectives for targeted research and policy actions. Journal of Comparative Pathology [Online]. 155 (1), pp.S27-S40. [Accessed 18 January 2018].

Data Protection Act 1998 [Online]. Chapter 29. (1998) GOV UK. Available from: http://www.legislation.gov.uk/ukpga/1998/29 [Accessed 25 September 2017].

Davey, H.M. (2011) Life, death, and in-between: meanings and methods in microbiology. Applied and Environmental Microbiology [Online]. 77 (16), pp.5571-5576. [Accessed 24 September 2018].

den Bakker, H.C., Warchocki, S., Wright, E.M., Allred, A.F., Ahlstrom, C., Manuel, C.S., Stasiewicz, M.J., Burrell, A., Roof, S., Strawn, L.K. and Fortes, E. (2014) Listeria floridensis sp. nov., Listeria aquatica sp. nov., Listeria cornellensis sp. nov., Listeria riparia sp. nov. and Listeria grandensis sp. nov., from agricultural and natural environments. International Journal of Systematic and Evolutionary Microbiology [Online]. 64 (6), pp.1882-1889. [Accessed 25 September 2018].

Deplazes, P., van Knapen, F., Schweiger, A. and Overgaauw, P.A. (2011) Role of pet dogs and cats in the transmission of helminthic zoonoses in Europe, with a focus on echinococcosis and toxocarosis. Veterinary Parasitology [Online]. 182 (1), pp.41-53. [Accessed 17 January 2018].

Dewhirst, F.E., Klein, E.A., Thompson, E.C., Blanton, J.M., Chen, T., Milella, L., Buckley, C.M., Davis, I.J., Bennett, M.L. and Marshall-Jones, Z.V. (2012) The canine oral microbiome. PloS one [Online]. 7 (4), pp. 1-12. [Accessed 18 January 2018].

de Vicente, F. and Hammond, G.(2017) Ultrasonographic identification of the dorsal atlantoaxial ligament in dogs. Veterinary Surgery [Online]. 46 (8), pp.1126-1130. [Accessed 28 September 2017].

Di Bonaventura, G., Piccolomini, R., Paludi, D., D’orio, V., Vergara, A., Conter, M. and Ianieri, A. (2008) Influence of temperature on biofilm formation by Listeria monocytogenes on various food‐contact surfaces: relationship with motility and cell surface hydrophobicity. Journal of applied microbiology [Online]. 104 (6), pp.1552-1561. [Accessed 08 March 2018].

Donofrio, R.S., Bechanko, R., Hitt, N., O’Malley, K., Charnauski, T., Bestervelt, L.L., Saha, R. and Saha, N. (2012) Are We Aware of Microbial Hotspots in Our Household? Journal of Environmental Health [Online]. 75 (2), pp. 12-19. [Accessed 27 September 2017].

Dortet, L., Bréchard, L., Poirel, L. and Nordmann, P., 2014. Impact of the isolation medium for detection of carbapenemase-producing Enterobacteriaceae using an updated version of the Carba NP test. Journal of Medical Microbiology [Online]. 63 (5), pp. 772-776. [Accessed 29 September 2017].

Dulo, F. and Pal, M. (2017) Emerging viral zoonoses and their implications on public health. World Applied Sciences Journal [Online]. 35 (2), pp.188-198. [Accessed 18 January 2018].

Earl, A.M., Losick, R. and Kolter, R. (2008) Ecology and genomics of Bacillus subtilis. Trends in Microbiology [Online]. 16 (6), pp.269-275. [Accessed 09 March 2018].

Eick, S., Glockmann, E., Brandl, B. and Pfister, W. (2004) Adherence of Streptococcus mutans to various restorative materials in a continuous flow system. Journal of Oral Rehabilitation [Online]. 31 (3), pp.278-285. [Accessed 09 March 2018].

Elliott, D.R., Wilson, M., Buckley, C.M. and Spratt, D.A. (2005) Cultivable oral microbiota of domestic dogs. Journal of Clinical Microbiology [Online]. 43 (11), pp.5470-5476. [Accessed 18 January 2018].

Faceira, A., Póvoa, S., Souteiro, P., Ceia, F. and Ferreira, S. (2017) Human infection by Pasteurella canis–A case report. Porto Biomedical Journal [Online]. 2 (2), pp.63-65. [Accessed 09 March 2018].

Faille, C., Jullien, C., Fontaine, F., Bellon-Fontaine, M.N., Slomianny, C. and Benezech, T. (2002) Adhesion of Bacillus spores and Escherichia coli cells to inert surfaces: role of surface hydrophobicity. Canadian Journal of Microbiology [Online]. 48 (1), pp.728-738. [Accessed 18 January 2018].

Faires, M.C., Tater, K.C. and Weese, J.S. (2009) An investigation of methicillin-resistant Staphylococcus aureus colonization in people and pets in the same household with an infected person or infected pet. Journal of the American Veterinary Medical Association [Online]. 235 (5), pp.540-543. [Accessed 18 February 2018].

Fatoba, O.S., Esezobor, D.E., Akanji, O.L., Fatoba, A.J., Macgregor, D. and Etubor, J. (2014) The Study of the Antimicrobial Properties of Selected Engineering Materials’ Surfaces. Journal of Minerals and Materials Characterization and Engineering [Online]. 2 (2), pp.78-87. [Accessed 09 March 2018].

Fernández-No, I.C., Guarddon, M., Böhme, K., Cepeda, A., Calo-Mata, P. and Barros-Velázquez, J. (2011) Detection and quantification of spoilage and pathogenic Bacillus cereus, Bacillus subtilis and Bacillus licheniformis by real-time PCR. Food Microbiology [Online]. 28 (3), pp.605-610. [Accessed 15 January 2018].

Fèvre, E.M., Bronsvoort, B.M.D.C., Hamilton, K.A. and Cleaveland, S. (2006) Animal movements and the spread of infectious diseases. Trends in Microbiology [Online]. 14 (3), pp.125-131. [Accessed 17 January 2018].

Flemming, H.C. (2011) Microbial biofouling: unsolved problems, insufficient approaches, and possible solutions [Online]. In: Biofilm highlights (pp. 81-109). Berlin Heidelberg: Springer. [Accessed 20 January 2018].

Flemming, H.C., Percival, S.L. and Walker, J.T. (2002) Contamination potential of biofilms in water distribution systems. Water Science and Technology: Water Supply [Online]. 2 (1), pp.271-280. [Accessed 20 January 2018].

Folorunso, O.R., Kayode, S. and Onibon, V.O. (2014) Poultry farm hygiene: microbiological quality assessment of drinking water used in layer chickens managed under the battery cage and deep litter systems at three poultry farms in southwestern Nigeria. Pakistan Journal of Biological Sciences [Online]. 17 (1), pp. 74-79. [Accessed 19 September 2017].

Frantz, L.A., Mullin, V.E., Pionnier-Capitan, M., Lebrasseur, O., Ollivier, M., Perri, A., Linderholm, A., Mattiangeli, V., Teasdale, M.D., Dimopoulos, E.A. and Tresset, A. (2016) Genomic and archaeological evidence suggest a dual origin of domestic dogs. Science [Online]. 352 (6290), pp.1228-1231. [Accessed 18 January 2018].

Fuster-Valls, N., Hernández-Herrero, M., Marín-de-Mateo, M. and Rodríguez-Jerez, J.J. (2008) Effect of different environmental conditions on the bacteria survival on stainless steel surfaces. Food Control [Online]. 19 (3), pp.308-314. [Accessed 02 March 2018].

Gebhardt, F., Seuss, S., Turhan, M.C., Hornberger, H., Virtanen, S. and Boccaccini, A.R. (2012) Characterization of electrophoretic chitosan coatings on stainless steel. Materials Letters [Online]. 66 (1), pp.302-304. [Accessed 08 March 2018].

Gharechahi, M., Moosavi, H. and Forghani, M. (2012) Effect of surface roughness and materials composition on biofilm formation. Journal of Biomaterials and Nanobiotechnology [Online]. 3 (4A), pp.541-546. [Accessed 20 January 2018].

Gould, L.H., Mody, R.K., Ong, K.L., Clogher, P., Cronquist, A.B., Garman, K.N., Lathrop, S., Medus, C., Spina, N.L., Webb, T.H. and White, P.L., (2013) Increased recognition of non-O157 Shiga toxin–producing Escherichia coli infections in the United States during 2000–2010: epidemiologic features and comparison with E. coli O157 infections. Foodborne Pathogens and Disease [Online]. 10 (5), pp.453-460. [Accessed 08 March 2018].

Gould, S.W., Fielder, M.D., Kelly, A.F., Morgan, M., Kenny, J. and Naughton, D.P. (2009) The antimicrobial properties of copper surfaces against a range of important nosocomial pathogens. Annals of Microbiology [Online]. 59 (1), pp.151-156. [Accessed 09 March 2018].

Greene, C., Wu, J., Rickard, A.H. and Xi, C. (2016) Evaluation of the ability of Acinetobacter baumannii to form biofilms on six different biomedical relevant surfaces. Letters in Applied Microbiology [Online]. 63 (4), pp.233-239. [Accessed 09 March 2018].

Guardabassi, L., Loeber, M.E. and Jacobson, A. (2004) Transmission of multiple antimicrobial-resistant Staphylococcus intermedius between dogs affected by deep pyoderma and their owners. Veterinary Microbiology [Online]. 98 (1), pp.23-27. [Accessed 20 January 2018].

Guardabassi, L., Schwarz, S. and Lloyd, D.H. (2004) Pet animals as reservoirs of antimicrobial-resistant bacteria. Journal of Antimicrobial Chemotherapy [Online]. 54 (2), pp.321-332. [Accessed 18 January 2018].

Gwida, M.M. and AL-Ashmawy, M.A. (2014) Culture versus PCR for Salmonella species identification in some dairy products and dairy handlers with special concern to its zoonotic importance. Veterinary Medicine International [Online]. 2014 (2014), pp. 1-5. [Accessed 18 January 2018].

Hall, C.W. and Mah, T.F. (2017) Molecular mechanisms of biofilm-based antibiotic resistance and tolerance in pathogenic bacteria. FEMS Microbiology Reviews [Online]. 41 (3), pp.276-301. [Accessed 20 January 2018].

Halsey, L.G., Curran-Everett, D., Vowler, S.L. and Drummond, G.B. (2015) The fickle P value generates irreproducible results. Nature Methods [Online]. 12 (3), pp.179-185. [Accessed 08 March 2018].

Hamadi, F., Asserne, F., Elabed, S., Bensouda, S., Mabrouki, M. and Latrache, H. (2014) Adhesion of Staphylococcus aureus on stainless steel treated with three types of milk. Food Control [online]. 38 (1), pp.104-108. [Accessed 08 March 2018].

Han, J.I., Yang, C.H. and Park, H.M. (2016) Prevalence and risk factors of Staphylococcus spp. carriage among dogs and their owners: A cross-sectional study. The Veterinary Journal [Online]. 212 (1), pp.15-21. [Accessed 12 February 2018].

Heikens, E., Fleer, A., Paauw, A., Florijn, A. and Fluit, A.C. (2005) Comparison of genotypic and phenotypic methods for species-level identification of clinical isolates of coagulase-negative staphylococci. Journal of Clinical Microbiology [Online]. 43 (5), pp.2286-2290. [Accessed 08 March 2018].

Hemraj, V., Diksha, S. and Avneet, G. (2013) A review on commonly used biochemical test for bacteria. Innovare Journal of Life Science [Online]. 1 (1), pp.1-7. [Accessed 28 September 2017].

Hočevar, M., Jenko, M., Godec, M. and Drobne, D. (2014) An overview of the influence of stainless-steel surface properties on bacterial adhesion. Materials and Technology [Online]. 48 (5), pp.609-617. [Accessed 08 March 2018].

Hoelzer, K., Switt, A.I.M. and Wiedmann, M. (2011) Animal contact as a source of human non-typhoidal salmonellosis. Veterinary Research [Online]. 42 (1), pp.34. [Accessed 21 February 2018].

Hoffmann, A.R., Patterson, A.P., Diesel, A., Lawhon, S.D., Ly, H.J., Stephenson, C.E., Mansell, J., Steiner, J.M., Dowd, S.E., Olivry, T. and Suchodolski, J.S. (2014) The skin microbiome in healthy and allergic dogs. PloS one [Online]. 9 (1), pp. 1-12. [Accessed 18 January 2018].

Hofs, B., Ogier, J., Vries, D., Beerendonk, E.F. and Cornelissen, E.R. (2011) Comparison of ceramic and polymeric membrane permeability and fouling using surface water. Separation and Purification Technology [Online]. 79 (3), pp.365-374. [Accessed 09 March 2018].

Holland, J.L., Louie, L., Simor, A.E. and Louie, M. (2000) PCR detection of Escherichia coli O157: H7 directly from stools: evaluation of commercial extraction methods for purifying fecal DNA. Journal of Clinical Microbiology [Online]. 38 (11), pp.4108-4113. [Accessed 15 January 2018].

Jensen, D.A., Friedrich, L.M., Harris, L.J., Danyluk, M.D. and Schaffner, D.W. (2013) Quantifying transfer rates of Salmonella and Escherichia coli O157: H7 between fresh-cut produce and common kitchen surfaces. Journal of Food Protection [Online]. 76 (9), pp.1530-1538. [Accessed 28 February 2018].

Jaeger, K., Linek, M., Power, H.T., Bettenay, S.V., Zabel, S., Rosychuk, R.A.W. and Mueller, R.S. (2010) Breed and site predispositions of dogs with atopic dermatitis: a comparison of five locations in three continents. Veterinary Dermatology [Online]. 21(1), pp.119-123. [Accessed 09 March 2018].

Jéquier, E. and Constant, F. (2010) Water as an essential nutrient: the physiological basis of hydration. European Journal of Clinical Nutrition [Online].  64 (2), p.115-123. [Accessed 20 January 2018].

Kataoka, Y., Ito, C., Kawashima, A., Ishii, M., Yamashiro, S., Harada, K., Ochi, H. and Sawada, T. (2013) Identification and antimicrobial susceptibility of enterococci isolated from dogs and cats subjected to differing antibiotic pressures. Journal of Veterinary Medical Science [Online]. 75 (6), pp.749-753. [Accessed 28 September 2017].

Katsikogianni, M. and Missirlis, Y.F. (2004) Concise review of mechanisms of bacterial adhesion to biomaterials and of techniques used in estimating bacteria-material interactions. European Cells and Materials [Online]. 8 (3), pp.37-57. [Accessed 09 March 2018].

Kenssington, M. (2014) Dog Care: How To Care For Your Dog: From Dog Health and Dog Nutrition To Dog Fitness, Dog Grooming, and more [Online]. 3rd Edition. Unknown: Michael Kenssington. [Accessed 18 January 2018].

Kundey, S.M., De Los Reyes, A., Taglang, C., Baruch, A. and German, R. (2010) Domesticated dogs’(Canis familiaris) use of the solidity principle. Animal Cognition [Online]. 13 (3), pp.497-505. [Accessed 24 September 2017].

Król, J., Bania, J., Florek, M., Pliszczak-Król, A. and Staroniewicz, Z. (2011) Polymerase chain reaction–based identification of clinically relevant Pasteurellaceae isolated from cats and dogs in Poland. Journal of Veterinary Diagnostic Investigation [Online]. 23 (3), pp.532-537. [Accessed 15 January 2018].

Kylie, J., McEwen, S.A., Boerlin, P., Reid-Smith, R.J., Weese, J.S. and Turner, P.V. (2017) Prevalence of antimicrobial resistance in fecal Escherichia coli and Salmonella enterica in Canadian commercial meat, companion, laboratory, and shelter rabbits (Oryctolagus cuniculus) and its association with routine antimicrobial use in commercial meat rabbits. Preventive Veterinary Medicine [Online]. 147 (1), pp.53-57. [Accessed 26 February 2017].

Lambertini, E., Buchanan, R.L., Narrod, C. and Pradhan, A.K. (2016) Transmission of bacterial zoonotic pathogens between pets and humans: The role of pet food. Critical Reviews in Food Science and Nutrition [Online]. 56 (3), pp.364-418. [Accessed 18 January 2018].

Lenz, J., Joffe, D., Kauffman, M., Zhang, Y. and LeJeune, J. (2009) Perceptions, practices, and consequences associated with foodborne pathogens and the feeding of raw meat to dogs. The Canadian Veterinary Journal [Online]. 50 (6), pp.637-643. [Accessed 22 February 2018].

Leonard, E.K. (2014) What Could Your Dog Be Carrying?–Zoonotic Enteric Bacteria in Pet Dogs in Ontario: Prevalence, Risk Factors, and Antimicrobial Resistance (Doctoral dissertation).

LeJeune, J.T., Besser, T.E., Merrill, N.L., Rice, D.H. and Hancock, D.D. (2001a) Livestock drinking water microbiology and the factors influencing the quality of drinking water offered to cattle. Journal of Dairy Science [Online]. 84 (8), pp.1856-1862. [Accessed 19 September 2017].

LeJeune, J.T., Besser, T.E. and Hancock, D.D. (2001b) Cattle Water Troughs as Reservoirs of Escherichia coli O157. Applied and Environmental Microbiology [Online]. 67 (7), pp.3053-3057. [Accessed 19 September 2017].

Li, Y., Afrasiabi, R., Fathi, F., Wang, N., Xiang, C., Love, R., She, Z. and Kraatz, H.B. (2014) Impedance based detection of pathogenic E. coli O157: H7 using a ferrocene-antimicrobial peptide modified biosensor. Biosensors and Bioelectronics [Online]. 58 (1), pp.193-199. [Accessed 08 March 2018].

Loeffler, A., Pfeiffer, D.U., Lindsay, J.A., Magalhães, R.S. and Lloyd, D.H. (2011) Prevalence of and risk factors for MRSA carriage in companion animals: a survey of dogs, cats and horses. Epidemiology & Infection [Online]. 139 (7), pp.1019-1028. [Accessed 10 February 2018].

Louie, M., De Azavedo, J., Clarke, R., Borczyk, A., Lior, H., Richter, M. and Brunton, J. (1994) Sequence heterogeneity of the eae gene and detection of verotoxin-producing Escherichia coli using serotype-specific primers. Epidemiology & Infection [Online]. 112 (3), pp.449-461. [Accessed 15 January 2018].

Lowden, P., Wallis, C., Gee, N. and Hilton, A. (2015) Investigating the prevalence of Salmonella in dogs within the Midlands region of the United Kingdom. BMC Veterinary Research [Online]. 11 (1), pp.239. [Accessed 21 February 2018].

Mafu, A.A., Massicotte, R., Pichette, G., Lafleur, S., Lemay, M.J. and Ahmad, D. (2013) Influence of surface and cloth characteristics on mechanical removal of meticillin-resistant Staphylococcus aureus (MRSA) attached to inanimate environmental surfaces in hospital and healthcare facilities. International Journal of Infection Control [Online]. 9 (3), pp.1-11. [Accessed 09 March 2018].

Majowicz, S.E., Musto, J., Scallan, E., Angulo, F.J., Kirk, M., O’brien, S.J., Jones, T.F., Fazil, A., Hoekstra, R.M. and International Collaboration on Enteric Disease “Burden of Illness” Studies. (2010) The global burden of nontyphoidal Salmonella gastroenteritis. Clinical Infectious Diseases [Online]. 50 (6), pp.882-889. [Accessed 09 March 2018].

Marks, S.L., Rankin, S.C., Byrne, B.A. and Weese, J.S. (2011) Enteropathogenic bacteria in dogs and cats: diagnosis, epidemiology, treatment, and control. Journal of Veterinary Internal Medicine [Online]. 25 (6), pp.1195-1208. [Accessed 09 March 2018].

Martineau, F., Picard, F.J., Ke, D., Paradis, S., Roy, P.H., Ouellette, M. and Bergeron, M.G. (2001) Development of a PCR assay for identification of staphylococci at genus and species levels. Journal of Clinical Microbiology [online]. 39 (7), pp.2541-2547. [Accessed 15 January 2018].

Martins, L.R.L., Simões, R.L.R., de Matos, A.J.F., da Costa, P.M.R. and Salazar, A. (2013) Common phenotypic and genotypic antimicrobial resistance patterns found in a case study of multiresistant E. coli from cohabitant pets, humans, and household surfaces. Journal of Environmental Health, 75 (6), pp.74-81. [Accessed 20 January 2018].

McGee, P., Bolton, D.J., Sheridan, J.J., Earley, B., Kelly, G. and Leonard, N. (2002) Survival of Escherichia coli O157: H7 in farm water: its role as a vector in the transmission of the organism within herds. Journal of applied microbiology [Online]. 93 (4), pp.706-713. [Accessed 21 February 2018].

Mehtar, S., Wiid, I. and Todorov, S.D. (2008) The antimicrobial activity of copper and copper alloys against nosocomial pathogens and Mycobacterium tuberculosis isolated from healthcare facilities in the Western Cape: an in-vitro study. Journal of Hospital Infection [Online]. 68 (1), pp.45-51. [Accessed 09 March 2018].

Mendonça, R.C.S., Morelli, A.M.F., Pereira, J.A.M., de Carvalho, M.M. and de Souza, N.L. (2012) Prediction of Escherichia coli O157: H7 adhesion and potential to form biofilm under experimental conditions. Food Control [Online]. 23 (2), pp.389-396. [Accessed 08 March 2018].

Morita, Y., Komoda, E., Ono, K. and Kumagai, S. (2011) Survival of biofilm-forming Salmonella on stainless steel bolt threads under dry conditions. Journal of the Food Hygienic Society of Japan [Online]. 52 (5), pp.299-303. [Accessed 20 January 2018].

Moyaert, H., De Graef, E.M., Haesebrouck, F. and Decostere, A. (2006) Acquired antimicrobial resistance in the intestinal microbiota of diverse cat populations. Research in Veterinary Science [Online]. 81 (1), pp.1-7. [Accessed 18 January 2018].

Nomura, F. (2015) Proteome-based bacterial identification using matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS): A revolutionary shift in clinical diagnostic microbiology. Biochimica et Biophysica Acta (BBA)-Proteins and Proteomics [Online]. 1854 (6), pp.528-537. [Accessed 03 March 2018].

Odonkor, S.T. and Ampofo, J.K. (2013) Escherichia coli as an indicator of bacteriological quality of water: an overview. Microbiology Research [Online].  4 (1), pp. 5-11. [Accessed 08 March 2018].

Oehler, R.L., Velez, A.P., Mizrachi, M., Lamarche, J. and Gompf, S. (2009) Bite-related and septic syndromes caused by cats and dogs. The Lancet Infectious Diseases [Online]. 9 (7), pp.439-447. [Accessed 09 March 2018].

Oliveira, K., Oliveira, T., Teixeira, P., Azeredo, J., Henriques, M. and Oliveira, R. (2006) Comparison of the adhesion ability of different Salmonella Enteritidis serotypes to materials used in kitchens. Journal of Food Protection [Online]. 69 (10), pp.2352-2356. [Accessed 09 March 2018].

Pahlow, S., Meisel, S., Cialla-May, D., Weber, K., Rösch, P. and Popp, J (2015) Isolation and identification of bacteria by means of Raman spectroscopy. Advanced Drug Delivery Reviews [Online]. 89 (1), pp.105-120. [Accessed 03 March 2018].

Pantosti, A. (2012) Methicillin-resistant Staphylococcus aureus associated with animals and its relevance to human health. Frontiers in microbiology [Online]. 3 (1), pp.127. [Accessed 21 February 2018].

Patnaik, S., Prasad, A. and Ganguly, S. (2014) Biochemical characterization and antibiogram of Staphylococcal microorganisms associated with subclinical mastitis in lactating crossbred cows. Animal Science [Online]. 8 (4), pp. 123-129. [Accessed 09 March 2018].

Pattison, K.F., Miller, H.C., Rayburn-Reeves, R. and Zentall, T. (2010) The case of the disappearing bone: Dogs’ understanding of the physical properties of objects. Behavioural Processes [Online]. 85 (3), pp.278-282. [Accessed 24 September 2017].

Paul, N.C., Bärgman, S.C., Moodley, A., Nielsen, S.S. and Guardabassi, L. (2012) Staphylococcus pseudintermedius colonization patterns and strain diversity in healthy dogs: a cross-sectional and longitudinal study. Veterinary Microbiology [Online]. 160 (3-4), pp.420-427. [Accessed 09 March 2018].

Pet Food Manufacturers’ Association (2018) Pet Food Manufacturers’ Association. Available from: https://www.pfma.org.uk/pet-population-2017 [Accessed 18 January 2018].

Peters, I.R., Helps, C.R., Hall, E.J. and Day, M.J. (2004) Real-time RT-PCR: considerations for efficient and sensitive assay design. Journal of Immunological Methods [Online]. 286 (1-2), pp.203-217. [Accessed 09 March 2018].

Pomba, C., Rantala, M., Greko, C., Baptiste, K.E., Catry, B., Van Duijkeren, E., Mateus, A., Moreno, M.A., Pyörälä, S., Ružauskas, M. and Sanders, P. (2017). Public health risk of antimicrobial resistance transfer from companion animals. Journal of Antimicrobial Chemotherapy [Online]. 72 (4), pp.957-968. [Accessed 20 January 2018].

Rawling, M.D., Merrifield, D.L. and Davies, S.J. (2009) Preliminary assessment of dietary supplementation of Sangrovit® on red tilapia (Oreochromis niloticus) growth performance and health. Aquaculture [Online]. 294 (1-2), pp.118-122. [Accessed 25 September 2018].

Reller, L.B., Weinstein, M., Jorgensen, J.H. and Ferraro, M.J. (2009) Antimicrobial susceptibility testing: a review of general principles and contemporary practices. Clinical Infectious Diseases [Online]. 49 (11), pp.1749-1755. [Accessed 09 March 2018].

Rendle, D.I. and Page, S.W. (2018) Antimicrobial resistance in companion animals. Equine Veterinary Journal [Online]. 50 (2), pp.147-152. [Accessed 20 January 2018].

Ribeiro, M., Monteiro, F.J. and Ferraz, M.P. (2012) Infection of orthopedic implants with emphasis on bacterial adhesion process and techniques used in studying bacterial-material interactions. Biomatter [Online]. 2 (4), pp.176-194. [Accessed 28 February 2018].

Rooney, N.J., Gaines, S.A. and Bradshaw, J.W. (2007) Behavioural and glucocorticoid responses of dogs (Canis familiaris) to kennelling: investigating mitigation of stress by prior habituation. Physiology & Behavior [Online]. 92 (5), pp.847-854. [Accessed 24 September 2017].

Rowson, C. and Townsend, R. (2016) Biofilms: prevention and treatment. British Journal of Hospital Medicine [Online]. 77 (12), pp.699-703. [Accessed 20 January 2018].

Rutland, B.E., Weese, J.S., Bolin, C., Au, J. and Malani, A.N. (2009) Human-to-dog transmission of methicillin-resistant Staphylococcus aureus. Emerging Infectious Diseases [Online]. 15 (8), pp.1328. [Accessed 12 February 2018].

Sargeant, J.M., Sanderson, M.W., Smith, R.A. and Griffin, D.D. (2003) Escherichia coli O157 in feedlot cattle feces and water in four major feeder-cattle states in the USA. Preventive Veterinary Medicine [Online]. 61 (2), pp.127-135. [Accessed 21 February 2018].

Sato, Y., Mori, T., Koyama, T. and Nagase, H. (2000) Salmonella virchow infection in an infant transmitted by household dogs. Journal of Veterinary Medical Science [Online]. 62 (7), pp.767-769. [Accessed 22 February 2018].

Sawian, P., Nongkynrih, K.J., Anand, U. and Charan, A.A. (2018) Biochemical tests performed for the identification of the isolates collected from local rice beer (Kiad). Journal of Pharmacognosy and Phytochemistry [Online]. 7 (1), pp.395-397. [Accessed 25 September 2018].

Schwarz, S., Loeffler, A. and Kadlec, K. (2017) Bacterial resistance to antimicrobial agents and its impact on veterinary and human medicine. Veterinary Dermatology [Online]. 28 (1), pp.82-97. [Accessed 21 January 2018].

Shannon, L.M., Boyko, R.H., Castelhano, M., Corey, E., Hayward, J.J., McLean, C., White, M.E., Said, M.A., Anita, B.A., Bondjengo, N.I. and Calero, J. (2015) Genetic structure in village dogs reveals a Central Asian domestication origin. Proceedings of the National Academy of Sciences [Online]. 112 (44), pp.13639-13644. [Accessed 18 January 2018].

Sinde, E. and Carballo, J. (2000) Attachment of Salmonella spp. and Listeria monocytogenes to stainless steel, rubber and polytetrafluorethylene: the influence of free energy and the effect of commercial sanitizers. Food Microbiology [Online]. 17 (4), pp.439-447. [Accessed 08 March 2018].

Sing, A., Tuschak, C. and Hörmansdorfer, S. (2008) Methicillin-resistant Staphylococcus aureus in a family and its pet cat. New England Journal of Medicine [Online]. 358 (11), pp.1200-1201. [Accessed 18 February 2018].

Sorrentino, R., Cochis, A., Azzimonti, B., Caravaca, C., Chevalier, J., Kuntz, M., Porporati, A.A., Streicher, R.M. and Rimondini, L. (2018) Reduced bacterial adhesion on ceramics used for arthroplasty applications. Journal of the European Ceramic Society [Online]. 38 (3), pp.963-970. [Accessed 08 March 2018].

Srinivasan, R., Karaoz, U., Volegova, M., MacKichan, J., Kato-Maeda, M., Miller, S., Nadarajan, R., Brodie, E.L. and Lynch, S.V. (2015) Use of 16S rRNA gene for identification of a broad range of clinically relevant bacterial pathogens. PloS one [Online]. 10 (2), pp.1-22. [Accessed 03 March 2018].

Steenackers, H., Hermans, K., Vanderleyden, J. and De Keersmaecker, S.C. (2012) Salmonella biofilms: an overview on occurrence, structure, regulation and eradication. Food Research International [Online]. 45 (2), pp.502-531. [Accessed 20 January 2018].

Sutton, S. (2011) Accuracy of plate counts. Journal of Validation Technology [Online]. 17 (3), pp.42. [Accessed 25 September 2017].

Talan, D.A., Citron, D.M., Abrahamian, F.M., Moran, G.J. and Goldstein, E.J. (1999) Bacteriologic analysis of infected dog and cat bites. New England Journal of Medicine [Online]. 340 (2), pp.85-92. [Accessed 09 March 2018].

Thomas, N. and Brook, I. (2011) Animal bite-associated infections: microbiology and treatment. Expert Review of Anti-infective Therapy [Online].  9 (2), pp.215-226. [Accessed 09 March 2018].

Tuson, H.H. and Weibel, D.B. (2013) Bacteria–surface interactions. Soft Matter [Online]. 9 (17), pp.4368-4380. [Accessed 09 March 2018].

Van Balen, J., Kelley, C., Nava-Hoet, R.C., Bateman, S., Hillier, A., Dyce, J., Wittum, T.E. and Hoet, A.E. (2013) Presence, distribution, and molecular epidemiology of methicillin-resistant Staphylococcus aureus in a small animal teaching hospital: a year-long active surveillance targeting dogs and their environment. Vector-Borne and Zoonotic Diseases [Online]. 13 (5), pp.299-311. [Accessed 14 February 2018].

Van Houdt, R. and Michiels, C.W. (2010) Biofilm formation and the food industry, a focus on the bacterial outer surface. Journal of Applied Microbiology [Online]. 109 (4), pp.1117-1131. [Accessed 21 January 2018].

van Veen, S.Q., Claas, E.C.J. and Kuijper, E.J. (2010) High-throughput identification of bacteria and yeast by matrix-assisted laser desorption ionization-time of flight mass spectrometry in conventional medical microbiology laboratories. Journal of Clinical Microbiology [Online]. 48 (3), pp.900-907. [Accessed 08 March 2018].

Vaseekaran, S., Balakumar, S. and Arasaratnam, V. (2010) Isolation and identification of a bacterial strain producing thermostable α-amylase. Tropical Agricultural Research [Online]. 22 (1), pp. 1-11. [Accessed 08 March 2018].

Vasoo, S., Stevens, J. and Singh, K. (2009) Rapid antigen tests for diagnosis of pandemic (Swine) influenza A/H1N1. Clinical Infectious Diseases [Online]. 49 (7), pp.1090-1093. [Accessed 08 March 2018].

Veluz, G.A., Pitchiah, S. and Alvarado, C.Z. (2012) Attachment of Salmonella serovars and Listeria monocytogenes to stainless steel and plastic conveyor belts. Poultry Science [Online]. 91 (8), pp.2004-2010. [Accessed 09 March 2018].

Verma, S., Sharma, M., Katoch, S., Verma, L., Kumar, S., Dogra, V., Chahota, R., Dhar, P. and Singh, G. (2013) Profiling of virulence associated genes of Pasteurella multocida isolated from cattle. Veterinary Research Communications [Online]. 37 (1), pp.83-89. [Accessed 09 March 2018].

Verma, A.K., Sinha, D.K. and Singh, B.R. (2011) Detection of Salmonella from clinical samples of dogs by PCR. Indian Journal of Animal Sciences [Online]. 81 (6), pp.552-555.

Verran, J., Airey, P., Packer, A. and Whitehead, K.A. (2008) Microbial retention on open food contact surfaces and implications for food contamination. Advances in Applied Microbiology [Online]. 64 (1), pp.223-246. [Accessed 09 March 2018].

Wang, G.D., Zhai, W., Yang, H.C., Wang, L., Zhong, L., Liu, Y.H., Fan, R.X., Yin, T.T., Zhu, C.L., Poyarkov, A.D. and Irwin, D.M. (2016) Out of southern East Asia: the natural history of domestic dogs across the world. Cell Research [Online]. 26 (1), pp.21-33. [Accessed 16 January 2018].

Watson, E., Jeckel, S., Snow, L., Stubbs, R., Teale, C., Wearing, H., Horton, R., Toszeghy, M., Tearne, O., Ellis-Iversen, J. and Coldham, N. (2012) Epidemiology of extended spectrum beta-lactamase E. coli (CTX-M-15) on a commercial dairy farm. Veterinary Microbiology [Online]. 154 (3-4), pp.339-346. [Accessed 21 February 2018].

Weese, J.S. (2010) Methicillin-resistant Staphylococcus aureus in animals. ILAR Journal [Online]. 51 (3), pp.233-244. [Accessed 18 February 2018].

Weese, J.S., Finley, R., Reid-Smith, R.R., Janecko, N. and Rousseau, J. (2010) Evaluation of Clostridium difficile in dogs and the household environment. Epidemiology & Infection [Online]. 138 (8), pp.1100-1104. [Accessed 09 March 2018].

Weese, J.S., Dick, H., Willey, B.M., McGeer, A., Kreiswirth, B.N., Innis, B. and Low, D.E. (2006) Suspected transmission of methicillin-resistant Staphylococcus aureus between domestic pets and humans in veterinary clinics and in the household. Veterinary Microbiology [Online]. 115 (1-3), pp.148-155. [Accessed 18 February 2018].

Weese, J.S. and Rousseau, J. (2006) Survival of Salmonella Copenhagen in food bowls following contamination with experimentally inoculated raw meat: Effects of time, cleaning, and disinfection. The Canadian Veterinary Journal [Online]. 47 (9), pp.887-889. [Accessed 24 February 2018].

Weese, J.S., Rousseau, J. and Arroyo, L. (2005) Bacteriological evaluation of commercial canine and feline raw diets. The Canadian Veterinary Journal [Online]. 46 (6), pp.513. [Accessed 01 March 2018].

Wingender, J. and Flemming, H.C. (2011) Biofilms in drinking water and their role as reservoir for pathogens. International Journal of Hygiene and Environmental Health [Online]. 214 (6), pp.417-23. [Accessed 20 January 2018].

Zambori, C., Cumpanasoiu, C., Mladin, B. and Tirziu, E. (2013) Biofilms in oral cavity of dogs and implication in zoonotic infections. Scientific Papers Animal Science and Biotechnologies [Online]. 46 (1), pp.155-158. [Accessed 28 February 2018].