Efl

HV

a

ARR2A

KPECRS

1

(itaft

0d

Preventive Veterinary Medicine 104 (2012) 101– 106

Contents lists available at SciVerse ScienceDirect

Preventive Veterinary Medicine

j ourna l ho me pag e: ww w.elsev i er .com/ locate /prev etmed

ndemic sheep scab: Risk factors and the behaviour of upland sheepocks

annah Rose ∗, Richard Walleterinary Parasitology and Ecology Group, School of Biological Sciences, Woodland Road, University of Bristol, BS8 1UG, United Kingdom

r t i c l e i n f o

rticle history:eceived 10 September 2010eceived in revised form6 September 2011ccepted 16 October 2011

eywords:soroptes ovisradicationommon grazingisk factorsheep scab

a b s t r a c t

Since the deregulation of sheep scab in 1992, the number of scab outbreaks per year in theUK has increased rapidly; there has been an estimated 60-fold increase in outbreak inci-dence between 1992 and 2007. How best to focus management or eradication efforts hastherefore been an issue of ongoing debate. Previous work has shown that scab outbreaksoccur repeatedly on some farms whereas others never or rarely experience outbreaks. Here,to consider why this pattern occurs, and to quantify the risk of scab associated with dif-ferent husbandry practices, face-to-face interviews with farmers at 16, randomly selected,repeat-outbreak farms and 16 matched farms with no or little history of scab over theprevious 10 years, were used to identify associated risk factors. This showed that farmsusing common grazing were at significantly higher risk of infestation than farms not usingcommon grazing, as were farms that had direct contact with neighbours’ sheep and thathad neighbours with scab. To further demonstrate the risk of common grazing, a survey ofsheep on one common in south Wales showed that there was significant mixing of sheepfrom three farms, highlighting the potential for scab transmission between flocks. Commongrazing, associated with upland regions, would appear to be the largest single obstacle toeffective national scab control in the UK. The results suggest that a cost-effective approach

to scab control may be to consider upland and lowland farms as separate epidemiologicalsystems, with upland regions requiring a co-ordinated, systematic approach to achieve anysignificant management of the problem. In contrast, on lowland farms outbreaks are alreadylargely contained effectively through good biosecurity and treated on a case-by-case basisif they occur.

© 2011 Elsevier B.V. All rights reserved.

. Introduction

The appropriate control of psoroptic mange in sheepscab) has proven to be a particularly intractable problemn the UK over the last 40 years. Scab is estimated to costhe UK sheep industry over £8 million per year (Nieuwhof

nd Bishop, 2005), with the costs of insecticide applicationsar exceeding the economic losses associated with mor-ality and reduced weight gain (ADAS, 2008). The initial

∗ Corresponding author. Tel.: +44 1179 287 489; fax: +44 1173 317 985.E-mail address: hannah.rose@bris.ac.uk (H. Rose).

167-5877/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.prevetmed.2011.10.010

failure to control scab after its reintroduction into the UKin 1973 (Loxam, 1974) can be attributed to a lack of initialclarity in the statutory response, compounded by subse-quent concerns and confusion associated with appropriateinsecticide use (Bates, 2004). This allowed the disease to re-establish itself widely within the UK (French et al., 1999).Deregulation of scab control in 1992 was then followedby an almost exponential increase in disease incidence,with the number of outbreaks increasing from less than

100 per year to an estimated 7000 outbreaks per year in2003/2004 (Bisdorff et al., 2006). In addition to the prob-lems associated with current and historic approaches toscab management, many aspects of the epidemiology of the

terinary

102 H. Rose, R. Wall / Preventive Ve

disease, such as the lengthy pre-patent period, existence incryptic sites in a non-pathogenic form, the duration of off-host survival and strain-specific variation in pathogenicity,all contribute to making it difficult to diagnose pre-patentinfection and prevent transmission, for example throughmarkets (Bates, 1996a, 1996b, 1999; O’Brien et al., 1994;O’Brien, 1999).

The appropriate management of scab is essential onboth welfare and economic grounds, however, identify-ing the most appropriate manner in which future effortsshould be focussed is an issue of ongoing debate (Wall,2007). Recent studies have shown that scab is more preva-lent on some farms; 85% of farms reporting a scab outbreakin 2007/2008 also reported having one or more outbreakseach year in the previous 10 years. In contrast, 76% offarms that had no scab in 2007/2008 had not had scab inthe previous 10 years (Rose et al., 2009). Spatial clusteringand an association between a focal outbreak and outbreakson neighbouring farms was also reported by French et al.(1999). These results suggest that there are likely to bespecific management or environmental risk factors whichdistinguish the farms on which persistent scab is reported.Poor fencing and bought in sheep were identified as impor-tant risk factors for scab in Ireland by O’Brien (1992) whosuggested that upland and mountain sheep remained theprimary focus of infection. However, to date no quan-tification of these risk factors has been attempted. Suchan analysis may be of considerable value, because it willallow management to be more clearly focussed, facilitat-ing the development of programmes which target high riskfarms, optimising the use of time and resource (Sargisonet al., 2007). In addition a more precise understanding mayenable individual farmers to better assess and manage riskappropriately.

The aims of this study therefore, were first to iden-tify and quantify the risk factors associated with repeatedsheep scab outbreaks, using face-to-face interviews withfarmers who had previously reported repeated outbreaksand then to follow this up with a detailed field investigationof the most significant of these factors.

2. Materials and methods

2.1. Farmer interviews

Sixteen focal farms that had reported repeated scabinfestations in a previous postal survey (Rose et al., 2009)were visited between November 2008 and February 2009.Each of these farms had reported an outbreak of scab in2007/2008 and a median of 8 outbreaks during the previ-ous 10 years. Each outbreak farm was matched as closely aspossible with a control farm that had never had scab duringthe previous 10 years. Farms were matched for geographi-cal location, with a median distance of 25 km between thepairs (range 2–60 km), height above sea level and flocksize. Previous studies had identified a regional variationin scab prevalence and an association between outbreaks

with both height above sea level and flock size (Corke andBroom, 2000; Bisdorff et al., 2006; Rose et al., 2009).

All farmers were asked a comprehensive series of ques-tions covering all aspects of their sheep husbandry and the

Medicine 104 (2012) 101– 106

history of scab outbreaks on their farm. For validation, theanswers to questions related to scab outbreak history werecompared with the same questions in the original postalsurvey undertaken a year earlier and from which they wereidentified as farms with a pattern of repeat outbreaks (Roseet al., 2009). Any farm which gave different answers on thetwo occasions would have been removed from the analysis;however none did so.

2.2. Risk factor analysis

Sign tests were used to examine differences betweenoutbreak and control farms in their use of common graz-ing (use of shared pastures), away grazing (the practice ofsending sheep to other farms, usually at lower latitudes oraltitudes, to graze over winter), winter housing and pro-phylactic treatment for scab. Wilcoxon signed ranks testswere used to test for a difference in the frequency withwhich neighbours had scab, the frequency of contact withneighbours’ sheep and the treatment and quarantine dura-tion of bought-in (purchased) sheep, between outbreak andcontrol farms.

The farms’ neighbours were recorded as having hadscab “never” or “sometimes” if the farmer knew thathis neighbours had had scab once or twice and “regu-larly” if the farmers reported that their neighbours hadscab more frequently. Similarly the farms were recordedas either “never” having contact with neighbours’ sheep,“sometimes” if they reported the occasional escapee, or“regularly” if they reported having more regular contactor had wire fences through which the sheep had contact.

Whether farmers treated all, some or none of thebought-in sheep with an appropriate ectoparasiticide wasdetermined. The length of time bought-in sheep were quar-antined was recorded for both rams and ewes and the meanquarantine time calculated. In some cases the farmer couldnot give a specific quarantine time, and this was estimatedfrom the available information, for example, if the farmerstated that he quarantined “until tupping”, the quarantinetime would be estimated from the time of purchase andtupping.

Univariable, exact conditional logistic regression wasperformed using LogXact (Cytel Inc. 2008, Massachusetts)to explore the relationships between the presence andabsence of scab outbreaks between March 2007 andFebruary 2008 and the use of common grazing, thefrequency of contact with neighbours’ sheep and thefrequency with which the neighbours had scab, and todetermine whether they were significant risk factors forsheep. Conditional logistic regression was employed toallow for the stratification of the data, introduced by thematched case–control pairs design. The exact method wasused as conditional logistic regression uses only discordantpairs, thus reducing the sample size. Independent variableswere input as ordered categorical variables resulting ina single coefficient estimate per variable. For each inde-pendent variable identified as a significant risk factor,

odds ratios were calculated from the antilogarithm of ˇ,the coefficient estimate (Dalgaard, 2008). Unlike logisticregression, where the odds ratio is the increase in the oddsof the event occurring for every one unit increase in the

terinary

iairwu

2

laPfmtnooAwermatwwnw

daedpahatoesdsresto

3

3

u5btf

H. Rose, R. Wall / Preventive Ve

ndependent variable, here the odds ratio is interpreteds the increase in odds of reporting sheep scab for anyncrease in the independent variable (Gould, 2000). To cor-ect for multiple testing a sequential Bonferroni correctionas used (Rice, 1989). Spearman rank correlations weresed to test for correlation between risk factors.

.3. Transmission potential on common grazing

Welsh Mountain and Brecknock Mountain ewes andambs (Ovis aries) from three farms were studied on anrea of common grazing in the Brecon Beacons Nationalark, UK. Sheep from each farm could be identified by theirarm-specific breed characteristics and flock-specific paint

arkings. Sheep from other farms were also present onhe common, but their owners were unknown or wishedot to participate in the study. All sheep were turned outnto the common grazing areas in spring 2009. A seriesf five surveys were then undertaken on foot betweenpril and September 2009. A minimum interval of 9 daysas allowed between surveys, to reduce autocorrelation

ffects on the observations. The locations of the sheep,eferred to as fixes, were determined relative to land-arks and recorded on an Ordnance Survey map to 100 m

ccuracy. The location of sheep from all farms, includinghose not participating in the study, were recorded. Sheepere grouped according to their proximity to others; sheepithin 100 m of each other were recorded as a group. Theumber of individuals from each farm within each groupas recorded for each fix.

Flocks of sheep occupy home ranges (Hunter, 1964),efined as the area utilised by an individual while for-ging, mating and caring for young, excluding occasionalxcursions outside of this area (Burt, 1943). Fixed kernelensity estimates of each flock’s home range were com-uted in R (R Development Core Team, 2009) using thedehabitat package (Calenge, 2006) and employing an adoc smoothing method. Fixed kernel density estimates are

nonparametric method of calculating the utilisation dis-ribution of a home range (the probability of an animalccurring at each location in the home range) (Marzlufft al., 2001). To determine whether significant mixing ofheep from different flocks occurs on the common theegree of spatial interaction was calculated for each of thetudy flocks as the percentage of 95% (home) and 50% (core)anges overlapped by at least one other flock (Kernohant al., 2001). These ranges represent an area where theheep were observed 95% and 50% of the time, respec-ively, or where 95% and 50% of sheep can be found at anyne time.

. Results

.1. Risk factor analysis

Significantly more of the outbreak farms (repeated scab)sed common grazing than the control farms (P = 0.016;

0% (n = 8) and 6.3% (n = 1), respectively). Amongst the out-reak farms that used common grazing 88% (n = 7) thoughthat the scab had come from the common. None of thearms using common grazing thought the scab had come

Medicine 104 (2012) 101– 106 103

from bought in sheep or neighbours. In contrast, 38% (n = 3)of the farms that did not use common grazing thoughtthat the scab had come from neighbours and 38% (n = 3)thought it had come from bought in sheep. Three remain-ing farms did not know where their scab had come fromand suggested contaminated transport, away grazing, com-mon grazing and escaped sheep as possible routes oftransmission. Farms that used common grazing often alsoaway-grazed. However, there was no significant differencein the number of outbreak and control farms using awaygrazing (P = 0.18) or winter housing (P = 1.0).

Outbreak group farms had significantly more con-tact with neighbours’ sheep than control farms (Z = 2.88,P = 0.004). All of the outbreak farms reported contact withneighbours’ sheep, and 68.8% (n = 11) reported regular con-tact, whereas 43.8% (n = 7) of the control farms reportednever having contact with neighbours’ sheep, and only18.8% (n = 3) reported regular contact. This did not includecontact with neighbours’ sheep on common grazing. Out-break farms also reported that their neighbours had scabmore often than the neighbours of control farms (Z = 2.92,P = 0.004). Amongst the outbreak farms, 62.5% (n = 10)reported that their neighbour had regular outbreaks ofscab and 18.8% (n = 3) reported that their neighbours some-times had scab. In contrast, 50% (n = 8) of control farmsreported that their neighbours never had a scab outbreakand only 12.5% (n = 2) reported that neighbours had regularoutbreaks.

All farms bought-in sheep, but there was no signifi-cant difference between the length of quarantine periodreported between outbreak and control farms (P = 0.31) orbetween whether they treated all, some or none of thebought-in sheep against possible infection (P = 0.52).

Exact conditional logistic regression with sequentialBonferroni correction identified common grazing, neigh-bours with scab, and contact with neighbours’ sheep as thesignificant risk factors associated with repeated scab out-breaks (Table 1). The use of common grazing increased theodds of reporting sheep scab 9.6 times. Similarly havingneighbours with more frequent outbreaks of scab and hav-ing more frequent contact with neighbours’ sheep than thematched farm increased the odds of disease 10.1 and 7.7times, respectively. Due to the large confidence intervalsthese odds may only provide an approximation of the sizeof the effect.

The risk associated with having neighbours with scaband having contact with a neighbours sheep was associ-ated with increasing frequency of outbreaks. Farms whoseneighbours had sheep scab regularly and who had regularcontact with neighbours’ sheep had outbreaks on averagein 9 and 5 of the previous 10 years, respectively. Farmsusing common grazing reported outbreaks on average eachyear for the previous 10 years.

There was a statistically significant, weak positive cor-relation between the use of common grazing and thefrequency of contact with neighbours’ sheep (r = 0.38,P = 0.03), and a statistically significant moderate posi-

tive correlation between the use of common grazingand frequency neighbours had scab (r = 0.55, P = 0.001).The weak correlation between the frequency with whichneighbours had scab and the frequency of contact with

104 H. Rose, R. Wall / Preventive Veterinary Medicine 104 (2012) 101– 106

Table 1Risk factors for sheep scab in UK flocks. The coefficient estimates (ˇ) and associated P-values derived from a conditional logistic regression for each riskfactor tested against the presence or absence of scab. The odds ratios represent the change in odds of reporting sheep scab for any increase in the riskfactor. The corrected significance level refers to the value calculated using the sequential Bonferroni technique, which imposes lower significance levelswhen multiple tests are used (Rice, 1989). Tests with P-values below this level are statistically significant.

Risk factor Unit coding Coefficient estimate (ˇ) Odds ratio (95% CI) P-value Correctedsignificance level

Contact with neighbours’ sheep 0 = never1 = sometimes2 = regularly

2.043 (0.56 − ˛) 7.714 (1.75 − ˛) 0.002 0.017

Neighbours with scab 0 = never1 = sometimes

2.308 (0.64 − ˛) 10.054 (1.90 − ˛) 0.002 0.025

0.37 − ˛

2 = regularly

Common grazing used 0 = no1 = yes

2.263 (

neighbours’ sheep was only just statistically significant(r = 0.35, P = 0.05).

3.2. Transmission potential on common grazing

During the surveys 79.7% of the sheep observedhad recognisable markings. Several completely unmarkedsheep were seen on the common amounting to 0.44% ofall observations. Notably, one unmarked ewe with a younglamb estimated to be several days old was observed beforethe ewes with lambs at foot had been turned out onto thecommon (put onto the common land to graze). As none ofthe fields adjacent to the common land had sheep in themat the time, this may indicate the existence of a feral pop-ulation of sheep which have escaped been missed duringthe annual gathering and are now breeding on the com-mon land. The remainder of the sheep were unrecognisablebecause of poor visibility or the animal’s orientation hid-ing the markings. Sheep from the farms which did not wishto participate in the study were included in the analysisof home range overlap, but their ranges were not mappedindividually. Marked sheep that could not be attributed toa farm, or sheep whose marking was not visible were notincluded in analysis.

Home range estimates (95% range) for the three farmsvaried between 2.4 and 13.2 km2. Core range estimates(50% range) varied between 0.5 and 3.4 km2. There wasconsiderable overlap of flock ranges with very little of thehome range exclusive to a single flock. A median of 97.7%(range 84.3–99.1%) of the home ranges and 51.4% (range41.1–70.7%) of the core ranges were shared with at leastone other flock (Fig. 1).

4. Discussion

Prevalence surveys for sheep scab have commonlyrecorded the number of outbreaks over a single year(Bisdorff et al., 2006). However, this survey method dis-guises the important fact that a relatively small numberof farmers appear to experience repeated scab outbreaksyear after year while a large proportion of farmers never,or rarely, see scab in their flocks (Rose et al., 2009).

Understanding why this is the case and determining therisk factors that produce this pattern is an importantstep towards designing more effective scab managementprogrammes.

) 9.612 (1.45 − ˛) 0.016 0.05

The results presented here, from face-to-face interviewswith 32 farmers, half of whom had a history of repeatedscab outbreaks, show that common grazing, having neigh-bours with scab, and contact with these neighbours’ sheepwere primary risk factors for endemic sheep scab on thesefarms, significantly increasing the risk of scab between 7and 10 times. Of these, common grazing was a particularlyimportant predictor of outbreaks; 8 out of 9 farms usingcommon grazing reported at least one scab outbreak dur-ing 2007/2008, and farms using common grazing had anoutbreak of scab, on average, once every year.

A similar pattern was found in a survey of Irish sheepfarms; those with scab were twice as likely as farms thathad never had scab to have poor fencing and outbreaksoccurred more often on upland farms with common graz-ing (O’Brien, 1999). The significance of common grazingfor the transmission of sheep scab has also been discussedpreviously (O’Brien, 1992; Winter, 1995; O’Brien, 1999;Armstrong and Davies, 2007). However, the risk associatedwith these practices has not previously been quantified.The results presented here will allow future scab manage-ment to be targeted where it is needed most at both thenational/regional level and at an individual farm level.

A larger sample size would have allowed the data forfarmers who did not use common grazing to be analysedseparately. This may have revealed other factors that sig-nificantly increased the risk for farms without commongrazing. For example, bought in sheep were shown to beimportant risk factors by O’Brien (1999), but this was notthe case in the current study. Such factors may have beenconcealed by the overwhelming risk posed by commongrazing and contact with neighbours’ sheep seen here.

Observations of sheep at one Brecon common clearlydemonstrate how sheep from different flocks overlapin their range use, increasing the potential for scabtransmission and making effective control difficult, if unco-ordinated. The three flocks in this study shared up to 99%of their home range with other flocks, greatly increasingthe risk of transmitting sheep scab between the flocks.Furthermore, sheep that are unfamiliar with the rangemay disperse widely (Warren and Mysterud, 1993) andmix with sheep from other flocks (Arnold and Pahl, 1974).

Here, the introduction of yearling sheep unfamiliar to thestudy common, prior to the introduction of older ewes withknowledge of the area may have increased the degree ofoverlap of ranges.

H. Rose, R. Wall / Preventive Veterinary Medicine 104 (2012) 101– 106 105

Fig. 1. Kernel density estimates for Farm 1 (a), Farm 2 (b) and Farm 3 (c), and the overlapping 95% home range estimates (d). The x and y axes give theBritish National Grid coordinates. The shade-gradient in (a–c) shows the probability (as a percentage) that flock members will be found within that regionof the home range. The area inside the outer ring (white) indicates where 100% of the sheep from this farm will be found and the area inside the inner ring( d. The

( s the areo

illru

ttaor

black) indicates the area where 10% of sheep from this farm will be foundotted line), Farm 2 (solid line) and Farm 3 (dashed line). The 95% range if the home range of the sheep.

Previous studies have noted that the prevalence of scabs significantly higher in Wales, Scotland and Northern Eng-and (Bisdorff et al., 2006; Rose et al., 2009). It would appearikely that the preponderance of common grazing in theseegions is strongly associated with these prevalence fig-res.

In addition to the increased potential for between-flockransmission, unmarked sheep, which may act as a persis-

ent source of scab, were observed on the study commonnd several farmers interviewed reported the presencef feral sheep as commonplace. One of the farmers withepeated scab outbreaks estimated that as many as 10% of

areas outlined in figure d represent the 95% range of sheep from Farm 1a where 95% of sheep will be found at any one time and is representative

his sheep may be missed during a single gathering and ina study of sheep on common grazing in Scotland, Hewsonand Verkaik (1981) reported that only 81% of marked sheepwere gathered for dipping.

It is clear that the difficulties farmers face trying tocontrol parasites in extensively grazed flocks are consid-erable, and the relative importance of the risk factorsidentified here differs between upland/hill and lowland

farms. On lowland farms, which do not use common graz-ing, repeat infestations are more likely to be acquired fromneighbours, with bought-in sheep probably accounting forsporadic outbreaks only. Therefore, regular prophylactic

terinary

106 H. Rose, R. Wall / Preventive Ve

treatments may not be necessary, particularly on farmswith no direct contact with neighbours. Strict farm biosecu-rity, such as stock-proof fencing, treatment and quarantineof all bought-in sheep are likely to be able to prevent theintroduction of scab onto lowland farms. Outbreaks canbe treated on a case-by-case basis and further transmis-sion prevented by employing biosecurity measures such asmovement restrictions and treatment of all neighbouringflocks.

On upland and hill farms common grazing is likely topose the greatest scab risk. None of the farms in the homerange study treated their sheep for scab prior to grazing onthe common, and two farmers admitted during interviewsto treating only visibly affected individual animals beforeturning them out onto the common, as the perceived riskof reinfestation was so great as to make prophylaxis seemuneconomic. Such gazing patterns present a significantwelfare issue; extensively grazed sheep may not be seen formonths (Winter, 1995) by which point scab infestation mayhave progressed to the point that welfare is severely com-promised. Several farmers in this study believed that sheepscab is a year-round disease but that they are unawareof outbreaks until sheep return from common grazing forwinter, possibly contributing to the increase in outbreaksreported in November.

Despite the difficulty of scab control in extensive sys-tems such as common grazing, the risk of infestation can bemitigated through cooperating with other common users.The Cheviot Sheep Scab Group and Dartmoor Commoner’sCouncil have independently demonstrated that sheep scabcan be eradicated from an area of common grazing throughcoordinating gathering and treatment with other commonusers and removing feral sheep (Lewis and Newton, 2005;ADAS, 2008).

Due to the differences in the relative importance of keyrisk factors on upland/hill and lowland farms, when con-sidering control it may be appropriate to treat upland/hilland lowland farms as separate epidemiological systems.Lowland farms should be able to adopt a reactive approachto scab control, treating outbreaks as they arise, avoid-ing overuse of the available treatment methods. In uplandand hill regions, however, where the use of common graz-ing allows frequent contact between sheep from differentfarms, coordinated, systematic treatment and gathering,possibly with participation enforced by legislation, with anemphasis on prophylactic treatment, would appear to benecessary to manage the problem. Research should nowfocus on farmer attitudes and the economics of sheep scabcontrol to identify why previous statutory and voluntaryattempts at scab control have failed and to ensure futurecontrol efforts succeed.

Acknowledgements

The authors would like to thank all of the farmers whoparticipated in this study. Thanks also to Dr Toby Knowlesand Dr Paul Craze for their invaluable statistical advice,

Dr Eric Morgan and Jane Learmount for their commentson the questionnaire, and Faith Smith, Harriett Holmesand Andrew Vineer for their assistance with interviewsand sheep observations. We would also like to thank two

Medicine 104 (2012) 101– 106

anonymous reviewers and Dr Marcus Doherr for their com-ments, and NERC and FERA for their ongoing support.

References

ADAS, 2008. Sheep scab controls in Scotland – a review of theevidence base. Available at: http://www.scotland.gov.uk/Resource/Doc/232166/0063533.pdf (accessed 30.04.10).

Armstrong, D., Davies, L., 2007. Sheep ectoparasites in Wales. Vet. Rec.160, 775–776.

Arnold, G.W., Pahl, P.J., 1974. Some aspects of social behaviour in domesticsheep. Anim. Behav. 22, 592–600.

Bates, P., 1996a. Epidemiology of subclinical ovine psoroptic otoacariasisin Great Britain. Vet. Rec. 138, 388–393.

Bates, P., 1996b. Ovine psoroptic otoacariasis: an abattoir survey. Vet. Rec.139, 235–236.

Bates, P., 1999. Inter- and intra-specific variation within the genusPsoroptes (Acari: Psoroptidae). Vet. Parasitol. 83, 201–217.

Bates, P., 2004. Therapies for ectoparasitism in sheep. In Pract. 26,538–547.

Bisdorff, B., Milnes, A., Wall, R., 2006. Prevalence and regional distributionof scab, lice and blowfly strike in Great Britain. Vet. Rec. 158, 749–752.

Burt, W.H., 1943. Territoriality and home range concepts as applied tomammals. J. Mammal. 24, 346–352.

Calenge, C., 2006. The package adehabitat for the R software: a tool forthe analysis of space and habitat use by animals. Ecol. Model. 197,516–519.

Corke, M.J., Broom, D.M., 2000. Sheep scab: a questionnaire survey of sheepfarmers. Proc. Sheep Vet. Soc. 24, 115–121.

Dalgaard, P., 2008. Introductory Statistics with R, 2nd ed. Springer Sci-ence + Business Media, New York, USA.

French, N.P., Berriatua, E., Wall, R., Smith, K., Morgan, K.L., 1999. Sheepscab outbreaks in Great Britain between 1973 and 1992: spatial andtemporal patterns. Vet. Parasitol. 83, 187–200.

Gould, W., 2000. Interpreting logistic regression in all its forms. Stata Tech.Bull. 53, 19–29.

Hewson, R., Verkaik, A.J., 1981. Body condition and ranging behaviourof Blackface hill sheep in relation to lamb survival. J. Appl. Ecol. 18,401–415.

Hunter, R.F., 1964. Home range behaviour in hill sheep. In: Crisp, D.J. (Ed.),Grazing in Terrestrial and Marine Environments. British EcologicalSociety Symposium No. 4. Blackwell Scientific Publications, Oxford.

Kernohan, B.J., Gitzen, R.A., Millspaugh, J.J., 2001. Analysis of animal spaceuse and movements. In: Millspaugh, J.J., Marzluff, J.M. (Eds.), RadioTracking and Animal Populations. Academic Press, San Diego, CA.

Lewis, A., Newton, P., 2005. The control of sheep scab and bovine tuber-culosis in the Dartmoor National Park. Gov. Vet. J. 15, 18–23.

Loxam, J.G., 1974. Sheep scab epidemic: January 1973. State Vet. J. 29,1–10.

Marzluff, J.M., Knick, S.T., Millspaugh, J.J., 2001. High-tech behaviouralecology: modeling the distribution of animal activities. In: Millspaugh,J.J., Marzluff, J.M. (Eds.), Radio Tracking and Animal Populations. Aca-demic Press, San Diego, CA.

Nieuwhof, G.J., Bishop, S.C., 2005. Costs of the major endemic diseasesof sheep in Great Britain and the potential benefits of reduction indisease impact. Anim. Sci. 81, 23–29.

O’Brien, D.J., 1992. Studies on the epidemiology of psoroptic mange ofsheep in Ireland. PhD thesis. National University of Ireland.

O’Brien, D.J., Gray, J.S., O’Reilly, P.F., 1994. Survival and retention of infec-tivity of the mite Psoroptes ovis off the host. Vet. Res. Commun. 18,27–36.

O’Brien, D.J., 1999. Treatment of psoroptic mange with reference to epi-demiology and history. Vet. Parasitol. 83, 177–185.

R Development Core Team, 2009. R: A Language and Environment for Sta-tistical Computing. R Foundation for Statistical Computing, Vienna,Austria.

Rice, W.R., 1989. Analyzing tables of statistical tests. Evolution 43,223–225.

Rose, H., Learmount, J., Taylor, M., Wall, R., 2009. Mapping risk foci forendemic sheep scab. Vet. Parasitol. 165, 112–118.

Sargison, N., Roger, P., Stubbings, L., Baber, P., Morris, P., 2007. Controllingsheep scab by eradication. Vet. Rec. 160, 419–492.

Wall, R., 2007. Eradication of sheep scab in Britain – will history repeat?

Vet. Times 37, 30–31.

Warren, J.T., Mysterud, I., 1993. Extensive ranging by sheep released ontoan unfamiliar range. Appl. Anim. Behav. Sci. 38, 67–73.

Winter, A., 1995. Problems of extensive sheep farming systems. In Pract.17, 217–220.