Journal of Roost selection by the brown long!eared bat PlecotusApplied Ecology0886\ 23\ auritus288Ð397

A[C[ ENTWISTLE\ P[A[ RACEY and J[R[ SPEAKMANDepartment of Zoology\ University of Aberdeen\ Aberdeen AB13 1TZ\ UK

Summary

0[ One of the key threats to bats in Britain is loss of suitable roost sites\ but little isknown about roost requirements for most species[1[ Roost selection in the brown long!eared bat Plecotus auritus was demonstrated bycomparison of buildings used as summer roosts in north!east Scotland with "i# randombuildings in the same area\ and "ii# a set of adjacent houses located in the same habitat[2[ Buildings containing roosts were situated closer to woodland and water relative tothe random houses\ and also had a greater area of woodland within a radius of 9=4 km\but not at distances beyond this[ This suggests that feeding habitat in the vicinity ofthe roost is important for roost selection[3[ When compared with adjacent houses\ roosts were older\ and had roof spacesdivided into more compartments\ which were more likely to be fully lined with roughwooden planking[4[ The temperatures inside summer roosts "mean 06=8 >C# were signi_cantly warmerthan those from random and adjacent houses "mean 05=6 >C#[5[ These results suggest that P[ auritus is selective of its roosts\ relative to the housesavailable[6[ The selection of speci_c types of roosts by P[ auritus has implications for themanagement advice provided when roosts are threatened\ and conservation actionsshould focus on e}orts to avert signi_cant change to\ or destruction of\ roost sites orthe woodland in their vicinity[

Key!words] Chiroptera\ roost requirements[

Journal of Applied Ecology "0886# 23\ 288Ð397

resulting in lower survival "Ransome 0878#[ Instead\Introduction

bats may minimize their thermoregulatory costs byselecting warmer roosts\ and thus remain home!Insectivorous bats\ in both temperate and tropical

zones\ use a variety of shelters\ generally termed roosts othermic for longer periods\ resulting in earlier birthdates and more rapid growth of young["Kunz 0871#[ Roosts provide protection from environ!

mental extremes "wind\ rain and temperature# and In summer\ bats may spend a considerable pro!portion of their time within their day roost sites andfrom predation\ and are the focus for most social

interactions within the colony[ For temperate zone females generally give birth and nurse their o}springthere[ Roost characteristics therefore are potentiallyspecies\ the thermal conditions of roosts may be of

critical survival value "Humphrey 0864#[ important to the survival and fecundity of bats[ How!ever\ not all potential roost sites may be suitable forThe selection of appropriate roosting conditions

during the day may be vital in balancing the energy long!term use by bats[ Rates of survival and fecundityof bats using unsuitable sites may be inadequate tobudget of bats "Kunz 0879#[ Under cool climatic con!

ditions\ when experiencing lower insect availability sustain a viable population "Brigham + Fenton 0875#[It has been suggested that the availability of suitable"Taylor 0852# coupled with increased ther!

moregulatory costs\ bats may enter torpor in order roosts may represent a primary constraint on thepopulation size and distribution of di}erent bat speciesto minimize daily energy expenditure "Kurta 0875#[

However\ use of torpor by females during the summer "Fenton 0869^ Humphrey 0864#[Population censuses of bat species that roost pre!may compromise successful reproduction\ by extend!

ing the gestation period "Racey 0862# and slowing the dominantly in houses suggest that bats occupy the loftÞ 0886 BritishEcological Society growth of young "Tuttle + Stevenson 0871#\ possibly spaces of only a small subsample of the available

288

399 houses in a given area[ For example a survey of all the monly found in buildings\ and bats are often visiblein the attic spaces of houses\ churches and barns "SwiftRoost selection by houses in Glen Lyon in central Scotland "45>N# found

that 12) of the buildings in the area were occupiedthe brown long! 0880#[ P[ auritus demonstrates a high degree of _delityto house roosts\ both within and between yearseared bat by bats "Pritchard + Murphy 0877#[ Such obser!

vations raise an important question[ Do bats occupy "Entwistle 0883#[ Other roosts that may be used bythis species include bat boxes "Boyd + Stebbings 0878#only a subset of the houses because the majority are

unsuitable or suboptimal as roost sites< If bats show and tree holes "Hana�k 0858#\ although the extent towhich such structures are used as maternity sites ina high level of selectivity\ then only a proportion of

available roost sites will ful_l their requirements\ Britain is unknown[which may ultimately limit population size[ Alter!natively\ if bats are not selective with respect to their STUDY AREAroosts\ then a random subset of the available roost

The primary study area was located to the west ofsites\ most of which are suitable\ will be used and roostAberdeen in north!east Scotland "46>N#\ and wasavailability is unlikely to be an important factor inenclosed by a rectangle that extended 59 km west fromdetermining bat population size or distribution[Aberdeen[ It measured 24 km north to south\ andMany studies have been carried out that describe incontained a land area of 1099 square kilometres "thedetail the features of roosts occupied by bats "Dalqu!balance of the rectangle enclosing the sea#[ A secondest + Walton 0869#[ These studies have revealed thatstudy site lay 34 km further west\ and enclosed 864bats often occupy a speci_c type of roosting site[square kilometres along the valley of the river Spey[Nevertheless these descriptive studies\ focusing onlyThe habitat in both areas included areas of agri!on the sites used by bats\ are generally inadequate tocultural land\ woodland in the valleys\ and open hea!demonstrate roost selection[ To demonstrate selec!ther moorland on the higher ground "× 099 m#[tivity\ it is necessary to characterize not only the sitesAlthough towns and villages occurred within the studyoccupied by bats\ but also those sites that are availablearea\ the city of Aberdeen "human population c[but are unoccupied "e[g[ Lunney et al[ 0877#[109 999# was not included in the study area sinceAcross Europe\ many bat species are consideredprevious investigations have failed to identify any con!to be threatened "Stebbings 0877#\ and in Britain alltemporary P[ auritus roosts within the city "Speakmanspecies are protected by the 0870 Wildlife andet al[ 0880#[Countryside Act "Hutson 0882#[ One of the main

threats to bats is considered to be loss of roost sites[ROOST SITESIn particular\ many species appear to be reliant on

man!made roosts\ either through the loss of natural A total of 43 summer day roosts of P[ auritus wereroost sites such as tree cavities\ or because man!made identi_ed within the two study areas\ and all of thesesites o}er preferable conditions[ Man!made sites\ par! buildings were visited[ Thirty!nine of these roosts hadticularly occupied houses\ are the source of most inter! been identi_ed during surveys carried out prior toactions between humans and bats[ These roosts are 0877 "Speakman et al[ 0880#\ and an additional _vesusceptible to disturbance\ pesticide applications for roosts had been reported to Scottish Natural Heritageremedial timber treatment\ and deliberate exclusions "the statutory nature conservation agency# after 0877["e[g[ Barclay\ Thomas + Fenton 0879^ Stebbings A poster survey was initiated in the primary study0877#[ area during 0880\ which led to the identi_cation of a

Where bats come into con~ict with people\ it is further 09 roosts used by P[ auritus[ Only those roostsoften assumed that bats will be able to relocate to where P[ auritus were observed and:or caught wereanother building[ However\ if bats are highly selective included in the sample[ Roosts were classi_ed fromof their roosts then roosting opportunities may be observations as] maternity "from regular capture ofrestricted and such selectivity needs to be considered breeding individuals#^ occasional "bats rarely foundbefore exclusion orders are granted[ In the present despite regular checks#^ or unknown "not enough visitsstudy we sought to evaluate whether the brown long! made to determine the status of the roost#[ To assesseared bat Plecotus auritus Linnaeus 0647 selects the the distribution of known roosts used by P[ auritus\houses in which it roosts\ or whether the roosts used the primary study area was divided into 117 squaresare only a random subset of the buildings available in "each of 2=2 km1#[ Squares through which a majorthe area[ river "the Dee or Don# passed "total � 76#\ were

deemed to be river valleys\ and those that containedareas of human settlement "× 049 inhabitants^Methodsn � 00# were considered to be villages[

STUDY SPECIES

Þ 0886 British RANDOM HOUSESPlecotus auritus is a small vespertilionid "body massEcological Society\5Ð01 g#\ distributed throughout northern Europe[ A random sample of 08 houses within the primaryJournal of Applied

Ecology\ 23\ 288Ð397 Summer day roosts used by P[ auritus are most com! study area was selected from the telephone directory\

390 for use as a control sample[ The directory was opened tape measure "Plasplugs Ltd\ Burton!on!Trent\Sta}s[#\ and the volume was calculated assuming aA[C[ Entwistle\ at random\ moved on 4 pages\ and the _rst name and

address in the right!hand column with a telephoneP[A[ Racey + triangular cross!section of the roof[Number of compartments in the attic] theseJ[R[ Speakman area code lying within the primary study area was

recorded[ The householders were then contacted and included small subsections of the attic space\ such asdormer windows which connected with the main atticvisits were arranged[ The use of telephone directories

to produce representative samples has been evaluated space[Attic lining materials] the presence of a full liningby Kviz "0873#\ who found little bias\ relative to postal

surveys\ in rural America[ of rough wooden planking "{sarking|# on the ceiling ofthe attic underneath the slates\ or alternatively otherboarding such as chipboard\ was recorded "coded as

ADJACENT HOUSESlacking sarking#[

Presence or absence of added insulation in the attic]Since some structural features of houses might beassociated with the locality of the building\ a second insulation was coded as present if the attic ~oor was

lined with either glass _bre or vermiculate grains\control sample was examined[ This included 06 housesthat were adjacent to known roosts "de_ned otherwise insulation was coded as absent[

Illumination in the attic space] this was classed asas ³ 149 m away#\ and therefore would share similarlocal habitat features[ light or dark\ con_rmed by the use of a light meter

"Gossen Mastersix\ Photex\ Tipton\ W[ Midlands#\with a zero reading indicating darkness[

HABITAT VARIABLESPresence or absence of other vertebrates\ except

bats\ in the attic space] this was recorded from corpsesThe habitat around 38 roosts in the primary studyarea\ and around the random houses\ was quanti_ed or droppings[ The type of vertebrate found was

recorded\ and attics were classi_ed for other ver!using 0]09 999 Ordnance Survey maps[ The habitatfeatures recorded included] "i# distance to woodland^ tebrates being present or absent[

Presence or absence of other bat species in the attic"ii# distance to water^ "iii# the area of woodland withinvarious distances from the building[ The habitat space] this was recorded from individuals seen or alter!

natively from droppings\ which could be identi_ed byaround the house was subdivided into concentric\ 9=4!km wide {distance bands|\ up to 2 km\ centred on the comparison with samples from known species[

Level of disturbance of the attic space by the house!house\ and the area of woodland in each distance bandwas then calculated[ holder] this was classed as either regular "used fre!

quently#\ occasional "seasonal or irregular storage#\ orDistances were measured directly from the maps\using a ruler\ and were converted by the appropriate rare disturbance[ This was assessed from questioning

the householder[scale\ but areas were calculated by photocopying themaps on to paper of a standard weight and then cut! Age of the building] this was recorded from the

owner of the property[ting out any woodland within each 9=4!km wide dis!tance band[ These pieces of paper were then weighed Number of storeys in the building] this was scored

as one storey "bungalow#\ or two or more storeyson a four!_gure top!loading balance "Sartorius Ltd\Goettingen\ Germany#[ The area of woodland in each "house#[

Whether or not the building was detached] housesdistance band was then calculated using a conversionfactor based on the average weight of paper repre! were coded as detached or non!detached "semi!

detached or terraced houses#[senting one square kilometre[ In addition\ the areas ofdeciduous and coniferous woodland in each distance Orientation of the building] this was classed as

northÐsouth\ eastÐwest\ north!eastÐsouth!west orband were calculated separately\ with mixed wood!land considered to be half deciduous and half conifer! north!westÐsouth!east[ous[ The woodland around seven roosts was investi!gated on foot\ and corresponded to the distribution

TEMPERATURE RECORDINGSshown on the maps[ This suggested that the habitat inthese areas had not changed substantially since the Temperatures inside roosts were recorded using ther!

mistor probes "CM type^ Grant Instruments "Cam!maps were last revised "0872#[bridge# Ltd\ Barrington\ UK^ accuracy 29=9954 >C at14 >C# attached to a data logger "Squirrel loggers^

STRUCTURAL VARIABLESGrant Instruments "Cambridge# Ltd#[ Temperaturedata were collected from 01 roosts during the summersStructural variables were recorded from all 43 roosts\

as well as from the 08 random and 06 adjacent houses[ of 0881 and 0882[ Temperatures were recorded everyThe structural variables included features of the attic 04 min over a 1!week period[ The data loggers were

Þ 0886 Britishspace and of the building itself[ removed at the end of each fortnight\ and downloadedEcological Society\

Volume of the attic space] the maximum length\ in the laboratory\ before being introduced to a di}er!Journal of AppliedEcology\ 23\ 288Ð397 breadth and height were measured using an ultrasonic ent roost site[ Two probes were attached in pre!

391 determined areas of the roost[ Both probes were pos! Results

Roost selection by itioned in the apex\ where the bats roosted[ The _rstAll the roosts were situated within buildings[ In most

the brown long! probe was positioned against the wooden planking incases\ the bats roosted in the apex of attic spaces\ or

eared bat the apex close to the site where the bats typicallyequivalent spaces in steadings "farm buildings of stone

roosted[ When bats were present they were alwaysconstruction#\ usually in direct contact with the

caught prior to installation of the temperature probes[wooden beams[ In addition\ the bats sometimes roos!

The second probe was positioned against the wood inted in smaller\ less accessible\ compartments that com!

the apex in the middle of the attic space\ away frommunicated with the primary roof area[ The majority

the bats[ Di}erences between the two probes wouldof roosts were found within occupied houses "n � 37#\

reveal whether the presence of the bats a}ected thebut two churches and four unoccupied steadings were

temperature recorded by the _rst probe\ since the pres!also used[ Of the 38 P[ auritus roosts within the pri!

ence of bats could arti_cially elevate the recorded tem!mary study area\ more occurred in the river valleys

peratures in roosts compared with random buildings[than was expected by chance "x1 � 10=08\ n � 117\

The temperature readings from the two probes ind[f[ � 0\ P ³ 9=990#[ In addition\ within the river

roosts were averaged[ This avoided disparities linkedvalleys\ roosts were also signi_cantly associated with

to the position of a particular probe[villages "x1 � 12=21\ n � 76\ d[f[ � 0\ P ³ 9=990#[

Temperatures were also recorded from two randomMost of the roosts "n � 32# were occupied by

houses and from nine houses not occupied by bats butmaternity colonies\ as determined by the regular pres!

situated adjacent to roosts[ The temperatures fromence of bats\ including pregnant or lactating females[

these houses\ which were unoccupied by bats\ wereNo di}erences in features of the building\ or in

combined as a {control| set of temperatures[ Againlocation\ were found when maternity "n � 32# and

two probes were positioned in the apex of the atticoccasional "n � 6# roosts were compared\ using

spaces\ and an average temperature from the two pro!MannÐWhitney U!tests for continuous variables and

bes was calculated[ The temperature was averaged forchi!squared tests for discontinuous variables[ All 43

each hour\ and this was used to examine di}erencesroosts were therefore included in further comparisons\

in the temperature of buildings with time "hour andindependent of their type of usage by bats[ The vari!

date# and also between roosts and control sites[ables measured from roosts are described in Table 0[

Data loggers were placed in houses in a oppor!tunistic way\ based upon when it was convenient forthe householder to allow us access[ To minimize any

HABITAT VARIABLESerrors linked to this experimental design\ it wasensured that temperature recordings were collected Compared with houses in the random sample\ build!

ings containing roosts were closer to woods and toequally and simultaneously from both roosts and con!trol buildings over the whole of the summer\ thus open water "Table 1#[ There was signi_cantly more

woodland within 9=4 km of roosts\ compared with theavoiding seasonal bias in the collection of data fromthe two sets of buildings[ random sample of houses "Table 2#[ However\ at dis!

Table 0[ A description of the roosts used by Plecotus auritus based on a series of quanti_able variables

Variable Description of roosts

Distance to woodland Median � 59 m "range 09Ð319 m#Distance to open water Median � 104 m "range 09Ð819 m#Area of woodland Median area woodland within 9=4 km � 9=086 km1 "range 9=994Ð9=533 km1 #^ see Table

2 for other distance bandsVolume of the roof space Median � 49=44 m1 "range 3=09Ð216=06 m1#Number of compartments 62) of roosts contained more than one compartment "range 0Ð7 compartments#Lining of the roof space All but two roosts were lined with sarking "the other two had wooden panelling#Insulation in the roof space 55) of the roosts contained added insulation "glass _bre in all but one case#Illumination in the roof space 78) of the roosts were completely dark\ while the others were illuminated by windowsOther vertebrates in roost 15) showed use by either mice "Mus musculus and Apodemus spp[#\ rats "Rattus spp[#

or birds "including swallows "Hirundo rustica#\ swifts "Apus apus#\ and blue tits "Paruscaeruleus##

Other bat species in the roost 13) of the roosts were also used by other bat species "including Pipistrellus pipistrellus\Myotis daubentonii\ and M[ nattereri#

Disturbance of the roof space None of the roosts were regularly disturbed\ and 17) were subject to occasionaldisturbance

Age of the building Median � 009 years "range 42Ð339 years#Number of storeys Of buildings containing roosts\ 79) had more than two storeys "range 0Ð4 storeys#

Þ 0886 British Detached:nondetached buildings All except two roosts were in detached buildingsEcological Society\ Orientation of the building 28) of the roosts were orientated northÐsouth\ 22) eastÐwest\ 04) north!eastÐsouth!Journal of Applied west\ and 02) north!westÐsouth!east[Ecology\ 23\ 288Ð397

Table 1[ The median distances to woodland and water "with sample sizes# for buildings containing roosts and those from a392random sample\ compared using MannÐWhitney U!testsA[C[ Entwistle\

P[A[ Racey + Median Median forJ[R[ Speakman for roosts random houses U

Distance to woodland "m# 59 "n � 37# 104 "n � 05# 0181���Distance to water "m# 104 "n � 37# 349 "n � 05# 0285�

�P ³ 9=94\ ��P ³ 9=90\ ���P ³ 9=990[

Table 2[ Comparison of the area of woodland within di}erent distance bands\ between roosts and control buildings usingMannÐWhitney U!tests

MedianMedian for random

Area of woodland "km1# for roosts houses U

Within 9=4 km band 9=086 "n � 37# 9=975 "n � 05# 0659�Within 9=4Ð0=9 km band 9=585 "n � 37# 9=533 "n � 05# 0486Within 0=9Ð0=4 km band 0=149 "n � 37# 0=217 "n � 05# 0441Within 0=4Ð1=9 km band 1=173 "n � 35# 1=282 "n � 04# 0224Within 1=9Ð1=4 km band 1=795 "n � 28# 2=341 "n � 04# 808��Within 1=4Ð2=9 km band 2=117 "n � 26# 2=864 "n � 04# 761

�P ³ 9=94\ ��P ³ 9=90\ ���P ³ 9=990[

Table 3[ Comparison of measured features of roosts "n � 43# with features from random buildings "n � 08# and to a sampleof adjacent buildings "n � 06#\ using MannÐWhitney U!tests

Median Median U Median UVariables roosts random roosts!random adjacent roosts!adjacent

Volume 42=23 17=58 1110�� 20=69 1973Age 096=49 12=99 1398��� 31=99 1041��

�P ³ 9=94\ ��P ³ 9=90\ ���P ³ 9=990[

tances beyond 9=4 km there were no signi_cant di}er! Compared with adjacent houses\ buildings con!taining roosts were older and had more roof com!ences in the areas of woodland between the two

samples\ except between 1=9 km and 1=4 km\ where partments\ which were more likely to be lined withmore woodland was recorded around the randomhouses "Table 2#[ Within 9=4 km of the roost there wasboth more deciduous and more coniferous woodland\ Table 4[ Comparison of discontinuous variables of roosts

"n � 43# with features from random buildings "n � 08# andrelative to the random houses ðarea of deciduousfrom adjacent buildings "n � 06# using chi!squared tests[woodland] median for roosts � 9=954 km1 "n � 37#\These tests have one degree of freedom unless otherwise

median for random houses � 9=913 km1 "n � 05#\ indicated[ Figures are x1 values\ and probabilities are]U � 0679\ P ³ 9=94^ area of coniferous woodland] �P ³ 9=94\ ��P ³ 9=90\ ���P ³ 9=990median for roosts � 9=003 km1 "n � 37#\ median for

Roosts Roostsrandom buildings � 9=940 km1 "n � 05#\ U � 0615\compared comparedP ³ 9=94Ł[

Variable with random with adjacenthouses houses

STRUCTURAL VARIABLESOne:more compartments 11=73��� 4=78�Sarking:not sarking 33=43��� 16=86���Buildings containing roosts had larger roof spaces andInsulation present:absent 9=22 1=26were older than those from the random sample "TableIlluminated:dark 9=90 9=49

3#[ In addition\ buildings used as roosts had on aver! Other bats present:absent 9=96 9=72age a greater number of storeys\ were more likely to Other vertebrates present: 9=70 1=69

absentbe detached houses\ and had on average more roofDisturbed:not disturbed 0=07 9=91compartments\ which were more likely to be fullyBungalow:house 6=03� 1=85lined with wooden planking than control buildings

Þ 0886 British Detached:not detached 17=21��� 2 =03"Table 4#[ These di}erences were also maintained ifEcological Society\ Orientation "NS:WE: 9=02 "d[f[ � 2# 1=50 "d[f[ � 2#the comparison was restricted to only those roosts inJournal of Applied NEÐSW:NWÐSE#

Ecology\ 23\ 288Ð397 the primary study area[

Table 5[ The interrelationships of the features of buildings demonstrated by use of MannÐWhitney U!tests and chi!squared393tests\ and relationships between these characteristics and temperature using t!tests and regressionRoost selection by

the brown long! Age Sarking Compartmentseared bat

Temperature F � 2=16 "d[f[ � 0# t � 0=32 "d[f[ � 03# t � 0=44 "d[f[ � 11#Compartments U � 0238=4� x1 � 0=10 "d[f[ � 0#Sarking U � 1014���

�P ³ 9=94\ ��P ³ 9=90\ ���P ³ 9=990[

planking "Tables 3 and 4#[ These three factors were between 09=6 and 15=5 >C\ with a mean of 06=8 >C"standard deviation � 2=95\ n � 409^ Fig[ 0#[ Meanfound to be interrelated\ when data from all buildings

"roosts\ control and adjacent houses# were combined hourly temperatures ranged from 5=2 to 39=5 >C[ Meandaily temperature in random and adjacent buildings"Table 5#[ However\ roost sites and adjacent buildings

did not di}er in whether or not the house was "{control buildings|# was 05=6 >C "standard deviation� 1=84\ n � 062^ Fig[ 0#[detached\ or in the number of storeys present[

A two!factor ANOVA with repeated measures"BMDP program 6d\ Dixon 0889# was used to analyse

TEMPERATURESthe temperature di}erences between roosts and con!trol buildings\ along with variation due to the date[There was no evidence that the temperature of probes

situated in roosts was in~uenced by the presence of Data were used from 01 roosts and 00 control houses\over a total of 028 days on which simultaneous record!bats[ The patterns of heating and cooling of the two

probes "near and far from the bats| roosting positions# ings were available from both types of building[ Asigni_cant e}ect was found due to both date and typewere similar\ and in half of the records the probe

further from the bats was signi_cantly warmer[ When of building "roost or control building# whether theanalysis used data from every hour of the day or wasroosts were inspected\ bats were never found roosting

in the direct vicinity of the probes[ restricted only to those hours in which bats were likelyto be present within the roost "Table 6#[ This was alsoTemperatures varied signi_cantly with both time of

day and date in each roost[ The average daily tem! the case if only one fortnight of data for each roostwas used\ to avoid errors linked to pseudoreplicationperature\ measured in the apex of roosts\ ranged

Þ 0886 BritishEcological Society\Journal of Applied Fig[ 0[ The average daily temperatures recorded in roosts and control buildings during the summer of 0882[ Ambient tem!

perature over the period was measured from an automatic weather station located in the study area[Ecology\ 23\ 288Ð397

Table 6[ The in~uence of date and type of building "roost or control house# on temperature inside the roof\ using data from394all hours of the day\ between 94]99 and 11]99 h\ and data from all hours of the day\ but only a single replicate from each roostA[C[ Entwistle\"two!way ANOVA#[ Numbers are F!values and probabilities are ���P ³ 9=9990

P[A[ Racey +J[R[ Speakman Restricted

Factors All data "94[99Ð11[99 h# Single replicate

Date "d[f[ � 027# 09=08��� 5=65��� 08=51���Roost:control "d[f[ � 0# 08=39��� 07=62��� 12=71���Date × building interaction 9=75 9=49 1=51���

due to unequal sample sizes from di}erent roosts studies which demonstrated that P[ auritus foragedpredominantly in woodland\ and that this foraging"Table 6#[

No relationships could be identi_ed between the was concentrated within a distance of 9=4 km from theroost "Entwistle\ Racey + Speakman 0885#[ Bats spenttemperature inside houses and structural features of

the buildings\ when all buildings for which tem! 59) of their foraging time within 9=4 km "signi_cantlymore than in any other distance band#\ with femalesperature records were available were included "Table

5#[ spending signi_cantly more time than males within9=4 km of the roost[ Roosts closer to woodland mightalso o}er the bene_ts of cover from aerial predators

Discussionsoon after the bats have emerged from the roost"Jones\ Duverge� + Ransome 0884#[ A relationshipThe non!random use of buildings by P[ auritus in

the present study suggests that roost selection was between the distance from the roost to woodland andthe bats| time of emergence has been demonstrated inoccurring[ Bats appeared to select their roosts at two

levels Ð for the location of the roost with respect to this species\ with bats at roosts closer to woodlandemerging earlier "Entwistle et al[ 0885#[surrounding habitat and for speci_c features of the

building[ The di}erence in habitat surrounding roosts couldalso be explained by the fact that houses adjacent towoodland may be more likely to be encountered by

LOCATION OF THE ROOSTthe bats[ Since bats appear to feed predominantly inwooded areas "Entwistle et al[ 0885#\ houses locatedThe association of roosts with river valleys may have

occurred because in north!east Scotland river valleys at some distance from woodland might not be enco!untered\ and such roosting opportunities would there!are often highly wooded\ relative to the surrounding

habitat "Jenkins 0874#[ P[ auritus forages in woodland fore be missed by the bats[habitats "Swift + Racey 0872^ Barataud 0889^ Swift0880# where insect prey\ particularly moths\ are avail!

FEATURES OF THE BUILDINGable[ Thus\ river valleys may provide valuable patchesof feeding habitat for P[ auritus[ The association of The use by P[ auritus of older houses lined with wood

is consistent with reports of roosts used by this speciesroosts with villages and towns is likely to re~ect theavailability of roosting opportunities in buildings from Britain and continental Europe "Brosset + Cau!

be�re 0848^ Harmata 0851^ Jones + Jayne 0877^ War!within areas of suitable habitat[ Thus\ areas with morehouses would be more likely to present more roosting dhaugh 0881#[ Three interrelated structural features

"age\ roof compartments and roof lining# were ident!opportunities[ The fact that no roosts are known fromwithin Aberdeen city can be explained by the lack of i_ed by comparisons with both random and adjacent

houses[su.cient woodland nearby to support colonies of P[auritus[ When exposed to heating from the sun\ di}erent

compartments of a roof space are likely to warm atWithin the river valleys\ P[ auritus did not occupyhouses at random[ Instead\ the houses used were di}erent rates and to di}ering extents\ providing a

range of temperatures within the attic "Saint Girons\located closer to trees and water\ and were surroundedby more woodland within 9=4 km than expected from Brosset + Saint Girons 0858^ Ransome 0889#[ Bats

frequently adjust their roosting areas both within andthe random sample[ Since P[ auritus forages in wood!land "Swift 0880#\ the use of roosts close to wooded between roost sites\ apparently repositioning them!

selves relative to thermal gradients "Licht + Leitnerareas may indicate selection for appropriate feedinghabitat[ The distribution of woodland around roosts\ 0856# to select the thermal conditions appropriate to

their energetic status[ Roosts with a high degree ofwhen compared with randomly chosen buildingswhich were not used as roosts\ suggests that woodland structural complexity\ such as the large\ compart!

Þ 0886 Britishwithin 9=4 km\ rather than at greater distances from mentalized attics used by bats in the present study\Ecological Society\the roost\ should be important foraging habitat for may thus o}er a wide range of thermal regimes to theJournal of Applied

Ecology\ 23\ 288Ð397 this species[ This was supported by radiotracking bats "Saint Girons et al[ 0858^ Studier + O|Farrell

395 0861#[ The range of temperatures in complex attic to retain heat over a period\ rather than solely on thebasis of their ambient temperature at any given time[Roost selection by spaces may therefore explain their selection in relation

to control buildings\ which had fewer roof compart!the brown long! The temperature within a building is likely to bein~uenced by its design and construction[ The conse!eared bat ments[ In addition\ the thick planks of wood "sarking#

which lined these roosts may provide considerable quent variation in internal temperatures betweenbuildings may contribute to roost site selection byinsulation to bats using them "Kurta 0874#[

The warmer temperatures recorded in roosts com! bats[ However\ this study did not identify any relation!ship between attic temperature and the structural fea!pared with control buildings were not due to heating

of the roost space by the bats\ since the similar tem! tures of the houses shown to in~uence roost selection"age\ number of compartments and roof lining#[ Thisperature pro_les from the two probes situated in

di}erent parts of the roof apex revealed no evidence may indicate that the structural features of the build!ing and the mean attic temperature are independentthat either was a}ected by the presence of bats[ Since

P[ auritus roosts in small clusters within large roost factors in~uencing roost selection in P[ auritus\although an e}ect of building structure on the rangespaces\ local heating would not be expected to occur

beyond the cluster[ Instead di}erences in temperature of temperatures available in the roost could not beexcluded[between roost sites and control buildings indicated

that roosts are located in inherently warmer buildings\and the use of warm attics is consistent with the selec!

ROOST SELECTION BY PLECOTUS AURITUStion of a suitable microclimate by bats duringsummer[ Non!random patterns of association between bats and

buildings with certain features may indicate roostIt is predicted that bats would select roosts whichprovide a microclimate suitable to their energetic selection[ However\ selection may not necessarily be

active since not all buildings are available or accessiblestatus[ For a bat that remained euthermic during thesummer\ a warm site would reduce energy expen! to bats and roosting may be a response to a speci_c

cue in certain buildings "Hutto 0874#[ Furthermore\diture[ Warm roost temperatures may be particularlyimportant during the summer since they may a}ect roosting patterns of bats may be in~uenced by roost

_delity or social cohesion[ Nevertheless\ the associ!the duration of pregnancy "Racey + Swift 0870#\ thegrowth rate of the young "Tuttle 0864# and their event! ation of P[ auritus with summer roosts close to abun!

dant feeding habitat\ and with warmer buildings com!ual survival "Ransome 0878#[ However\ warmer roostsdo not provide suitable conditions for the use of pared to others that are both available and accessible\

suggests that bats select roosts with features that maytorpor[ Radiotracking studies "Entwistle 0883# havedemonstrated that following nights of low tempera! enhance their _tness[ It has been shown that di}er!

ences in roost characteristics are re~ected in _tness!tures\ when insect availability is decreased "Taylor0852#\ non!reproductive individuals of P[ auritus related variables in this species "Entwistle 0883#

The selection of speci_c types of roosts by P[ auritusmoved to peripheral day roosts that were signi_cantlycooler than the main roosts examined in the present suggests that not all buildings will be equally suitable

as roost sites\ and that roosting opportunities maypaper[ It is likely that such alternative roost sites o}erconditions suitable for the use of torpor over much of even be restricted[ Given the high roost _delity evident

in this species "Entwistle 0883#\ and the importance ofthe day roosting period "Audet + Fenton 0877#[The average di}erence in temperature between speci_c roost characteristics\ loss of a speci_c roost

site may have detrimental e}ects on the colony of batsroosts and control buildings was only 0=1 >C[ A pre!diction was made for the di}erence in daily energy using it\ since other sites in the vicinity may not be

suitable "or may be sub!optimal#[expenditure for bats roosting at temperatures foundin roosts and control buildings\ using equations for P[ Conservation advice provided when roosts of P[

auritus are threatened should take into account theauritus established using respirometry "Speakman +Racey 0876#\ and assuming a bat of body mass 7 g degree of selectivity shown by this species in relation

to its roost sites\ rather than assuming that any otherroosting for 07=4 h "assuming a ~ight time of 4=4 h^Entwistle et al[ 0885#[ The di}erence in temperature house might be a suitable roost[ Signi_cant changes

to the roost or exclusion of colonies from such sitesbetween roosts and control buildings results in energysavings of 0=9 kJ per day[ This represents a saving of should be avoided whenever possible[ In addition\

changes to woodland in the vicinity of the roost thatjust over 3) of the estimated total roosting energyexpenditure "14=4 kJ#\ and shows that selection of may a}ect the available foraging habitat need to be

taken into account[ Such factors should be consideredwarmer roost sites results in measurable energeticbene_t for the bats[ However\ the records of sig! carefully when applications are made to planning

authorities or statuatory organizations in the UKni_cantly higher temperatures in buildings used asroosts may re~ect the fact that heat is retained to a regarding disturbance to roosts under the 0870 Wil!

Þ 0886 Britishgreater extent by these buildings and they thus provide dlife and Countryside Act[ Thus protection of theEcological Society\a higher degree of insulation for the bats that use roost and surrounding woodland should be the pri!Journal of Applied

Ecology\ 23\ 288Ð397 them[ Bats may be selecting buildings for their ability mary conservation advice provided for this species[

diversity of nearctic bats[ Journal of Mammalogy\ 45\ 210Ð396 Furthermore\ since the number of roosting sites avail!235[A[C[ Entwistle\ able to P[ auritus may decline in future\ as older houses

Hutson\ A[M[ "0882# Action Plan for the Conservation ofP[A[ Racey + are renovated or demolished\ and older trees are Bats in the United Kingdom[ The Bat Conservation Trust\J[R[ Speakman felled\ the erection of alternative roosts\ such as bat London[

Hutto\ R[L[ "0874# Habitat selection in nonbreeding\boxes\ may be an important factor in increasing roost!migratory land birds[ Habitat Selection in Birds "ed[ M[ L[ing opportunities\ close to suitable habitat\ for thisCody#\ pp[ 344Ð362[ Academic Press\ NY[species[

Jenkins\ D[ "0874# The Biology and Management of the RiverDee[ Institute of Terrestrial Ecology "NERC#[

Jones\ G[ + Jayne\ A[F[ "0877# Bats in Avon] their distri!bution in relation to the urban environment[ ProceedingsAcknowledgementsof the Bristol Naturalist Society\ 37\ 20Ð40[

We are very grateful to all the householders who Jones\ G[\ Duverge�\ P[L[ + Ransome\ R[D[ "0884# Con!servation biology of an endangered species] _eld studiesallowed us access to their homes\ and who showed usof greater horseshoe bats[ Symposium of the Zoologicala great deal of hospitality while we carried out thisSociety\ 56\ 298Ð213[work[ Roosts were visited under licence from Scottish

Kunz\ T[H[ "0879# Daily energy budgets of free!living bats[Natural Heritage and the work was funded by a Natu! Proceedings of the 4th Internation Bat Research Conferenceral Environment Research Council studentship sup! "eds D[ E[ Wilson + A[ L[ Gardner#\ pp[ 258Ð281[ Texas

Tech[ Press\ Lubbock[porting A[ C[ Entwistle[ The Automatic WeatherKunz\ T[H[ "0871# Roosting ecology of bats[ Ecology of BatsStation was provided from the NERC Scienti_c Ser!

"ed[ T[ H[ Kunz#\ pp[ 0Ð44[ Plenum\ NY[vices equipment pool[Kurta\ A[ "0874# External insulation available to a non!

nesting mammal\ the little brown bat "Myotis lucifugus#[Comparative Biochemistry and Physiology 71 A\ 302Ð319[

Kurta\ A[ "0875# Factors a}ecting the resting and post~ightReferencesbody temperature of little brown bats\ Myotis lucifugus[

Audet\ D[ + Fenton\ M[B[ "0877# Heterothermy and the Physiological Zoology\ 48\ 318Ð327[use of torpor by the bat Eptesicus fuscus "Chiroptera] Kviz\ F[J[ "0873# Bias in a directory sample for a mail surveyVespertilionidae#] a _eld study[ Physiological Zoology\ 50\ of rural households[ Public Opinion Quarterly\ 37\ 790Ð086Ð193[ 795[

Barataud\ M[ "0889# Ele�ments sur le comportement ali! Licht\ P[ + Leitner\ P[ "0856# Behavioral responses to highmentaire des oreillards brun et gris Plecotus auritus "Lin! temperatures in three species of California bats[ Journal ofnaeus\ 0647# et Plecotus austriacus "Fischer\ 0718#[ Le Rhi! Mammalogy\ 37\ 41Ð50[nolophe\ 6\ 2Ð09[ Lunney\ D[\ Barker\ J[\ Priddel\ D[ + O|Connell\ M[ "0877#

Barclay\ R[M[R[\ Thomas\ D[W[ + Fenton\ M[B[ "0879# Roost selection by Gould|s long!eared bat\ NyctophilusComparison of methods used for controlling bats in build! gouldi Tomes "Chiroptera] Vespertilionidae#\ in logged for!ings[ Journal of Wildlife Management\ 33\ 491Ð495[ est on the south coast of New South Wales[ Austalian

Boyd\ I[L[ + Stebbings\ R[E[ "0878# Population changes of Wildlife Research\ 04\ 264Ð273[brown long!eared bats "Plecotus auritus# in bat boxes at Pritchard\ S[ + Murphy\ F[ "0877# A Highland bat surveyThetford forest[ Journal of Applied Ecology\ 15\ 090Ð001[ 0875Ð76] Glen Lyon and Rannoch[ Batchat\ 09\ 8Ð03[

Brigham\ R[M[ + Fenton\ M[B[ "0875# The in~uence of Racey\ P[A[ "0862# Environmental factors a}ecting theroost closure on the roosting and foraging behaviour of length of gestation in the pipistrelle bat\ Pipistrellus pip!Eptesicus fuscus "Chiroptera] Vespertilionidae#[ Canadian istrellus[ Journal of Reproduction and Fertility Supplement\Journal of Zoology\ 53\ 0017Ð0022[ 08\ 064Ð078[

Brosset\ A[ + Caube�re\ B[ "0848# Contribution a l|etude Racey\ P[A[ + Swift\ S[M[ "0870# Variations in gestationecologique des chiropteres de l|ouest de la France et du length in a colony of pipistrelle bats "Pipistrellus pip!Bassin Parisien[ Mammalia\ 12\ 079Ð127[ istrellus# from year to year[ Journal of Reproduction and

Dalquest\ W[W[ + Walton\ D[W[ "0869# Diurnal retreats of Fertility\ 50\ 012Ð018[bats[ About Bats "eds R[ H[ Slaughter + D[ W[ Walton#[ Ransome\ R[D[ "0878# Population changes of greater horse!Southern Methodist University Press\ Dallas[ shoe bats studied near Bristol over the past 15 years[ Bio!

Dixon\ W[J[ "0889# BMDP Statistical Software[ University logical Journal of the Linnaen Society\ 27\ 60Ð71[of California Press\ Berkeley[ Ransome\ R[D[ "0889# The Natural History of Hibernating

Entwistle\ A[C[ "0883# Roost ecology of the brown long!eared Bats[ Christopher Helm\ London[bat "Plecotus auritus\ Linnaeus 0647# in north!east Scot! Saint!Girons\ H[A[\ Brosset\ A[ + Saint!Girons\ M[C[ "0858#land[ PhD thesis\ University of Aberdeen[ Contibution a� la connaissance du cycle annuel de la

Entwistle\ A[C[\ Racey\ P[A[ + Speakman\ J[R[ "0885# Habi! chauve!souris Rhinolophus ferrumequinum "Schreber\tat exploitation by a gleaning bat\ Plecotus auritus[ Philo! 0663#[ Mammalia\ 22\ 246Ð369[sophical Transactions of the Royal Society\ 240\ 810Ð820[ Speakman\ J[R[ + Racey\ P[A[ "0876# The energetics of preg!

Fenton\ M[B[ "0869# Population studies of Myotis lucifugus nancy and lactation in the brown long!eared bat\ Plecotus"Chiroptera] Vespertilionidae# in Ontario[ Life Sciences auritus[ Recent Advances in the Study of Bats "eds M[ B[Contributions Royal Ontario Museum\ 66\ 0Ð23[ Fenton\ P[ A[ Racey + J[ M[ V[ Rayner#[ Cambridge

Hana�k\ V[ "0858# OÝkologische bemerkungen zur verbreitung University of Press\ Cambridge[der langohren "Gattung Plecotus Geo}roy\ 0707# in der Speakman\ J[R[\ Racey\ P[A[\ Catto\ C[M[C[\ Webb\ P[I[\Tschechoslwakei[ Lynx\ 09\ 24Ð28[ Swift\ S[M[ + Burnett\ A[M[ "0880# Minimum population

Harmata\ W[ "0851# Seasonal rhythmicity of behaviour and estimates and densities of bats in NE Scotland\ near to theÞ 0886 British the ecology of bats "Chiroptera# living in some old build! northern borders of their distributions\ during summer[Ecological Society\ ings in the district of Krako�w[ Zeszyty Naukowe Uniwer! Journal of Zoology\ London\ 114\ 216Ð234[Journal of Applied sytetu Jagiellonskiego\ 47\ 038Ð068[ Stebbings\ R[E[ "0877# Conservation of European Bats[ Chri!

stopher Helm\ London[Ecology\ 23\ 288Ð397 Humphrey\ S[R[ "0864# Nursery roosts and community

Studier\ E[H[ + O|Farrell\ M[J[ "0861# Biology of Myotis Tuttle\ M[D[ "0864# Population ecology of the gray bat397thysanodes and M[ Lucifugus "Chiroptera] Vesper! "Myotis grisescens#] factors in~uencing early growth andRoost selection bytilionidae#[ I[ Thermoregulation[ Comparative Bio! development[ Occasional Papers\ Museum of Natural

the brown long! chemistry and Physiology A\ 30\ 456Ð484[ History\ University of Kansas\ 25\ 0Ð13[eared bat Swift\ S[M[ "0880# Genus Plecotus "long!eared bats#[ The Tuttle\ M[D[ + Stevenson\ D[ "0871# Growth and survival

Handbook of British Mammals\ 2rd edn "eds G[ B[ Corbet of bats[ Ecology of Bats "ed[ T[ H[ Kunz#\ pp[ 094Ð049[+ S[ Harris#[ Blackwell\ Oxford[ Plenum\ NY[

Swift\ S[M[ + Racey\ P[A[ "0872# Resource partitioning in Wardhaugh\ A[A[ "0881# Bats and their roosts in Clevelandtwo species of vespertilionid bats "Chiroptera# occupying and north east Yorkshire[ Naturalist\ 006\ 88Ð097[the same roost[ Journal of Zoology\ London\ 199\ 138Ð148[

Taylor\ L[R[ "0852# Analysis of the e}ect of temperature oninsects in ~ight[ Journal of Animal Ecology\ 21\ 88Ð006[ Received 03 November 0884^ revision received 02 June 0885

Þ 0886 BritishEcological Society\Journal of AppliedEcology\ 23\ 288Ð397