13
Room for nature? Conservation management of the Isle of Rum, UK and prospects for large protected areas in Europe Brian Wood* Ecology and Conservation Unit, Department of Biology, University College London, Gower Street, London WC1E 6BT, UK Received 10 December 1998; accepted 13 September 1999 Abstract The Isle of Rum is one of the largest protected areas in UK and has been owned and managed by national conservation agencies since 1957. Detailed management plans have been prepared. The restoration of more natural conditions has been a principal objective for over 40 years. Even so, commercial forestry techniques have been used to reinstate native woodlands and domestic livestock were introduced to assist in the diversification of grasslands on the island. Many other management practices have been similar to ones employed in Highland deer forests, rather than specific to the needs of conservation and the restoration of a more natural state. The present population of red deer were introduced to Rum, but it is not certain that deer are truly native to the island. Nevertheless, conservation managers are now attempting to integrate deer into the replanted woodland. This is unlikely to replicate conditions that formerly existed on Rum since the interpretation of a pollen core suggests that, almost 4000 years ago, the native woodland declined when grazing animals arrived on the island. Managers also wish to preserve mementoes of earlier occu- pation and sustain and expand its present human population. Rum is one of a series of prime conservation sites selected under the Habitats Directive of the European Union. However, in seeking to protect the wildlife that is special to Europe, this legislation may promote the preservation of plant and animal communities that are principally cultural artefacts, rather than encourage a reliance on natural processes and the ecosystems they engender. Whilst this may be a reasonable approach to the conservation management of small and isolated reserves, it is evident that similar practices are used in some of the largest and most secure sites. There is a danger that, in seeking to preserve the parts (species and communities), conservation in Europe may lose sight of the entity that includes natural variability and the processes that sustain wild ecosystems. # 2000 Elsevier Science Ltd. All rights reserved. Keywords: Conservation management; Isle of Rum; Red deer; Restoration of nature; Habitats Directive 1. Introduction In North America, where protected areas are often extensive and quite natural, conservation biologists accept that these sites alone will be inadequate to pre- serve wildlife resources. This recognition has led to the growth of the concept of ecosystem management, which includes attempts to forge links between protected areas and with surrounding lands that are managed for pur- poses other than conservation (Agee, 1996). Ecosystem management has evolved because it is evident that even very large and predominantly natural sites, such as National Parks and Wilderness Areas, suer losses of spe- cies when they become isolated from their surroundings (Newmark, 1987). Essentially, it is an attempt to restore some of the linkages that would have existed before Eur- opean colonisation of North America and the widespread development of commercial agriculture, ranching and forest management, which has reduced the capacity of managed land to support wildlife. If important linkages are disrupted by an increase in isolation or a reduction in the area available to wildlife then it can be predicted, from island biogeography theory, that the number of species that the system can sustain will decline. Many of the observed losses from North American protected areas closely fit the predictions of this theory (Shafer, 1991). Over 7% of total land area in North and Central America is protected as strict nature reserves, national parks and natural monuments of at least 1000 ha in extent (IUCN protected area categories I–III), com- pared to <0.8% of Europe. This contrast is partly related to the very dierent histories of land-use in these two regions, but also to dierences in the approach to 0006-3207/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved. PII: S0006-3207(99)00160-3 Biological Conservation 94 (2000) 93–105 www.elsevier.com/locate/biocon *Corresponding author. E-mail address: [email protected] (B. Wood).

Room for nature? Conservation management of the Isle of Rum, UK and prospects for large protected areas in Europe

Embed Size (px)

Citation preview

Room for nature? Conservation management of the Isle of Rum,UK and prospects for large protected areas in Europe

Brian Wood*

Ecology and Conservation Unit, Department of Biology, University College London, Gower Street, London WC1E 6BT, UK

Received 10 December 1998; accepted 13 September 1999

Abstract

The Isle of Rum is one of the largest protected areas in UK and has been owned and managed by national conservation agencies

since 1957. Detailed management plans have been prepared. The restoration of more natural conditions has been a principalobjective for over 40 years. Even so, commercial forestry techniques have been used to reinstate native woodlands and domesticlivestock were introduced to assist in the diversi®cation of grasslands on the island. Many other management practices have been

similar to ones employed in Highland deer forests, rather than speci®c to the needs of conservation and the restoration of a morenatural state. The present population of red deer were introduced to Rum, but it is not certain that deer are truly native to theisland. Nevertheless, conservation managers are now attempting to integrate deer into the replanted woodland. This is unlikely to

replicate conditions that formerly existed on Rum since the interpretation of a pollen core suggests that, almost 4000 years ago, thenative woodland declined when grazing animals arrived on the island. Managers also wish to preserve mementoes of earlier occu-pation and sustain and expand its present human population. Rum is one of a series of prime conservation sites selected under theHabitats Directive of the European Union. However, in seeking to protect the wildlife that is special to Europe, this legislation may

promote the preservation of plant and animal communities that are principally cultural artefacts, rather than encourage a relianceon natural processes and the ecosystems they engender. Whilst this may be a reasonable approach to the conservation managementof small and isolated reserves, it is evident that similar practices are used in some of the largest and most secure sites. There is a

danger that, in seeking to preserve the parts (species and communities), conservation in Europe may lose sight of the entity thatincludes natural variability and the processes that sustain wild ecosystems. # 2000 Elsevier Science Ltd. All rights reserved.

Keywords: Conservation management; Isle of Rum; Red deer; Restoration of nature; Habitats Directive

1. Introduction

In North America, where protected areas are oftenextensive and quite natural, conservation biologistsaccept that these sites alone will be inadequate to pre-serve wildlife resources. This recognition has led to thegrowth of the concept of ecosystem management, whichincludes attempts to forge links between protected areasand with surrounding lands that are managed for pur-poses other than conservation (Agee, 1996). Ecosystemmanagement has evolved because it is evident that evenvery large and predominantly natural sites, such asNational Parks andWilderness Areas, su�er losses of spe-cies when they become isolated from their surroundings(Newmark, 1987). Essentially, it is an attempt to restore

some of the linkages that would have existed before Eur-opean colonisation of North America and the widespreaddevelopment of commercial agriculture, ranching andforest management, which has reduced the capacity ofmanaged land to support wildlife. If important linkagesare disrupted by an increase in isolation or a reduction inthe area available to wildlife then it can be predicted, fromisland biogeography theory, that the number of speciesthat the system can sustain will decline. Many of theobserved losses from North American protected areasclosely ®t the predictions of this theory (Shafer, 1991).Over 7% of total land area in North and Central

America is protected as strict nature reserves, nationalparks and natural monuments of at least 1000 ha inextent (IUCN protected area categories I±III), com-pared to <0.8% of Europe. This contrast is partlyrelated to the very di�erent histories of land-use in thesetwo regions, but also to di�erences in the approach to

0006-3207/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.

PI I : S0006-3207(99 )00160-3

Biological Conservation 94 (2000) 93±105

www.elsevier.com/locate/biocon

*Corresponding author.

E-mail address: [email protected] (B. Wood).

the conservation of wildlife (Henderson, 1992). In Eur-ope, most nature reserves are small and quite intensivelymanaged. Almost all have detailed management plansthat prescribe the actions needed in order to preservevalued species and communities. For the most threatenedwildlife, detailed species and habitat action plans havebeen prepared. These usually seek to maintain andincrease the numbers or extent of wildlife that may havedeclined in recent times, but often do not fully consider theconsequences for features that are not the speci®c target ofthese action plans (Wood, 1996). Nevertheless, the hopeof wildlife scientists is to be able to fully understand thefactors which may limit populations of threatened species,so as to be able to prescribe the precise remedial actionsnecessary for their recovery (Dempster, 1977).For small sites that are the remnants of near-natural

systems left aside from intensive agriculture, forestryand more destructive land-uses, conservation manage-ment may provide the only real hope of saving valuedwildlife. It is less clear that this is the only option thatcan be applied to the largest protected areas in Europe(Bibelriether, 1998). The most valued sites are protectedby domestic legislation that varies from country tocountry. They may also be recognised under a variety ofinternational designations. Recent European Union(EU) legislation attempts to perpetuate wildlife speciesand habitats that are special to Europe through theidenti®cation, protection and management of sites inwhich these still occur (Juniper, 1994; Hopkins, 1995). Italso emphasises the value of large sites, but does notspecify a minimum area for quali®cation as a prime site(Natura 2000 Site). Many of the reserves selected forinclusion in this European-wide series of protected areasare drawn from those that have long been recognisedunder earlier conservation designations. Few containextensive natural environments and many may haveundergone quite intensive conservation managementsince they were ®rst designated as protected areas.Ifmany of the largest European protected areas continue

to be managed for speci®c conservation objectives, it isimportant to determine how much wildlife will be pre-served for future generations in comparison with the typeof approach that prevails in North America. Ultimately,only timewill tell if any particular strategy ismore succesfulthan another one. However, in the meantime, it is worth-while examining the outcomes of the conservation man-agement that has been applied to prominent Europeanprotected areas since their designation, so as to assess thebene®ts inherent in a strategy of intensive conservationmanagement and to identify any potential shortcomings.

2. Isle of Rum NNR

The Isle of Rum, situated in the Inner Hebrides o�the north-west coast of Scotland, was recommended as

a Nature Conservation Area in the report to govern-ment on ``Nature Reserves in Scotland'' that was pub-lished as Command 7814 in 1949 (Magnusson, 1997). Itwas purchased by the governmental conservationagency, the Nature Conservancy (NC), in 1957 anddeclared as a National Nature Reserve (NNR) thatsame year. With a land area of 10,684 ha, it forms thethird largest NNR in Scotland and, since it encompassesthe whole island, must be one of the most securely pro-tected areas in western Europe. It is also designated,under domestic legislation, as a Site of Special Scienti®cInterest (SSSI) and forms a part of the Small IslesNational Scenic Area. Under EU legislation it is a Spe-cial Protection Area (SPA) for birds and has been pro-posed as a Special Area of Conservation (SAC) andcandidate Natura 2000 Site. It is recognised byUNESCO as a Biosphere Reserve. For over 40 years,since the time of its acquisition, it has been managed bythe principal government conservation agencies in the UKand a series of plans have been prepared, which detail theconservation management that has been applied (ScottishNatural Heritage, 1998). These also provide an excellenthistoric record of conservation on Rum. Britain has beenin the forefront of the development of conservation man-agement in Europe and much of the developing world andhas been prominent in the promotion of Europe-wideconservation legislation. Thus, the Isle of Rum providesan excellent example of the European approach to natureconservation within protected areas.Rum is a mountainous island with several major

peaks rising to over 800 m. A long history of humanoccupation has left it almost treeless, like much of themainland Scottish Highlands. However, there is a smallarea of woodlands, of both native and exotic trees, plan-ted around the castle, which was erected on the island byits former owner Sir George Bullough. Its principal fea-tures of conservation importance are set out in Ratcli�e(1977) and include the complex geology of the remains ofa volcanic cone that forms the core of the island. The topsof its highest mountains provide nesting places for a col-ony of over 60,000 manx shearwaters (Pu�nus pu�nus)which constitute over 20% of the world population of thisseabird (Wormell, 1976). They are home to several rareplant species, many of which are relics of the last glacia-tion in Europe. It is also noted for its herd of red deer(Cervus elaphus) which provides a special opportunity forpopulation studies because of the isolation of the island.

3. History of conservation management

Six management plans cover most of the period of con-servation management on Rum from 1960 to the presentday (Eggeling, 1964; Ball, 1970; Nature ConservancyCouncil, 1974, 1987; Boyd, 1977; Scottish Natural Heri-tage, 1998). The few years that are not included in these

94 B. Wood / Biological Conservation 94 (2000) 93±105

published plans (1975±6, 1983±6, 1997) are partially cov-ered by unpublished interim documents. In any case, con-servation management has continued uninterrupted sincethe island was purchased in 1957. However, during thistime, the emphasis of management has varied (Table 1).Several principal themes of management are apparent

throughout the 40-year history of conservation on theisland. All are interlinked and cover not only speci®cgoals for the maintenance of biological and physio-graphic attributes of the reserve but also social andeconomic objectives.

3.1. Research

Although research is usually regarded as a meansrather than an end in itself, in the ®nal report of theScottish Wild Life Conservation Committee, submittedto Parliament in 1949, Rum was recommended as aNature Conservation Area and described as ``Isolated,yet within easy reach of the mainland, it would make anoutstanding station for research, and indeed is the mostsuitable island for this purpose in Scotland''. In the ®rst10 years, following its purchase and designation as aNNR, sta� of the NC studied both the populationdynamics of red deer and their response to a novel pro-gramme of management (Lowe, 1966, 1969, 1971; Charles

et al., 1977) However, this research faltered when theoutcome of the management experiments producedresults which were contrary to those desired by the NC.Fortunately, research on the population biology andsocial interactions of red deer by Cambridge Universityexpanded soon after the demise of the work undertakenby NC researchers. Rum has become world famous as thelocation for this very long-term and informative socio-biological study (Clutton-Brock et al., 1982).A great deal of internationally important research has

also been undertaken on the special geology of theisland. The vegetation of Rum has been mapped indetail and the response of plant communities to a vari-ety of grazing regimes has been investigated. Never-theless, much research on the island has not had directimplications for its conservation management and agreat deal could have equally well been undertakenelsewhere. The most recent management plan (ScottishNatural Heritage, 1998) speci®es that, in future, onlyresearch that will improve understanding of the featuresof interest and their management will be permitted.

3.2. Deer and estate management

There is a long-held view that most of the ScottishHighlands consists of impoverished landscapes that

Table 1

Conservation objectives for the Isle of Rum

Period of plan Objectives

1960±64 Obtain basic information about island.

Initiate research on red deer, deer management and moorland productivity.

Provide shelter and increase variety of habitats by re-creating woodland.

1965±69 Restore vegetation cover that has been lost.

Bring island to a higher level of biological production.

Conserve ¯ora, fauna and other features of scienti®c interest.

Enhance diversity of habitats, provide shelter, and increase the biological turnover by means of rea�orestation.

Undertake research necessary to attain primary aims of management.

Manage reserve in best traditions of estate management.

1970±74 Restore, in selected parts of island, vegetation, animal communities and soils thought to have existed before

being destroyed by man's activities.

Demonstrate that this is compatible with sustained cropping of native and domestic herbivores and eventually

of forest products.

Carry out research, to identify biological trends and to ®nd best methods of managing the land.

Use best standards of estate management.

1977±82 Allow for natural development by reducing management input.

Retain opportunity to foster research.

Keep options open for future.

1987±96 Allow natural development under minimum management to maintain and enhance nature conservation interest,

the latter by habitat and species manipulation.

Encourage use for research and education.

Move away from management typical of a Scottish Highland estate.

1998±2008 Conserve and, where appropriate, enhance Rum's outstanding natural and cultural heritage, whilst providing a

demonstration of sustainable land use and facilitating compatible use of the island for study and enjoyment.

Have a thriving (human) community on Rum that also contributes to the socioeconomic development of the

Small Isles as a whole, and which functions in a way that is compatible with the other objectives of management.

B. Wood / Biological Conservation 94 (2000) 93±105 95

have resulted from generations of unsustainable grazingby livestock, burning and felling of woodlands (McVeanand Lockie, 1969). The conservation managers of Rumhave evidently considered that the island provides agreat opportunity for the demonstration of moreappropriate rural practices. In seeking to develop newways of managing the Highlands they clearly believethat more sustainable practices would also provideconsiderable conservation bene®ts.It is traditional to cull approximately one tenth of the

red deer on Scottish deer forests each year (Clutton-Brock and Albon, 1989). Land owners have been reluc-tant to take a greater crop than this because the value oftheir estate is intimately linked with the number of stagseach can sustain (Watson, 1989). This fear seems tohave been a major factor in the gradual rise in deernumbers throughout Scotland (Clutton-Brock andAlbon, 1992). The Isle of Rum had been managed as atraditional deer forest for >100 years before it becamean NNR. Soon after it was acquired by the NC a cull ofapproximately one sixth of all deer was instigated andthis demonstrated that this level of cropping could besustained without causing a progressive populationdecline. It also reduced mortality of deer by naturalcauses to very low levels, thereby avoiding what couldbe viewed as unnecessary wastage (Lowe, 1969).More recently, following detailed population modelling

that resulted from the intensive studies of an uncroppedsegment of the red deer on Rum by researchers fromCambridge University, an island-wide experiment hasbeen undertaken into techniques that may maximise thepro®t from deer forest management. It is hoped that thiswill entail a reduction in stocking density, whichmay haveconservation bene®ts. This experiment has used con-trolled manipulation of sex ratios in order to determinethe resultant changes in growth rates of each segment ofthe population. Since trophy hunting is the most pro®t-able aspect of deer management, these experiments seek toachieve an ideal balance between the provision of largestags and the total yield of venison from both sexes of deer(Clutton-Brock and Lonergan, 1994).Other aspects of the management of Highland estates

were also considered in the plans for the Isle of RumNNR. The practice of muirburn, to regenerate woodyericaceous vegetation and make it more palatable tolivestock, was terminated when the reserve was estab-lished because of the risk of continuing degradation andsoil erosion. At the same time, sheep were removedfrom Rum. However, a small herd of Highland ponies,of a bloodstock considered to be unique to Rum, werekept in order to provide pack animals for the removal ofdeer carcases from remote hill land. Highland cattlehave been introduced to the island and the introductionof Hebridean sheep is also under consideration, becausethey consume grasses that are largely unpalatable todeer (Scottish Natural Heritage, 1998).

Rum supports a herd of about 200 feral goats (Caprasp.) on its western sea-cli�s and adjacent grasslands(Boyd, 1980). Initial conservation management planssought to regularly cull this population, so as to preventits expansion. However, the cull was abandoned when itproved di�cult to implement and was evidently largelyunnecessary. Nevertheless, if goat numbers increase inthe future there are proposals to limit their numbers anddistribution on the island. Early plans also sought todevelop a strategy to control crows (Corvus corone),which are widely viewed as vermin (Eggeling, 1964). Thiswas soon dropped from the management programmeand, in the plan for 1987-96, it was advocated thatmanagement should move away from that typical of aScottish Highland estate. Even so, it is still possible forvisitors to Rum to purchase a permit to ®sh for trout(Salmo trutta) in the few streams on the island.

3.3. Woodland restoration

Apart from the plantations around the castle at Kin-loch on Rum and some shelterbelts in the nearby farm-land, in 1957 there were only a few fragments of nativewoodland growing on sea cli�s and in gullies that areinaccessible to deer. However, a survey of pollen andmacroscopic plant remains quickly established thatthere was once a quite extensive woodland cover on thelower parts of the island, especially in the north and eastthat is less exposed to the prevailing winds and saltspray (Ball, 1987). The restoration of this woodlandbecame one of the principal objectives of conservationmanagers. A previous Scottish director of the NCC sta-ted that ``the main objective of the nature reserve is torecreate a habitat resembling that which existed in Rumbefore the island was made treeless by man, and toachieve this, the managers must have a reliable pictureof the island as it was 2500 years ago, at the beginningof the Iron Age'' (Boyd and Boyd, 1990). In e�ect, thisamounts to a desire to return the island to a more nat-ural state, in contrast to the state it was in whenacquired by NCC, which was the outcome of manygenerations of human settlement and pastoralism.Experience with commercial forestry plantations on

the Scottish mainland had established that tree regen-eration is only feasible if deer are excluded from wood-land areas. Consequently, after a few early trialplantations had demonstrated that tree growth waspossible in several parts of Rum, large blocks of land inthe east of the island were surrounded by deer fencesand planted with seedlings of native trees. These weremostly derived from seed sources on Rum, or occasion-ally from the nearby parts of the mainland. However, inorder to attain acceptable rates of seedling establish-ment, commercial forestry techniques were applied(Wormell, 1968). These included deep ploughing ofpeat-covered land, the application of fertilizers and the

96 B. Wood / Biological Conservation 94 (2000) 93±105

use of nurse crops of non-native lodgepole pines (Pinuscontorta). Over one million trees have been planted onRum in the past 40 years or so.It appears that it was always expected that, even-

tually, the protective fences would be removed fromaround the new woodlands and deer permitted to enter.Successive conservation managers on Rum have heldthe belief that deer and woodland happily coexisted onthis island in the past, although it was recognised thatthis would only be feasible if deer numbers are verysubstantially reduced. Formerly, this may have resultedfrom human predation of the deer, as it is believed that``Neolithic man penetrated the island's scrub forest tohunt deer'' (Love, 1987). This view is supported by thediscovery of an arrowhead of white siliceous stone on theslopes of Rum's tallest peak and two more arrowheadson the coast, made from the local bloodstone.The most recent management plan (Scottish Natural

Heritage, 1998) clearly recognises that browsing by deerimposes severe restrictions on the regeneration of nativetree species. It advocates a substantial reduction in deernumber over the next few years, from the current levelof about 1500 to only 3±400 animals. This, it is hoped,will enable much more widespread natural regenerationof trees to occur. However, there is concern that thedevelopment of the coarse herbaceous vegetation thatpresently occupies much of the lower ground on Rummay e�ectively prevent the successful germination oftree seeds. Thus, the use of Highland cattle and Hebri-dean sheep is advocated, as a means of opening up thevegetation and to create regeneration niches.The managers are not speci®c about the need for further

control of deer numbers once the woodlands are fullyestablished and the protective fences are removed. Othershave suggested that a wild predator should be introducedto the island in order to limit deer numbers and permitcontinuing tree regeneration and growth (Nevard andPenfold, 1978; Yalden, 1986). However, there is currentlyno evidence to suggest that predatory wild mammals haveever occurred on Rum in the past and population model-ing has shown that wolves, if introduced, would quicklydecline to extinction (Spinney, 1995).

3.4. Maintenance and restoration of diversity

The cessation of the practice of muirburn and thetotal removal of sheep from Rum, soon after it wasacquired as a NNR, brought to an end a period of pas-toral agriculture that had begun during the 1820s,immediately after almost the entire human populationof the island was evicted and made to emigrate to NovaScotia. It was quickly apparent to NC scientists that thisreduction in grazing pressure had a substantial e�ect onthe vegetation of the island. Exclusion plots were estab-lished in order to study the e�ect that deer have ongrasslands and these too showed that, in general, plant

species richness was declining (Ball, 1974). In turn, a fewrelatively unpalatable herbs and grasses became domi-nant and this was implicated in a general decline in bodycondition of the deer (Lowe, 1971). So as to reverse theloss of plant diversity and also attempt to improve thegrazing for deer, a herd of Highland cattle was intro-duced to Rum in 1970. Where these animals were con-centrated, at Harris in the south of the island, it quicklybecame apparent that this management enhanced plantdiversity and led to improvements in condition of reddeer hinds which also increased their birth rates (Gor-don, 1988, 1989a,b,c). Cattle and other domestic herbi-vores are widely used in conservation management ofimportant grasslands throughout Europe (Bakker et al.,1983; de Vries, 1995; van Weiren, 1995). Scottish Nat-ural Heritage (SNH) are convinced of their value forthat purpose and also to keep open areas considered tobe of cultural interest. It is their intention to increasecattle numbers at Harris in the near future, whererecently the population has fallen, as a result of severalpoor calving seasons (Scottish Natural Heritage, 1998).Close-cropped grasslands along the western seaboard

of the UK are home to the chough (Pyrrhocorax pyr-rhocorax), a declining member of the crow family(Bignal and McCrachen, 1996). This bird becameextinct on Rum by 1871 but is now being considered forre-introduction to the island. Although there is no evi-dence to show that they have ever naturally lived there,other species that inhabit the Scottish Highlands havealso been proposed as candidates for introduction to theIsle of Rum. These include roe deer (Capreolus capreolus)and blue hare (Lepus timidus) (Nature ConservancyCouncil, 1987). One of the strongest arguments for there-establishment of woodland on the island is that itenables the conservation managers to restore ``animalcommunities and soils that are thought to have existedthere before they were destroyed by man's activities''(Ball, 1970). It was quickly shown that the new plantingshad a dramatic e�ect on the diversity and abundance ofwoodland birds (Batten and Pomeroy, 1969).

3.5. Sustainable human use and community support

A public ferry service connects the Isle of Rum withthe Scottish mainland at the port of Mallaig. This boatalso serves the other Small Isles (Eigg, Muck andCanna). Throughout the period of conservation man-agement it has been considered essential to maintain aworkforce on Rum at all times. At present the entirepopulation, with the one exception of the school teacher,is supported by Scottish Natural Heritage.To provide for this workforce, the dwellings in the

one remaining settlement at Kinloch, in the east of theisland, are maintained and new buildings have beenerected to serve the community. Some are also availablefor hire to visitors to the island. Rum castle, which is a

B. Wood / Biological Conservation 94 (2000) 93±105 97

listed building and must therefore be maintained by law,also provides hostel accommodation and is home to anostentatious collection of artefacts assembled there byits former owner, Sir George Bullough.In practice, the maintenance of the castle, houses and

other buildings, together with equipment, vehicles androads, requires a substantial annual expenditure andmakes the Isle of Rum the most expensive NNR tooperate (McKie, 1987). There have been times when themanaging agency has sought to reduce or eliminatethese costs by passing on its responsibility. Thus, theNational Trust for Scotland, a non-governmentalagency, took over the castle for a 2-year period. How-ever, they relinquished their interest at the end of thattime because it proved to be so ®nancially demanding.More recently, there was discussion betweeen ScottishNatural Heritage and another non-governmental body,the John Muir Trust, about the possibility of the lattertaking over the entire management of the NNR. In the1970s, following the disbanding of the NC and its repla-cement by the Nature Conservancy Council, a period of®nancial austerity meant that the working populationwas reduced to just six sta� and management of theisland dropped to a minimum.With ®nancial assistance from the EU there are cur-

rently plans to improve the port facilities on Rum andthereby better facilitate access to the island. ScottishNatural Heritage (1998) have expressed an increasedcommittment to ``have a thriving community on Rumthat also contributes to the socioeconomic developmentof the Small Isles as a whole, and which functions in away that is compatible with the other objectives of man-agement''. They are patently aware of the need to con-sider the aspirations and traditions of the people of thisregion, even when these may not primarily relate towildlife conservation. For example, it is considered that``psychological bene®ts would accrue from the coste�ective intensi®cation of cattle (on Rum), as this wouldbe regarded as a positive form of land use by other localcommunities'' (Scottish Natural Heritage, 1998). Thelong-term vision for Rum is as ``an island which has acommunity living and working in harmony with theirenvironment. An island which is economically and eco-logically dynamic and sustainable. An island which willbe inherited by each successive generation with prideand a committment to carry on working with nature''(Ritchie, 1997).

4. A natural heritage?

Whilst there has been a succession of governmentalconservation bodies in charge of the management ofRum NNR during its existence and `nature' is includedin each of their names, the management that they haveimposed has been far more all-encompassing than the

conservation of nature alone may require. Rackham(1994) suggests that one principal function of a con-servation management plan is to bu�er the reserveagainst the vagaries of conservation fashion. However,it is apparent that the conservation management of theIsle of Rum has, in fact, been strongly in¯uenced by anapparent need to conform to changing practices anddesires (Table 2).A re-ordering of management emphasis may have

been essential in order to survive the periodic strin-gencies imposed on conservation agencies by the gov-ernment of the day (Evans, 1997). Alternatively, it couldgenerously be interpreted as adaptive management, eventhough not engendered by an improved understandingof natural systems. Nevertheless, most of the majorthemes that were apparent in 1960 have persisted untiltoday. They include a vision of an original natural stateof the island, before the time of damaging human in¯u-ence. However, they also clearly encompass a desire toperpetuate cultural artefacts that have generated sys-tems that are liked by conservation practitioners, forexample the diverse grassland at Harris, which was oncecultivated by former island peoples. There is now a clearwish to also restore species such as the chough, whoseecology is intimately connected with human pastoralsystems. The wish to include mementoes of earlieroccupation extends to domestic livestock; the specialbloodline of Rum ponies, Highland cattle and Hebri-dean sheep. It also extends to wild red deer, for whichRum is now considered important as a refuge of puregenotypes, unpolluted by interbreeding with sika deer(Cervus nippon), as in much of mainland Britain (Scot-tish Natural Heritage, 1998). It further extends to theduty to maintain Rum castle, as a relic of Victorianextravagance, and to the need to sustain the island'shuman community and its role in the wider communityof the Small Isles. As well as giving support to people inthis sparsely populated area of Scotland, there has been acontinuing desire to show that good estate managementmay potentially generate conservation bene®ts.Although there is no indication that some of the plant

communities that are now found on Rumwere necessarilypresent there when the island was in a natural state,conservation managers have been willing to utilise agri-cultural techniques to sustain these for their perceivedimportance. The meadowlands of the former farm nearKinloch and the pasturelands at Harris are both cul-tural artefacts. In seeking to restore the woodlandswhich are known to have been quite widespreadthroughout the lower ground on Rum in the past, theyhave shown a desire for rapid achievements. So as toproduce demonstrable results within a few decades,commercial forestry techniques have been employed(Wormell, 1968) and the distribution and mix of treespecies that have resulted cannot be considered to trulyre¯ect natural patterns. Woodland herbs, which may

98 B. Wood / Biological Conservation 94 (2000) 93±105

eventually recolonise the new woodlands, have beentranslocated by conservation managers (Ball, 1987), sothat plant communities can now be found in isolatedblocks of new woodland that are apparently morecomplete than could otherwise be expected.Yet, one major ecological consideration remains

apparently unresolved on Rum: the desire to eventuallyhave a well-wooded island inhabited by deer. It isrecognised that deer numbers must be very substantiallyreduced before they are permitted access to the newwoodlands (Boyd, 1977; Scottish Natural Heritage,1998). Indeed, a remote plantation on the slopes of ArdNev, near the centre of Rum, has recently been totallydestroyed by deer and goats that gained access to it. If itproves that constant management of deer numbers isneeded to permit their coexistence with the newlyrestored woodlands, then this would be very di�cult tomatch with the vision of the island in a more naturalstate. In that eventuality, only a belief that people wereat all times a part of the natural ecology of deer andtrees on Rum, or that some unidenti®ed natural pre-dator of deer was formerly present on the island, wouldallow this vision of a former state to be sustained, asopposed to an arti®cial construct.

5. Evidence of the natural state of Rum

5.1. Vegetation changes

Our knowledge of early events in the history of theisland is incomplete, but greatly enhanced by ®ndingsfrom recent excavations at Kinloch on the eastern shoreof Rum (Wickham-Jones and Pollock, 1987). Theseshow that there was a human presence on the island asmuch as 8500 years ago, but do not indicate how per-manent this was, nor what impact people may have hadon the island's ecosystem. However, analysis of a pollenpro®le reveals that a drastic and permanent decline inwoodland took place between about 4000 and 3500years BP (before the present) (Hirons and Edwards,1990). This decline in woodland was accompanied byother vegetation changes, some of which are shown inTable 3.Even at Kinloch, on the most sheltered part of Rum

and near to the centre of the area that has recently beenreplanted with trees, the native woodland was appar-ently always quite open in structure. The change in pol-len concentrations have all been interpreted as due to lossof this woodland, perhaps caused by human clearance.

Table 2

Possible in¯uences on the principal objectives within conservation management plans for the Isle of Rum

Period of plan In¯uencing factors

1960±64 Rum acquired by the Nature Conservancy (NC) in 1957 and declared as a National Nature Reserve.

The Deer (Scotland) Act of 1959 established the Red Deer Commission with responsibility for all aspects of

the conservation and control of red and sika deer.

1965±69 In 1964 the International Council of Scienti®c Unions established the International Biological Programme with

the theme of ``the biological basis of productivity and human welfare''.

Peter Wormell, a trained forester, was Chief Warden of Rum NNR.

1970±74 UNESCO launched it Man and the Biosphere programme in 1971, to investigate problems arising from human

uses of the environment.

Research by Cambridge University on the red deer on Rum began in 1970.

In 1972 the NC agreed to terminate the annual cull of deer in the north block of the island in order to facilitate

long-term investigations of population dynamics and sociobiology of the deer.

1977±82 The NC was disbanded by Act of Parliament in 1973. Its functions were divided between two new organisations.

The Nature Conservancy Council (NCC) became responsible for conservation designations and the management

of National Nature Reserves. The Institute of Terrestrial Ecology incorporated the research branch of the former NC

and became responsible for environmental research. This reorganisation led to reductions in sta� and funding for

conservation work by NCC.

Rum was designated as a Biosphere Reserve by UNESCO in 1976.

1987±96 There was continuing concern about the mismanagement of red deer in Scotland which had led to overgrazing of

upland areas and the prevention of natural regeneration of woodland. The Royal Society for the Protection of Birds

(RSPB) on its reserve at Abernethy and the NCC on Creag Meagaidh NNR both instigated massive reductions in deer

numbers in order to encourage the regrowth of native woodlands.

Rum was declared a Special Protection Area (SPA) under the EU Birds Directive in 1982.

In 1987 Rum was re-designated as a Site of Special Scienti®c Interest under the Wildlife and Countryside Act.

1998±2008 The NCC was disbanded and three separate conservation agencies were formed for England, Scotland and Wales.

Scottish Natural Heritage (SNH) was created in 1992 by the amalgamation of the Countryside Commission for

Scotland and the Scottish part of the former NCC. The combined agency has responsibilities for landscape protection

and enjoyment of the countryside as well as wildlife conservation.

Rum was proposed as a Special Area of Conservation under the EC Habitats Directive in 1995. It was on the list of

candidate SACs sent to the EC in 1997.

B. Wood / Biological Conservation 94 (2000) 93±105 99

Charcoal concentrations suggest that this clearance mayhave partially resulted from the use of ®re. However, thechange in abundance of certain plant species indicatesthat other factors were also important in the eliminationof woodland and its replacement by acidic grasslandand heath. Both meadowsweet (Filipendula ulmaria) androyal fern (Osmunda regalis) are very sensitive to graz-ing. They are still present on Rum, but are con®ned tothe few places that are ecologically suitable and yetinaccessible to deer. Elsewhere in the Hebrides they arecommon components of the vegetation of ungrazedislands within freshwater lochs. In contrast, tormentil(Potentilla erecta) is unpalatable and is commonlyfound in grazed acidic grassland. In the Hebrides,modern woodlands that are lightly grazed have a ferny¯oor, dominated by sweet mountain fern (Oreopterislimbosperma), male fern (Dryopteris sp.), lady fern(Athyrium sp.) and beech fern (Phlegopteris connectilis).When heavily grazed, the woodland ¯oor is dominatedeither by bracken (Pteridium aquilinum) or by a commu-nity including sweet vernal grass (Anthoxanthemum odor-atum), bent grass (Agrostis sp.), fescues (Festuca spp.),pignut (Conopodium majus), yellow pimpernel (Lysi-machia nemorum), bugle (Ajuga reptans) and enchantersnightshade (Circaea lutetiana) (Boyd and Boyd, 1990). Itis notable that, at about the time that woodlanddeclined on Rum, there was a striking decline in fernsand a marked increase in grasses (Table 3).Some of the vegetation changes indicated by the pol-

len record could have resulted from the clearance ofwoodland by human settlers. However, the substantialand almost contemporaneous decline in palatable plantspecies and the rise in unpalatable ones is highly sugges-tive that this coincided with the ®rst appearance ofmammalian herbivores on Rum. That this was synchro-nised with the demise of the island's woodlands, indicatesthat mammalian herbivores and native woodland maybe mutually incompatible in this situation. However,changes in the abundance of palatable and grazing

resistant plant species can not tell us which herbivorespecies may have been involved, nor if they reached theisland unaided or were deliberately taken there byhumans.Although there are at present large populations of red

deer on the nearby mainland and larger islands of theInner Hebrides, there are no historical records of anyhaving succeeded in swimming the 15 km wide stretch ofsea between Skye and Rum, against the prevailingwinds. Since the retreat of the last glaciers there hasalways been a deep water channel between the island ofRum and the mainland. Red deer probably returned tomainland Scotland around 9500 BP (Yalden, 1982).They are now present on many of the Hebridean islandsbut are mostly believed to have been introduced byman. It is entirely possible that man may have intro-duced them to Rum about 4000 years ago, but we willnever know if that was the case, or if they managed tonaturally colonise this island. What is important is that,once herbivores arrived, the forest apparently declinedrapidly.There is strong evidence that other mammal species

were introduced to Rum by people. Wood mice (Apo-demus sylvaticus) evidently came via Eigg (Berry, et al.,1967), which is a relatively fertile island and, conse-quently, better able to sustain a human community.Yalden (1982) is of the opinion that almost all islandso� the west coast of Britain owe all their mammal faunato accidental introductions by man.

5.2. Red deer

The earliest historical account of red deer on Rumappears to be that of Dean Monro in the 16th century,who recorded ``an abundance of little deire''' (®rstpublished 1549; new edition, Monro, 1934). These weresubsequently driven to extinction as the human popula-tion of the island rose to a peak of almost 450 people atthe end of the 18th century. The present population of

Table 3

Some vegetation changes on Rum as indicated by pollen analysis

Vegetation Changes in status

a. Decreases

Woodland Declined sustantially after 4000 BP. Present abundance attained by about 3000 BP.

Meadowsweet (Filipendula ulmaria) Abundant until 4800 BP. Present continually since then, but only local in distribution and not numerous.

Bracken (Pteridium aquilinum) Common until about 4000 BP. Present, but not common since that time.

Polypody fern (Polypodium sp.) Common until 3300 BP. Occasional since then.

Royal fern (Osmunda regalis) Abundant until 5000 BP. Common from then until about 4000 BP. Disappeared from pollen pro®le

after 3400 BP. Still occurs on Rum in places inaccessible to deer.

Other ferns (Filicales) Constituted 20±70% of total land pollen until about 4000 BP. Uncommon in pollen record since then.

b. Increases

Heather (Calluna vulgaris) Present throughout the pollen pro®le. Common since 4000 BP.

Tormentil? (Potentilla sp.) Present but not common until 4000 BP. Widespread and abundant since then.

Grasses (Gramineae) Showed a brief peak between about 6000 and 5800 BP when they represented 50±60% of total pollen.

Otherwise, made up about 20% of total pollen until about 4000 BP, then increased to around 50% of total.

100 B. Wood / Biological Conservation 94 (2000) 93±105

red deer is descended from animals introduced by theMarquiss of Salisbury in 1845 and later supplementedby deer from mainland estates, including Woburn Parkin southern England. Scottish Highland red deer, whichgenerally live in open, treeless uplands, are much smal-ler than the same species in southern England and con-tinental Europe. On Rum, stags may have an averagelive weight of about 121 kg and hinds around 85 kg.This may be compared with body masses of 152 and 133kg for stags and hinds, respectively, in southern Scot-land, where they inhabit forestry plantations at lowdensity. Red deer stags that are descended from Scottishanimals introduced to New Zealand may weigh over 200kg (Corbet and Harris, 1991).If mammals are con®ned to islands where they have

no e�ective predators, the strongest selection pressurebecomes the need to be thermodynamically e�cient.There is an optimum body size for energy acquisition atabout 1 kg. Thus, on islands, small mammal speciestend to evolve towards larger body size and largermammals evolve to become smaller (Damuth, 1993).Lister (1989, 1995) has shown that on Jersey, situatedo� the northern coast of France, its red deer werereduced to less than one sixth of their former body massover a period of <6000 years when sea levels rose tocreate the island during the Eemian interglacial.During the ®rst 10 years of conservation management

on Rum there was a signi®cant decline in body mass ofdeer (Lowe, 1971) and the body size of animals in thenorth block, where no cull has taken place since 1972because of the long-term sociobiological study, hascontinued to decline. It seems entirely feasible that DeanMonro's ``little deire'' were red deer that had dramati-cally declined in size on Rum since they ®rst becameisolated there. Small deer are now protected by con-servation managers in the Florida Keys where they arevalued for their uniqueness (Burt and Grossenheider,1952). If the deer on Rum were a comparable dwarfvariety, it is a pity that they were extinguished byhuman predation and have been replaced, subsequently,by animals that do not di�er signi®cantly from those inmainland UK.

6. Discussion

6.1. Natural states

Much of the philosophy for the conservation man-agement of the Isle of Rum appears to be based on thebelief that humans have progressively degraded its eco-systems and considerably modi®ed its natural featuresduring the course of their occupation and use of theisland (Ball, 1970; Boyd, 1977; Boyd and Boyd, 1990).To a certain extent, it is a matter of personal opinion asto whether or not people can be regarded as a part of

nature. However, many authors would agree withPeterken (1996), that it is safer to ``retain the idea of`nature' as separate from people''. In seeking to undothe apparent harm done by past human inhabitants, thisaim of conservation management can be regarded asgoing some way towards the restoration of nature.Two, alternative, visions of the natural state of the

Isle of Rum appear possible. First, it seems highly likelythat when woodland clothed the lower slopes of thisisland, over 4000 years ago, there were very few mam-mal species on the island. Certainly, it is probable thatno large herbivores existed on Rum before that time. Ifthe woodland was quite open, as is indicated by the onepollen pro®le so far constructed (Hirons and Edwards,1990), there would also have been glades that would bedominated by ferns and woodland herbs. Elsewhere,there is likely to have been grassland on sand dunes, seacli�s and other coastal areas subject to strong salt spray.However, the pattern of dominance of the plant speciesin these vegetation types would have been quite unlikethose presently found on Rum. Some would be muchmore akin to the grasslands now developing withinexperimental plots from which grazing animals areexcluded (Ball, 1974). In contrast, the vegetation on theupper slopes of the hills would have been quite similarto that to be found there today, where climatic factorsare a more important determinant of plant speciescomposition than is grazing pressure.Second, within the last 4000 years, when woodlands

disappeared and deer numbers rose, the island wouldhave been transformed until its appearance becamequite similar to that at the time when it ®rst became anature reserve. Trees and woodland herbs would havebeen restricted to inaccessible crags and sea cli�s. Thevegetation on the lower parts of the island would havecontained a high proportion of relatively unpalatableplants, such as mat grass (Nardus stricta). The deerwould probably be of great interest to the conserva-tionist, being numerous and of much smaller body sizethan even those found in the poorest and most over-grazed parts of Scotland at present. But, strictly speak-ing, this second state should be only considered asnatural if deer were able to successfully colonise Rumunaided.Both the natural woodland and the possibly unique

small deer are now gone from Rum and the island is, inmany ways, very like large areas of the Scottish High-lands. However, whichever of the two states outlinedabove prevailed, the island would have still possessed adistinctive and internationally important geology, andalmost certainly nesting shearwaters in internationallyimportant numbers. Trollval, one of the highest peakson Rum, was supposedly so named by Norse invadersover 1500 years ago because of the noises emanatingfrom shearwaters nesting in mountain-top burrows, asthey still do today (Love and Wormell, 1987).

B. Wood / Biological Conservation 94 (2000) 93±105 101

6.2. Room for nature?

There are very few areas in Europe where naturalecosystems still persist. So, the restoration of the Isle ofRum to a more natural state would, indeed, appear tobe a worthy objective. It could provide a rare examplewith which to compare more modi®ed ecosystems andwould also satisfy at least one major requirement of theconcept of biosphere reserves (Goodier and Je�ers,1981). However, it is not easy to equate many of themanagement activities over the past 40 years with thisparticular conservation objective. Nor is it certain thatthe restoration of nature in parts of Rum could be suc-cessfully accommodated alongside active conservationmanagement, that seeks to retain plant and animalassemblages, because they have high biodiversity or canrepresent derived communities that are becoming increas-ingly rare in Europe, even though they are probably notnative to this island.There also remains the dilemma as to whether or not

red deer should form part of a more natural state for theIsle of Rum, although the long-term vision of managersincludes both woodlands and deer on the island. Tomaintain this ``balance'' will require a regular cull ofdeer. That may be more di�cult to achieve in woodlandthan on the open hill. It will certainly require substantialexpenditure on trained stalkers. If the deer carcases areto be utilised, as tradition in Scotland dictates, there willalso be a need for ponies, vehicles, roads, larder andshipping facilities. It is di�cult to incorporate theshooting of deer as a part of an ecosystem in its naturalstate. There will also be a continuing need to restrictaccess to the nature reserve whilst culling is in progress.Few would disagree that the ®rst goal of conservation

should be to prevent the loss of wild species. To do this,it is also necessary to protect the places where thesespecies occur. Because Europe has a long history ofhuman settlement and land-use, the wild species that wevalue are now principally to be found within ecosystemsthat have resulted from commercial uses of the land. Bystudying the ecological requirements of each species andby an understanding of the ways in which land wasmanaged in the past, we have acquired the ability toundertake conservation management that can favourthe wild species we particularly wish to preserve. Theknowledge that has been acquired during conservationmanagement of the Isle of Rum over the past 40 yearshas played a part in enhancing our ability to prevent theloss of species through the application of suitable man-agement practices. However, I believe that it is alsoimportant to ask if the preservation of wild species withinmanaged ecosystems is a su�cient goal for Europeanconservation agencies.The wild species that we value have evolved in largely

unmanaged, natural ecosystems where processes beyondthe control of humanswere responsible for the distribution

and abundance of wildlife and unpredictable temporalchanges to both of these. If natural processes are largelyreplaced by conservation management, we will not onlydetermine which species and how many will occurwithin protected areas but will also inevitably in¯uencetheir future evolution. However, conservation manage-ment that is derived from extensive experience andexperimentation has the security of generating pre-dictable outcomes. A reliance on nature is often far lesspredictable and could risk the loss, at least locally, of spe-cies and communities that are highly valued. Even so, ourundertanding of nature will be incomplete if we have nonatural systems to learn from. Moreover, it could also beargued that nature conservation itself is incomplete if itsucceeds in preserving species alone, without examples ofthe natural systems from which these species were derivedand which helped to shape the species themselves.The conservation management of the Isle of Rum

NNR has, for the past 40 years, been characterised by adesire to maximise the bene®ts which this reserve maygenerate. The high cost of maintaining Rum as areserve, the aspirations of commercial land-users andthe stringencies imposed by successive governmentshave only reinforced this desire. So as to quicklydemonstrate the potential bene®ts of conservationmanagement, intensive, commercial techniques havesometimes been applied. There is a danger that mis-understandings about the natural state of the island, anunwillingness to let artefacts of past management dis-appear, and the adoption of conservation legislationthat, when interpreted in certain ways, can lead to thepreservation of features and a supression of naturalchange, will severely limit future possibilities. The out-come will be a further separation between those featuresthat are preserved and the natural state from which theywere originally derived. If this pattern of conservationmanagement is widely adopted, future generations inEurope will ®nd it ever harder to appreciate that naturalsystems evolved and can be self-sustaining without theneed for continual management.Whilst seeking to generate new knowledge that can be

applied to resolve management problems both on thisisland and elsewhere in Europe, managers and policymakers have also expressed a wish to return Rum to amore natural state. However, if direct manipulation ofwildlife is inimical to the restoration of unmanaged,natural conditions, it is perhaps rather ambitious toattempt to achieve both within a single site. The re-creation of conditions that prevailed before the island'secosystems were modi®ed by humans requires that, ®rst,there must be a clear and accurate vision of what thoseconditions were. Then, secondly, it requires the removalof all species and activities that can be shown to havederived from human actions. Third, it is vital thatmanagers subsequently refrain from interfering with theways in which the system develops. Fourth, they must

102 B. Wood / Biological Conservation 94 (2000) 93±105

accept that continuing change will be an inevitable out-come of a return to a more natural state and that this statewill, consequently, not be static or readily predictable.It should be possible to mimic either of the two,

alternative, visions of the natural state of Rum, thathave been outlined above. In the absence of deer anddomestic herbivores it would probably take severalcenturies for extensive woodlands to develop. It is likelythat these would never be as widespread and luxuriantas those that existed over 4000 years ago, because sig-ni®cant changes in climate and soil have occurred sincethen. However, some woodland would regrow quiterapidly: at least one experimental exclosure now hasnative trees >4 m tall. These have developed in <40years without the need for any planting. It is almostcertain that plant species will have been introduced toRum in the past, some deliberately by people, othersinadvertantly in animal fodder, clothing and by othermeans. Some plant species will also have been lost.However, it is unlikely that a change in the present graz-ing regime on Rum would totally eliminate species thatoccur there now. If herbivores were removed, the greatestchanges that would occur would be in the distributionand abundance of plant species.So long as deer numbers are controlled by conserva-

tion managers on Rum, the survivors will never experi-ence the strong selection pressure exerted by very highlevels of intraspeci®c competition. Consequently, theyare unlikely to become dwarfed in response to a need toincrease their thermodynamic e�ciency. Scottish reddeer are small compared to those in mainland Europeand some further decline in size may occur if the qualityof herbage on Rum deteriorates. However, it wouldprobably require several thousand years of total non-interference with deer populations, for them to showgenetic, as well as phenotypic changes commensuratewith declining body mass. Thus, native woodland mayrequire less time to reinstate than would a unique islandpopulation of deer.Of course, both of these two alternatives could be

created side-by-side. However, this would require thetwo to be completely and permanently separated (byfencing). Such a management strategy would be expen-sive to ful®l and would inevitably divide the island in avery arti®cial way. Moreover, the subdivision of theisland, so as to accommodate multiple managementobjectives, would be the direct antithesis of one of theprincipal goals of ecosystem management, as it is beingapplied to extensive and still quite natural ecosystems inNorth America. Europe has long been divided into apatchwork of di�erent land uses. This presents a frag-mented ecological landscape that is very di�erent to theone originally occupied by most of our wild species. If itis eventually demonstrated that deer and woodlandscannot successfully coexist on Rum without permanentseparation or continual culling of the deer, what is likely

to be generated by conservation managers will at best,be little more than a ``cameo'' of the features that arenow found throughout the Scottish Highlands.

6.3. Implications for conservation in Europe

Particularly in Britain, but also in many other parts ofEurope, nature reserves have been selected because theycontain species and communities that were formerlyquite widespread but have now been largely eliminatedby commercial land uses (Ratcli�e, 1977). These formoften the last refuges for the remnants of nature inEurope. So long as protected areas remain small andunconnected it will be essential to manage them in orderto preserve these remnants. In recent years, much e�orthas gone into developing and re®ning managementtechniques that can produce the desired results (Suther-land and Hill, 1995). This attitude to conservationseems also to have permeated the minds of some con-servation practitioners (Bibelreither, 1998), so that theyare willing to apply quite intensive management to eventhe largest of Europe's protected areas.The adoption of the Habitats Directive (Council

Directive on the Conservation of Natural Habitats andWild Fauna and Flora Ð 92/43/EEC) by member statesof the European Union requires the preservation ofparticular habitats that are examples of those peculiarto Europe (Hopkins, 1995). Rum has been selected as acandidate SAC for ®ve habitats of special importance.These include grasslands whose diversity is at least par-tially maintained by grazing herbivores and which aresusceptible to invasion by trees. On Rum, it will benecessary to retain Highland cattle and perhaps alsointroduce sheep to the island once again, if some ofthese preferred grasslands are to be promoted. Treegrowth will have to be prevented in such areas. It is alsoadvocated that trees will not be tolerated on sites whichmust be kept open as cultural mementoes (ScottishNatural Heritage, 1998). The promotion of goals suchas these will inevitably discourage a return to a naturalstate.For Europe as a whole, an optimistic long-term vision

would be that species and habitats which are now pre-served in managed reserves will eventually be utilised torestore nature in much bigger sites than we can pre-sently a�ord to set aside from commercial land-uses.Perhaps, because their ®rst priority is the perpetuationof declining species, it appears that rather few con-servation managers are willing to look beyond the spe-cies themselves and the communities which theycollectively form. Whilst acknowledging that certainnatural processes are vital to the creation of valuedconservation features, many managers seem reluctant toallow su�cient scope for these to operate. Instead, theyhave a need to tightly control nature in order not to losewhat they identify as valuable. At the same time, this

B. Wood / Biological Conservation 94 (2000) 93±105 103

approach to conservation is in danger of losing sight ofthe natural processes and unpredictability that arefound in wild ecosystems, which are essential parts ofnature and the basis of its continuing evolution.There are additional potential disadvantages of the

controlled management of wildlife features. In utilisingquite intensive techniques to preserve species and gen-erate habitats which it is feared may otherwise be lostforever, conservation managers can give the impressionthat nature needs to always be managed. In turn, thisundervalues the often unpredictable outcome of longperiods of natural development of ecosystems. It mayalso hasten the elimination of the last remnants of nat-ure in Europe, that have survived despite the intensiveattention of mankind over several millenia. If somethingthat closely resembles natural systems can, apparently,be created by appropriate management, it is less easy toresist developments that may signi®cantly harm theremaining natural areas.

Acknowledgements

I wish to thank the sta� of Rum NNR who havehelped in many ways during my visits to the island,especially Ray Collier, Martin Curry, Chris Eatoughand the late Wilf Nelson. My colleagues Barrie Gold-smith and Max Hooper, together with students from theMSc Conservation Course, shared in fruitful discussionsabout the management of Rum. The comments ofreferees and the journal editor encouraged me to moreprecisely de®ne the concerns that I have addressed inthis paper.

References

Agee, J.K., 1996. Ecosystem management: an appropriate concept for

parks? In: Wright, R.G. (Ed.), National Parks and Protected Areas:

Their Role in Environmental Protection. Blackwell Science, Cam-

bridge, MA, pp. 31±44.

Bakker, J.P., de Bie, S., Dallinga, J.H., Tjaden, P., de Vries, M.F.W.,

1983. Sheep grazing as a management tool for heathland conserva-

tion and regeneration in the Netherlands. Journal of Applied Ecol-

ogy 20, 541±560.

Ball, M.E. 1970. Isle of Rhum NNR: Plan of Operations 1970±74. The

Nature Conservancy.

Ball, M.E., 1974. Floristic changes on grasslands and heaths on the

Isle of Rhum after a reduction or exclusion of grazing. Journal of

Environmental Management 2, 299±318.

Ball, M.E. 1987. Botany, woodland and forestry. In: Clutton-Brock,

T.H., Ball, M.E. (Eds.), Rhum: The Natural History of an Island.

Edinburgh University Press, Edinburgh, pp. 43±62.

Batten, L.A., Pomeroy, D.E., 1969. E�ects of rea�orestation on the

birds of Rhum, Scotland. Bird Study 16, 13±16.

Berry, R.J., Evans, I.M., Sennitt, B.F.C., 1967. The relationships and

ecology of Apodemus sylvaticus from the Small Isles of the Inner

Hebrides, Scotland. Journal of Zoology (London) 152, 333±346.

Bibelriether, H., 1998. Natural heritage conservation in Europe: a

review. In: Synage, H. (Ed.), Parks for Life: Proceedings of the

IUCN/WCPA European Regional Working Session on Protecting

Europe's Natural Heritage. IUCN, Switzerland, pp. 33±36.

Bignal, E.M., McCracken, D.I., 1996. Low-intensity farming systems

in the conservation of the countryside. Journal of Applied Ecology

33, 413±424.

Boyd, I.L., 1980. Population changes and the distribution of a herd of

feral goats (Capra sp.) on Rhum, Inner Hebrides, 1960±78. Journal

of Zoology (London) 193, 287±304.

Boyd, J.M., 1977. Management Plan for the Isle of Rhum NNR 1977±

82. Nature Conservancy Council.

Boyd, J.M., Boyd, I.L., 1990. The Hebrides. Collins, London.

Burt, W.H., Grossenheider, R.P., 1952. A Field Guide to the Mam-

mals. Houghton Mi�in, Boston.

Charles, W.N., McKown, D., East, K., 1977. Selection of upland

swards by red deer (Cervus elaphus) on Rhum. Journal of Applied

Ecology 14, 55±64.

Clutton-Brock, T.H., Albon, S.D., 1989. Red Deer in the Highlands.

BSP Professional Books, London.

Clutton-Brock, T.H., Albon, S.D., 1992. Trial and error in the High-

lands. Nature 358, 11±12.

Clutton-Brock, T.H., Lonergan, M.E., 1994. Culling regimes and sex

ratio biases in Highland red deer. Journal of Applied Ecology 31,

521±527.

Clutton-Brock, T.H., Guinness, F.E., Albon, S.D., 1982. Red Deer:

Behaviour and Ecology of Two Sexes. Edinburgh University Press,

Edinburgh.

Corbet, G.B., Harris, S., 1991. The Handbook of British Mammals,

3rd Edition. Blackwell, Oxford.

Damuth, J., 1993. Cope's rule, the island rule and the scaling of

mammalian population density. Nature 365, 748±750.

Dempster, J.P., 1977. The scienti®c basis of practical conservation:

factors limiting the persistence of populations and communities of

animals and plants. Proceedings of the Royal Society of London B

197, 69±76.

Eggeling, W.J., 1964. A nature reserve management plan for the Island

of Rhum, Inner Hebrides. Journal of Applied Ecology 1, 405±419.

Evans, D., 1997. A History of Nature Conservation in Britain, 2nd

Edition. Routledge, London.

Goodier, R., Je�ers, J.N.R., 1981. Biosphere reserves. Advances in

Applied Biology 6, 279±317.

Gordon, I.J., 1988. Facilitation of red deer grazing by cattle and its

impact on red deer performance. Journal of Applied Ecology 25, 1±

10.

Gordon, I.J., 1989a. Vegetation community selection by ungulates on

the Isle of Rhum. I. Food supply. Journal of Applied Ecology 26,

35±51.

Gordon, I.J., 1989b. Vegetation community selection by ungulates on

the Isle of Rhum. II. Vegetation community selection. Journal of

Applied Ecology 26, 53±64.

Gordon, I.J., 1989c. Vegetation community selection by ungulates on

the Isle of Rhum. III. Determinants of vegetation community selec-

tion. Journal of Applied Ecology 26, 65±79.

Henderson, N., 1992. Wilderness and the nature conservation ideal:

Britain, Canada, and the United States contrasted. Ambio 21, 394±

399.

Hirons, K.R., Edwards, K.J., 1990. Pollen and related studies at Kin-

loch, Isle of Rhum, Scotland, with particular reference to possible

early human impacts on vegetation. New Phytologist 116, 715±727.

Hopkins, J.J., 1995. The Habitats Directive Ð selecting the UK sites.

British Wildlife 6 (5), 297±306.

Juniper, T., 1994. Comment: conservation on the global stage Ð the

Habitats Directive, the Biodiversity Convention and the UK. British

Wildlife 6 (2), 99±105.

Lister, A.M., 1989. Rapid dwar®ng of red deer on Jersey in the Last

Interglacial. Nature 342, 539±542.

Lister, A.M., 1995. Sea-levels and the evolution of island endemics: the

dwarf red deer of Jersey. In: Preece, R.C. (Ed.), Island Britain: A

104 B. Wood / Biological Conservation 94 (2000) 93±105

Quaternary Perspective. Geological Society Special Publication No.

96, pp. 151±172.

Love, J.A., 1987. Rhum's human history. In: Clutton-Brock, T.H.,

Ball, M.E. (Eds.), Rhum: The Natural History of an Island, Edin-

burgh University Press, Edinburgh, pp. 27±42.

Love, J.A., Wormell, P., 1987. The birds of Rhum. In: Clutton-Brock,

T.H., Ball, M.E. (Eds.), Rhum: The Natural History of an Island,

Edinburgh University Press, Edinburgh, pp. 78±94.

Lowe, V.P.W., 1966. Observations on the dispersal of red deer on

Rhum. Symposia of the Zoological Society of London 18, 211±228.

Lowe, V.P.W., 1969. Population dynamics of the red deer (Cervus

elaphus) on Rhum. Journal of Animal Ecology 38, 425±457.

Lowe, V.P.W., 1971. Some e�ects of a change in estate management

on a deer population. In: Du�ey, E., Watt, A.S. (Eds.), The Scien-

ti®c Management of Animal and Plant Communities for Conserva-

tion, 11th Symposium of the British Ecological Society. Blackwell

Scienti®c Publications, Oxford, pp. 437±456.

Magnusson, M., 1997. Rum: Nature's Island. Luath Press, Edinburgh.

McKie, R., 1987. Gamekeeper turned housekeeper. New Scientist

1579, 79±80.

McVean, D.N., Lockie, J.D., 1969. Ecology and Land Use in Upland

Scotland. Edinburgh University Press, Edinburgh.

Monro, D., 1934. A Description of the Western Isles of Scotland

called Hybrides. Stirling (1549) (new edition).

Nature Conservancy Council, 1974. Isle of Rhum National Nature

Reserve: Reserve Handbook.

Nature Conservancy Council, 1987. Isle of Rhum NNR: Management

Plan 1987±96.

Nevard, T.D., Penfold, J.B., 1978. Wildlife conservation in Britain: the

unsatis®ed demand. Biological Conservation 14, 25±44.

Newmark, W.D., 1987. A land-bridge island perspective on mamma-

lian extinctions in western North American parks. Nature 325, 430±

432.

Peterken, G.F., 1996. Natural Woodland: Ecology and Conservation

in Northern Temperate Regions. Cambridge University Press,

Cambridge.

Rackham, O., 1994. The Illustrated History of the Countryside. Wei-

denfeld and Nicholson, London.

Ratcli�e, D.A., 1977. A Nature Conservation Review: The Selection

of Biological Sites of National Importance to Nature Conservation

in Britain, 2 vols. Cambridge University Press, Cambridge.

Ritchie, W., 1997. Rum Ð building on the dream. Scotland's Natural

Heritage 11, 4±5.

Scottish Natural Heritage, 1998. Isle of Rum National Nature

Reserve: Draft Management Plan 1998±2008.

Shafer, C.L., 1991. Nature Reserves: Island Theory and Conservation

Practice. Smithsonian Institute Press, Washington, DC.

Spinney, L., 1995. Return to the wild. New Scientist 199, 35±38.

Sutherland, W.J., Hill, D.A. (Eds.), 1995. Managing Habitats for

Conservation. Cambridge University Press. Cambridge.

Vries, M.F.W.de., 1995. Large herbivores and the design of large-scale

nature reserves in Western Europe. Conservation Biology 9, 25±33.

Watson, D., 1989. Red deer in Scotland Ð a resource out of control.

Ecos 10 (3), 22±24.

Wickham-Jones, C.R., Pollock, D., 1987. Excavations at farm ®elds,

Kinloch, Rhum 1984-85: a preliminary report. Glasgow Archae-

ological Journal 12, 19±29.

Wieren, S.E.van., 1995. The potential of large herbivores in nature

conservation and extensive land use in Europe. Biological Journal

of the Linnean Society 56 (supplement), 11±23.

Wormell, P., 1968. Establishing woodland on the Isle of Rhum. Jour-

nal of the Royal Scottish Forestry Society 22, 207±220.

Wormell, P., 1976. The manx shearwaters of Rhum. Scottish Birds 9,

103±118.

Wood, B., 1996. Species and habitats: room for more perceptive

management. In: Lawson, T.G., Reed, T.M. (Eds.), Nature

reserves Ð who needs them? Joint Nature Conservation Committee,

Peterborough, pp. 47±55.

Yalden, D.W., 1982. When did the mammal fauna of the British Isles

arrive? Mammal Review 12 (1), 13±56.

Yalden, D.W., 1986. Opportunities for reintroducing British mam-

mals. Mammal Review 16, 53±63.

B. Wood / Biological Conservation 94 (2000) 93±105 105