Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Holocene climate changes and warm man‐made refugia may explain why a sixth of British butterflies possess unnatural early‐successional habitats

Holocene climate changes and warm man‐made refugia may explain why a sixth of British butterflies... Analyses of their habitats indicate that 18% of British butterfly species are restricted to the earliest serai stages of ecosystems, whereas the same species occupy later serai stages in central Europe, where spring and summer temperatures are warmer. The microclimates of their British habitats are exceptionally warm, compensating for the cooler climate. Most of these British habitats are also ephemeral, and have long depended on man for their creation and regeneration This poses the question of where these species lived before man created their habitats, roughly 6000 BP, I suggest that they are relics from a period when British summers were warmer than today, and that they avoided extinction when the climate cooled hy moving into warm refugia created hy prehistoric man within three types of ecosystem If summer temperatures become warmer, these species should return to later serai stages that are commoner and less dependent on man. Introduction Most of the 55 non-migratory species of butterfly that breed in Britain have abrupt northern limits to their ranges which correlate closely with summer isotherms (Dennis 1977, Heath et al 1984, Dennis and Wilhams 1986, Turner et al 1987) Within these ranges, c 80% ot species live in predominantly closed populations, each restricted m its local distribution by the distribution of Its larval foodplant(s) (Thomas 1984) However, few species occupy more than a small fraction of the sites that fulfil these basic requirements (Heath et al 1984, Thomas 1984) Research on 43% of Bntish butterfly species has shown that, with three exceptions, this IS because temales are selective during oviposition, and restrict egglaying to foodplants of a particular growthform or micro-habitat, species are confined to localities where plants of the required growth-forms exist m sufficient abundance, and continuity, for a population to survive (Thomas 1991) The narrowest and most ephemeral niches recorded among Bntish butterflies belong to the 10 (18% of) species that breed near the ground in the earliest stages of seres (Thomas 1991) Grassland examples include Macuhnea anon L (which is an obligate parasite of the ant Myrmtca sabulett Meinert) (Thomas 1991), Lxsandra bellargus Rott (Thomas 1983), Hesperta comma L (Thomas et al 1986), Melitaea ctnxta L (Thomas and Simcox 1982, Thomas 1991) and Htpparchta semele L (Heath et al 1984) Boloria euphrosyne L , B selene Denis & Schiff , Argynnts adippe Denis & Schiff and Mellicta athalta Rott also breed in the earliest serai stages of British grassland, but most populations occur in recent woodland clearings (Warren 1987a, Warren and Thomas 1992) At the northern limit of its range, on Welsh heathland, Plebejus argus L is restncted to patches where the vegetation has been burnt or cleared in the previous five years (Thomas, C D 1983, 1985) The ephemeral British habitats of seven of these species depend entirely on man for their regeneration, whereas a minority of populations of P argus and H semele also breed in "natural" pnmary habitats, such as dunes and undercliffs, Melitaea ctnxta is wholly dependent on the latter in Britain (Heath et al 1984, Thomas 1991) This IS no recent phenomenon since reeords began, most or all populations of these species (apart from M cinxta) bred on land that was being managed by the traditional methods that have prevailed on British heaths, grasslands and woods from roughly 6000 BP until the present century (Harris 1766, Barrett 1893, Heath et al 1984, Thomas 1984, 1991, Rackham 1986, Warren and Key 1991) indeed, it is because modern forms of management have reduced both the abundance and continuity of early serai stages within all three ecosystems, that there have been exceptionally severe recent declines of the nine species that inhabit them (Heath et al 1984) populations typically survive 2-10 ECOGRAPHY yr after traditional management is abandoned (Thomas 1984, 1991) This raises the question as to where these butterflies lived between c 95CX) BP, when modern insects reinvaded Britain after the last ice age (Osborne 1976, Dennis 1977, 1992) and c 6000 BP, when man first created their early serai habitats (Rackham 1986) It is a question relevant to other taxa in Britain, and perhaps elsewhere in northern Europe The suggestion that British butterfly populations are more recent colonists from southern Europe is untenable, due to their distribution pattern on the remoter islands, and to the subspeciation and other adaptations that have evolved (Dennis 1977, 1992) Moreover all seven species in this group whose dispersal has been studied were found to be unusually sedentary (Thomas and Simcox 1982, Thomas 1983, 1991, Thomas, C D 1983, 1985, Thomas et al 1986, Warren 1987b, Thomas and Snazell 1989, 1985, Thomas, C D etal 1992),itisextremely unlikely that they could have crossed the English Channel after the land hnk was broken, roughly 1000 yr before man altered the British landscape The sedentariness and population structure of these butterflies also make it improbable that even the four species that inhabit modern woodland could have utilised their narrow niches when these occasionally existed, for periods of 1 to 10 yr, in small isolated patches in the ancient wildwood following the death of a large tree (Warren 1987 a,b, Thomas and Snazell 1989) I drew attention to this conundrum in a review of the conservation of European butterflies (Thomas 1991), and speculated that these species might once have inhabited broader or different niches to those of today In this paper I bring together the published evidence supporting this idea, and present new data on lntraspecific variation in niche breadth, and on the warm microclimates inhabited by the young stages of some butterflies I also extend my hypothesis, by suggesting how these species may have survived m Britain when summer temperatures became cooler, after the land link with the continent was broken height on Old Winchester Hill, Hampshire, S England Annual counts were made on 30 randomly chosen Hippocrepis comosa plants m each of 15 adjoining compartments that were grazed on rotation in 1982-86 Plant height (Thomas 1983) and the number of eggs was recorded in each of the 2250 samples Variation in the abundance of the ant Myrmica sabuleti was recorded by baiting or by pit-fall trapping beneath Thymus spp and Origanum vulgare L (the early larval foodplants of Maculinea anon), in swards of varying height on land with varying aspects Ants were sampled during the peak evening foraging period of M sabuleti beneath plants containing M arton eggs, or beneath suitable oviposition plants on extinct sites Pooled data are presented from one occupied and six extinct M anon sites in SW England (N = 347), from SX occupied and two unoccupied sites in Dordogne, I France (N = 227), and from five occupied sites in Oland, Sweden (N = 97) Munguira (pers comm ) descnbed the sward heights and aspects inhabited by L bellargus and H comma in Spain and south France, and I recorded the occupancy of shaded woods in Dordogne and the Rhone Valley, France by the four species of fntillary Microclimate measurements I measured temperature in the niches occupied by larvae, and in similar situations beneath foodplants rejected for oviposition, in a range of turf heights in south-facing grassland in south England Measurements were made in cloudy-bright weather in Apnl and May, using a thermocouple, for L bellargus (Thomas 1983. N = 67) and M anon (new data, N = 143), and in October using a laser thermometer for H comma (new data, N = 58) Thomas, C D (1983) used the former technique for P argus on south-facing heathland in north Wales Integrated temperatures were measured m the niches of B euphrosyne (N = 16 x 12) and B selene (N = 16 X 12), dunng the 12 wk feeding period of their larvae, using Pallman's sucrose inversion method (Lee 1969) Measurements were made m woodland blocks m S England that had been eut 3 months (B euphrosyne) or 18 months (B selene) earher Integrated temperatures were also measured beneath unsuitable Viola (N = 32 x 12) growing under shrubs that had not been cut for c 15 Methods and sources Niche breadth of butterflies and ants at different latitudes Precise definitions of the current British habitats of seven species were obtained from publications, theses or conservation reports Hesperia comma (Thomas et al 1986), Melitaea cinxia (Thomas and Simcox 1982, Thomas 1991), Boloria euphrosyne, B selene, and y4rgynnis adtppe (Thomas et al in press), Mellicta athalta (Warren 1987a), Plebejus argus in north Wales (Thomas, C D 1983) and south England (Read 1986) In addition, the distribution of 2nd generation Lysandra bellargus eggs was recorded in swards of varying ECOGRAPHY 16 3 (1993) Climate change and land management in the Holocene A simplified reconstruction of trends in mean summer temperatures m Britain over the past 8000 yr was made from the evidence of Sawyer 1966, Pearson 1978, Wigley et al 1986, Grove 1988, Huntley and Prentice Results and discussion Intraspecific variation in habitats at different latitudes In 1972-92, British populations oi Maculinea anon, Lysandra bellargus, Hespena comma, and Melttaea ctnxia occurred only on southern south-facing hillsides, and there only on sites where many foodplants grew in turf that was cropped to < 3 cm tall in the cases of M anon (Fig la) and L bellargus (Fig lb), or that was in the first one to three years of a succession from bare —r —\—1—1—1—1—1 1 2 3 4 5-9 10- 2 1 - 3 1 - 4 1 - > 5 0 20 30 40 5 0 1—r- — 1 — 1 " 28 26 - \ O24 Ny •••••-• ^20 VH»., •• \ •• +-: 10 2 3 4 5 6 Sward height (cm) 8 9 >9 8 1 1 1 1 1 1 1 I 1 1 1 2 3 4 5 6 7 8 9 10 11 V e g e t a t i o n height Ccm] Fig (la) Variation in the abundance of Myrmica sabuleti (the obhgate host of Macutinea anon) beneath Thymus and Origanum spp (the early larval foodplants) in swards of varying height at different latitudes Over 51% of foodplants must coincide with M sahuleti to support a population of M anon (Thomas 1991) (•—•) SW England. (D- -D) Dordogne. France Fig (lb) Distribution of Lysandra beltargus eggs {0—•. N = 2372) compared with that of its foodplant Hippocrepis comosa (D- -D N = 2250) m swards of varying height within one site in Hampshire. S England 1988. Wigley 1988, Wigley and Kelly 1990, Folland et al 1990 The timing and extent to which prehistoric man cleared woods and managed land to create modern habitats was reconstructed from Smith (1980, calcareous grassland), Rackham (1986, woods and most biotopes) and Webb (1986, heathland) May June July Fig 2 Microclimate measurements of the niches occupied by larvae of six species of butterfly in Bntish eeosystems (a) South-facing southern grassland, (—) Maculinea anon. {—) Lysandra bettargus, { ) Hespena comma, south-facing north Welsh heathland ( ) Ptebejus argus (from Thomas. C D 1983) Bold lines and symbols indicate the turf range ( = niche) inhabited by each butterfly, open circles indicate air temperature when measurements were made For clarity. SD included for M anon only (b) Integrated temperatures of the mam niches of Boloria euphrosyne ( • — • ) and B selene (A- -A), and of unsuitable shady Viola beneath tall shrubs (•- - • ) For clarity, 95% confidence limits omitted for B selene habitat ECOGRAPHY 16 3 (1993) hottest niches currently available in British grassland (Fig 2a), yet occupy 20 cm to 40 em tall swards on slopes of all aspects in lowland central France and Spam (Munguira pers comm ) And in the extreme south of England, P argus breeds in both the pioneer and building phases of heathland, again on all aspects (Read 1986), compared with its restriction to the hottest patches of south-facing pioneer heathland that exist in north Wales (Thomas, C D 1983, Fig 2a), where summer temperatures are c 2°C cooler In the woodland ecosystem, Bolorta euphrosyne feeds on Viola spp in only the first five years after a clearance, and B selene, A adtppe and M athalta are shaded out before 10 years' regrowth, although their foodplants persist for much longer (Warren 1987a, Warren and Thomas 1992, Thomas et al in press) These also represent the warmest microclimates available on British woodland floors (Fig 2b), whereas all four species occur in semi-shaded woodland in central France Past summer climates and the availability of anthropogenic habitats Published evidence of climatic change over the past 8000 yr is conflicting, but not in details that affect my Fig 3 July isotherms (1931-60) showing 3°C warmer and 3°C hypothesis There is a consensus that 8000-6000 BP was cooler mean temperatures than in southern England, where a period of relative warmth (Wigley et al 1986, Folland most butterfly habitat and microclimate measurements were made ( • ) English M anon sites, ( • ) English L bellargus, et al 1990), when average British summer temperatures (O) English Boloria spp, (A) north Welsh P argus, (A) south were up to 2°C higher than today (although winters may English P argus, (D) study area in Dordogne Isotherms from have been colder), especially towards the end of this Anonymous (1952), Arlery (1970), and Wallen (1970) penod (Huntley and Prentice 1988, Wigley and Kelly 1990, Folland et al 1990) This optimum may have lasted until 5300 or 5000 BP, before the summer climate ground, in the cases of H comma (Thomas et al 1986), cooled to something similar to or colder than that of M ctnxta (Thomas and Simcox 1982, Thomas 1991) and today (Folland et al 1990) Since then there may have Htpparchia semele (Heath et al 1984) been nine century-timescale periods, including the LittThe niches occupied by these species represent the warmest microclimates that are available in British grassland (Fig 2a) For example, an increase in mean turf height from 2 cm to 5 cm on British M arton sites ^ ^ anthropogenic creation of downs, causes the ground to become c 3°C cooler (Fig 2a), 5_ heaths and early and results in a rapid decline of Myrmica sabuleti folserai stages lowed by the local extinction of the butterfly (Thomas 1984, 1991) However in Dordogne, where mean summer tempeiatures are currently c 3°C warmer than in south Britain (Fig 3), M sabulett is abundant only on north-facing slopes and on flat land where the sward is tall (Fig la) Thus in central Europe, M anon is confined to a locally cool, late serai stage withm grassland this IS more widespread, less ephemeral, and much less dependent on man than its habitat in modern Bntain In Oland, where mean July temperatures are 1°C-2°C '000 Years before present warmer than in south Bntain, M sabulett is abundant enough to support M anon on flat terrain and south- Fig 4 Approximate changes in mean Bntish summer temperatures, the date when the land-link with Europe was broken, facing slopes, in swards ranging from 3 em to 10 cm tall and dates when man created downland, heathland, and early This too IS a much more widespread form of grassland serai stages withm both these and woodland Vertical scale for than the British habitat of M anon habitat creation is approximate and unquantifiable Sources of Lysandra bellargus and H comma also breed in the data are given under "methods" ECOGRAPHY 16 3 (1993) le Ice Age, when mean summer temperatures were up to 0 9°C cooler than today (Grove 1988, Wigley 1988, Wigley and Kelly 1990) In intervening centuries, summer temperatures were similar to or temporarily warmer than today, the main warm period being the Medieval Warm Epoch of 1100-700 BP A simplified reconstruction of these changes is given in Fig 4, omitting some cool intervals postulated by Wigley and Kelly (1990), which do not affect my hypothesis, and showing in dashed lines the periods for which evidence is conflicting or obscure Figure 4 also shows the approximate dates when the land link between Britain and the continent disappeared, and when man created warm early serai habitats within three ecosystems Hypothesis for the survival of thermophilous butterflies since 6000 BP I suggest that the ten species of butterfly that currently depend on early successional habitats in Bntain are relics from a period when summer temperatures were warmer (Fig 4), and that their pnmary habitats before 5000-6000 BP were m later, longer-lived successional stages, corresponding to their modern habitats in Oland or central Europe (e g Fig la) These habitats were probably quite common m pnmaeval Bntish woodland, given the extensive clearings made by beavers (Warren and Key 1991), and the presence of herbivores to browse gaps (Rackham 1986) Moreover, most of such permanent grassland and heathland as existed would have been habitable under a warmer climate, in contrast to the very small proportion that is south-facing, southerly and close-cropped, and hence warm enough today It follows that all except P argus, H semele and M cinxia of this group should have become extinct roughly 5000 yr ago, when the climate entered the first extended cool period after the land hnk with the continent was broken However, by then man had been making significant woodland clearings for at least a thousand years (Fig 4), creating heaths and grasslands and, through domestic grazing, early successional habitats within these, as well as plots of freshly cut woodland that were easy to crop for fuel (Smith 1980, Rackham 1986, Webb 1986) The warm microclimates that these inevitably contained (Fig 2) should have more than compensated for the cooling of the climate I suggest that these seven (13%) of British butterfly species survived by moving into warm man-made refugia, where they lived until the next warm period allowed them temporarily to reoccupy later serai stages, and perhaps to spread into warm anthropogenic habitats further north By the same reasoning, these species were probably more restricted than today dunng the coldest periods, such as the Little Ice Age Maculinea anon was perhaps confined to southern south-facing sites with exceptionally short (0 5 cm - 1 cm) swards, where temperatures at the soil surface are 1°C - 4°C higher than under the longest (2 cm) grassland that can support M anon today (Fig 2a) There is no fossil record of butterflies with which to test this theory, but data for Coleoptera are broadly consistent with it Around 10000 BP, the Bntish Midlands experienced an abrupt extinction of species now typical of northern latitudes or high ground, and by 9500-9000 BP had been colonised by assemblages similar to those of southern Bntain today, plus a few thermophilous species now restricted to the continent (Osborne 1976) Twenty-three of these species have since disappeared, most around 3000 BP, insofar as this can be dated (Osborne 1976, Girling 1984) Most extinctions were of arboreal species, '"many of which are now only found further south on the continent" (Osborne 1976), both Girling (1984) and Osborne (1976) implicate forest clearance and climatic cooling as contributory factors My theory for butterflies also implies a disproportionately high extinction rate among thermophilous arboreal insects after 5000 BP, because the warm microhabltats created by traditional land management exist mainly at ground level, not in shrubs and trees It is also unlikely that prehistoric man created warm habitats within aquatic systems, in which a thermophilous Gyrinid became extinct during the Little Ice Age (Girling 1984) On the other hand, the prehistoric extinction of a thermophilous ground-living species of dung beetle is inconsistent with my theory (Osborne 1976) Stronger support comes from the sub-fossil record of one of the few "arboreal" beetles to have warm habitats provided by man (Buckland 1975) Xestobium rufovillosum Deg oviposits in the fungal-infested timbers of dwellings, as well as m dead wood m the wild Wild British populations existed well north of their current range during the early Medieval Warm Epoch, but became confined to the Midlands and south England during the Little Ice Age and subsequently Yet X rufovillosum survived locally throughout its former northern range "in the artificially warm habitats created by man" in urban centres, such as York (Buckland 1975) It is easy to envisage a slightly more thermophilous species than X rufovillosum disappearing completely in the wild when the climate cooled, yet surviving as a British species in southern cities Conservation I conclude that seven species of British butterfly avoided extinction by moving into warm man-made refugia c 5000 yr ago But extinction was merely postponed in the present cerrtury British farmers and foresters have largely abandoned traditional forms of management, in favour of regimes that produce shadier woods, tall swards in unfertilised grassland, and mature or degenerate heathland (Smith 1980, Thomas 1984, 1990, 1991, Webb 1986, Warren and Key 1991, Warren ECOGRAPHY 16 3 (1993) and Thomas 1992) This has resulted in cooler microclimates in the ground layers of most sites (Fig 2), and caused the extinction of most populations of thermophilous butterfly species (Heath et al 1984) Many climatologists predict that mean summer temperatures will rise over the next 50 yr, through the greenhouse effect or as a further recovery from the Little Ice Age If so, I predict that these rare British butterflies will once again expand into later successional niches These habitats are still locally common in seminatural ecosystems m northern Europe, and require less intensive conservation management This interpretation offers an incentive and justification for conservationists, who are sometimes criticised for preserving unnatural obsolete forms of land-use in ecosystems It is particularly important that populations of scarce thermophilous species are maintained in Britain under the present climate, because these will be the founders of most future expansions, only the most mobile species are likely to recolonise Britain from across the English Channel Acknowledgements-I XhdnkR G Snazell and 1 Moy for help with integrated temperature measurements on Boloria sites, D J Simcox for help with M anon measurements m Oland and C D Thomas for permission to quote unpublished data from his M Sc thesis R L H Dennis encouraged me to publish this idea and he, P C Buckland, A J Gray C D Thomas, M G Morns G W Elmes and M S Warren made helpful comments on the ms This research was partly funded by a NERC TIGER award http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Ecography Wiley

Holocene climate changes and warm man‐made refugia may explain why a sixth of British butterflies possess unnatural early‐successional habitats

Thomas, J. A
Ecography , Volume 16 (3) – Jul 1, 1993
Download PDF
8 pages

Loading next page...
 
/lp/wiley/holocene-climate-changes-and-warm-man-made-refugia-may-explain-why-a-o0xxULHGiA

References (27)

Publisher
Wiley
Copyright
Copyright © 1993 Wiley Subscription Services, Inc., A Wiley Company
ISSN
0906-7590
eISSN
1600-0587
DOI
10.1111/j.1600-0587.1993.tb00217.x
Publisher site
See Article on Publisher Site

Abstract

Analyses of their habitats indicate that 18% of British butterfly species are restricted to the earliest serai stages of ecosystems, whereas the same species occupy later serai stages in central Europe, where spring and summer temperatures are warmer. The microclimates of their British habitats are exceptionally warm, compensating for the cooler climate. Most of these British habitats are also ephemeral, and have long depended on man for their creation and regeneration This poses the question of where these species lived before man created their habitats, roughly 6000 BP, I suggest that they are relics from a period when British summers were warmer than today, and that they avoided extinction when the climate cooled hy moving into warm refugia created hy prehistoric man within three types of ecosystem If summer temperatures become warmer, these species should return to later serai stages that are commoner and less dependent on man. Introduction Most of the 55 non-migratory species of butterfly that breed in Britain have abrupt northern limits to their ranges which correlate closely with summer isotherms (Dennis 1977, Heath et al 1984, Dennis and Wilhams 1986, Turner et al 1987) Within these ranges, c 80% ot species live in predominantly closed populations, each restricted m its local distribution by the distribution of Its larval foodplant(s) (Thomas 1984) However, few species occupy more than a small fraction of the sites that fulfil these basic requirements (Heath et al 1984, Thomas 1984) Research on 43% of Bntish butterfly species has shown that, with three exceptions, this IS because temales are selective during oviposition, and restrict egglaying to foodplants of a particular growthform or micro-habitat, species are confined to localities where plants of the required growth-forms exist m sufficient abundance, and continuity, for a population to survive (Thomas 1991) The narrowest and most ephemeral niches recorded among Bntish butterflies belong to the 10 (18% of) species that breed near the ground in the earliest stages of seres (Thomas 1991) Grassland examples include Macuhnea anon L (which is an obligate parasite of the ant Myrmtca sabulett Meinert) (Thomas 1991), Lxsandra bellargus Rott (Thomas 1983), Hesperta comma L (Thomas et al 1986), Melitaea ctnxta L (Thomas and Simcox 1982, Thomas 1991) and Htpparchta semele L (Heath et al 1984) Boloria euphrosyne L , B selene Denis & Schiff , Argynnts adippe Denis & Schiff and Mellicta athalta Rott also breed in the earliest serai stages of British grassland, but most populations occur in recent woodland clearings (Warren 1987a, Warren and Thomas 1992) At the northern limit of its range, on Welsh heathland, Plebejus argus L is restncted to patches where the vegetation has been burnt or cleared in the previous five years (Thomas, C D 1983, 1985) The ephemeral British habitats of seven of these species depend entirely on man for their regeneration, whereas a minority of populations of P argus and H semele also breed in "natural" pnmary habitats, such as dunes and undercliffs, Melitaea ctnxta is wholly dependent on the latter in Britain (Heath et al 1984, Thomas 1991) This IS no recent phenomenon since reeords began, most or all populations of these species (apart from M cinxta) bred on land that was being managed by the traditional methods that have prevailed on British heaths, grasslands and woods from roughly 6000 BP until the present century (Harris 1766, Barrett 1893, Heath et al 1984, Thomas 1984, 1991, Rackham 1986, Warren and Key 1991) indeed, it is because modern forms of management have reduced both the abundance and continuity of early serai stages within all three ecosystems, that there have been exceptionally severe recent declines of the nine species that inhabit them (Heath et al 1984) populations typically survive 2-10 ECOGRAPHY yr after traditional management is abandoned (Thomas 1984, 1991) This raises the question as to where these butterflies lived between c 95CX) BP, when modern insects reinvaded Britain after the last ice age (Osborne 1976, Dennis 1977, 1992) and c 6000 BP, when man first created their early serai habitats (Rackham 1986) It is a question relevant to other taxa in Britain, and perhaps elsewhere in northern Europe The suggestion that British butterfly populations are more recent colonists from southern Europe is untenable, due to their distribution pattern on the remoter islands, and to the subspeciation and other adaptations that have evolved (Dennis 1977, 1992) Moreover all seven species in this group whose dispersal has been studied were found to be unusually sedentary (Thomas and Simcox 1982, Thomas 1983, 1991, Thomas, C D 1983, 1985, Thomas et al 1986, Warren 1987b, Thomas and Snazell 1989, 1985, Thomas, C D etal 1992),itisextremely unlikely that they could have crossed the English Channel after the land hnk was broken, roughly 1000 yr before man altered the British landscape The sedentariness and population structure of these butterflies also make it improbable that even the four species that inhabit modern woodland could have utilised their narrow niches when these occasionally existed, for periods of 1 to 10 yr, in small isolated patches in the ancient wildwood following the death of a large tree (Warren 1987 a,b, Thomas and Snazell 1989) I drew attention to this conundrum in a review of the conservation of European butterflies (Thomas 1991), and speculated that these species might once have inhabited broader or different niches to those of today In this paper I bring together the published evidence supporting this idea, and present new data on lntraspecific variation in niche breadth, and on the warm microclimates inhabited by the young stages of some butterflies I also extend my hypothesis, by suggesting how these species may have survived m Britain when summer temperatures became cooler, after the land link with the continent was broken height on Old Winchester Hill, Hampshire, S England Annual counts were made on 30 randomly chosen Hippocrepis comosa plants m each of 15 adjoining compartments that were grazed on rotation in 1982-86 Plant height (Thomas 1983) and the number of eggs was recorded in each of the 2250 samples Variation in the abundance of the ant Myrmica sabuleti was recorded by baiting or by pit-fall trapping beneath Thymus spp and Origanum vulgare L (the early larval foodplants of Maculinea anon), in swards of varying height on land with varying aspects Ants were sampled during the peak evening foraging period of M sabuleti beneath plants containing M arton eggs, or beneath suitable oviposition plants on extinct sites Pooled data are presented from one occupied and six extinct M anon sites in SW England (N = 347), from SX occupied and two unoccupied sites in Dordogne, I France (N = 227), and from five occupied sites in Oland, Sweden (N = 97) Munguira (pers comm ) descnbed the sward heights and aspects inhabited by L bellargus and H comma in Spain and south France, and I recorded the occupancy of shaded woods in Dordogne and the Rhone Valley, France by the four species of fntillary Microclimate measurements I measured temperature in the niches occupied by larvae, and in similar situations beneath foodplants rejected for oviposition, in a range of turf heights in south-facing grassland in south England Measurements were made in cloudy-bright weather in Apnl and May, using a thermocouple, for L bellargus (Thomas 1983. N = 67) and M anon (new data, N = 143), and in October using a laser thermometer for H comma (new data, N = 58) Thomas, C D (1983) used the former technique for P argus on south-facing heathland in north Wales Integrated temperatures were measured m the niches of B euphrosyne (N = 16 x 12) and B selene (N = 16 X 12), dunng the 12 wk feeding period of their larvae, using Pallman's sucrose inversion method (Lee 1969) Measurements were made m woodland blocks m S England that had been eut 3 months (B euphrosyne) or 18 months (B selene) earher Integrated temperatures were also measured beneath unsuitable Viola (N = 32 x 12) growing under shrubs that had not been cut for c 15 Methods and sources Niche breadth of butterflies and ants at different latitudes Precise definitions of the current British habitats of seven species were obtained from publications, theses or conservation reports Hesperia comma (Thomas et al 1986), Melitaea cinxia (Thomas and Simcox 1982, Thomas 1991), Boloria euphrosyne, B selene, and y4rgynnis adtppe (Thomas et al in press), Mellicta athalta (Warren 1987a), Plebejus argus in north Wales (Thomas, C D 1983) and south England (Read 1986) In addition, the distribution of 2nd generation Lysandra bellargus eggs was recorded in swards of varying ECOGRAPHY 16 3 (1993) Climate change and land management in the Holocene A simplified reconstruction of trends in mean summer temperatures m Britain over the past 8000 yr was made from the evidence of Sawyer 1966, Pearson 1978, Wigley et al 1986, Grove 1988, Huntley and Prentice Results and discussion Intraspecific variation in habitats at different latitudes In 1972-92, British populations oi Maculinea anon, Lysandra bellargus, Hespena comma, and Melttaea ctnxia occurred only on southern south-facing hillsides, and there only on sites where many foodplants grew in turf that was cropped to < 3 cm tall in the cases of M anon (Fig la) and L bellargus (Fig lb), or that was in the first one to three years of a succession from bare —r —\—1—1—1—1—1 1 2 3 4 5-9 10- 2 1 - 3 1 - 4 1 - > 5 0 20 30 40 5 0 1—r- — 1 — 1 " 28 26 - \ O24 Ny •••••-• ^20 VH»., •• \ •• +-: 10 2 3 4 5 6 Sward height (cm) 8 9 >9 8 1 1 1 1 1 1 1 I 1 1 1 2 3 4 5 6 7 8 9 10 11 V e g e t a t i o n height Ccm] Fig (la) Variation in the abundance of Myrmica sabuleti (the obhgate host of Macutinea anon) beneath Thymus and Origanum spp (the early larval foodplants) in swards of varying height at different latitudes Over 51% of foodplants must coincide with M sahuleti to support a population of M anon (Thomas 1991) (•—•) SW England. (D- -D) Dordogne. France Fig (lb) Distribution of Lysandra beltargus eggs {0—•. N = 2372) compared with that of its foodplant Hippocrepis comosa (D- -D N = 2250) m swards of varying height within one site in Hampshire. S England 1988. Wigley 1988, Wigley and Kelly 1990, Folland et al 1990 The timing and extent to which prehistoric man cleared woods and managed land to create modern habitats was reconstructed from Smith (1980, calcareous grassland), Rackham (1986, woods and most biotopes) and Webb (1986, heathland) May June July Fig 2 Microclimate measurements of the niches occupied by larvae of six species of butterfly in Bntish eeosystems (a) South-facing southern grassland, (—) Maculinea anon. {—) Lysandra bettargus, { ) Hespena comma, south-facing north Welsh heathland ( ) Ptebejus argus (from Thomas. C D 1983) Bold lines and symbols indicate the turf range ( = niche) inhabited by each butterfly, open circles indicate air temperature when measurements were made For clarity. SD included for M anon only (b) Integrated temperatures of the mam niches of Boloria euphrosyne ( • — • ) and B selene (A- -A), and of unsuitable shady Viola beneath tall shrubs (•- - • ) For clarity, 95% confidence limits omitted for B selene habitat ECOGRAPHY 16 3 (1993) hottest niches currently available in British grassland (Fig 2a), yet occupy 20 cm to 40 em tall swards on slopes of all aspects in lowland central France and Spam (Munguira pers comm ) And in the extreme south of England, P argus breeds in both the pioneer and building phases of heathland, again on all aspects (Read 1986), compared with its restriction to the hottest patches of south-facing pioneer heathland that exist in north Wales (Thomas, C D 1983, Fig 2a), where summer temperatures are c 2°C cooler In the woodland ecosystem, Bolorta euphrosyne feeds on Viola spp in only the first five years after a clearance, and B selene, A adtppe and M athalta are shaded out before 10 years' regrowth, although their foodplants persist for much longer (Warren 1987a, Warren and Thomas 1992, Thomas et al in press) These also represent the warmest microclimates available on British woodland floors (Fig 2b), whereas all four species occur in semi-shaded woodland in central France Past summer climates and the availability of anthropogenic habitats Published evidence of climatic change over the past 8000 yr is conflicting, but not in details that affect my Fig 3 July isotherms (1931-60) showing 3°C warmer and 3°C hypothesis There is a consensus that 8000-6000 BP was cooler mean temperatures than in southern England, where a period of relative warmth (Wigley et al 1986, Folland most butterfly habitat and microclimate measurements were made ( • ) English M anon sites, ( • ) English L bellargus, et al 1990), when average British summer temperatures (O) English Boloria spp, (A) north Welsh P argus, (A) south were up to 2°C higher than today (although winters may English P argus, (D) study area in Dordogne Isotherms from have been colder), especially towards the end of this Anonymous (1952), Arlery (1970), and Wallen (1970) penod (Huntley and Prentice 1988, Wigley and Kelly 1990, Folland et al 1990) This optimum may have lasted until 5300 or 5000 BP, before the summer climate ground, in the cases of H comma (Thomas et al 1986), cooled to something similar to or colder than that of M ctnxta (Thomas and Simcox 1982, Thomas 1991) and today (Folland et al 1990) Since then there may have Htpparchia semele (Heath et al 1984) been nine century-timescale periods, including the LittThe niches occupied by these species represent the warmest microclimates that are available in British grassland (Fig 2a) For example, an increase in mean turf height from 2 cm to 5 cm on British M arton sites ^ ^ anthropogenic creation of downs, causes the ground to become c 3°C cooler (Fig 2a), 5_ heaths and early and results in a rapid decline of Myrmica sabuleti folserai stages lowed by the local extinction of the butterfly (Thomas 1984, 1991) However in Dordogne, where mean summer tempeiatures are currently c 3°C warmer than in south Britain (Fig 3), M sabulett is abundant only on north-facing slopes and on flat land where the sward is tall (Fig la) Thus in central Europe, M anon is confined to a locally cool, late serai stage withm grassland this IS more widespread, less ephemeral, and much less dependent on man than its habitat in modern Bntain In Oland, where mean July temperatures are 1°C-2°C '000 Years before present warmer than in south Bntain, M sabulett is abundant enough to support M anon on flat terrain and south- Fig 4 Approximate changes in mean Bntish summer temperatures, the date when the land-link with Europe was broken, facing slopes, in swards ranging from 3 em to 10 cm tall and dates when man created downland, heathland, and early This too IS a much more widespread form of grassland serai stages withm both these and woodland Vertical scale for than the British habitat of M anon habitat creation is approximate and unquantifiable Sources of Lysandra bellargus and H comma also breed in the data are given under "methods" ECOGRAPHY 16 3 (1993) le Ice Age, when mean summer temperatures were up to 0 9°C cooler than today (Grove 1988, Wigley 1988, Wigley and Kelly 1990) In intervening centuries, summer temperatures were similar to or temporarily warmer than today, the main warm period being the Medieval Warm Epoch of 1100-700 BP A simplified reconstruction of these changes is given in Fig 4, omitting some cool intervals postulated by Wigley and Kelly (1990), which do not affect my hypothesis, and showing in dashed lines the periods for which evidence is conflicting or obscure Figure 4 also shows the approximate dates when the land link between Britain and the continent disappeared, and when man created warm early serai habitats within three ecosystems Hypothesis for the survival of thermophilous butterflies since 6000 BP I suggest that the ten species of butterfly that currently depend on early successional habitats in Bntain are relics from a period when summer temperatures were warmer (Fig 4), and that their pnmary habitats before 5000-6000 BP were m later, longer-lived successional stages, corresponding to their modern habitats in Oland or central Europe (e g Fig la) These habitats were probably quite common m pnmaeval Bntish woodland, given the extensive clearings made by beavers (Warren and Key 1991), and the presence of herbivores to browse gaps (Rackham 1986) Moreover, most of such permanent grassland and heathland as existed would have been habitable under a warmer climate, in contrast to the very small proportion that is south-facing, southerly and close-cropped, and hence warm enough today It follows that all except P argus, H semele and M cinxia of this group should have become extinct roughly 5000 yr ago, when the climate entered the first extended cool period after the land hnk with the continent was broken However, by then man had been making significant woodland clearings for at least a thousand years (Fig 4), creating heaths and grasslands and, through domestic grazing, early successional habitats within these, as well as plots of freshly cut woodland that were easy to crop for fuel (Smith 1980, Rackham 1986, Webb 1986) The warm microclimates that these inevitably contained (Fig 2) should have more than compensated for the cooling of the climate I suggest that these seven (13%) of British butterfly species survived by moving into warm man-made refugia, where they lived until the next warm period allowed them temporarily to reoccupy later serai stages, and perhaps to spread into warm anthropogenic habitats further north By the same reasoning, these species were probably more restricted than today dunng the coldest periods, such as the Little Ice Age Maculinea anon was perhaps confined to southern south-facing sites with exceptionally short (0 5 cm - 1 cm) swards, where temperatures at the soil surface are 1°C - 4°C higher than under the longest (2 cm) grassland that can support M anon today (Fig 2a) There is no fossil record of butterflies with which to test this theory, but data for Coleoptera are broadly consistent with it Around 10000 BP, the Bntish Midlands experienced an abrupt extinction of species now typical of northern latitudes or high ground, and by 9500-9000 BP had been colonised by assemblages similar to those of southern Bntain today, plus a few thermophilous species now restricted to the continent (Osborne 1976) Twenty-three of these species have since disappeared, most around 3000 BP, insofar as this can be dated (Osborne 1976, Girling 1984) Most extinctions were of arboreal species, '"many of which are now only found further south on the continent" (Osborne 1976), both Girling (1984) and Osborne (1976) implicate forest clearance and climatic cooling as contributory factors My theory for butterflies also implies a disproportionately high extinction rate among thermophilous arboreal insects after 5000 BP, because the warm microhabltats created by traditional land management exist mainly at ground level, not in shrubs and trees It is also unlikely that prehistoric man created warm habitats within aquatic systems, in which a thermophilous Gyrinid became extinct during the Little Ice Age (Girling 1984) On the other hand, the prehistoric extinction of a thermophilous ground-living species of dung beetle is inconsistent with my theory (Osborne 1976) Stronger support comes from the sub-fossil record of one of the few "arboreal" beetles to have warm habitats provided by man (Buckland 1975) Xestobium rufovillosum Deg oviposits in the fungal-infested timbers of dwellings, as well as m dead wood m the wild Wild British populations existed well north of their current range during the early Medieval Warm Epoch, but became confined to the Midlands and south England during the Little Ice Age and subsequently Yet X rufovillosum survived locally throughout its former northern range "in the artificially warm habitats created by man" in urban centres, such as York (Buckland 1975) It is easy to envisage a slightly more thermophilous species than X rufovillosum disappearing completely in the wild when the climate cooled, yet surviving as a British species in southern cities Conservation I conclude that seven species of British butterfly avoided extinction by moving into warm man-made refugia c 5000 yr ago But extinction was merely postponed in the present cerrtury British farmers and foresters have largely abandoned traditional forms of management, in favour of regimes that produce shadier woods, tall swards in unfertilised grassland, and mature or degenerate heathland (Smith 1980, Thomas 1984, 1990, 1991, Webb 1986, Warren and Key 1991, Warren ECOGRAPHY 16 3 (1993) and Thomas 1992) This has resulted in cooler microclimates in the ground layers of most sites (Fig 2), and caused the extinction of most populations of thermophilous butterfly species (Heath et al 1984) Many climatologists predict that mean summer temperatures will rise over the next 50 yr, through the greenhouse effect or as a further recovery from the Little Ice Age If so, I predict that these rare British butterflies will once again expand into later successional niches These habitats are still locally common in seminatural ecosystems m northern Europe, and require less intensive conservation management This interpretation offers an incentive and justification for conservationists, who are sometimes criticised for preserving unnatural obsolete forms of land-use in ecosystems It is particularly important that populations of scarce thermophilous species are maintained in Britain under the present climate, because these will be the founders of most future expansions, only the most mobile species are likely to recolonise Britain from across the English Channel Acknowledgements-I XhdnkR G Snazell and 1 Moy for help with integrated temperature measurements on Boloria sites, D J Simcox for help with M anon measurements m Oland and C D Thomas for permission to quote unpublished data from his M Sc thesis R L H Dennis encouraged me to publish this idea and he, P C Buckland, A J Gray C D Thomas, M G Morns G W Elmes and M S Warren made helpful comments on the ms This research was partly funded by a NERC TIGER award

Journal

Ecography – Wiley

Published: Jul 1, 1993

There are no references for this article.