Patterns of roost use by bats in a neotropical savanna: implications for conservation

Neotropic bats comprise a highly species-rich group and are considered important regulators of complex ecological processes. Resource partitioning of roosts and food is considered a key element in shaping bat communities. In this paper we examine the roosting patterns of 10 bat species belonging to a neotropical bat community in a Bolivian savanna. Roosts were more common in open woodlands than in forest islands and human settlements. In open woodlands, Tabebuia heptaphylla comprised the largest proportion of roosting trees, whereas in forest islands, Gallesia integrifolia was the most important roosting tree for phyllostomid bats. By comparing active roosts with control holes in Tabebuia heptaphylla, we found that roost site selection by Noctilio albiventris was mainly based on tree diameter and inner width of the cavities. The earlier results are discussed in view of resource partitioning within savanna bat communities and their practical conservation.     

Location of roosts of the lesser long-eared bat Nyctophilus geoffroyi and Gould's wattled bat Chalinolobus gouldii in a fragmented landscape in south-eastern Australia

The location of roosts of two species of vespertilionid bats, the lesser long-eared bat (Nyctophilus geoffroyi) and Gould’s wattled bat (Chalinolobus gouldii), was investigated in a remnant vegetation-farmland mosaic and adjacent floodplain forest in south-eastern Australia. A total of 45 individuals of N. geoffroyi and 27 C. gouldii were fitted with radio transmitters, which resulted in the location of 139 and 89 roosts respectively. Selection of roosting habitat showed both inter- and intra-specific differences. At the landscape level, locations of roosts used by male and female N. geoffroyi differed markedly. Most males roosted in the farmland mosaic within 3 km (mean 1.9±2.9 km) of where they were captured while foraging in remnant vegetation. In contrast, roost sites of females were predominantly in the floodplain forest, 6-12 km from their capture site in farmland (mean for all females, 6.7±2.9 km). All maternity roosts were in the extensive floodplain forest, 4-10 km from foraging areas. Distances moved by this species were greater than previously recorded and further than predicted by flight morphology. Most C. gouldii roosted in the floodplain forest, 4-10 km from their capture site (6.9±1.6 km). Within the floodplain forest, both species roosted in areas that had higher densities of hollow-bearing trees than generally available; dead hollow-bearing trees for N. geoffroyi and large, live trees for C. gouldii. The scale of movements undertaken by these species means that they can obtain resources from multiple landscape elements. Both species used different habitats for roosting and foraging despite the energetic costs of commuting relatively large distances. Conservation of bat populations in rural environments needs to be considered at the landscape scale, with particular attention to identifying landscape elements that provide key resources.     

Conservation of bats in suburban landscapes: roost selection by Myotis yumanensis in a residential area in California

Protection of roosting habitat is essential to the conservation of bats in human-dominated landscapes. To help define bat roosting needs in suburban settings, we used radio telemetry to locate day roosts of a common North American species (Myotis yumanensis) within a residential area in California. Between June and August 2000, we tracked 16 bats to 20 roosts in two buildings and 18 trees. We used multiple logistic regression to assess roost selectivity at multiple spatial scales. Of 15 tree, plot, and site characteristics considered, only three helped distinguish roosts from random comparisons: tree diameter, distance to water, and forest cover. Myotis yumanensis preferred large trees (mean diameter 115 cm), and roosted only in the five species of largest mean diameter (Sequoia sempervirens, Pseudotsuga menziesii, Quercus agrifolia, Quercus lobata, and Acer macrophyllum). At the site level, these bats selected roosts that were close to water and had substantial forest cover in the surrounding 100-m radius. Unlike other North American bats, they often roosted in live trees (89% live). Relatively high roost fidelity (mean 4.8 days) and large travel distances between consecutive roosts (mean 1130 m) and between capture sites and roosts (mean 2007 m) may indicate a greater degree of roost limitation at this site versus other research sites. We recommend the preservation of large trees and forested parkland, particularly along stream corridors, to help maintain bat populations in urbanizing landscapes.     

Habitat selection of endangered and endemic large flying-foxes in Subic Bay, Philippines

Large flying-foxes in insular Southeast Asia are the most threatened of the Old World fruit bats due to high levels of deforestation and hunting and effectively little local conservation commitment. The forest at Subic Bay, Philippines, supports a rare, large colony of vulnerable Philippine giant fruit bats (Pteropus vampyrus lanensis) and endangered and endemic golden-crowned flying-foxes (Acerodon jubatus). These large flying-foxes are optimal for conservation focus, because in addition to being keystone, flagship, and umbrella species, the bats are important to Subic Bay’s economy and its indigenous cultures. Habitat selection information streamlines management’s efforts to protect and conserve these popular but threatened animals. We used radio telemetry to describe the bats’ nighttime use of habitat on two ecological scales: vegetation and microhabitat. The fruit bats used the entire 14,000 ha study area, including all of Subic Bay Watershed Reserve, as well as neighboring forests just outside the protected area boundaries. Their recorded foraging locations ranged between 0.4 and 12 km from the roost. We compared the bats’ use to the availability of vegetative habitat types, riparian areas, and bat trees. The fruit bats’ locations showed a preference for undisturbed forest types and selection against disturbed and agricultural areas. Bat locations also showed selection for particular fruiting/flowering bat trees. The bats showed strong preference for riparian areas; locations were in riparian areas over four times more than expected. From these results we recommend that management focus flying-fox conservation efforts on undisturbed forest and riparian areas.     

Species-specific barrier effects of a motorway on the habitat use of two threatened forest-living bat species

It is generally recognized that roads can adversely affect local animal populations but little is known how roads affect bats. In particular, no study compared the response of bats that differ in foraging ecology to motorways that cut through the breeding habitat. As bats are key species in conservation, such data are urgently needed for designing management plans. Using radio-telemetry, mist netting, and mark-recapture data we investigated the effects of a motorway with heavy traffic on the habitat use of two threatened forest-living bats. We compared barbastelle bats (Barbastella barbastellus), which forage in open space, to Bechstein’s bats (Myotis bechsteinii), which glean prey from the vegetation. Five of six radio-tracked barbastelle bats crossed the motorway during foraging and roost switching, flying through underpasses and directly over the motorway. In contrast, only three of 34 radio-tracked Bechstein’s bats crossed the motorway during foraging, all three using an underpass. Bechstein’s bats, unlike barbastelle bats, never crossed the motorway during roost switching. Moreover, only in Bechstein’s bats individuals foraging close to the motorway had smaller foraging areas than individuals foraging further away, whereas other forest edges had no such effect. Our data show that motorways can restrict habitat accessibility for bats but the effect seems to depend on the species’ foraging ecology and wing morphology. We suggest that motorways have stronger barrier effects on bats that forage close to surfaces than on bats that forage in open space, and discuss the implications of our findings for bat conservation during road construction.     

Habitat associations of bats in Northern Ireland: implications for conservation

Agricultural intensification in Northern Ireland has brought about large-scale changes to the landscape with a detrimental effect on biodiversity. Between 1996 and 1998, we surveyed a stratified random sample of 1 km squares for bats using a spot-sample technique and time expansion bat detector to establish linear and area habitat associations. Bats strongly selected water bodies with bankside vegetation, treelines, and deciduous and mixed woodland edge, avoiding open areas such as upland/unimproved grassland and improved grassland. Of three sympatric pipistrelle species (Pipistrellus pipistrellus, P. pygmaeus and P. nathusii), two were shown to forage in different habitats; P. pipistrellus being more of a ‘generalist’, foraging in a wider variety of habitats than P. pygmaeus, a ‘specialist’ which foraged in a smaller range of habitats. Nyctalus leisleri selected parkland/amenity grassland, deciduous woodland edge and rivers/canals and avoided improved grassland, hedgerows, whilst Myotis spp. selected deciduous woodland centre, lake/reservoir margins and rivers/canals. Reduction in area and quality of inland water, deciduous woodlands and field boundaries in Northern Ireland may impact on bat populations. Habitat management should focus on improving those habitats which are selected by bats, including the maintenance and enhancement of connecting linear habitats.     

Roost availability may constrain shorebird distribution: Exploring the energetic costs of roosting and disturbance around a tropical bay

High tides force shorebirds from their intertidal feeding areas to refuges known as roosts. This paper explores the energetic costs of roost disturbance of great knot (Calidris tenuirostris) and red knot (C. canutus) at Roebuck Bay, North-western Australia, assessing disturbance levels at different roost sites through direct observation and automatic radio-telemetry, and applying physiological equations and predictive roost choice models to estimate energetic costs of disturbance through a complete tidal cycle. The study area had a variety of roosts, but use of each was constrained by conditions of tide and time. The roost most suitable for shorebirds on daytime high tides of intermediate height experienced high levels of disturbance from both natural sources (birds of prey) and humans. Flight costs caused by disturbance at this site exceeded the costs of flying to and roosting at the nearest alternative roost, 25 km away. However, shorebirds did not roost at the alternate site, possibly because of the risk of heat stress in a prolonged flight in tropical conditions. Increases in disturbance levels at just one of the roost sites of Roebuck Bay would increase energetic costs substantially, and could easily reach the point at which feeding areas accessed from this roost cannot be used without incurring a net energy deficit. Roost availability can therefore limit access to feeding areas and hence limit population size. Adequate provision and management of roost sites is accordingly an important consideration in conservation of sites used by coastal shorebirds.     

Habitat selection by the Mediterranean horseshoe bat, Rhinolophus euryale (Chiroptera: Rhinolophidae) in a rural area of southern Italy and implications for conservation

We studied habitat selection by Rhinolophus euryale in a rural area of southern Italy in 1998-2000 by radio-tracking. Two comparisons were carried out, one between habitat occurrence within individual home ranges and within the study area, the other between time spent in each foraging habitat and habitat occurrence within the home range. The first analysis showed that olive groves and conifer plantations were, respectively, the most and the least important habitats. The second analysis highlighted the importance of woodland for R. euryale, while urban sites, open areas and conifer plantations were avoided. We recommend that clearing of continuous, large areas of woodland for tree harvesting should be avoided. Conifers should not be used for reforestation. Urbanisation should be limited in the areas of greatest importance for the species, and linear landscape elements such as tree lines and hedgerows should be maintained.     

Conservation of Pteropus livingstonii based on roost site habitat characteristics on Anjouan and Moheli, Comoros islands

Livingstone's fruit bats, Pteropus livingstonii, (Pteropodidae: Chiroptera) are endemic to only two islands, Moheli and Anjouan of the Comoros Islands in the West Indian Ocean. With an estimated population of less than 2000 individuals, P. livingstonii is listed as Critically Endangered on the International Union for the Conservation of Nature's Red List. Yet the habitat characteristics of these giant fruit bats, or flying foxes, are poorly known. I conducted field research during 5 months in 1998 focused on finding new roost sites, measuring characteristics of both roost sites and roost trees at new as well as previously identified roost sites, and recording the same characteristics for a set of absence sites. I identified six new roost sites allowing me to measure the characteristics of 15 known roost sites and seven absence sites. The study spanned both dry and rainy seasons and I found a significant increase in the number of individuals at roost sites during the rainy season. Bat roost sites were associated with presence of water, southeast aspect, steep slopes, and presence of several native tree species. Roost habitat was characterized based on biotic, abiotic and human factors. This research on habitat characteristics is contributing to the development of a conservation action plan for the Livingstone's fruit bats. However, further research on dry season roost sites and distribution as well as feeding sites is necessary in order to better understand their seasonal habitat requirements and therefore ensure the conservation of P. livingstonii.     

Agricultural landscape modification increases the abundance of an important food resource: Mistletoes, birds and brigalow

Agricultural landuse alters landscape pattern, with important consequences for native species. Complex responses to changes in landscape pattern may be observed for species with strong inter-specific relationships. We investigate how the abundance of an ecologically important food resource, the grey mistletoe (Amyema quandang), is influenced by the interactive influences of human-modified landscape pattern, bird disperser abundance and host condition. Mistletoe abundance was estimated in brigalow (Acacia harpophylla) dominated remnants in three separate study areas with differing landscape patterns. Negative binomial generalised linear models of mistletoe abundance were tested with combinations of seven explanatory variables: study area, landscape connectivity and edge contrast, patch shape, bird disperser abundance, brigalow foliage cover and basal area of standing dead trees. Ranking of model averaged parameter estimates showed that study area, bird disperser abundance, patch shape, dead trees and connectivity had the strongest influence on A. quandang abundance. Mistletoe abundance apparently increased with landscape modification, particularly in narrow linear patches, but may be limited by the availability of seed dispersers and host condition. Given high landscape modification and ongoing degradation, management should be targeted towards maintaining brigalow remnant condition to ensure bird and mistletoe populations can be supported in the long-term.     

Gap-crossing by gliding marsupials: thresholds for use of isolated woodland patches in an agricultural landscape

Single trees and small patches of trees in farmland are conspicuous components of agricultural landscapes around the world, but their value for the conservation of biodiversity is not well known. In this study, arboreal mammals were censused by using hair-sampling tubes in small patches of woodland (single trees to patches <1.0 ha) in cleared farmland adjacent to a linear network of woodland known to support resident populations of arboreal mammals. Ninety-one small isolates were stratified by size (single trees or small patches) and distance from the linear network to test the capacity of animals to cross habitat ‘gaps’. The genus Petaurus (small gliding marsupials), the most commonly detected taxon, was recorded in 31% of hair-tubes (98 of 316). It occurred in 21% of sites in isolated trees and patches, and in all linear strips. A logistic regression model demonstrated that Petaurus sp. was most likely to occur in isolates in close proximity to linear strips and other patches of woodland. Ninety-five per cent of sites at which this taxon occurred were within 75 m of the linear network. This threshold corresponds with the maximum distance that animals can glide in a single movement between trees. The size of isolates did not influence utilisation rates. Such isolates are smaller than a single home range and were probably used to supplement home ranges centred on the linear network, by providing additional foraging habitat and den sites. Protection and restoration of isolated trees and small woodland clumps in cleared landscapes contributes to mammal conservation and this study provides quantitative data that can assist landscape design and habitat restoration in rural environments.     

Adaptive selection of foraging and nesting habitat by black kites (Milvus migrans) and its implications for conservation: a multi-scale approach

Black kites (Milvus migrans) are vulnerable and in decline within Europe. Here, we investigate selection of foraging and breeding habitat in a high-priority population in the Italian pre-Alps. Compared to a random distribution, kites foraged preferentially near water, over extensively managed grassland and within 1 km of nest-sites. Urban areas were positively selected near lakes but otherwise avoided. Foraging performance was higher over water than over terrestrial habitats. Kites nested on cliffs and trees and preferentially near water, far from paths and villages and in rugged and steep micro-sites. Tree-nests were located in the most mature tree in the stand. Productivity was positively related to the availability of water bodies. Therefore, food availability and human disturbance/persecution limited foraging and breeding performance. Guidelines to maintain or enhance current population levels include: (1) setting up reserves covering 10% of the areas within 1 km of large lakes; (2) converting current derelict coppice-woodland to high forest; and (3) enhancing subsidies for extensively managed grassland. Our results highlight the importance of cross-scale models integrating selection of foraging and breeding habitat and reinforce the importance of the spatial configuration of key resources for more realistic conservation management in mosaic landscapes.     

Effects of landscape-scale broadleaved woodland configuration and extent on roost location for six bat species across the UK

Although forest fragmentation can greatly affect biodiversity, responses to landscape-scale measures of woodland configuration in Europe have been examined for only a limited range of taxa. Almost all European bat species utilise woodland, however little is known about how they are affected by the spatial arrangement of woodland patches. Here we quantify landscape structure surrounding 1129 roosts of six bat species and a corresponding number of control locations across the UK, to examine associations between roost location and landscape composition, woodland proximity and the size of the nearest broadleaved woodland patch. Analyses are performed at two spatial scales: within 1 km of the roost and within a radius equivalent to the colony home-range (3–7 km). For four species, models at the 1 km scale were better able to predict roost occurrence than those at the home-range scale, although this difference was only significant for Pipistrellus pipistrellus. For all species roost location was positively associated with either the extent or proximity of broadleaved woodland, with the greatest effect of increasing woodland extent seen between 0% and 20% woodland cover. P. pipistrellus, Pipistrellus pygmaeus, Rhinolophus hipposideros, Eptesicus serotinus and Myotis nattereri all selected roosts closer to broadleaved woodland than expected by chance, with 90% of roosts located within 440 m of broadleaved woodland. Roost location was not affected by the size of the nearest broadleaved patch (patches ranged from 0.06–2798 ha ± 126 SD). These findings suggest that the bat species assessed here will benefit from the creation of an extensive network of woodland patches, including small patches, in landscapes with little existing woodland cover.     

Compound-specific δC and δH analyses of plant and soil organic matter: A preliminary assessment of the effects of vegetation change on ecosystem hydrology

Here we present δ¹³C and δ²H data of long-chained, even-numbered (C₂₇-C₃₁) n-alkanes from C3 (trees) and C4 (grasses) plants and from the corresponding soils from a grassland-woodland vegetation sequence in central Queensland, Australia. Our data show that δ¹³C values of the C4 grassland species were heavier relative to those of C3 tree species from the woodland (Acacia leaves) and woody grassland (Atalaya leaves). However, n-alkanes from the C4 grasses had lighter δ²H values relative to the Acacia leaves, but showed no significant difference in δ²H values when compared with C3 Atalaya leaves. These results differ from those of previous studies, showing that C4 grasses had heavier δ²H values relative to C3 grasses and trees. Those observations have been explained by C4 plants accessing the more evaporation-influenced and isotopically heavier surface water and tree roots sourcing deeper, isotopically lighter soil water (“Two-layered soil-water system”). By comparison, our data suggest that ecosystem changes (vegetation “thickening”) can significantly alter the soil hydrological characteristics. This is shown by the heavier δ²H values in the woodland soil compared with lighter δ²H values in the grassland soil, implying that the recent vegetation change (increased tree biomass) in the woodland had altered soil hydrological conditions. Estimated δ²H values of the source-water for vegetation in the grassland and woodland showed that both trees and grasses in open settings accessed water with lighter δ²H values (avg. −46‰) compared with water accessed by trees in the woodland vegetation (avg. −7‰). These data suggest that in semi-arid environments the “two-layer” soil water concept might not apply. Furthermore, our data indicate that compound-specific δ²H and δ¹³C analyses of n-alkanes from soil organic matter can be used to successfully differentiate between water sources of different vegetation types (grasses versus trees) in natural ecosystems.     

Nest site selection of the great spotted woodpecker Dendrocopus major in two oak woods in southern England and its implications for woodland management

Nest site selection of the great spotted woodpecker Dendrocopus major in two mature oak woods in southern England over an 11-year period is examined. Data describing the characteristics of 106 nest trees and sample surveys of available trees in the woods are used in the analyses. Although approximately equal numbers of nests were found in living and dead trees the latter were so rare in the study woods that this result represented an extremely strong selection for dead trees. Birch Betula spp. and elm Ulmus sp. were the major dead trees used. The smallest diameter tree used for nesting was 18 cm at breast height, although overall 90% of the nests were in trees of diameter 28 cm or greater. Apart from the minimum diameter, there was no strong selection for the use of large diameter trees for nesting, although this would probably not be expected in mature woodlands such as those in this study. The implications of the results for the management of woodlands for nesting woodpeckers are discussed.     

The habitat requirements of the rufous treecreeper (Climacteris rufa). 1. Preferential habitat use demonstrated at multiple spatial scales

Determining the habitat requirements of a species is fundamental to effective conservation, particularly if the species is declining in areas where its habitat is being modified. Multi-scaled investigations of habitat use are essential because different selection processes may operate at different scales. I examined the habitat use of a declining woodland passerine, the rufous treecreeper (Climacteris rufa), at three spatial scales (landscape, woodland and territory) in the wheatbelt of Western Australia. Preferential habitat use was exhibited at all scales. At the landscape scale, wandoo (Eucalyptus wandoo) woodland was used at a significantly greater rate than three other common vegetation types. Territory use within woodlands was positively related to the density of hollow-bearing logs, the density of nest sites, and tree age. Within an individual territory, nest sites (hollows) were favoured if they had a spout angle of ?50° to the horizontal and an entrance size of between 5 and 10 cm. The rufous treecreeper preferentially used habitat with traits characteristic of old-growth wandoo woodland. Degradation of wandoo through habitat modification (e.g. grazing, logging, fire and removal of deadwood) represents a significant threat to the persistence of treecreepers.     

Dead wood threshold values for the three-toed woodpecker presence in boreal and sub-Alpine forest

Predicting species' responses to habitat loss is a significant challenge facing conservation biologists. We examined the response of both European three-toed woodpecker subspecies Picoides tridactylus tridactylus and P. tr. alpinus to different amounts of dead wood in a boreal and a sub-Alpine coniferous forest landscape in central Sweden and Switzerland, respectively. Habitat variables were measured by fieldwork in forests with breeding woodpeckers (n=10+12) and in control forests without breeding woodpeckers (n=10+12) in the same landscape. Logistic regression analyses revealed steep thresholds for the amount of dead standing trees and the probability of three-toed woodpecker presence in both Sweden and Switzerland. The probability of the presence of three-toed woodpeckers increased from 0.10 to 0.95 when snag basal area increased from 0.6 to 1.3 m² ha⁻¹ in Switzerland and from 0.3 to 0.5 m² ha⁻¹ in central Sweden. In Switzerland, a high road network density was negatively correlated to the presence of woodpeckers (r=−0.65, p=0.0007). The higher volumes of dead wood in Switzerland, where population trends are more positive, than in central Sweden, where the population is declining, would suggest that the volumes of dead wood in managed forests in Sweden are too low to sustain three-toed woodpeckers in the long-term. In terms of management implications, we suggest a quantitative target of at least 5% of standing trees in older forests being dead over at least 100 ha large forest areas. This corresponds about to ?1.3 m² ha⁻¹ (basal area) or ?15 m³ ha⁻¹ (volume), still depending on site productivity.     

The effects of illuminating the roost entrance on the emergence behaviour of Pipistrellus pygmaeus

In an attempt to increase the accuracy of roost emergence counts for a monitoring programme, the exits of two Pipistrellus pygmaeus roosts were illuminated with light of different colours and intensities. Light intensity affected bat emergence more than light colour. At one roost there was no significant difference in the bat emergence pattern between when the roost exit received no illumination and when it was illuminated with red light. The use of the latter is proposed to increase the accuracy of bat roost emergence counts.     

Changes in a remnant of salmon gum Eucalyptus salmonophloia and York gum E. loxophleba woodland, 1978 to 1997. Implications for woodland conservation in the wheat-sheep regions of Australia

The condition of salmon gums Eucalyptus salmonophloia with large hollows in them in a 15-ha patch of remnant salmon gum-York gum E. loxophleba woodland in the northern wheatbelt of Western Australia was examined in 1978. The patch was an important breeding area for six species of cockatoo, including two endangered species. The patch was revisited in 1981 when the condition of all 682 salmon gums and York gums in the patch was examined and each was measured and photographed. A further visit was made in 1997 when the condition of the surviving trees was examined and each was again measured and photographed. The condition of the trees at each visit was classified as “good”, “staghorn”, “broken top”, “dead” or “fallen.” Over the period of the study there was a serious decline in the condition of the trees, with few large trees in the “good” category by 1997. The decline was particularly marked between 1978 and 1981 after a period of well-below average annual rainfall. Using data based on the rate of decline over the period 1978-1997, predictions were made of the fate of the trees in the patch. By 2125 only 46 (11% of the 1981 total) salmon gums were predicted to be alive with only one in the “good” category. Only 16 (17%) York gums were predicted to be alive by 2125, with only one in the “good” category. There was no evidence of any regeneration of woodland trees since 1929 when the patch was isolated by clearing for agriculture, and domestic livestock allowed to graze the patch. This deterioration of the dominant trees in the patch is symptomatic of remnant native vegetation over vast areas of Australia's extensively cleared wheat-sheep regions. The future of woodland patches like the one studied is bleak, as is the future of animals dependent on them for food, breeding sites and shelter. Active management, including fencing to exclude domestic livestock and measures to encourage regeneration of native plant communities, is necessary to counter the present regime of benign neglect that characterises most of Australia's management of native vegetation in agricultural landscapes.     

Parasitoids (Hymenoptera, Ichneumonoidea) of Saproxylic beetles are affected by forest successional stage and dead wood characteristics in boreal spruce forest

The habitat requirements and effects of forest management on insects belonging to higher trophic levels are relatively unknown in forest ecosystems. We tested the effect of forest successional stage and dead wood characteristics on the saproxylic parasitoid (Hymenoptera, Ichneumonoidea) assemblage in boreal spruce-dominated forests in northern Sweden. Within each of nine areas, we selected three sites with different management histories: (1) a clear-cut (2) a mature managed forest and (3) an old-growth forest. Parasitoids were collected in 2003 using eclector traps mounted on fresh logs, which were either untreated (control), burned, inoculated with fungi, or naturally shaded, and on artificially-created snags.Both forest type and dead wood characteristics had a significant effect on parasitoid assemblages. Grouped idiobionts and some species, such as Bracon obscurator and Ontsira antica, preferred clear-cuts, while others, such as Cosmophorus regius (Hym., Braconidae) and other koinobionts, were associated with older successional forest stages. No single dead wood substrate was sufficient to support the entire community of parasitoids in any forest type, even when the regular host was present. In particular, snags hosted a different assemblage of species from other types of dead wood, with parasitoids of Tetropium spp. such as Rhimphoctona spp. (Hym., Ichneumonidae) and Helconidea dentator (Hym., Braconidae) being abundant. These results indicate that a diversity of dead wood habitats is necessary to support complete assemblages of beetle-associated parasitoids from early successional stages of dead wood and that parasitoids may be more sensitive to habitat change than their hosts.