Contributed Papers

Distribution of Carabid Beetles (Coleoptera,

Carabidae) across a Boreal Forest–Clearcut Ecotone
JANNE HELIÖLÄ, MATTI KOIVULA, AND JARI NIEMELÄ*
Department of Ecology and Systematics, Division of Population Biology, P.O. Box 17,
FIN-00014 University of Helsinki, Helsinki, Finland

Abstract: We studied the occurrence of carabid beetles (Coleoptera, Carabidae) in boreal forest fragments,
their edges, and adjacent clearcuts in central Finland. Beetles were collected with pitfall traps along transects
extending 60 m from the edge into clearcuts and 60 m into forest interior. Our main findings were that (1)
species richness was significantly higher in the clearcut than in the forest fragments, (2) clearcuts hosted
many open-habitat species, which increased overall species richness in these sites, (3) carabid assemblages in
the edges were more similar to forest assemblages than to those found in the clearcuts, (4) no edge specialists
were found, and (5) open-habitat species did not penetrate into the forest fragments from the clearcut. Be-
cause forest specialists occurred all the way to the edge on the forest side, it seems that edge effects per se do
not adversely affect these species, at least in the short term. In the long term, however, habitat conditions in
the edges may deteriorate for interior species because of trees falling over in strong winds, thereby reducing
the size of the fragments and widening the edge zone.

Distribución de Escarabajos Carábidos (Coleoptera, Carabidae) a lo Largo de un Ecotono de Bosque Boreal - Zona
Talada
Resumen: Estudiamos la presencia de escarabajos carábidos (Coleoptera, Carabidae) en fragmentos de un
bosque boreal, de sus bordes y de zonas taladas adyacentes en Finlandia central. Los escarabajos fueron
colectados usando trampas cubiertas a lo largo de transectos que se extendían desde 60 m del borde hacia el
área de tala y 60 m hacia el interior del bosque. Nuestros resultados principales fueron: (1) la riqueza de es-
pecies fue significativamente mayor en la zona talada que en los fragmentos de bosque, (2) las zonas taladas
hospedaron muchas especies de hábitat abierto, lo cual incrementó la riqueza general de especies en estos si-
tios, (3) los ensamblajes de carábidos en los bordes fueron más similares a los ensamblajes del bosque, que
aquellos en las zonas taladas, (4) no se encontraron especialistas de borde, y (5) las especies de hábitat abi-
erto no penetraron de las zonas taladas hacia los fragmentos de bosque. Debido a que los especialistas de
bosque aparecieron a todo lo largo del borde en el lado del bosque, aparentemente los efectos de borde en sí
no afectan adversamente a estas especies, al menos a corto plazo. Sin embargo, a largo plazo las condiciones
del hábitat en los bordes podrían deteriorar para las especies del interior, debido a la caída de árboles oca-
sionada por vientos fuertes, lo cual reduce el tamaño de los fragmentos, ampliando la zona de borde.

Introduction et al. 1994; Murcia 1995; Didham et al. 1996; Didham

1997a, 1997b). In many parts of the boreal region, log-
Fragmentation and habitat loss are among the most im- ging has transformed the forest landscape into a patch-
portant causes of species decline worldwide (e.g., Haila work of forest remnants of various sizes that are isolated
from one another by clearcuts (Esseen et al. 1997). In
Finland, for instance, forest fragmentation has already
caused species declines and probably local extinctions
*Address correspondence to J. Niemelä, email [email protected]
Paper submitted September 27, 1999; revised manuscript accepted (Rassi et al. 1992; Siitonen & Martikainen 1994; Niemelä
May 24, 2000. 1997).
370

Conservation Biology, Pages 370–377

Volume 15, No. 2, April 2001
Heliölä et al. Carabid Beetles in Boreal Forest Edges 371

One consequence of fragmentation and associated Methods

habitat loss is decreasing fragment size, which leads to a
decreasing proportion of interior habitat and increasing The study area is located in Northern Häme (southcen-
proportion of edge (Ranney et al. 1981; Harris 1984; tral Finland, approximately lat 61⬚N, long 24⬚E) in the
Murcia 1995; Ranta et al. 1998; Haila 1999; Matlack & municipalities of Keuruu, Kuorevesi, Längelmäki, and
Litvaitis 1999). Changes in abiotic and biotic conditions Orivesi. The study sites are dominated by 80- to 100-
in the edge compared with the intact habitat are collec- year-old spruce (Picea abies), with some pines (Pinus
tively called edge effects. These effects are many, includ- sylvestris) and birches (Betula spp.) among the spruce.
ing changes in wind, humidity, radiation, predation, par- The field layer is dominated by blueberry (Vaccinium
asitism, and species interactions (Saunders et al 1991; myrtillus) and lingonberry (Vaccinium vitis-idaea), and
Andrén 1995; Chen et al. 1995; Murcia 1995; Risser the ground layer is dominated by mosses.
1995; Donovan et al. 1997; Voller 1998). Increased lev- The study sites of approximately 2 ha each were estab-
els of human-induced disturbance at habitat edges also lished to examine the effects of modified forest harvest-
affect species composition (Kruger & Lawes 1997). ing methods on biodiversity. We used six of the sites,
To counteract the negative effects of edges, it is im- which represented two of the five harvesting methods
portant to understand how species respond to condi- examined: (1) three sites were treated with traditional
tions in habitat edges (Haila et al. 1994). Forest edges are clearcutting (all merchantable trees removed) and (2)
the preferred habitat of many vertebrate species, such as three sites were treated by modified clearcutting in
ungulates (Voller 1998). Furthermore, forest edges har- which three or four groups of 20–30 trees each were re-
bor a rich invertebrate fauna (e.g., Helle & Muona 1985; tained per hectare. The traditionally cut sites faced west
Jokimäki et al. 1998), but few of these species appear to and the modified ones east, but this had no effect on car-
be edge specialists (Didham 1997b). Invertebrate assem- abid assemblages. The distances between the six sites
blages in edges are mixtures of species found on either were 500 m or more. The sites were harvested in the win-
side of the edge zone (Kotze & Samways 1999). This im- ter of 1995–1996, and our edge study was conducted in
plies that edges of forest fragments are easily invaded by the summer of 1997. The contrast between the clearcut
invertebrate species from the surrounding matrix and and the mature forest was strong, and the edge could be
that some species may continue through the edge “fil- considered a hard edge (Demaynadier & Hunter 1998).
ter” into the forest interior (Spence et al. 1996). Small Pitfall traps were used to collect beetles (Greenslade
forest fragments with a high proportion of edge habitat 1964; Spence & Niemelä 1994). Traps were plastic jars
are particularly vulnerable to invasion by species from (diameter 65 mm, depth 70 mm) partly filled with 20%
the surrounding matrix (Bauer 1989; Halme & Niemelä ethylene glycol solution. A plastic roof of 10 ⫻ 10 cm was
1993; Usher et al. 1993; Burke & Goulet 1998). Further- placed a few centimeters above each trap to prevent
more, edge habitat is unsuitable for species requiring in- flooding from rainwater. The trapping period covered
terior habitat (Demaynadier & Hunter 1998; Stevens & most of the growing season (15 May–17 September), and
Husband 1998; von Sacken 1998), and consequently the traps were emptied and serviced once a month. In
such species may be lost if fragments become too small spite of the 4-week servicing intervals, the traps and their
(Haila 1999). Forest managers should minimize adverse catches remained in good condition, and few traps were
edge effects by, for example, leaving fragments large lost due to flooding by water or debris. Furthermore, few
enough to maintain specialists of the forest interior small mammals or frogs fell into the traps.
(Spence et al. 1996; Burke & Goulet 1998; von Sacken We studied edges as gradients that included, in addition
1998). to the edge, disturbed (clearcut) and undisturbed (mature
We examined the distribution of carabid beetles in forest) habitat, as recommended by Meiners and Pickett
“hard edges” between mature spruce forest and adjacent (1999). Our gradient extended from a point 60 m from the
clearcuts in central Finland. As recommended by Mein- edge into the forest interior and through the edge to a
ers and Pickett (1999), we studied edges as gradients point 60 m into the clearcut. Carabid samples from four
that included, in addition to the edge zone, both dis- traps 4–5 m apart arranged in a line (trapline) parallel to the
turbed (clearcut) and undisturbed (mature forest) habi- edge were pooled and used in the analyses. One trap line
tat. We focused on the following questions: (1) Are for- of four traps was placed along the edge between the ma-
est-interior species sensitive to the forest edge, as shown ture forest and the clearcut. The other trap lines ran from
for salamanders and anurans (Demaynadier & Hunter the edge into the forest and similarly into the clearcut, with
1998) and small mammals (Stevens & Husband 1998)? traps at 15, 30, 45, and 60 m from the edge. Each of the six
(2) Do open-habitat species invade forest fragments, as study sites had nine trap lines, for a total of 36 traps per site
suggested by Halme and Niemelä (1993)? (3) What are (four lines in the forest, one line in the edge, and four lines
the differences in carabid assemblages between mature in the clearcut). In the modified clearcuts, we avoided
forest patches and clearcut patches? groups of retained trees when we placed the trap lines.

Conservation Biology
Volume 15, No. 2, April 2001
372 Carabid Beetles in Boreal Forest Edges Heliölä et al.

We used a 60-m distance from the edge for two rea-

sons. First, at this distance into the forest most edge ef-
fects are dampened (Harris 1984; Kapos 1989; Murcia
1995; Demaynadier & Hunter 1998). Second, in both the
clearcut and the forest, the last trap line was close to the
center of the habitat, and going farther would have
brought the opposite forest edge too close.
We used Kruskall-Wallis nonparametric analysis of
variance (ANOVA) to examine the species richness and
abundance of carabids between the two treatments (tra-
ditional clearcutting and modified clearcutting) and
among different distances along the gradient from forest
interior through the edge to the clearcut. Species rich-
nesses were standardized by rarefaction to remove the Figure 1. Average number of carabid individuals and
effects of differing beetle abundances (Ludwig & Rey- species with 95% confidence interval in the forest inte-
nolds 1988; Krebs 1989). We used Scheffe’s test for rior (traps 30–60 m inside the forest fragment), edge
post-hoc pairwise comparisons. For the ANOVA we di- (traps 15 m inside the forest, at the edge, and 15 m
vided the trap lines into three distance groups (forest in- into the clearcut), and in the clearcut (30–60 m into
terior, trap lines 30, 45, and 60 m inside the forest; edge, the clearcut).
trap lines 15 m inside the forest, at the edge, and 15 m
into the clearcut; and clearcut, trap lines 30, 45, and 60 m
into the clearcut). or richness (U ⫽ 30.00, df ⫽ 1, p ⫽ 0.354) between the
A cluster analysis (group-average linking algorithm traditionally clearcut sites and the retention tree cutting.
with Czekanowski-Sørensen similarity metric) was used Furthermore, there were no differences in species rich-
to compare carabid assemblage structure among the ness or abundance in relation to the exposure of the
three distance groups described above (Ludwig & Rey- edge. Thus, we hereafter use the pooled data from the
nolds 1988; Krebs 1989). The value of the similarity met- two harvesting types.
ric ranges between 0, for no species shared, and 1, for Species richness differed significantly between forest
identical samples. We used the BIODIV software pack- interior, edge, and clearcut (U ⫽ 8.56, df ⫽ 2, p ⫽
age to perform the cluster analysis (Baev & Penev 1995). 0.014) (Fig. 1). According to Scheffe’s post hoc test, spe-
Ground and field-layer vegetation cover, litter, and log- cies richness was significantly higher in the clearcuts
ging residue were estimated within a circle 3-m in diam- than in the forest interiors and edges, whereas species
eter that was placed in the center of each four-trap line. richness in edges and forest interiors did not differ ( p ⬎
These measurements were used in a redundancy analy- 0.05). There were no statistically significant differences
sis (RDA; Jongman et al. 1995) to study the relationship in carabid abundance between forest interior, edge, and
between carabid occurrence and their environment. We clearcut (U ⫽ 3.71, df ⫽ 2, p ⫽ 0.157) (Fig. 1).
excluded from the RDA analysis environmental variables According to the cluster analysis, edges grouped to-
with more than 75% 0 values and those with some (usu-
ally one to three) extreme values because of their poten-
tially strong and misleading effect on the analysis. We
also excluded from the analysis carabid species with
only one individual. After these amendments we in-
cluded in the analysis ten environmental variables and
27 carabid species, and we used data transformed by
ln(x ⫹ 1). We centered by species when running the
analysis and did the RDA analysis using the CANOCO 4.0
software package (ter Braak & Smilauer 1998).

Results Figure 2. Cluster analysis with Czekanowski-Sørensen

index (group-average linking algorithm) of the cara-
Changes in Community Structure across the Edge bid assemblages in the forest interior (traps 30–60 m
inside the forest), at the edge (traps from 15 m into the
There were no statistically significant differences in cara- forest, at the edge, and 15 m into the clearcut), and in
bid species abundance (U ⫽ 45.00, df ⫽ 1, p ⫽ 0.691) the clearcut (traps 30–60 m into the clearcut).

Conservation Biology
Volume 15, No. 2, April 2001
Heliölä et al. Carabid Beetles in Boreal Forest Edges 373

gether with forest interiors, indicating that carabid as- tus terminatus), none of the species was restricted to
semblages in the edges were more similar to forest as- the edge zone or was clearly more abundant there than
semblages than to those in the clearcuts (Fig. 2). elsewhere (Table 1). Nevertheless, many species that
Similarity within these clusters was high (⬎73%). were abundant in the clearcut occurred in low numbers
in the edge zone (e.g., Bembidion lampros, Harpalus
quadripunctatus, Pterostichus adstrictus, P. nigrita)
Species Occurrence across the Edge
( Table 1). Most of these species require open habitat
A total of 5609 carabid beetles representing 34 species (Lindroth 1985, 1986). Because only single individuals
was captured. Calathus micropterus was the most of these species were found in the forest interior (30–60
abundant species, with 3373 individuals (60% of the to- m into the forest), it is evident that open-habitat species
tal sample), and Pterostichus oblongopunctatus was the entered the forest only sporadically from the clearcut.
second most abundant species, with 1363 individuals For instance, the most abundant colonizer, P. adstrictus,
(24% of the total sample) (Table 1). decreased drastically from the center of the clearcut
Except for two scarce species (Amara famelica, Leis- toward the edge and did not penetrate the forest at all
(Fig. 3).

Table 1. Pooled numbers of carabid individuals collected in the Environmental Variables and Carabid Occurrence
six sites divided into forest trap lines, edge trap lines, and clearcut
trap lines. There were dramatic differences in the field and ground-
layer vegetation between the forest and the clearcut (Fig.
Number of individuals* 4). Most changes occurred right in the edge, although
Scientific name (abbreviation) forest edge clearcut there was some logging residue and bare soil in the for-
Agonum fuliginosum (Agonfuli) 45 37 132 est close to the edge. In the field layer, dwarf shrubs, es-
A. mannerheimii (Agonmann) 16 5 7 pecially blueberry (Vaccinium myrtillus) and lingon-
A. sexpunctatum (Agonsexp) 0 1 11 berry (V. vitis-idaea), decreased from a cover of 30–40%
Amara brunnea (Amarbrun) 3 4 8 in the forest to approximately 10% in the clearcut.
A. famelica 0 1 0
Grasses (Deschampsia flexuosa, Calamagrostis sp.) and
A. lunicollis 0 0 5
Bembidion bruxellense 0 0 1 fireweed (Epilobium angustifolium), on the other hand,
B. lampros (Bemblamp) 0 4 18 were more common in the clearcut than in the forest. In
Calathus micropterus (Calamicr) 1253 1051 1069 the ground layer, mosses (Pleurozium spp., Dicranum
C. cancellatus (Caracanc) 1 2 9 spp.) were common in the uncut forest, whereas in the
C. glabratus (Caraglab) 15 34 36
exposed clearcuts they did not thrive. Also the mois-
C. hortensis (Carahort) 12 3 3
Cicindela campestris 0 0 1 ture-dependent Sphagnum mosses were scarce in the
Cychrus caraboides (Cychcara) 18 5 6 clearcuts. Exposed soil, logging residue, and needle lit-
Dyschirius globosus 1 0 0
Harpalus quadripunctatus
(Harpquad) 0 9 12
Leistus ferrugineus 0 0 2
L. terminatus 0 1 0
Miscodera arctica 0 0 1
Notiophilus biguttatus (Notibigu) 13 13 37
N. palustris 0 0 5
Patrobus assimilis (Patrassi) 17 7 20
Pterostichus adstrictus (Pteradst) 0 7 110
P. cupreus (Ptercupr) 0 1 4
P. diligens (Pterdili) 6 6 3
P. niger (Pternige) 0 4 8
P. nigrita (Pternigr) 0 1 26
P. oblongopunctatus (Pteroblo) 171 370 822
P. strenuus (Pterstre) 1 3 17
P. vernalis 0 0 3
P. versicolor (Ptervers) 0 3 2
Trechus rubens (Trecrube) 1 0 1
T. secalis (Trecseca) 9 26 49
Trichocellus placidus 0 0 1
Number of individuals 1582 1598 2429
Number of species 16 24 31
Figure 3. Number of Pterostichus adstrictus in the trap
* Forest, traps 30–60 m into the forest interior; edge, traps 15 m into
the forest, lines on the edge, and traps 15 m into the clearcut; lines at different distances from the edge. Average
clearcut traps 30–60 m into the clearcut. catch from the six sites with 95% confidence interval.

Conservation Biology
Volume 15, No. 2, April 2001
374 Carabid Beetles in Boreal Forest Edges Heliölä et al.

forest (right side of the ordination space: tall spruces,

mixed litter, mosses, Vaccinium spp.) and those located
in the clearcut (left side: logging residue, stumps, needle
litter). Also, the cover of grasses, an indication of early
successional stage, and the cover of exposed soil, indi-
cating amount of mechanical soil preparation, were lo-
cated toward the left side of the ordination space (Fig. 5).
The second axis (axis 2) seems to indicate moisture: the
percent cover of Sphagnum mosses correlated strongly
and positively with this axis. On the contrary, the cover
of forest mosses correlated slightly negatively with Sphag-
num, indicating somewhat drier conditions.
In the RDA ordination, almost every carabid species
was located left of the origin, indicating that most of the
species were positively associated with habitat variables
in the clearcut (Fig. 5). For instance, Pterostichus ad-
strictus and P. oblongopunctatus correlated highly posi-
tively with logging residue and needle litter. Thus, it is
possible that some carabid species benefit from the in-
creased structural elements on the forest floor following
clearcutting. In contrast, forest specialist species were
located in the lower part of the ordination space. These
species (e.g., C. hortensis and C. caraboides) did not
correlate with variables associated with clearcuts, but
they correlated positively with, for instance, the cover
of forest mosses, Vaccinium dwarf shrubs, and the
amount of large spruce trees. The third group of cara-
bids, located in the upper part of the plot, consisted of
species most often caught from moist patches character-
ized by the occurrence of Sphagnum mosses (e.g., Ago-
num mannerheimii, A. fuliginosum, Pterostichus dili-
gens, and Patrobus assimilis). These species occurred
Figure 4. Percent cover of dwarf shrubs (mainly Vac- in the forest and clearcut (Table 1; Fig. 4). Carabus gla-
cinium myrtillus and Vaccinium vitis-idea) and grasses bratus, C. micropterus, and P. niger correlated negatively
in the field layers, and Sphagnum mosses, other with Sphagnum.
mosses, needle litter, logging residue, bare soil, and
mixed needle and leaf litter in the bottom layer. All six
sites are combined.
Discussion
Carabid Responses to Forest Edges
ter from the residue covered a high proportion of the
clearcut (Fig. 4). The main findings of our study were that (1) carabid as-
In the redundancy analysis based on environmental semblages in the edges were more similar to assem-
variables and the trap line–specific samples, the carabid blages in the forest interior than to those in the
species had eigenvalues in the first four axes of 0.195, clearcuts; (2) carabids did not avoid the edge zone on
0.098, 0.030, and 0.021, respectively (Fig. 5). The first the side of the forest, but there were no edge specialists;
two axes explained 29.3% of the total variance in occur- (3) although species richness and carabid abundance
rence of carabid species and 76.5% of the variance in the were high in the clearcuts, open-habitat species only oc-
relationship between carabid species and the environ- casionally penetrated into the forest interior from the
ment. We analyzed the statistical significance of the axes clearcut; and (4) some forest species were less abundant
by Monte-Carlo randomization. The trace value of all in the clearcuts and may have been suffering from frag-
axes was 0.383 (F ⫽ 2.665, p ⫽ 0.005, 199 permuta- mentation of mature boreal forest.
tions), indicating that the environmental variables and The finding that carabid samples from the edges were
carabid species had a highly significant relationship. similar to those from the forest trap lines corroborates
The first axis of the ordination (axis 1, Fig. 5) indi- results from forest-grassland edges in South Africa (Kotze
cated a division between trap lines located in the mature & Samways 1999). Furthermore, both our results and those

Conservation Biology
Volume 15, No. 2, April 2001
Heliölä et al. Carabid Beetles in Boreal Forest Edges 375

Figure 5. The RDA ordination for

carabids and 10 environmental
variables. Carabid species are
marked with black circles and the
environmental variables with ar-
rows. Trees-picea is mature spruce
individuals (minimum 20 m
height) per 78.5 m2. The other vari-
ables are percentage coverages per
28.3 m2. Label explanations:
Grasses, Calamagrostis, Deschamp-
sia, etc.; logging r, logging residue;
mixed l, mixed deciduous and nee-
dle litter; mosses, forest-inhabiting
mosses; needle l, needle litter; ex-
posed soil, visible sand layer cre-
ated by mechanical soil preparation;
sphagnum, Sphagnum mosses; Vac-
cinium, V. vitis-idaea and V. myrtil-
lus dwarf shrubs. Box includes five
carabid species: Agonum sexpunc-
tatum, Amara lunicollis, Leistus fer-
rugineus, Notiophilus palustris, and
Pterostichus vernalis. Abbreviations
of the carabid species are listed in
Table 1.

of Kotze and Samways (1999) show that the edge was were collected from clearcuts than from forests. Most of
abrupt for carabids. In our study, forest species occurred the species found in the clearcuts were specialists of
right to the edge on the forest side, and it appears that open habitats (Lindroth 1985, 1986), but some forest
the physical conditions in the edges did not negatively generalists, notably P. oblongopunctatus and T. secalis,
affect forest-dwelling carabids. As our analyses showed, were more common in the clearcuts than in the forests.
field and ground-layer vegetation remained virtually in- The open, dry, and warm clearcuts are favored by many
tact close to the forest edge, which may have contrib- carabid species occurring in grasslands and other similar
uted to the maintenance of populations of forest cara- habitats; only a few species have adapted to the dark
bids near the edge. This result, however, may not be and cool spruce forest (Niemelä 1993). In addition to
applicable to other taxa. For instance, some forest-dwell- carabids (Niemelä et al. 1988, 1993, 1994; Halme & Nie-
ing salamanders and anurans in the United States (De- melä 1993; Haila et al. 1994; Spence et al. 1996), the
maynadier & Hunter 1998) and small mammals in Brazil phenomenon of high abundance in clearcuts appears to
(Stevens & Husband 1998) avoid forest edges. be the rule for spiders (Pajunen et al. 1995), ants (Punt-
As in boreal Canada (Spence et al. 1996), no carabid tila et al. 1991, 1994), and butterflies (Väisänen 1995).
species were restricted to the edge in our study. Thus, it But Kotze and Samways (1999) reported from Afromon-
seems that there are no edge species among boreal for- tane forest-grassland ecotones in South Africa that,
est carabids. In temperate oak-hornbeam forest in Hun- whereas ants were more abundant in the grassland, cara-
gary, however, several carabid species occurred exclu- bids were more abundant in adjacent forests, perhaps
sively at the forest edge, indicating that these habitats because of competitively superior ants. In our study,
are important for the maintenance of carabid diversity there was a negative correlation between the number of
(Magura & Tothmeresz 1997). The different responses red wood ants and carabids in the traps (r ⫽ ⫺0.31, p ⫽
of carabids to edges may be attributed to differences in 0.02). This correlation does not necessarily indicate
vegetation. Our edges and those studied by Spence et al. competition; it may be related to different microhabitat
(1996) were hard, abrupt, and created by clearcutting, preferences among ants and carabids.
whereas those studied by Magura and Tothmeresz (1997) Although many open-habitat carabid species were
were gradual, with abundant bushes. common in our clearcuts, they rarely penetrated into
In our study, more carabid species and individuals the forest. Similarly, in Switzerland only 1 of 15 carabid

Conservation Biology
Volume 15, No. 2, April 2001
376 Carabid Beetles in Boreal Forest Edges Heliölä et al.

species occurred commonly in both a grassland patch edge zone, thereby increasing the permeability of the
and the surrounding forest edge (Hänggi & Baur 1998). edge for open-habitat species. This process has already
Contrary to these studies, some carabid species classi- started in our forest fragments.
fied in literature as open-habitat specialists (e.g., Bem-
bidion grapei and P. adstrictus) occurred 20–40 m into
the forest in a Canadian study (Spence et al. 1996). Acknowledgments

Conservation Implications This study was funded by the Academy of Finland and
the Finnish Biodiversity Research Programme. We thank
Populations of forest-dwelling carabids appeared not to J. Kotze for comments.
be affected negatively by the edge zone, because beetle
numbers did not decrease near the edge. A similar obser-
vation was made in Canada (Spence et al. 1996). Al- Literature Cited
though these findings suggest that edges do not have Andrén, H. 1995. Effects of landscape composition on predation rates
much negative effect on forest carabids, continued frag- at habitat edges. Pages 225–255 in L. Hansson, L. Fahrig, and G.
mentation and decreases in the size of fragments may in Merriam, editors. Mosaic landscapes and ecological processes.
the long run threaten currently viable populations. For Chapman & Hall, London.
instance, small forest fragments are more vulnerable to Baev, P. V., and L. D. Penev. 1995. BIODIV: program for calculating bi-
ological diversity parameters, similarity, niche overlap, and cluster
invasion by open-habitat species from the surrounding analysis. Exeter Software, New York.
regenerating forests (Halme & Niemelä 1993; Burke & Bauer, L. C. 1989. Moorland beetle communities on limestone ‘habitat
Goulet 1998). Furthermore, other taxa may be more sen- islands’. I. Isolation, invasion and local species diversity in carabids
sitive to edge. Bark beetles (Coleoptera, Scolytidae) and staphylinids. Journal of Animal Ecology 58:1077–1098.
were less common in forest edges than in interior forest Burke, D., and H. Goulet. 1998. Landscape and area effects on beetle
assemblages in Ontario. Ecography 21:472–479.
in Finland (Peltonen & Heliövaara 1998). Similarly, in Chen, J., J. F. Franklin, and T. A. Spies. 1995. Growing-season microcli-
Amazonian rain forest, beetle abundances were lower in matic gradients from clearcut edges into old-growth Douglas-fir for-
the edge and peaked 26–105 m from the edge toward ests. Ecological Applications 5:74–86.
forest interior (Didham 1997b); however, that study did Demaynadier, P. G., and M. L. Hunter Jr. 1998. Effects of silvicultural
not examine the fragments’ surroundings. In our study edges on the distribution and abundance of amphibians in Maine.
Conservation Biology 12:340–352.
there appeared to be an edge effect on the side of the Didham, R. K. 1997a. An overview of invertebrate responses to forest
clearcuts, because some open-habitat species decreased fragmentation. Pages 304–319 in A. D. Watt, N. E. Stork, and M. D.
near the edge. Hunter, editors. Forests and insects. Chapman & Hall, London.
Although forest species were not negatively affected Didham, R. K. 1997b. The influence of edge effects and forest frag-
by edge effects, they may suffer from isolation effects mentation on leaf litter invertebrates in central Amazonia. Pages
55–70 in W. F. Laurance and R. O. Bierregaard Jr., editors. Tropical
caused by fragmentation (see also Burke & Goulet forest remnants: ecology, management, and conservation of frag-
1998). Flightless forest specialists, such as Carabus mented communities. The University of Chicago Press, Chicago.
hortensis and Cychrus caraboides, may perish when at- Didham, R. K., J. Ghazoul, N. E. Stork, and A. J. Davis. 1996. Insects in
tempting to cross clearcuts. We found some individuals fragmented forests: a functional approach. Trends in Ecology and
of these species in the clearcuts, but Halme and Niemelä Evolution 11:255–260.
Donovan, T. M., P. W. Jones, E. M. Annand, and F. R. Thompson III.
(1993) reported that C. caraboides was virtually absent 1997. Variation in local-scale edge effects: mechanisms and land-
from open habitat and from forest fragments of ⬍20 ha scape context. Ecology 78:2064–2075.
and that C. hortensis was scarce in fragments of ⬍3 ha. Esseen, P-A., E. Ehnström, L. Ericson, and K. Sjöberg. 1997. Boreal for-
Similarly, in Canada, the local cychrine species Scaphi- ests. Ecological Bulletins 46:16–47.
notus marginatus was restricted to mature forest Greenslade, P. J. M. 1964. Pitfall trapping as a method for studying
populations of Carabidae (Coleoptera). Journal of Animal Ecology
(Spence et al. 1996). To maintain populations of such 33:301–310.
species in the landscape, it is imperative to preserve Haila, Y. 1999. Islands and fragments. Pages 234–264 in M. L. Hunter
large, intact forest tracts (Halme & Niemelä 1993). Jr., editor. Maintaining biodiversity in forest ecosystems. Cam-
These conclusions are supported by findings from the bridge University Press, Cambridge, United Kingdom.
Amazonian rainforest, where a forest area of 500–1000 Haila, Y., I. K. Hanski, J. Niemelä, P. Punttila, S. Raivio, and H. Tukia.
1994. Forestry and boreal fauna: matching management with natu-
ha is required for maintaining an intact terrestrial inver- ral forest dynamics. Annales Zoologici Fennici 31:187–202.
tebrate assemblage (Didham 1997b). Halme, E., and J. Niemelä. 1993. Carabid beetles in fragments of conif-
We found that numbers of carabids did not decrease erous forest. Annales Zoologici Fennici 30:17–30.
toward the edge in the forest fragment. This pattern may Hänggi, A., and B. Baur. 1998. The effect of forest edge on ground-liv-
change over the years, because forest fragmentation ing arthropods in a remnant of unfertilized calcareous grassland in
the Swiss Jura mountains. Mitteilungen der Schweizerischen Ento-
started by humans appears to continue through natural mologischen Gesellschaft, Bulletin de la Société Entomologique Su-
forces. Trees at edges tend to fall over in strong winds, isse 71:343–354.
reducing the area of core habitat and opening up the Harris, L. D. 1984. The fragmented forest: island biogeography theory

Conservation Biology
Volume 15, No. 2, April 2001
Heliölä et al. Carabid Beetles in Boreal Forest Edges 377

and the preservation of biotic diversity. University of Chicago Ground-dwelling spiders (Arachnida, Araneae) in fragments of old
Press, Chicago. forest and surrounding managed forests in southern Finland. Ecog-
Helle, P., and J. Muona. 1985. Invertebrate numbers in edges between raphy 18:62–72.
clear-fellings and mature forests in Northern Finland. Silva Fennica Peltonen, M., and K. Heliövaara. 1998. Incidence of Xelechinus pilo-
19:281–294. sus and Cryphalus saltuarius (Scolytidae) in forest-clearcut edges.
Jokimäki, J., E. Huhta, J. Itämies, and P. Rahko. 1998. Distribution of ar- Forest Ecology and Management 103:141–147.
thropods in relation to forest patch size, edge, and stand character- Punttila, P., Y. Haila, T. Pajunen, and H. Tukia. 1991. Colonisation of
istics. Canadian Journal of Forest Research 28:1068–1072. clearcut forests by ants in the southern Finnish taiga: a quantitative
Jongman, R. H. G., C. J. F. Ter Braak, and O. F. R. Van Tongeren, edi- survey. Oikos 61:250–262.
tors. 1995. Data analysis in community and landscape ecology. Punttila, P., Y. Haila, J. Niemelä, and T. Pajunen. 1994. Ant communi-
Cambridge University Press, Cambridge, United Kingdom. ties in fragments of old-growth taiga and managed surroundings.
Kapos, V. 1989. Effects of isolation on the water status of forest Annales Zoologici Fennici 31:131–144.
patches in the Brazilian Amazon. Journal of Tropical Ecology 5: Ranney, J. W., M. C. Bruner, and J. B. Levenson. 1981. The importance
173–185. of edge in the structure and dynamics of forest islands. Pages 67–
Kotze, D. J., and M. J. Samways. 1999. Invertebrate conservation at the 96 in R. L. Burgess and D. M. Sharpe, editors. Forest islands dynam-
interface between the grassland matrix and natural Afromontane ics in man-dominated landscapes. Springer Verlag, New York.
forest fragments. Biodiversity and Conservation 8:1339–1363. Ranta, P., T. Blom, J. Niemelä, E. Joensuu, and M. Siitonen. 1998. The
Krebs, C. J. 1989. Ecological methodology. Harper & Row, New York. fragmented Atlantic rain forest of Brazil: size, shape and distribu-
Kruger, S. C., and M. J. Lawes. 1997. Edge effects at an induced forest- tion of forest fragments. Biodiversity and Conservation 7:385–403.
grassland boundary: forest birds in the Ongoye Forest Reserve, Rassi, P., H. Kaipiainen, I. Mannerkoski, and G. Ståhls. 1992. Report on
KwaZulu-Natal. South African Journal of Zoology 32:82–91. the monitoring of threatened animals and plants in Finland. Ko-
Lindroth, C. H. 1985. The Carabidae (Coleoptera) of Fennoscandia and miteamietintö 1991: 30. Ympäristöministeriö, Helsinki (in Finnish
Denmark. Fauna Entomologica Scandavica 15(1):1–225. with English summary).
Lindroth, C. H. 1986. The Carabidae (Coleoptera) of Fennoscandia and Risser, P. G. 1995. The status of the science of examining ecotones.
Denmark. Fauna Entomologica Scandinavica 15(2):233–497. BioScience 45:318–325.
Ludwig, J. A., and J. F. Reynolds. 1988. Statistical ecology. Wiley, New Saunders, D. A., R. J. Hobbs, and C. R. Margules. 1991. Biological con-
York. sequences of ecosystem fragmentation: a review. Conservation Bi-
Magura, T., and B. Tothmeresz. 1997. Testing edge effect on carabid ology 5:18–32.
assemblages in an oak-hornbeam forest. Acta Zoologica Academiae Siitonen, J., and P. Martikainen. 1994. Occurrence of rare and threat-
Scientarum Hungaricae 43:303–312. ened insects living on decaying Populus tremula: a comparison be-
Matlack, G. R., and J. A. Litvaitis. 1999. Forest edges. Pages 210–233 in tween Finnish and Russian Karelia. Scandinavian Journal of Forest
M. L. Hunter Jr., editor. Maintaining biodiversity in forest ecosys- Research 9:185–191.
tems. Cambridge University Press, Cambridge, United Kingdom. Spence, J. R., and J. Niemelä. 1994. Sampling carabid assemblages with
Meiners, S. J., and S. T. A. Pickett. 1999. Changes in community and pitfall traps: the madness and the method. Canadian Entomologist
population responses across a forest-field gradient. Ecography 22: 126:881–894.
261–267. Spence, J. R., D. Langor, J. Niemelä, H. Cárcamo, and C. Currie. 1996.
Murcia, C. 1995. Edge effects in fragmented forests: implications for Northern forestry and carabids: the case for concern about old-
conservation. Trends in Ecology and Evolution 10:58–62. growth species. Annales Zoologici Fennici 33:173–184.
Niemelä, J. 1993. Mystery of the missing species: species-abundance Stevens, S. M., and T. P. Husband. 1998. The influence of edge on
distribution of boreal ground-beetles. Annales Zoologici Fennici small mammals: evidence from Brazilian Atlantic forest fragments.
30:169–172. Biological Conservation 85:1–8.
Niemelä, J. 1997. Invertebrates and boreal forest management. Conser- ter Braak, C. J. F., and P. Smilauer. 1998: CANOCO for Windows. Ver-
vation Biology 11:601–610. sion 4.0. Centre for Biometry, Wageningen, The Netherlands.
Niemelä, J., Y. Haila, E. Halme, T. Lahti, T. Pajunen, and P. Punttila. Usher, M. B., J. P. Field, and S. E. Bedford. 1993. Biogeography and di-
1988. The distribution of carabid beetles in fragments of old conif- versity of ground-dwelling arthropods in farm woodlands. Biodiver-
erous taiga and adjacent managed forest. Annales Zoologici Fennici sity Letters 1:54–62.
25:107–119. Väisänen, R. 1995. Perhoset. Pages 103–108 in S. Raivio, editor. Talous-
Niemelä, J., D. Langor, and J. R. Spence. 1993. Effects of clear-cut har- metsien luonnonsuojelu—yhteistutkimushankkeen väliraportti. Met-
vesting on boreal ground-beetle assemblages (Coleoptera: Cara- sähallinnon Luonnonsuojelujulkaisuja, sarja A, 43 (in Finnish).
bidae) in western Canada. Conservation Biology 7:551–561. Voller, J. 1998. Managing for edge effects. Pages 215–233 in J. Voller
Niemelä, J., J. R. Spence, D. Langor, Y. Haila, and H. Tukia. 1994. Log- and S. Harrison, editors. Conservation biology principles for for-
ging and boreal ground-beetle assemblages on two continents: im- ested landscapes. University of British Columbia Press, Vancouver,
plications for conservation. Pages 29–50 in K. Gaston, M. Samways, Canada.
and T. New, editors. Perspectives in insect conservation. Intercept von Sacken, A. 1998. Interior habitat. Pages 130–145 in J. Voller and S.
Publications, Andover, Massachusetts. Harrison, editors. Conservation biology principles for forested land-
Pajunen, T., Y. Haila, E. Halme, J. Niemelä, and P. Punttila. 1995. scapes. University of British Columbia Press, Vancouver, Canada.

Conservation Biology
Volume 15, No. 2, April 2001