Biodiversity and Conservation 12: 1815–1831, 2003.

 2003 Kluwer Academic Publishers. Printed in the Netherlands.

Sampling the rarest: threatened beetles in boreal

forest biodiversity inventories

PETRI MARTIKAINEN* and JARI KOUKI

Faculty of Forest Sciences, University of Joensuu, P.O. Box 111, FIN-80101 Joensuu, Finland; * Author
for correspondence (e-mail: petri.martikainen@ joensuu.fi; fax: 1358 -13 -2514444)

Received 30 November 2001; accepted in revised form 25 September 2002

Key words: Biodiversity inventories, Boreal forest, Coleoptera, Finland, Representativeness, Reserve
selection, Sampling methods, Saproxylic beetles, Threatened species, Trapping efficiency

Abstract. Presence or absence of threatened species in samples is information that is widely used in
designing and implementing conservation actions. We explored the effectiveness of beetle (Coleoptera)
inventories and contribution of different sampling methods in revealing occurrences of threatened and
near threatened species in boreal forests. The number of species caught using traps in a particular area
proved to be a useful indicator of the representativeness of data, the relationship between total number of
species and the number of threatened and near threatened species being almost exponential. Samples
containing less than 200 trapped species (or 2000 individuals) are almost useless in surveying threatened
and near threatened species. The probability of finding such species increases considerably when the
number of trapped species exceeds 400. Window traps attached directly on the trunks of dead trees proved
to be the most efficient sampling method in trapping threatened beetles, whereas many other standard
methods gave relatively poor results. We suggest that the best alternative in surveying threatened species
in boreal forests is a combination of intensive direct searching and trunk window traps. Finding
threatened beetles with rigorous probability requires very large sample sizes, even if the most effective
sampling methods are used. For example, ranking 10 boreal forest areas to be protected according to the
occurrence of threatened species with some reliability may require trapping of over 100 000 beetle
individuals. Collecting and identifying these large samples routinely in conservation actions is not
feasible, which means that shortcuts (indicators etc.) are necessary. However, a lot of good-quality
inventories with appropriate sampling efforts are needed before these shortcuts can be identified and
elaborated. Such inventories are also crucial for the improvement of the classification of threatened
species and full assessment on how past forest management has eventually affected the biota.

Introduction

Worldwide patterns of biodiversity are still rather poorly documented. Most of the
studies have been able to assess the general patterns from the taxonomic groups with
large-sized species, such as birds, mammals and vascular plants (e.g. ICBP 1992;
Myers et al. 2000). Yet these groups form only a marginal fraction of the global
species richness (Heywood 1995), and assuming that they reflect species richnesses
also of other, less well studied but definitely more species-rich groups may be
doubted (May 1988; Lawton et al. 1998). The analyses on global biodiversity
patterns have pointed out biodiversity hotspots (Myers et al. 2000), and sophisti-
cated procedures for designing and selecting conservation areas, both within the
hotspots and other areas, have been developed (Margules et al. 1988). However,
1816

Table 1. Number of regionally extinct (RE), threatened and near threatened (NT) forest beetles in
Finland (Anonymous 2000; Rassi et al. 2000).
Regionally extinct Threatened Near threatened Total
(RE) CR EN VU (NT)
Saproxylic species 12 33 48 56 65 214
Other species 9 6 13 27 23 78
Total 21 39 61 83 88 292
Threatened species are further divided into critically endangered (CR), endangered (EN) and vulnerable
(VU) species. Saproxylic species (species dependent on dead wood) and other species are treated
separately.

their true applicability critically depends on the underlying data on the occurrence of
species in the potential conservation areas (Cabeza and Moilanen 2001).
Beetles (Coleoptera) make up a large proportion of species diversity in several
ecosystems from the tropics to the boreal zone. They also have a special role in the
recent estimates of global biodiversity patterns. The first new but rough estimates of
global species diversity were mainly based on tropical beetles (Erwin 1982). These
estimates showed that the species richness must be remarkably higher than what was
previously thought. Unfortunately the high species diversity imposes challenges on
revealing the diversity patterns: a large proportion of the beetle species are rarities in
the species assemblages. Their detection requires considerable sampling efforts
(McArdle 1990; Novotny´ and Basset 2000). Yet these rarities are often also the most
threatened, requiring special attention from the viewpoint of species conservation.
Although the tropical forest beetles have received much attention, the beetles
occurring in boreal forests are often better known, and thus they have potential to
shed new light on the problems involved in inventories of species-rich taxonomic
groups. In Fennoscandia beetles have been a subject of intensive research for a long
time, and particularly so during the past 10–20 years. This has been mainly due to
the fact that forest beetles form the largest single group of organisms that are
Red-listed in this region (Rassi et al. 2000; see Table 1).The knowledge on forest
beetles has improved significantly recently, partly because of the increased use of
intensive and quantitative trapping methods (e.g. Siitonen 1994; Økland et al. 1996;
Kaila et al. 1997; Jonsell et al. 1998; Muona 1999; Martikainen 2000; Kouki et al.
2001).
Forest beetles have been trapped for a variety of purposes, using variable
methods. In many surveys rare and threatened species have received most attention.
Presence or absence of such species in the samples has been used e.g. to indicate the
conservation value of a certain forest area, or to demonstrate the effects of intensive
forest management on biodiversity. However, it is not always easy to estimate how
the quality and the quantity of the data, sample size in particular, may affect the
results. It is clear that a small sample usually contains fewer threatened species than
does a large sample, but what would be a sufficient sample size and how should the
sample be collected? A considerable amount of new inventory data has accumulated
recently, but the data have been collected and reported in highly variable ways.
In this paper we estimate the effectiveness of beetle inventories in revealing the
 1817

occurrences of rarities: threatened and near threatened species, as these are the most
important ones in designing and implementing conservation actions. For this
purpose we compiled results from published surveys that have been conducted in
more or less natural forests in Finland during the past few decades. We compare
these data to a large amount of beetle material collected from a reference forest area
with the highest number of forest-dwelling beetle species recorded in Finland so far.
For practical reasons we focus only on studies dealing with the whole beetle fauna.
Although beetle inventories do not always cover all groups of beetles and species
pools, and lists of threatened species vary from one country to another even within
the boreal zone, we believe that our results can be helpful and used as a rough
guideline in many regions and sampling situations dealing with threatened forest
beetles. On the other hand, our findings cannot be applied to studies that are not at
all concerned with threatened species.

Materials and methods

We first reviewed all recently published inventories dealing with forest beetles in
Finland. Two inventories made in Russian Karelia just next to the Finnish border
were also included. To make the data more comparable, we only accepted studies
where all or nearly all beetle families had been sampled and identified to species
level, and which originated from natural or semi-natural forests that currently are the
main concern in the conservation actions. However, we included samples not only
from old-growth forests, but also from disturbed habitats, such as burned stands
located near the Russian border or clearcuts with retained trees adjacent to valuable
forest areas. Managed forests in managed forest landscape were rejected. In some
studies only part of the data fulfilled the criteria and were included. Despite these
restrictions, the data are still rather heterogeneous. For example, the size of the area
sampled varies considerably, from individual stands to whole national parks. In
some cases we combined data collected from separate stands, if the distance
between stands was not large (usually ,5 km). The studies often included several
distant areas that were treated separately. The studies included are summarised in
Table 2.
From the studies we recorded the following variables for each area studied:
• total number of individuals caught (not always available),
• total number of species recorded and how many species were caught in traps (if
this was not indicated, we assumed that 90% of the species were trapped),
• number of threatened and near threatened species (called NT species below)
found and methods how they were captured,
• sampling methods.

Almost all studies were made prior to the latest evaluation of threatened species in
Finland (Rassi et al. 2000). We re-classified the observed species as threatened and
NT beetle species according to the latest national Red List, which follows the
general principles and procedures developed by the IUCN (1994). Additional
 1818

Table 2. Summary of the studies included in analyses.

Description of forests Latitude Sampling Number of Number of Number of Number of Number of NT Reference
(8N) methods a areas individuals/ species/area threatened species
area species

Old forest 61 pf, df, ds 1 1951 154 0 1 Bistrom

¨ (1978)
Old forest 63 pf, df, ds 1 ? 410 16 4 Bistrom ¨ ¨
¨ and Vaisanen (1988)
Burned areas 63 pf, twt, ds 2 3568–10931 296–536 1–2 2–7 Rutanen (1994b)
Old forests 60–61 pf, twt, cn, ds 6 ? 265–644 0–8 0–7 Rutanen (1994a)
Old forests and burned area 61–62 pf, twt, cn, ds 6 ? 142–573 0–7 0–11 Rutanen (1995)
Forests of various ages 62 fwt 1 14829 507 1 5 Siitonen et al. (1995)
Timberline forests 68–69 fwt 4 475–2233 78–118 0 0 Sippola et al. (1995)
Old forests 61–62 fwt 4 55–232 39–74 0 0 Raivio and Mannerkoski (1997)
Old forests and burned areas 64–65 pf, twt, light, bait, ds 8 ? 217–371 0–2 1–5 Rutanen and Kashevarov (1997)
Old forest 63 pf, twt 1 31607 373 1 1 Bistrom
¨ et al. (1998)
Mainly old forests 66 pf, fwt, twt, cn, ds 1 .200000 630 28 24 Muona (1999)
Old forests 62–63 pf, twt, ds 16 121–554 45–110 0–1 0–2 ¨
Paivinen et al. (1999)
Old forests 62 twt 9 1097–2844 145–193 0–1 0–3 Martikainen et al. (2000)
Old forest 66 pf, fwt, ds 1 7000 215 5 4 Muona (2000)
Old forests and clearcuts 62 and 66 twt 2 6271–17558 403–676 4–11 7–16 Martikainen (2001)
Old forest 66 twt, ds 1 .5000 .300 3 4 Martikainen and Rassi (2001)
Forests of all ages 63 fwt 4 131–3081 58–260 0–4 0–4 Simila¨ et al. (2002a)
Total 68 66 46
a
pf 5 pitfall trap; df 5 dry funnel; fwt 5 freely hanging window trap; twt 5 trunk window trap; cn 5 car net; light 5 light trap; bait 5 bait trap (for moths); ds 5
direct searching.
 1819

information on sampling methods and occurrence of threatened species was asked

directly from the authors when necessary.
Sampling methods and procedures varied among the studies, and this gave us
opportunities also to evaluate these methodological differences on survey results.
We consider here as direct searching sieving, netting, peeling of bark and all other
searching of beetles by the entomologists themselves, and as trapping we count
different kinds of window traps, pitfall traps, light traps, bait traps, dry funnels (for
soil samples) and car net. For a general description of the methods, see e.g.
Southwood (1978). Window traps we divide further into traps that are supposed to
catch beetles randomly and those that are placed so that attraction is likely
(Southwood 1978). In the former group traps were mostly freely hanging window
traps, distributed randomly or systematically to the study area; in the latter group
traps were intentionally placed directly on or very close to the trunks of dead trees,
these we call trunk window traps (e.g. Bakke and Kvamme 1993; Kaila 1993;
Økland 1996).
The results of these surveys are compared with data collected from a reference
forest area containing old growth with plenty of aspen, located in Savonranta, SE
Finland. Two extensive inventories of the beetle fauna have been conducted in this
area during the 1990s (Rutanen 1995; Martikainen 2001; see Table 2). More than
25 000 beetle individuals, representing 838 species, have been collected and
identified from this area (Table 3). This is the highest number of forest species
recorded from a single forest area in Finland so far, despite the fact that the
inventory is still far from complete. Most of this material is collected from a 21 ha
old-growth stand and from a 14 ha clearcut with plenty of retained aspens just next
to the old growth. A minor part of the data (4380 individuals), which added 38
species to the list, come from another similar clearcut, located ca. 4 km from the
other stands. Rutanen (1995) used several sampling methods including car net in the
reference area, whereas Martikainen (2001) collected his data using 60 large trunk
window traps on dead aspen trees. A few species have been recorded after these
studies and are included as well (see Table 3). Data from the two studies carried out
in the reference area have been kept separately in the analyses and figures, if not
otherwise indicated. In addition, a data point for the pooled material of the reference
area is presented in some figures (Figures 1–3). Accumulation of threatened and NT
species in the 60 trunk window traps used by Martikainen (2001) in the reference
area is shown in Figures 2b and 3b, calculated from 200 random combinations of the
traps.
Information on the sizes of threatened and NT species was gathered from the
literature, the main source being Freude et al. (1965–1983).

Results

Accumulation of threatened species

The relationship between total sample size and number of species found followed
1820

Table 3. Comparison data from the reference area. Number of species not caught in any kind of trap in
parentheses.
Data source Trap Number of Total Number of Number of
days individuals number threatened NT species
in traps of species
species
Rutanen (1995) 3140 .7157 572 (?) 7 (2) 11 (3)
Martikainen (2001) 8100 17558 676 (0) 11 (0) 15 (0)
New species collected later 135 501 7 (4) 2 (2) 1 (1)
Total 11 375 .25 216 838 (,50) 16 (3) 19 (1)

the same curve (not drawn) in most of the studies, indicating that the forests studied
have a rather similar species–abundance distribution of beetles (Figure 1). The
number of species in Puijo (Bistrom¨ et al. (1998) in Table 2) is low because some
abundant staphylinid species in pitfall traps dominated the whole sample. Un-
fortunately, most of the data points for areas with more than 200 species sampled are
missing because only the number of species found was given by the authors (Table
2). Species richness typically decreases towards the north and the few areas situated
in the northern boreal zone (including Oulanka national park; Muona 1999; see
Table 2) are below the curve. In most areas the number of species sampled is clearly
less than 200 and the sample size is below 2000 individuals. A noteworthy pattern is
that the accumulation curve does not seem to reach an asymptote, indicating that the
true species richness may be clearly higher than could be revealed with even the
largest samples. This shows that true species diversity in the study areas may still be
relatively poorly known.
Numbers of threatened and NT species increase rather slowly when the total

Figure 1. Relationship between total sample size and number of species found in the areas where total
sample size is known. Data points for the reference area are, from left to right, Rutanen (1995),
Martikainen (2001) and pooled data (see Table 3). Data points named Oulanka and Puijo refer to
studies Muona (1999) and Bistrom¨ et al. (1998) in Table 2.
 1821

number of species found increases (Figures 2a and 3a), a result that is expected since
these species are typically very rare. Pyha-Hakki¨ ¨ and Oulanka national parks
¨ and Vaisanen
(Bistrom ¨ ¨ (1988) and Muona (1999) in Table 2) are, however,
exceptions. A large proportion of data from these two areas have been sampled with
extensive direct searching during the 1900s, and the difference is likely to reflect the
methodological difference. When only species caught in traps are included, these
two areas no longer deviate from the others (Figures 2b and 3b). It is noteworthy
that the accumulation curves based on samples collected from dead aspens in the
reference area (Martikainen 2001) run above almost all other data points, forming
apparently an upper limit that can be reached through the use of traps only if the
locality is exceptionally good.
Although the relationship between total species richness and number of threat-
ened and NT species in the reference area is almost linear (Figures 2b and 3b), the
situation is quite different in the other areas. When similar curves are drawn for the
other studies, based on moving average of nine areas, it becomes clear that the
relationship is almost exponential (Figure 4). This pattern remains essentially
similar even if studies where trunk window traps have not been used are excluded
(data not shown here). In the light of these data it seems that samples must contain
on average more than 200 trapped species (corresponding to roughly 2000 in-
dividuals) to reveal the first threatened or NT species. The probability of finding
such species seems to increase considerably when the number of trapped species
exceeds 400.
General species–abundance distribution models dictate that normally increasing
the number of species recorded from a particular site is fairly easy at first, but
becomes more and more laborious when the number of species found increases.

Figure 2. (a) Relationship between total number of species and number of threatened species found in
different areas. (b) same as in (a) but only species caught using traps are included. Filled circles
represent different data sets from the reference area, from left to right: Rutanen (1995), Martikainen
(2001) and pooled data (see Table 3). Small black dots show the accumulation of threatened species in
trunk window traps on dead aspens in the reference area when the number of traps is increased. Data
¨ ¨
points named Oulanka and Pyha-Hakki refer to the studies Muona (1999) and Bistrom¨ and Vaisanen
¨ ¨
(1988) in Table 2.
1822

Figure 3. (a) Relationship between total number of species and number of NT species found in different
areas. (b) same as in (a) but only species caught using traps are included. Filled circles represent
different data sets from the reference area, from left to right: Rutanen (1995), Martikainen (2001) and
pooled data (see Table 3). Small black dots show the accumulation of NT species in trunk window
traps on dead aspens in the reference area when the number of traps is increased. The data point named
Oulanka refers to the study Muona (1999) in Table 2.

Figure 4. Number of threatened and NT species observed as a function of number of species trapped.
All studies and areas are included, except data of Martikainen (2001) from the reference area. The
curves are based on moving average of nine areas to smooth out variation between individual areas. To
calculate this, all the areas were first ranked according to the number of species trapped. Means were
then calculated for every combination of nine consecutive areas.

This is especially true in sampling threatened species, as demonstrated with the data
from aspens in the reference area (Figure 5). At first almost each new trap gives
records of new threatened species, but fairly soon the effort required for observation
of new species becomes very big. In these data, the eleventh threatened species
required the addition of 13 new traps and ca. 4000 identified individuals (and ca. 50
species). The first 13–14 traps and ca. 4000 individuals yielded altogether ca. 400
species, of which six were threatened ones.
 1823

Figure 5. Accumulation of threatened species in relation to number of species and individuals caught in
trunk window traps on dead aspens in the reference area [original data from Martikainen (2001)].
Numbers above symbols indicate number of threatened species and number of traps required.

Effectiveness of different sampling methods

We also compared how the different sampling methods can reveal threatened
species. Because the documentation of sampling details varies considerably in the
studies, the results presented in Table 4 can only be used to reveal rough differences
between the methods. Especially the role of direct searching is undervalued here,
because studies based solely on direct searching were omitted. Furthermore, it is not
possible to estimate the time spent in direct searching in the studies included
because this information is not given in the data sources. Keeping these limitations
in mind, the pitfall traps seem to be poor in catching threatened and NT species:
1200 traps operated over one season yielded only one threatened species, a longhorn
beetle Monochamus urussovii (Fischer v. Waldheim) (Cerambycidae) and three
NT species: Cercyon emarginatus Baranowski (Hydrophilidae), Catops longulus Kell-
ner (Catopidae) and a rove beetle Paranopleta inhabilis (Kraatz) (Staphylinidae).
Results concerning car net are fairly similar, although the data are rather scattered:
over 100 hours of driving gave only two forest-dwelling NT species: a featherwing
beetle Micridium halidaii (Matthews) (Ptiliidae) and a false click beetle Rhacopus
sahlbergi (Mannerheim) (Eucnemidae). The role of dry funnel and light and bait
traps remains rather unclear, because these methods were used only accidentally in
these studies. Tragosoma depsarium (Linnaeus), a vulnerable longhorn beetle
(Cerambycidae) was once caught in a light trap.
Window traps were the most effective trap type in sampling threatened and NT
beetles. Locating these traps close to or directly on trunks of dead trees increased
the effectiveness markedly (Table 4), especially in the case of species ranked high
in endangerment. Randomly placed freely hanging window traps gave a relatively
poor result. Although a comparison between direct searching and trunk window
 1824

Table 4. Number of observations of threatened and NT species made using different sampling methods
Sampling method Number Number Threatened species Near threatened species
of studies of trap
(total n 5 17) seasons
or samples,
rough
estimates
CR EN VU NT
Direct searching 10 ? 10 17 43 59
Trunk window traps 10 500 seasons 1 14 27 80
Freely hanging window traps 6 850 seasons – 2 13 29
Pitfall traps 10 1200 seasons – – 1 3
Car net 3 150 hours – – – 2
Dry funnel (soil samples) 2 150 samples – – – –
Light trap 1 20 seasons – – 1 –
Bait trap 1 low – – – –
 1825

traps based on this material may not be very quantitative (though the number of
studies is quite similar), it still shows that trunk window traps can be a comparable
method to direct searching in surveying the occurrence of threatened and NT
species in many cases. The most remarkable difference was in critically en-
dangered species, which were practically absent in all kinds of traps but were
detected by direct searching.
Small threatened and NT species were proportionally more often caught in trunk
window traps than found by direct searching, whereas the opposite was true for
larger species (Figure 6). Data from freely hanging window traps were too scarce
to warrant any conclusions. Most of the observations (7 out of 11) of critically
endangered species concerned species larger than 6 mm; the only critically
endangered species caught in trunk window traps was Cyrtanaspis phalerata
(Germar) (Anaspidae), which is a fairly small species (3.7 mm). This pattern is
expected, as direct searching relies on visual detection of specimens, and small-
sized species easily escape collector’s attention.

Discussion

Sampling threatened species

Threatened and near threatened species are almost by definition rare, and sampling
rare species is always challenging. Theoretical models based on empirical data
(e.g. Fisher et al. 1943; Preston 1948; Hughes 1986) have shown that the
abundance among species varies by several orders of magnitude, which in practice
means that rare and threatened species are inevitably scarce in small samples, and
the presence of such species is largely dictated by chance alone. What makes the
situation even more complicated is that threatened forest beetles can be twice as
difficult to catch than other beetle rarities in the forests (Muona 1999). It is thus

Figure 6. Proportion of records of threatened and near threatened species obtained using different
sampling methods in different size classes of beetles. Numbers above symbols indicate total number of
records in each size class.
1826

not surprising that, according to our data, catches containing less than 200 trapped
species (or less than 2000 individuals) are almost useless in surveying threatened
and NT beetle species in boreal forests. The number of trapped species should be
over 300–400 (corresponding to .4000 individuals) to find some threatened
species with certainty, but still most of the threatened species present in the study
area will be missed. Increasing the number of species (i.e. sample size) further to
increase the probability of catching threatened species would thus be highly
desirable, but often impossible in practice (see McArdle 1990). As Muona (1999)
has stated, even a sample of 20 000 individuals is far from adequate in surveying
threatened species.
The total number of species recorded in a particular area proved to be a fairly
good, albeit robust estimate on the representativeness of data, whenever the main
aim is to compare numbers of threatened species. Number of individuals is another
potential estimate for the representativeness of data. However, sample size can be
seriously affected by one or few abundant species such as Zyras humeralis
(Gravenhorst) (Staphylinidae), which sometimes make up a large proportion, even
up to 70% of the total catch of forest beetles (e.g. in Simila¨ et al. 2002b). Thus,
two comparable samples containing the same number of species but differing in
the abundance of one or a few species by 10 000 individuals may be rather equal in
revealing occurrences of threatened species, whereas a sample containing 100
species more than another, otherwise similar, sample is likely to give a con-
siderably better picture of the threatened species occurring in the study area.
Number of species has several advantages compared to many other estimates: it
is almost always readily available and if not, easy to calculate for almost all kinds
of inventory data. Superabundant species do not affect it in any way, and it does
not necessarily require standardisation as regards trapping period, number of traps,
trap sizes etc., because an increase in these factors is likely to increase the number
of species caught. Furthermore, the species richness of forest beetles is rather
similar over relatively large geographical areas (such as Fennoscandia) within the
subzones of the boreal zone (Stokland 1992; Siitonen 2001), which further
increases the usefulness of this measure. No matter what is the total sample size or
what trapping methods have been used, a small number of trapped species
immediately reveals that the sample is not sufficient for catching threatened
species.
Although the number of species gives a fairly good picture of the representative-
ness of data, it is still wise to examine how the data were collected, because there
are huge differences in the efficiency of trap types in catching threatened forest
species. The fact that ca. 3 / 4 of threatened species in the boreal forest of Finland
are saproxylic (dependent on dead trees) (Anonymous 2000) explains some of the
differences. For instance, pitfall traps mainly catch ground beetles (Carabidae),
rove beetles (Staphylinidae) and other species living on the ground, most of which
are not dependent on dead wood and not threatened. Thus, the probability of
catching threatened species using pitfall traps on the forest floor is relatively small
and the same applies to dry funnelling of soil samples. However, these methods
may give good results when used in surveying threatened beetles in tree cavities
 1827

(Ranius and Jansson 2002). Car net is an effective method in catching certain
species that are difficult to capture otherwise (Kronblad and Lundberg 1978;
Rutanen and Muona 1982), but threatened species seem to be rather scarce in car
net samples. The role of light and bait traps remains largely unclear due to the
small amount of data. The observation that trunk window traps (including traps
located close to but not directly on dead trees) captured several times more
threatened species than randomly located, freely hanging window traps clearly
demonstrates that the samples are fundamentally affected by the trap location (see
also Martikainen 2000). However, this effect may be difficult to demonstrate in
many cases, especially if all species are included and the scale is not fine enough
(i.e. a few metres) (cf. Siitonen 1994; Økland 1996; Økland et al. 1996).
Our results also show that experienced coleopterists cannot be substituted in
faunal surveys, particularly if the aim is to detect critically endangered species. On
the other hand, use of traps is advantageous in obtaining data on small threatened
species. Small species are more difficult to find by direct searching than large
ones, and their abundance per unit area is higher than that of larger species (Stork
and Blackburn 1993), which could explain their better perceptivity using traps.
Furthermore, as trunk window traps yielded a fairly high number of observations
of threatened and NT species, comparable to that of direct searching, it can be
concluded that this trapping method can be used as complementary and potentially
even as an alternative method to direct searching in surveying threatened species,
unlike many other standard methods commonly used in beetle studies in forests
(see Muona 1999). However, the fact that the efficiency of trunk window traps is
more or less based on attraction introduces a significant source of error when
different kinds of stands (e.g. with varying amounts of dead wood) are compared
(Southwood 1978; Martikainen 2000), but similar errors are present in direct
searching, too.
How, then, should threatened boreal forest beetles be sampled using a
reasonable effort? The answer is naturally highly dependent on the objectives of
each study, but some general prescriptions can be given. Several authors have
suggested that in order to get a good coverage of the species present in a particular
area, many trapping methods should be used (e.g. Kohler ¨ 1996; Økland 1996;
Muona 1999). Because most standard methods are poor in catching threatened
species, we suggest that the best alternative in surveying threatened beetles might
be a combination of intensive direct searching and trunk window traps. As Figure
5 shows, traps attached on one substrate type only fairly soon reveal most of the
threatened species possible to obtain with a finite number of traps on that substrate.
The best trapping result may be attained if the available traps are placed on several
potential habitat types preferred by different threatened species. Besides different
combinations of tree species and type of decay, different environments should also
be included, because recent studies have shown that dead wood in sun-exposed
habitats is favoured by many threatened species (e.g. Jonsell et al. 1998; Kouki et
al. 2001; Martikainen 2001). It may also be beneficial to concentrate on the groups
of beetles which contain the majority of threatened species and to continue
¨
trapping for several years (Kohler 1996; Martikainen et al. 1998).
1828

Practical implications

The number of beetles which have been trapped from forests for various purposes
during the past few decades in Finland and elsewhere in the boreal zone is
probably several millions. Quite often the motivation is somehow connected to the
conservation of threatened species, but in many cases it is obvious from the
beginning that the sampling scheme is not appropriate in producing sufficient data
on such species. Conclusions based on limited sampling efforts may easily be
misleading. Most of the abundant beetle species, non-saproxylic species in
particular, occurring in old boreal forests are the same regardless of the
management history (Martikainen et al. 2000), whereas the important differences
from the conservation viewpoint are expressed in the numbers of rare and
threatened (usually saproxylic) species. As the majority of threatened species will
be missed in smaller samples, it means that such samples mainly consist of
common species actually present in most of the forests. Using this kind of data to
rank conservation priorities may give almost random results. Finding threatened
species with a rigorous probability requires very large sample sizes if traps are
used (McArdle 1990; Muona 1999). For example, ranking 10 boreal forest areas to
be protected according to the occurrence of threatened and NT beetle species with
some reliability (e.g. .500 observed species per area) would require trapping of
100 000 beetle individuals, even if the most effective sampling methods were used.
Processing such large samples is very time-consuming, especially because the
number of scientist-hours required is inversely related to the mean body length of
the study organisms and beetles are relatively small (Lawton et al. 1998).
Our results demonstrate that obtaining sufficient data on the occurrence of
threatened beetles in focal areas by trapping may be too difficult to be routinely
used in conservation actions such as reserve design. On the other hand, as beetles
form one of the major groups among threatened forest species, they cannot be
totally ignored either. Shortcuts are apparently necessary for practical purposes. It
may be possible to find e.g. indicator taxa (Nilsson et al. 1994, 2000; Jonsson et al.
2001) or umbrella species (Lambeck 1997; Martikainen et al. 1998) indicating
beetle hotspots, but their wider applicability can nevertheless be limited (Kerr
1997; Lawton et al. 1998; Simberloff 1998). Certain stand characteristics, such as
the volume (Martikainen et al. 2000) and diversity (Simila¨ et al. 2003) of dead
wood tend to correlate positively with the species richness of beetles, but
especially without knowledge on the previous availability of dead wood (continui-
ty) these indicators may tell little about the occurrence of threatened species
(Jonsson et al. 2001; Stokland 2001).
There is no doubt that biodiversity patterns in the tropical areas are often poorly
known and that their detailed inventories would require huge scientific efforts
(Lawton et al. 1998). Our findings, however, show that the biodiversity patterns in
the boreal forests may also be relatively poorly documented. A lot of good-quality
inventories and experiments with appropriate sampling efforts are needed also in
boreal forests, before we can use any shortcuts with confidence. More large-scale
inventories are also needed to further improve the classification of threatened
 1829

species and to fully understand how the past intensive forest management in the
boreal zone has eventually affected the biota.

Acknowledgements

We are grateful to all those coleopterologists who collected the original beetle
¨
materials included in this study. Jyrki Muona, Jussi Paivinen and Ilpo Rutanen
kindly provided additional information concerning their data. This study was
funded by the Academy of Finland (Finnish Centre of Excellence Programme
2000–2005, project no. 64308).

References
¨
Anonymous 2000. Metsien suojelun tarve Etela-Suomessa ja Pohjanmaalla [Forest Protection in
Southern Finland and Ostrobothnia]. The Finnish Environment 437. Ministry of the Environment,
Helsinki, Finland.
Bakke A. and Kvamme T. 1993. Beetles attracted to Norway spruce under attack by Ips typographus.
Meddelelser fra Skogforsk 45: 1–24.
¨ O. 1978. Coleoptera and Heteroptera in natural forest in Mantyharju
Bistrom ¨ (Southern Finland).
Notulae Entomologicae 58: 95–100.
¨ O. and Vaisanen
Bistrom ¨ ¨ ¨
R. 1988. Ancient-forest invertebrates of the Pyhan-Hakki ¨ national park in
Central Finland. Acta Zoologica Fennica 185: 1–69.
¨ O., Savolainen E. and Savolainen R. 1998. Puijon vanhojen metsien kuoriaiset Coleoptera in
Bistrom
the old forests of Puijo (central Finland). Kulumus 13: 37–56.
Cabeza M. and Moilanen A. 2001. Design of reserve networks and the persistence of biodiversity.
Trends in Ecology and Evolution 16: 242–248.
Erwin T.L. 1982. Tropical forests: their richness in Coleoptera and other Arthropod species. The
Coleopterists Bulletin 36: 74–75.
Fisher R.A., Corbet A.S. and Williams C.B. 1943. The relation between the number of species and the
number of individuals in a random sample of an animal population. Journal of Animal Ecology 12:
42–58.
Freude H., Harde K.W. and Lohse G.A. 1965–1983. Die Kafer ¨ Mitteleuropas Vols. 1–11. Goecke &
Evers Verlag, Krefeld, Germany.
Heywood V.H. (ed.) 1995. Global Biodiversity Assessment. UNEP, Cambridge University Press,
Cambridge, UK.
Hughes R.G. 1986. Theories and models of species abundance. The American Naturalist 128: 879–899.
ICBP 1992. Putting Biodiversity on the Map: Priority Areas for Global Conservation. International
Council for Bird Preservation, Cambridge, UK.
IUCN 1994. IUCN Red List Categories. IUCN Species Survival Commission, IUCN, Gland,
Switzerland.
Jonsell M., Weslien J. and Ehnstrom ¨ B. 1998. Substrate requirements of red-listed saproxylic
invertebrates in Sweden. Biodiversity and Conservation 7: 749–764.
Jonsson M., Jonsell M. and Nordlander G. 2001. Priorities in conservation biology: a comparison
between two polypore-inhabiting beetles. Ecological Bulletins 49: 195–204.
Kaila L. 1993. A new method for collecting quantitative samples of insects associated with decaying
wood or wood fungi. Entomologica Fennica 4: 21–23.
Kaila L., Martikainen P. and Punttila P. 1997. Dead trees left in clear-cuts benefit saproxylic Coleoptera
adapted to natural disturbances in boreal forest. Biodiversity and Conservation 6: 1–18.
1830

Kerr J.T. 1997. Species richness, endemism, and the choice of areas for conservation. Conservation
Biology 11: 1094–1100.
¨
Kohler ¨
F. 1996. Kaferfauna in Naturwaldzellen und Wirtschaftswald, Band 6. Landesanstalt fur ¨
¨
Okologie, Bodenordnung und Forsten / Landesamt fur ¨
¨ Agrarordnung NRW, LOBF-Schriftenreihe.
¨
Kouki J., Lofman S., Martikainen P., Rouvinen S. and Uotila A. 2001. Forest fragmentation in
Fennoscandia: linking habitat requirements of wood-associated threatened species to landscape and
habitat changes. Scandinavian Journal of Forest Research Suppl. 3: 27–37.
˚
Kronblad W. and Lundberg S. 1978. Bilhavning ˚
– en intressant fangstmetod ¨ skalbaggar och andra
for
insekter. Entomologisk Tidskrift 99: 115–118.
Lambeck R.J. 1997. Focal species: a multi-species umbrella for nature conservation. Conservation
Biology 11: 849–856.
Lawton J.H., Bignell D.E., Bolton B., Bloemers G.F., Eggleton P., Hammond P.M. et al. 1998.
Biodiversity inventories, indicator taxa and effects of habitat modification in tropical forest. Nature
391: 72–76.
Margules C.R., Nicholls A.O. and Pressey R.L. 1988. Selecting networks of reserves to maximise
biological diversity. Biological Conservation 43: 63–76.
Martikainen P. 2000. Effects of forest management on beetle diversity, with implications for species
conservation and forest protection, Ph.D. Thesis, Faculty of Forestry, University of Joensuu,
Finland.
Martikainen P. 2001. Conservation of threatened saproxylic beetles: significance of retained aspen
Populus tremula on clearcut areas. Ecological Bulletins 49: 205–218.
¨
Martikainen P. and Rassi P. 2001. Syote-Life-projektin kovakuoriaislajiston perusselvitys. In: Lehtonen
¨
H. (ed.), Luontoa ja historiaa Syotteen ¨ Life-projektin perusselvitykset. Metsahal-
alueelta – Syote ¨
litus, Pohjanmaan-Kainuun luontopalvelut, pp. 119–129.
Martikainen P., Kaila L. and Haila Y. 1998. Threatened beetles in white-backed woodpecker habitats.
Conservation Biology 12: 293–301.
Martikainen P., Siitonen J., Punttila P., Kaila L. and Rauh J. 2000. Species richness of Coleoptera in
mature managed and old-growth boreal forests in southern Finland. Biological Conservation 94:
199–209.
May R.M. 1988. How many species are there on Earth? Science 241: 1441–1449.
McArdle B.H. 1990. When are rare species not there? Oikos 57: 276–277.
Muona J. 1999. Trapping beetles in boreal coniferous forest – how many species do we miss? Fennia
177: 11–16.
Muona J. 2000. Kovakuoriaisinventointi. In: Virkkala R. and Anttila I. (eds), Nature Inventory of
¨ ¨ ¨
Old-Growth Forests and Mires of Southern Kuusamo – Pajupuronsuo, Romevaara, Naranganvaara,
¨ ¨
Virmajoki. Alueelliset ymparistojulkaisut 153, Oulu, Finland, pp. 181–191.
Myers N., Mittermeier R.A., Mittermeier C.G., da Fonseca G.A.B. and Kent J. 2000. Biodiversity
hotspots for conservation priorities. Nature 403: 853–858.
Nilsson S.G., Arup U., Baranowski R. and Ekman S. 1994. Tree-dependent lichens and beetles as
indicators in conservation forests. Conservation Biology 9: 1208–1215.
Nilsson S.G., Baranowski R., Ehnstrom ¨ B., Eriksson P., Hedin J. and Ljungberg H. 2000. Svartoxen,
¨
Ceruchus chrysomelinus (Coleoptera, Lucanidae), en forsvinnande urskogsrelikt? [Ceruchus
chrysomelinus (Coleoptera, Lucanidae), a disappearing virgin forest relict species?]. Entomologisk
Tidskrift 121: 137–146.
Novotny´ V. and Basset Y. 2000. Rare species in communities of tropical insect herbivores: pondering
the mystery of singletons. Oikos 89: 564–572.
Økland B. 1996. A comparison of three methods of sampling saproxylic beetles. European Journal of
Entomology 93: 195–209.
Økland B., Bakke A., Hagvar˚ S. and Kvamme T. 1996. What factors influence the diversity of
saproxylic beetles? A multiscaled study from a spruce forest in southern Norway. Biodiversity and
Conservation 5: 75–100.
Preston F.W. 1948. The commonness and rarity of species. Ecology 29: 254–283.
¨
Paivinen ¨
J., Suomi T., Ahlroth P., Hyvarinen E., Korkeamaki¨ E., Mattila J. et al. 1999. Keski-Suomen
 1831

¨
suojeltujen vanhojen metsien hyonteiset ¨ ¨ ¨ Metsahallituksen
ja kaavat. ¨ luonnonsuojelujulkaisuja,
Vantaa, Sarja A, No. 100.
Raivio S. and Mannerkoski I. 1997. Kovakuoriaiset. In: Raivio S. (ed.), Nature Conservation in
¨
Managed Forests – The First Year After Management Practices. Metsahallituksen luonnonsuojelu-
julkaisuja, Vantaa, Sarja A 87, pp. 55–70.
Ranius T. and Jansson N. 2002. A comparison of three methods to survey saproxylic beetles in hollow
oaks. Biodiversity and Conservation 11: 1759–1771.
Rassi P., Alanen A., Kanerva T. and Mannerkoski I. 2000. Suomen lajien uhanalaisuus 2000.
¨
Uhanalaisten lajien II seurantatyoryhma,¨ esipainos. Ymparistoministerio,
¨ ¨ ¨ Helsinki, Finland.
¨
Rutanen I. 1994a. Etela-Suomen vanhojen metsien kovakuoriaiset I [Beetles of the old forests in
South-Finland I]. Helsinki, Publications of the Water and Environment Administration – Series A
175.
¨
Rutanen I. 1994b. Metsapalon vaikutuksesta kovakuoriaislajistoon Patvinsuon kansallispuistossa [The
impact of forest fire on the beetle fauna in the National Park of Patvinsuo in Eastern Finland].
Helsinki, Publications of the Water and Environment Administration – Series A 196.
¨
Rutanen I. 1995. Etela-Suomen vanhojen metsien kovakuoriaiset II [Beetles of the Old Forests in
South-Finland II]. Helsinki, Publications of the Water and Environment Administration – Series A
215.
Rutanen I. and Kashevarov B. 1997. Coleoptera of the Nature Reserve Friendship with adjacent
primeval forests. In: Lindholm T., Heikkila¨ R. and Heikkila¨ M. (eds), Ecosystems, Fauna and Flora
of the Finnish–Russian Nature Reserve Friendship. The Finnish Environment 124, Helsinki, pp.
257–294.
Rutanen I. and Muona J. 1982. Coleoptera collected with a car net in Finland. Notulae Entomologicae
62: 69–72.
Siitonen J. 1994. Decaying wood and saproxylic Coleoptera in two old spruce forests: a comparison
based on two sampling methods. Annales Zoologici Fennici 31: 89–95.
Siitonen J. 2001. Forest management, coarse woody debris and saproxylic organisms: Fennoscandian
boreal forests as an example. Ecological Bulletins 49: 11–41.
Siitonen J., Martikainen P., Kaila L., Nikula A. and Punttila P. 1995. Kovakuoriaislajiston
¨
monimuotoisuus eri tavoin kasitellyilla ¨ metsaalueilla
¨ Suomessa ja Karjalan tasavallassa. In:
Hannelius S. and Niemela¨ P. (eds), Monimuotoisuus metsien hoidossa. Vantaa Metsantutkimus-
¨
laitoksen tiedonantoja 564., pp. 43–63.
Simberloff D. 1998. Flagships, umbrellas, and keystones: is single-species management passe´ in the
landscape era? Biological Conservation 83: 247–257.
Simila¨ M., Kouki J., Martikainen P. and Uotila A. 2002a. Conservation of beetles in boreal pine forests:
the effects of forest age and naturalness on species assemblages. Biological Conservation 106:
19–27.
Simila¨ M., Kouki J., Monkkonen
¨ ¨ M. and Sippola A.-L. 2002b. Species richness along the forest
productivity gradient in northern Finland. Ecography 25: 42–52.
Simila¨ M., Kouki J. and Martikainen P. 2003. Saproxylic beetles in managed and seminatural Scots
pine forests: quality of dead wood matters. Forest Ecology and Management 174: 365–381.
Sippola A.L., Siitonen J. and Kallio R. 1995. Faunistics of Coleoptera in subarctic pine forests in
Finnish Lapland. Entomologica Fennica 6: 201–210.
Southwood T.R.E. 1978. Ecological Methods. 2nd edn. Chapman & Hall, London.
Stokland J.N. 1992. Altitude and latitude gradients in species richness of beetles (Coleoptera) in
Scandinavia. In: Biological diversity and conservation strategies in Scandinavian boreal forests,
Ph.D. Thesis, Department of Biology, University of Oslo, Oslo, Norway.
Stokland J.N. 2001. The coarse woody debris profile: an archive of recent forest history and an
important biodiversity indicator. Ecological Bulletins 49: 71–83.
Stork N.E. and Blackburn T.M. 1993. Abundance, body size and biomass of arthropods in tropical
forest. Oikos 67: 483–489.