Ecological Entomology (2001) 26, 302±311
The spatial distribution of beetles within the canopies
of oak trees in Richmond Park, U.K.
N . E . S T O R K , 1 P . M . H A M M O N D , 2 B . L . R U S S E L L 1 , 3 and
W . L . H A D W E N 1 , 4 1Cooperative Research Centre for Tropical Rainforest Ecology and Management,
James Cook University, Australia and 2Department of Entomology, The Natural History Museum, London
Abstract. 1. 5054 adult beetles of 144 species were collected in a total of 696
1-m2 collecting trays by knockdown insecticide fogging of 36 different oak trees in
closed canopy woodland at Richmond Park, U.K., with three of the trees sampled
on each of 12 dates, at 2- to 3-week intervals, between April and October 1984.
2. In late spring (April/May), more individuals and species of beetles were
collected in trays close to the trunks of trees than in trays more distant from the
trunk. The reverse was the case in late September/October. Neither pattern
prevailed in the intervening months.
3. Individual species exhibited a variety of patterns, with some species more
abundant near the trunk, e.g. Leiopus nebulosus (L.), Strophosoma melanogrammum
(Forster), Cylindronotus laevioctostriatus (Goeze), and Dromius agilis (Fabricius),
and some less abundant near the trunk, e.g. Curculio pyrrhoceras (Marsham) and
Rhynchaenus signifer (Creutzer). For Adalia decempunctata (L.), this preference
changed with season. The observed species preferences for parts of a tree crown
near or distant from the main trunk are discussed with reference to their known
biologies.
4. No pronounced pattern of preference for north- or south-facing aspects of
trees in closed canopy woodland was observed, however populations of some
species exhibited patterns of within-tree distribution that correlate with compass
angle; for one species, the ladybird Adalia decempunctata, this distribution changed
with season and between colour morphs.
Key words. Beetles, canopy fogging, insecticides, oak, spatial distribution.
Introduction
The way in which insects form a community on particular
plant species has been examined extensively (e.g. Southwood,
1960, 1961; Kennedy & Southwood, 1984), typically for nonwoody biennials and/or annuals. In the last 15 years, many of
the physical and logistical problems of studying communities
associated with woody trees seem to have been overcome,
especially those dif®culties involved in obtaining adequate
quantitative samples of insects. Recent advances have come
Correspondence: N. E. Stork, Cooperative Research Centre for
Tropical Rainforest Ecology and Management, James Cook
University, PO Box 6811, Cairns, Queensland, Australia 4870. E-mail:
[email protected]
3
Current address: CSIRO, Division of Entomology, Australia.
4
Current address: School of Environmental Sciences, Grif®th
University, Australia.
302
primarily through the adoption of knockdown insecticide
sampling (e.g. Moran & Southwood, 1982; Southwood et al.,
1982a,b; Erwin, 1983a,b, 1990; Adis et al., 1984; Stork,
1987a,b, 1988, 1991; Basset, 1988, 1989, 1990, 1991a,b,c,
1992; Morse et al., 1988; Stork & Brendell, 1990, 1993; Stork
et al., 1997). Despite this, with a few notable exceptions
(Nielsen & Ejlersen, 1977; Barnard et al., 1986; Basset, 1992;
Basset et al., 1992; Koponen et al., 1997; Reynolds &
Crossley, 1997; Richardson et al., 1997), knowledge of the
basic spatial distribution of insects within trees is scant,
coming largely from anecdotal observations, coupled with
knowledge of resource requirements of individual species.
European deciduous oaks Quercus are among the best
known trees in terms of their associated invertebrate faunas
(Morris & Perrin, 1974; PatoeÁka et al., 1999; Hammond &
Owen, in press). The beetle fauna of Richmond Park near
London, U.K., and in particular of its oak trees, is extremely
# 2001
Blackwell Science Ltd
Spatial distribution of beetles in oak trees
well known, with more than a quarter of all 4000+ British
species having been recorded there in the last 20 years
(Hammond & Owen, in press). In 1984, an intensive arthropod
sampling programme using knockdown insecticides was
carried out from April to November to examine the arthropod
community structure of oak trees in Richmond Park (Stork,
1988; Stork & Hammond, 1997). Barnard et al. (1986)
examined the spatial distribution of species of neuropteroids
collected as part of this sampling programme and found that
Nothochrysa capitata, Sympherobius pellucidus, and
Hemerobius humulinus were more abundant in collecting trays
positioned near the trunk of fogged trees than in trays
positioned towards the edge of the crown.
The work reported here was designed to investigate spatial
patterns of the canopy beetle fauna on Richmond Park oak
trees, examining whether some species of beetles are
associated more closely with the trunk region or outer parts
of trees, whether beetles prefer a particular sector of a tree, e.g.
the northern or southern part, and whether these preferences
change during the year. Where possible, explanations for these
patterns are suggested based on the known biology of the
beetle fauna.
303
canopy edge, as tree crowns were of irregular shape. A plan
diagram was drawn for each tree showing the arrangement of
the trays (Fig. 1).
Statistical analysis
Distance from the trunk of the tree. In order to examine the
relationship between beetle abundance and distance from the
Methods
Field site and sampling
Three different mature oak trees were fogged every 2 or
3 weeks (depending on suitable weather conditions) on 12
separate occasions between April and October 1984, with a
total of 36 different trees sampled overall [26 April (fog 1),
15 May (2), 30 May (3), 12 June (4), 26 June (5), 11 July (6),
25 July (7), 14 August (8), 1 September (9), 11 September
(10), 26 September (11), 16 October (12)]. Oak trees sampled
at the study site were situated in woodland with semicontinuous canopy broken by small to large gaps, both natural
and man-made, the most signi®cant being the 5±15 m wide ride
running NW±SE across the NW of the wood. Fuller details of
the sampling programme, including sampling dates, site map,
tree information, and weather conditions, were given by Stork
and Hammond (1997).
Arthropods were collected by applying a 1% solution of
knockdown insecticide (a synthetic pyrethroid, Reslin E),
using an insecticide fogger (Swing Fog SN11), to the canopies
of mature oak trees Quercus robur in Richmond Park.
Specimens were collected on a series of circular, funnelshaped, 1-m2 trays distributed below the canopy of the sample
trees. Ropes were strung at chest-height between available tree
trunks from which 12±21 trays were suspended. The trays were
not spaced regularly but an attempt was made to ensure that
they were positioned such that collections were made from
under most areas of the focal tree's crown. All beetles were
extracted from the samples and identi®ed to species.
The horizontal distance from the centre of each tray to the
tree trunk (measured to the nearest 0.1 m) and the compass
position of each tray in relation to the tree trunk were recorded.
No attempt was made to measure the distance of trays from the
# 2001
Blackwell Science Ltd, Ecological Entomology, 26, 302±311
Fig. 1. Spatial distribution of collecting trays (open circles) beneath
a representative fogged tree in terms of (a) proximity of tray to tree
trunk/outer canopy (distance classes indicated by concentric rings)
and (b) compass orientation with respect to the tree trunk (with the
tree trunk as centre point).
304 N. E. Stork et al.
trunk of the tree, the collecting trays were allocated to the
following classes according to the distance measured from the
centre of the tray to the nearest point on the trunk: class 1 =
0±1 m, class 2 = 1.1±2 m, class 3 = 2.1±3 m, class 4 = 3.1±4 m,
class 5 = 4.1±5 m, class 6 = > 5.1 m (Fig. 1a).
The logarithms of overall adult beetle abundances and of
beetle species number for each tray were correlated against
their distribution in the six distance classes. While this
approach has some limitations, the data generated in this
study were not suited to analyses using univariate t-tests or
ANOVA because of the high level of variation observed in
abundances and distribution among sampled trees. This was
established through multivariate analyses (not presented here)
using PATN: Pattern Analysis Package (Belbin, 1995).
Patterns of spatial distribution were assessed by plotting
relative beetle abundance (log ni/nm; where ni = number of
beetles in a tray, nm = mean number of beetles in all trays for a
given tree) against tray distance class to determine whether
beetles were distributed uniformly throughout the crown.
Beetle preference for compass sectors of the canopy: angle
of preferred distribution. The compass angle of each tray with
respect to the tree trunk was measured to the nearest 5° and
each tray was allocated to one of the 12 30° sectors on the
compass (Fig. 1b). The number of trays in each sector was
uneven (mean n = 58), with a minimum number of 35 trays in
any compass sector across the 12 sampling dates. To ensure
even sampling of trays across compass sectors, the beetle
frequency of 10 replicate samples of 35 trays was selected
randomly for each sector from the data set using SPSS 8.0
(1997). Circular distribution mean angle analyses, as described
by Zar (1996) and summarised in the Appendix, were used to
determine the distribution of beetles with respect to compass
sectors in the canopy.
A mean angle (i.e. compass direction) was determined for
each of the 10 replicate data sets (Appendix: eqns 1±8) and a
second-order mean of means was calculated (eqns 9±32). The
mean of means angles were tested to determine whether the
populations of beetles displayed signi®cant mean direction
using a parametric one-sample second-order analysis of angles
(eqns 33±36). In this way, the angle of preferred direction was
calculated for all beetles and individually for the seven
commonest species: Dromius quadrimaculatus, Dromius
quadrinotatus, Dalopius marginatus, Phyllobius argentatus,
Strophosoma melanogrammum, Adalia bipunctata, and Adalia
decempunctata. Adalia decempunctata was collected in large
enough numbers for analyses of seasonal variation in compass
distribution and variation in the compass distribution of its
three main colour morphs. The morphs recognised are the
typical form (t) ± reddish-yellow with small dark spots, the
dark or melanic form (d) ± largely black with one or more
small pale patches, and the intermediate or socks form (s) ±
many pale spots on a dark background (see Majerus, 1994).
For seasonal analyses, samples were grouped (e.g. fogs 1
and 2, fogs 3 and 4, etc.). Because there were insuf®cient data
for replicated analysis of means as described above, for each of
the six resultant groups, a single mean angle, angular
dispersion value, and 95% con®dence limits for the mean
were calculated for each group. Rayleigh's test was performed
to test whether beetle populations were distributed uniformly
around the tree (eqns 7±8).
Results
Beetle abundance
A total of 5054 adult beetles (144 species) and 575 beetle
larvae was collected on 696 1-m2 trays under the 36 trees
sampled on the 12 occasions from April to November,
representing an average of 8.09 adult and larval beetles per
tray (SD = 7.5), with a maximum of 51 in any 1-m2 tray. There
were no beetles in 37 of the 696 trays. There was an average of
4.24 species per tray (SD = 3.00, maximum number in a
tray = 17), 18.88 species per tree (for an average 13-m2 catch),
and 33.25 species per fogging occasion (39-m2 catch from
three trees). Seasonal variation in abundance of individuals and
species was considerable. In addition, there were sometimes
substantial differences among trees fogged on the same
occasion, in part because there were sometimes large
differences in the ¯owering, leaf, and fruiting phenology of
the oak trees sampled on each occasion.
Abundance and species number in relation to distance from
the tree trunk
On average, more adult beetles were found in trays close to
the trunks of trees (distance class 1 mean = 11.14) than in trays
closer to the edge of the canopy (distance class 6 mean = 7.78)
(Table 1). When relative beetle abundance for each of the 696
trays used throughout the sampling period was plotted against
tray distance class, the plot had a signi®cant negative slope
(regression coef®cient = ±0.009, SD = 0.002, correlation coef®cient = ±0.205, P < 0.001), indicating greater numbers of
beetles in trays closer to the trunk of the tree. When all trays
in distance class 1 were excluded, however, the slope was
reduced and was not signi®cant (regression coef®cient =
±0.002, SD = 0.002, P > 0.05). This suggests that for the
6-month sampling period as a whole there were, on average,
Table 1. Mean number (M) of adult beetles and beetle species per
tray in different distance classes (D) from trunk of tree to outer
canopy. n = number of trays for each distance class, SD = standard
deviation of the mean.
Individuals
Species
D
n
M
SD
M
SD
1
2
3
4
5
6
111
167
146
129
85
58
11.14
7.01
7.80
7.52
7.81
7.78
9.25
5.69
7.10
6.78
7.00
7.68
5.47
3.88
4.21
3.92
4.09
3.88
3.56
2.42
3.04
2.87
2.69
3.43
# 2001
Blackwell Science Ltd, Ecological Entomology, 26, 302±311
Spatial distribution of beetles in oak trees
signi®cantly more beetles in trays close to the trunk of the
sampled trees than in trays away from the tree trunks.
For nine of the 36 trees sampled, the regression coef®cient
of the mean number of beetles per tray against distance class
was signi®cant. The regression coef®cient for 29 trees was
negative, of which seven were signi®cant (i.e. a preference for
the tree trunk), while seven trees exhibited positive slopes, of
which two were signi®cant (i.e. a preference for the outer
canopy) (Fig. 2a). Plotting the regression coef®cients for each
tree for the log number of beetles per tray against distance
class (using the log number of beetles per tray overcomes
problems of difference in abundance among trees and at
different times of the year) indicated a preference for the outer
canopy in summer and a preference for the trunk in autumn.
The pattern for the regression coef®cients for the log number
of species of beetles per tray against distance class showed a
similar trend (Fig. 2b). The regression coef®cients for log
species number and for log abundance were highly correlated
(correlation coef®cient = 0.84, P < 0.001).
305
main trunk of the focal trees. Both the adults and larvae of this
species are known to forage on trunks at night and conceal
themselves on trunks during the day. Similarly, most of the
large beetles collected were distributed close to the trunk of the
focal trees [e.g. the carabid Agonum assimile (Paykull), the
clerid Opilo mollis (L.), the elaterid Stenagostus villosus
(Fourcroy), and the lymexylonid Lymexylon navale (L.)].
Agonum assimile (Paykull) is a nocturnal woodland-¯oor
predator that is known to rest during the day under loose bark
and in crevices on tree trunks (Cook, 1987). Opilo mollis,
Abundance of individual species in relation to distance from
the tree trunk
The relationship of abundance to distance from the trunk of
the tree was investigated for adults of 27 of the most common
species of beetle collected throughout the year and for the most
common larvae (Table 2).
There were no clear relationships between abundance and
distance from the trunk for many species, including representatives of the leaf-chewing phytophage guild [e.g.
Cryptocephalus pusillus Fabricius, Phyllobius argentatus
(L.), P. pyri (L.)], and other foliage or ¯ower-associated
species [e.g. Curculio glandium Marsham, C. venosus
(Gravenhorst), Malthinus ¯aveolus (Herbst), Malthodes
marginatus (Latreille), Anaspis lurida Stephens, A. maculata
Fourcroy, and A. regimbarti Schilsky].
A number of species was found to have a signi®cant positive
association with the trunk (see Table 2), i.e. their abundances
were highest in the trays closest to the trunk. These include the
cerambycid Leiopus nebulosus (L.), the weevil Strophosoma
melanogrammum (Forster), the tenebrionid Cylindronotus
laevioctostriatus (Goeze), and the carabid Dromius agilis
(Fabricius). The positive association of adult L. nebulosus with
the trunk is consistent with their bark-like cryptic markings.
Strophosoma melanogrammum adults feed on shoots and
leaves of a wide range of deciduous and coniferous trees
(Bevan, 1987), moving at certain times of the year to different
tree species when their leaves are most palatable. Adults are
completely apterous, so tree to tree movements may entail
ground crossings to other tree trunks. This species is also rather
bark-like in its colouration. Another ¯ightless species, the
tenebrionid C. laevioctostriatus, is a nocturnal epiphyte grazer/
scavenger, commonly found under loose bark of logs and tree
trunks during the day. The highly signi®cant relationship found
for this species probably re¯ects a real preference for the tree
trunk where the algae on which the adults feed are mostly
found. Dromius agilis appeared to be virtually con®ned to the
# 2001
Blackwell Science Ltd, Ecological Entomology, 26, 302±311
Fig. 2. Regression coef®cients of (a) log total number of individual
beetles per tray and (b) log total number of beetle species per tray,
against distance from trunk, and against time of year, for each of 36
trees sampled (on 12 dates between April and October). Signi®cant
regression coef®cients are indicated with circles.
306 N. E. Stork et al.
S. villosus, and L. navale, are all saproxylic species and are
intimately associated with tree trunks or decaying wood.
Many of the rarer saproxylic species regarded as indicators
of ancient woodland conditions (Harding & Rose, 1986),
while being collected in small numbers, were also only found
close to tree trunks [e.g. the elaterid Agrilus pannonicus
(Piller & Mitterpacher), the latridiid Corticaria alleni
Johnson, the scraptiid Scraptia testacea Allen, the colydiid
Synchita humeralis (Fabricius), and the tetratomid Tetratoma
desmaresti Latreille].
For the coccinellid Adalia decempunctata (L.), there were
suf®cient data to carry out a seasonal analysis of adult beetle
trunk/canopy preferences (Table 2), this species preferring the
outer canopy in April and May and being spread more evenly
at other times of the year.
The weevils Curculio pyrrhoceras Marsham and
Rhynchaenus signifer (Creutzer) were both present in reasonable numbers in all parts of the tree but were signi®cantly more
abundant in trays nearer the canopy edge (Table 2). Another
weevil, Phyllobius pyri (L.), on the other hand, was not
Table 2. Regression coef®cients (R) for the commonest species of beetles for number of individuals in trays against each distance class from the
trunk of the sampled trees. *P < 0.05, **P < 0.01, ***P < 0.001. T = trunk, OC = outer canopy.
Species
Family
Adalia decempunctata (L.)
Fog 1
Fog 2
Fog 3
Fogs 4±7
Fogs 8±12
All fogs (1±12)
Larvae
Anaspis lurida Stephens
A. maculata Fourcroy
A. regimbarti Schilsky
All Anaspis larvae
Cis vestitus MellieÂ
Cortinicara gibbosa (Herbst)
Cryptocephalus pusillus Fabricius
Curculio glandium Marsham
Curculio pyrrhoceras Marsham
Curculio venosus (Gravenhorst)
Cylindronotus laevioctostriatus (Goeze)
Dalopius marginatus (L.)
Dasytes aeratus Stephens
Dromius agilis (Fabricius)
D. angustus BrulleÂ
D. quadrimaculatus (L.)
D. quadrinotatus (Zenker)
All Dromius larvae
Hemicoelus fulvicornis (Sturm)
Leiopus nebulosus (L.)
Malthinus ¯aveolus (Herbst)
Malthodes marginatus (Latreille)
Phyllobius argentatus (L.)
P. pyri (L.)
Rhinosimus planirostris (Fabricius)
Rhynchaenus signifer (Creutzer)
R. quercus
Fogs 1±2
Fogs 5±7
Fogs 10±12
Strophosoma melanogrammum (Forster)
Coccinellidae
Scraptiidae
Scraptiidae
Scraptiidae
Scraptiidae
Ciidae
Latridiidae
Chrysomelidae
Curculionidae
Curculionidae
Curculionidae
Tenebrionidae
Elateridae
Melyridae
Carabidae
Carabidae
Carabidae
Carabidae
Carabidae
Anobiidae
Cerambycidae
Cantharidae
Cantharidae
Curculionidae
Curculionidae
Salpingidae
Curculionidae
Curculionidae
Curculionidae
Individuals
Trays
R
Signi®cance
Area
34
160
72
110
87
465
277
121
117
56
46
76
73
249
83
280
30
62
420
24
43
20
263
231
153
94
134
56
25
351
163
182
71
216
44
88
55
413
63
64
64
259
246
696
696
696
696
696
696
696
696
696
696
696
696
696
696
696
696
696
696
696
696
696
696
696
696
696
696
696
696
696
127
194
120
696
0.194
0.154
0.021
±0.006
±0.016
0.044
0.020
0.000
0.011
±0.006
±0.187
0.008
±0.007
0.003
±0.009
0.038
0.008
±0.042
0.000
0.010
±0.032
±0.017
±0.102
0.007
0.007
0.006
±0.030
0.006
±0.003
0.006
0.007
±0.017
0.027
±0.061
0.017
±0.111
±0.152
±0.137
***
OC
*
OC
**
T
*
OC
***
T
*
***
***
***
OC
T
T
T
**
T
***
***
OC
T
***
***
***
T
T
T
Fig. 3. Mean compass angle of distribution of beetles within oak trees with 95% con®dence limits (north at top of page).
# 2001
Blackwell Science Ltd, Ecological Entomology, 26, 302±311
Spatial distribution of beetles in oak trees
# 2001
Blackwell Science Ltd, Ecological Entomology, 26, 302±311
307
308 N. E. Stork et al.
associated signi®cantly with either the trunk area or the canopy
edge but appeared to be more abundant at an intermediate
distance between the trunk and the outer canopy.
Beetle preference for compass sectors of the canopy: angle
of preferred distribution
Of the seven species examined, all except A. bipunctata
showed signi®cant population directionality (Fig. 3). For most
of these, the con®dence limits for the mean were narrow,
signifying a narrow distribution of individuals around that
compass sector. All three morphs of A. decempunctata, both
separately and combined, and larvae of A. decempunctata, also
exhibited signi®cant population directionality but the mean
direction of these species and groups varied considerably. For
example, D. quadrimaculatus, D. marginatus, S. melanogrammum, and A. decempunctata morph (d) all displayed signi®cant
north-west through to north-easterly mean population directionality, whereas A. decempunctata morph (t), Adalia larvae,
and total A. decempunctata adults all had signi®cant southerly
mean population directionality.
The distribution of individuals of A. decempunctata within
the canopy appeared to vary during the year, as indicated by
different mean directionalities (Fig. 3b): fogs 1 and 2 (26 April/
15 May) south-westerly, fogs 3 and 4 (30 May/12 June)
easterly, fogs 5 and 6 (26 June/11 July) westerly, fogs 7 and 8
(25 July/14 August) easterly, fogs 9±12 (1 September/
11 September/26 September/16 October) no signi®cant mean
direction. Adalia decempunctata larvae also displayed a
variety of mean population directions at different seasons:
fogs 3 and 4 north-easterly, fogs 5 and 6 south-westerly. At
other sampling times, no signi®cant mean directionality was
found, indicating a more uniform larval distribution.
Discussion
The data presented here indicate several patterns of beetle
distribution within British oak trees in closed canopy woodland
at the time of greatest insect abundance and activity from April
to October: (1) some beetle species are more closely associated
with the trunk of a tree; (2) others are associated with the outer
parts of a tree crown; (3) some species of beetle prefer
particular compass sectors of a tree; (4) the distribution
patterns (and hence trunk/canopy/compass preferences) for
some species change during the year.
As Basset (1992) has shown, there is an enormous range
of variables that ultimately in¯uence the spatial distribution of
beetles within the crown of an individual tree. The distribution
patterns reported here may re¯ect preferences for the surface
of the trunk and large branches or for dense foliage, for warmer
or cooler, lighter or shadier parts of the tree. These preferences
may in turn re¯ect particular microclimatic and microhabitat
requirements and availability of food resources such as pollen,
¯owers, fresh leaves, lichens/algae, particular prey species or
even availability of enemy-free space.
The complicated patterns of variation in the distribution of
individuals of a species at different times of the year is best
exempli®ed by the results for one of the most abundant
species, the ladybird A. decempunctata. In the early part of the
year, adults of this species showed a preference for the outer
canopy and for west-facing parts of trees, presumably seeking
the warmer parts of trees in search of aphids. Later in the year,
there was no clear preference for the outer canopy and the
compass preference moved to the east. It appears, however,
that the three colour morphs that make up the adult population
have different preferences for parts of the tree and that these
change during the year: morphs s and t at the beginning of the
sampling programme both showed a preference for the west;
morph t changed to a south-east preference in fogs 3±4, while
morph d exhibited a north-east preference at this time,
changing to a south-east preference in fogs 5±6. Nonsigni®cant mean directions at other times of the year either
indicate no real preferred direction or that if such a preference
exists there were insuf®cient data to reveal such patterns.
Adalia decempunctata larvae also appeared to exhibit changes
in their distribution in the trees observed, as the mean
directions changed from north-east to south-east from fogs
3±4 to fogs 5±6. Geographical variation in the relative
incidence of individual morphs of polymorphic ladybird
species is well known, and a correlation between high
incidence of melanic or dark morphs and low sunshine levels
has been reported for several species (Majerus, 1994).
Differences in the within-tree distribution of the principal
morphs of A. decempunctata may also be explained by a
greater ability of the darkest individuals to operate in
conditions of lower temperature and/or less sunshine.
Variation over time may re¯ect the pattern of adult foraging,
seeking out concentrations of aphids that are most likely to be
found near points of active growth on the focal trees.
This study demonstrates that careful attention to sampling
design is required when carrying out ecological studies
of arthropod communities on trees. The arthropod composition
of catches will re¯ect only the arthropod faunas of the parts of
the tree sampled, as discussed above. In addition, sampling
artefacts, such as leaning of a few tree trunks, asymmetric
distribution of foliage, variation in epiphyte load, or the chance
occurrence of a single large bracket fungus on a tree, may also
distort the picture. Some variation may be due to the genetic
variability of the trees or variation in their environment (e.g.
soil quality, exposure, etc.). Some individual oaks are more
advanced, seasonally, than others (see also Hunter, 1992),
sometimes by a matter of several weeks. Surrounding
vegetation may also affect faunal load and distribution,
particularly for some trees where birch formed an undercanopy. Some beetles, such as members of the broad-nosed
weevil phyllophage guild (e.g. Phyllobius, etc.), move
naturally from one species of tree to another (e.g. from birch
to oak) at particular times of the year, depending on the
palatability of leaves [see Bale (1981) for similar movements
between different species of trees for Rhynchaenus fagi (L.)].
Birch loses its leaves before oak, and some species therefore
move to oak in the autumn. Such movements can alter the
abundance and distribution of some species greatly. For
# 2001
Blackwell Science Ltd, Ecological Entomology, 26, 302±311
Spatial distribution of beetles in oak trees
example, in the October/November fogs, the commonest
species of Hemiptera collected were species that normally
feed on birch (W. R. Dolling, pers. comm.). Stork and
Hammond (1997) have also shown that the arthropod fauna
collected from trees changes with time of day; early evening
samples have more transient (tourist) species than morning
samples.
Another possible confounding factor is the in¯uence of edge
effects. Murcia (1995) predicted, on theoretical grounds, that
there should be bimodal patterns in edge effects. This was
con®rmed by Didham (1996), who found that in Amazonian
lowland rainforest there are two peaks in abundance for leaf
litter beetles, one within the ®rst few metres from the edge, the
other less than 100 m from the edge. With many trees having
canopy widths of tens of metres, within-tree distribution may
be complicated by such broad-scale edge effects, particularly
in an open woodland with frequent tree gaps such as that at
Richmond Park. Finally, it is to be expected that within-tree
variation in the abundance of various beetle species is likely to
be both more marked and more consistent in pattern in freestanding trees than in continuous woodland. The individual
trees sampled at Richmond Park varied substantially in their
degree of isolation from neighbouring trees and the extent to
which parts of their crowns were exposed. Such differences,
rather than seasonal effects per se, may explain some of the
differences between sampling dates that are reported above.
In north temperate forests, the low angle of the sun,
particularly early in the year, means that the southern parts of
the tree are likely to be sunnier and warmer than northern
parts. In an earlier unpublished study, N. E. Stork and
P. M. Hammond found that almost twice as many beetles
were collected after application of knockdown insecticides on
the south side of an isolated oak tree than on the north side.
Richardson et al. (1997) also found that there were such north/
south differences in insect abundance in small trees sampled in
forests outside Sydney, Australia. The shift in beetle
preference for the more northerly parts of oak trees in May
and early June may be related to changes in the light and heat
regime such that more beetles were seeking darker, cooler
parts of the canopy.
Acknowledgements
We thank M. B. Brown, the Richmond Park Superintendent,
for permission to carry out our studies in Richmond Park and
we are grateful to many of our colleagues, particularly Martin
Brendell and Campbell Smith, for assistance in sampling. We
also thank Clive Moncrieff, David Westcott and Jamie
Seymour for statistical advice and Daiqin Li, Simon Grove,
John Lawton and an anonymous referee for useful comments
on the manuscript.
References
Adis, J., Lubin, Y.D. & Montgomery, G.G. (1984) Arthropods from the
canopy of inundated and terra ®rme forests near Manaus, Brazil,
# 2001
Blackwell Science Ltd, Ecological Entomology, 26, 302±311
309
with critical considerations on the Pyrethrum-fogging technique.
Studies on the Neotropical Fauna and Environment, 19, 223±236.
Bale, J.S. (1981) Seasonal distribution and migratory behaviour of the
beech leaf-mining weevil Rhynchaenus fagi L. Ecological
Entomology, 6, 109±118.
Barnard, P.C., Brooks, S.J. & Stork, N.E. (1986) The seasonality and
distribution of Neuroptera, Raphidioptera and Mecoptera on oaks in
Richmond Park, Surrey, as revealed by insecticide knock-down
sampling. Journal of Natural History, 20, 1321±1331.
Basset, Y. (1988) A composite interception trap for sampling
arthropods in tree canopies. Journal of the Australian
Entomological Society, 27, 213±219.
Basset, Y. (1989) The temporal and spatial distribution of arboreal
arthropods associated with the rainforest tree Argyrodendron
actinophyllum. PhD thesis, Grif®th University, Brisbane, Australia.
Basset, Y. (1990) The arboreal fauna of the rainforest tree
Argyrodendron actinophyllum as sampled with restricted canopy
fogging. Composition of the fauna. Entomologist, 109, 173±183.
Basset, Y. (1991a) The seasonality of arboreal arthropods foraging
within an Australian rainforest tree. Ecological Entomology, 16,
265±278.
Basset, Y. (1991b) The taxonomic composition of the arthropod fauna
associated with an Australian rainforest tree. Australian Journal of
Zoology, 39, 171±190.
Basset, Y. (1991c) The spatial distribution of herbivory, mines and
galls within an Australian rainforest tree. Biotropica, 23, 271±281.
Basset, Y. (1992) In¯uence of leaf traits on the spatial distribution of
arboreal arthropods within an overstorey rainforest tree. Ecological
Entomology, 17, 8±16.
Basset, Y., Aberlenc, H.-P. & Delvare, G. (1992) Abundance and
strati®cation of foliage arthropods in a lowland rain forest of
Cameroon. Ecological Entomology, 17, 310±318.
Belbin, L. (1995) PATN: Pattern Analysis Package. Division of
Wildlife and Ecology, CSIRO, Canberra.
Bevan, D. (1987) Forest Insects. A Guide to Insects Feeding on Trees
in Britain. Forestry Commission Handbook 1. HMSO, London.
Cook, A.A. (1987) Ground beetle activity in a Northamptonshire
wood. Entomologist's Monthly Magazine, 123, 61±67.
Didham, R.K. (1996) The effects of forest fragmentation on leaf-litter
invertebrates in Central Amazonia. PhD thesis, Imperial College,
University of London.
Erwin, T.L. (1983a) Beetles and other insects of tropical forest
canopies at Manaus, Brazil, sampled by insecticidal fogging.
Tropical Rain Forest Ecology and Management (ed. by S. L.
Sutton, T. C. Whitmore and A. C. Chadwick), pp. 59±75. Blackwell
Scienti®c Publications, Oxford.
Erwin, T.L. (1983b) Tropical forest canopies: the last biotic frontier.
Bulletin of the Entomological Society of America, 29, 14±19.
Erwin, T.L. (1990) Canopy arthropod biodiversity: a chronology of
sampling techniques and results. Revista Peruana de Entomologia,
32, 71±77.
Hammond, P.M. & Owen, J.A. (in press) The Saproxylic and other
Beetles of Richmond Park SSSI. A Case Study. English Nature
Science.
Harding, P.T. & Rose, F. (1986) Pasture-woodlands in Lowland
Britain. A Review of their Importance for Wildlife Conservation.
Institute of Terrestrial Ecology, Huntingdon, U.K.
Hunter, M.D. (1992) A variable insect±plant interaction: the relationship between tree budburst phenology and population levels of
insect herbivores among trees. Ecological Entomology, 16, 91±95.
Kennedy, C.E.J. & Southwood, T.R.E. (1984) The number of species
of insects associated with British trees: a re-analysis. Journal of
Animal Ecology, 53, 455±478.
310 N. E. Stork et al.
Koponen, S., Rinne, V. & Clayhills, T. (1997) Arthropods on oak
branches in SW Finland, collected by a new trap type. Entomologica
Fennica, 3, 177±183.
Majerus, M.E.N. (1994) Ladybirds. Harper Collins, London.
Moran, V.C. & Southwood, T.R.E. (1982) The guild composition of
arthropod communities in trees. Journal of Animal Ecology, 51,
289±306.
Morris, M.G. & Perrin, F. (1974) The British Oak: its History and
Natural History. Faringdon Classey for the Botanical Society of the
British Isles.
Morse, D.R., Stork, N.E. & Lawton, J.H. (1988) Species number,
species abundance and body length relationships of arboreal beetles
in Bornean lowland rain forest trees. Ecological Entomology, 13,
25±37.
Murcia, C. (1995) Edge effects in fragmented forests: implications for
conservation. Trends in Ecology and Evolution, 10, 58±62.
Nielsen, B.O. & Ejlersen, A. (1977) The distribution pattern of
herbivory in a beech canopy. Ecological Entomology, 2, 293±299.
PatoeÁka, J., KrisÏtõÂn, A., Kulfan, J. & Zach, P. (1999) Die
eichenschaÈdlinge und ihre feinde. Institut fuÈr WaldoÈkologie der
Slowakischen Akademie der Wissenschaften Zvolen, Zvolen.
Reynolds, B.C. & Crossley, D.A., Jr (1997) Spatial variation in
herbivory by forest canopy arthropods along an elevation gradient.
Entomological Society of America, 26, 1232±1239.
Richardson, B.J., Burgin, S., Azarbayjani, F.F. & Lutubula, S. (1997)
Distinguishing the woods from the trees. Canopy Arthropods (ed. by
N. E. Stork, J. Adis and R. K. Didham), pp. 497±509. Chapman &
Hall, London.
Southwood, T.R.E. (1960) The abundance of Hawaiian trees and the
number of their associated insect species. Proceedings of the
Hawaiian Entomological Society, 17, 299±303.
Southwood, T.R.E. (1961) The number of species associated with
various trees. Journal of Animal Ecology, 30, 1±8.
Southwood, T.R.E., Moran, V.C. & Kennedy, C.E.J. (1982a) The
assessment of arboreal insect fauna: comparisons of knockdown
sampling and faunal lists. Ecological Entomology, 7, 331±340.
Southwood, T.R.E., Moran, V.C. & Kennedy, C.E.J. (1982b) The
richness, abundance and biomass of the arthropod communities on
trees. Journal of Animal Ecology, 51, 635±649.
Stork, N.E. (1987a) Guild structure of arthropods from Bornean rain
forest trees. Ecological Entomology, 12, 69±80.
Stork, N.E. (1987b) Arthropod faunal similarity of Bornean rain forest
trees. Ecological Entomology, 12, 219±226.
Stork, N.E. (1988) Insect diversity; facts, ®ction and speculation.
Biological Journal of the Linnean Society, 35, 321±337.
Stork, N.E. (1991) The composition of the arthropod fauna of Bornean
lowland rain forest trees. Journal of Tropical Ecology, 7, 161±180.
Stork, N.E., Adis, J. & Didham, R.K. (eds) (1997) Canopy Arthropods.
Chapman & Hall, London.
Stork, N.E. & Brendell, M.J.D. (1990) Variation on the insect fauna of
Sulawesi trees with season, altitude and forest type. Insects and the
Rain Forests of South East Asia (Wallacea) (ed. by W. J. Knight and
J. D. Holloway), pp. 173±190. Royal Entomological Society of
London, London.
Stork, N.E. & Brendell, M.J.D. (1993) Arthropod diversity studies in
lowland rain forest of Seram: Indonesia. The Natural History of
Seram (ed. by I. Edwards, A. A. MacDonald and J. Proctor),
pp. 115±130. Intercept, Andover, U.K.
Stork, N.E. & Hammond, P.M. (1997) Sampling arthropods from treecrowns by fogging with knockdown insecticides: lessons from
studies of oak tree beetle assemblages in Richmond Park (U.K.).
Canopy Arthropods (ed. by N. E. Stork, J. Adis and R. K. Didham),
pp. 3±26. Chapman & Hall, London.
Zar, J.H. (1996) Biostatistical Analysis. Prentice Hall, Englewood
Cliffs, New Jersey.
Accepted 20 October 2000
# 2001
Blackwell Science Ltd, Ecological Entomology, 26, 302±311
Spatial distribution of beetles in oak trees
Appendix: Summary of the calculations required to
determine the mean angle for beetle distribution in the
canopy of oak trees (see Methods: after Zar, 1996)
Part A
Equations 1±8 - Mean angle of each beetle taxon or morph:
The mean angle for each beetle taxon analysed was calculated
as follows where a is each angle in degrees and n is the number
of `angles' (i.e. trays).
n
X
Xˆ
cos ai /n
…1†
iˆ1
Yˆ
n
X
sin ai /n
…2†
iˆ1
r = Ö(X2 + Y2)
(3)
cos a = X/r
(4)
sin a = Y/r
(5)
Angular dispersion/deviation (analogous to a linear standard
deviation). Value range is from a minimum of 0o to a
maximum of 81.03o
311
(15)
Sxy = SXjYj ± [SXjYj/k]
2
(16)
A = (k±1) / Sx
(17)
B = [(k±1)Sxy] / (Sx2Sy2)
2
(18)
C = (k±1) / Sy
2
2
2
D = <[2(k±1)]{1 ± [(Sxy) / Sx Sy ]}Fa(1),2,k±2> / [k(k±2)](19)
H = AC ± B2
(20)
G = A X 2 + 2B X Y + C Y 2 ± D
(21)
2
U = HX ± CD
p
V ˆ DGH
(22)
…23†
W = H X Y + BD
(24)
b1 = (W + V) / U
(25)
b2 = (W ± V) / U
(26)
The quantities b1 and b2 are then examined separately, each
yielding one of the con®dence limits:
For b1
M = Ö(1 + b12)
(27)
For b1
sine = b1/M
(28)
Rayleigh's R ± Rayleigh test. Ho = the sampled population is
uniformly distributed around the circle (canopy):
For b1
cosine = 1/M
(29)
R = nr
For b2
M = Ö(1 + b22)
(30)
For b2
sine = b2/M
(31)
For b2
cosine = 1/M
(32)
s = (180°/p)Ö2(1 ± r)
(6)
(7)
Rayleigh z test ± H0 = no population mean direction:
z = R2/n or z = (nr)2/n
(8)
Equations 9±32 - Second-order analysis: The mean of mean
angles:
k
X
Xˆ
X j /k
…9†
jˆ1
Yˆ
k
X
Y j /k
…10†
jˆ1
Having obtained X and Y, they are substituted for X and Y
respectively into eqns 3, 4, and 5 to determine a, which is the
grand mean
95% con®dence limits for the second-order mean angle (mean
of a set of mean angles):
Part B
Equations 33±36 - Second order mean of means testing:
2
!2 3
k
k
k
X
X
X
x2 ˆ
X2 j ÿ 4
Xj =k5
…33†
jˆ1
k
X
(11)
k
X
Xj = rjsinaÅj
(12)
jˆ1
Sx2 = SX2j ± [(SXj)2/k]
Sy2 = SY2j ± [(SYj)2/k]
# 2001
(13)
(14)
Blackwell Science Ltd, Ecological Entomology, 26, 302±311
y2 ˆ
jˆ1
Xj = rjcosaÅj
Where Xj and Yj are the quantities X and Y respectively
applying eqns 1 and 2 to sample j; k is the total number of
replicates or samples.
jˆ1
k
X
jˆ1
2
k
X
Y2 j ÿ 4
jˆ1
xy ˆ
k
X
jˆ1
3
!2
Yj
=k5
…34†
jˆ1
"
Xj Yj ÿ
k
X
jˆ1
Xj
k
X
!
#
Yj =k
…35†
jˆ1
Test H0 = no mean direction:
F = {[k(k ± 2)]/2}{(X 2 Sy2 ± 2 X Y Sxy + Y 2 Sx2)/[Sx2Sy2 ±
(36)
(Sxy)2]}
With the critical value being the one-tailed F with 2 and k-2
degrees of freedom.