Content uploaded by Christopher R. Dickman
Author content
All content in this area was uploaded by Christopher R. Dickman on Jun 28, 2023
Content may be subject to copyright.
Oecologia (1991) 85: 464471
Oecologia
9 Springer-Verlag 1991
Mechanisms of competition among insectivorous mammals
C.R. Dickman
School of Biological Sciences, University of Sydney, N.S.W. 2006, Australia
Received May 28, 1990 / Accepted in revised form September 9, 1990
Summary. This study investigates the mechanisms of
competition between congeneric pairs of insectivorous
mammals in two communities in Australia and England.
Direct field observations showed that physical interac-
tions between species do not occur, whereas conspecifie
encounters are frequent. In field enclosures the smaller,
subordinate species in each community
(Antechinus stuar-
tii:
Marsupialia:
Sorex minutus:
Eutheria) remained alert
in the presence of the dominant species
(A. swainsonii,
S. araneus),
and moved quickly away when the latter
approached. The rate of prey capture by subordinate
individuals also increased immediately after removal of
the dominants. Hourly removals of some individuals of
the dominant species in each community over 24 h
produced hourly increases in the numbers of subordinate
individuals that were captured. The rapidity of these
responses suggests strongly that the dominant insec-
tivores in each community interfered with the resource
use of the subordinate species. Biomass of invertebrates
increased inconsistently or slowly within 3 months of
removal of the dominant insectivores; hence the rapid
responses by subordinate individuals in experiments were
not due to simple exploitation or tracking of resource
levels. The subordinate insectivores probably detected
and avoided contact with dominants instantaneously
using auditory or olfactory cues; reciprocal avoidance of
congeneric odours was demonstrated using odour-scent-
ed traps. Insectivorous mammals may usually compete
by interference (or encounter competition,
sensu
Schoener 1983). For dominant species within com-
munities the cost of interference is minimal and the bene-
fit of gaining exclusive access to resource-rich microhabi-
tats is high. Conversely for subordinate species the be-
nefit of temporarily exploiting the same rich microhabi-
tats may exceed the small costs of vigilance and move-
ment to nearby refugia.
Key words: Competition
Antechinus - Sorex -
Inter-
ference - Insectivores
Although interspecific competition probably occurs in
many ecological systems and may have important effects
on species distribution and abundance (Schoener 1983),
we have relatively little understanding of the mechanisms
by which competition occurs. Yet, knowledge of the
mechanisms of competition is crucial, both for predicting
individual behaviour and resource use and for under-
standing community processes (Tilman 1987).
Traditionally, two mechanisms of competition have
been distinguished. "Interference" refers to situations
where individuals directly reduce each others' access to
shared resources, as by fighting, whereas "exploitation"
refers to situations where individuals affect each other
indirectly by utilizing resources and depriving others of
their benefits (Park 1954; Miller 1967). Case and Gilpin
(1974) argued that interference is the most widespread
form of competition in nature, while Miller (1969) con-
cluded further that interference is the "more highly
evolved" of the two mechanisms. More recently,
Schoener (1983) noted that the terms "interference" and
"exploitation" are often confused in the ecological litera-
ture, and proposed a new terminology that distinguishes
six mechanisms of competition.
Among mammals, interference competition (includ-
ing encounter and territorial competition
sensu
Schoener
1983) has been inferred from observations of interspecific
territoriality and fighting (Brown 1971 ; Frye 1983; Wolff
et al. 1983), from microhabitat segregation at high but
not at a low population density (Dickman and Woodside
1983; Abramsky and Pinshow 1989) or from rapid re-
sponses by species in experiments where putative com-
petitor species have been removed (Lemen and Freeman
1983, 1986; Fox and Pople 1984). Further inferences
about interference have been drawn from observations
that some species avoid congeneric or confamilial odours
(Daly et al. 1980; Stoddart 1980; Randall 1981); how-
ever, the effectiveness of this form of interference is not
known.
In contrast, exploitation or consumptive competition
(Schoener 1983) has been inferred mostly from observa-
tions of differential food or microhabitat use among
465
coexisting species (Gliwicz 1981; Montgomery 1981;
Schroder 1987). Relatively slow responses of species to
competitor removal may also suggest some form of ex-
ploitation because there is presumed to be a time lag
between depletion and recovery of resource levels (Mun-
ger and Brown 1981). However, this suggestion has not
been critically tested, and it is unclear how slow a re-
sponse must be before exploitation, rather than interfe-
rence, can be implicated.
In this study, I attempt to distinguish the importance
of exploitative and interference mechanisms of com-
petition in two simple communities of insectivorous
mammals. In each community competition is asymmetri-
cal, with the larger of two species restricting the numbers
and microhabitat use of the smaller; competition prob-
ably occurs for food (Dickman 1986a, b, 1988a, 1989).
Five major questions are addressed:
1. Do direct (encounter or territorial) interactions occur
between the species?
2. Do indirect (olfactory) interactions occur between the
species?
3. Does the dominant species affect the rate of capture
of prey of the subordinate species?
4. How quickly does the subordinate species respond to
changes in the numbers of the dominant?
5. How quickly do resource (prey) levels respond to
changes in the numbers of the dominant species?
These questions are addressed by observing animals
under natural and enclosed conditions in the field, by
observing the responses of animals to traps scented with
conspecific and interspecific odours, and by assessing the
responses of the subordinate species and prey popula-
tions to fluctuations in the numbers of the dominant
species.
Materials and methods
Animal communities and study areas
The first community comprised two species of dasyurid marsupials,
Antechinus swainsonii
(ca. 50 g) and
A. stuartii
(ca. 20 g). Both
species occur in forest and heath over much of south-eastern Aus-
tralia, with
A. stuartii
extending slightly further inland into wood-
land (Dickman 1982). Both species take a variety of invertebrates,
although
A. swainsonii
also feeds occasionally on small skinks and
berries (Green 1972; Hall 1980; Dickman 1986a).
Observations on
Antechinus
were made in seven study areas
(1.72 - 3.70 ha) in the Brindabella Range, near Canberra, Australian
Capital Territory, between May 1978 and March 1982. The habitat
in each study area is a tall dense forest of
Eucalyptus viminalis,
E. radiata
and
E. fastigata,
with understory shrubs
Acacia dealbata,
A. melanoxylon, Pomaderris aspera
and
Rubus fruticosus,
and a
dense understory of ferns
Blechnum nudum
and
Pteridium esculen~
turn
(Florence 1973). All study areas are situated in valley floors
along the banks of creeks at altitudes of 740-825 m.
The second community comprised two species of eutherian
shrews,
Sorex araneus
(ca. 7 g) and
S. minutus
(ca. 4 g). These
species are widespread throughout much of Europe in grassland,
scrub and woodland habitats. Both species eat a wide range of
invertebrates (Pernetta 1976; Churchfield 1982).
Observations on shrews were made in eight study areas
(0.4-2.0 ha) within the City of Oxford, England, between November
1982 and July 1984. Each study area is rough grassland domi-
nated by
Danthonia
sp.,
Phleum pratense, Agropyron
spp. and
Arrhenatherum elatius,
with patches of herbs
Urtica dioica, Ranun-
culus
spp.,
Cirsium
spp.,
Vicia
spp. and isolated clumps of black-
berry
Rubus fruticosus.
The study areas are enclosed by roads,
buildings or watercourses, which appear to severely limit the move-
ments of shrews (Dickman and Doncaster 1987).
Trapping
Live-traps were used to carry out short - and long-term manipula-
tions of animal numbers in each study area, and to investigate
olfactory interactions. In Australia, Elliott aluminium live-traps
were set on grids or lines at intervals of 5-15 m. Most traps were
set on the ground, but, to capture
A. stuartii,
small numbers were
also placed in the lower limbs of trees. Traps were baited with rolled
oats, peanut butter, honey and bacon, provided with grass or cotton
wool for bedding, and checked in the morning and sometimes
evening for 3-12 days a month. In England, shrews were trapped
in Longworth traps baited with meat or fly pupae, after a pre-bait-
ing period of 24-36 h. Traps were set 10 m apart on grids, and
checked twice daily for three consecutive days a month. All cap-
tured mammals were identified, weighed, sexed, toe-clipped and
released at the point of capture. Population estimates in each study
area were expressed as the minimum numbers of animals known to
be alive each month (Krebs 1966).
I investigated the speed of response of the subordinate species
in each community to changes in the numbers of the dominant
species by effecting short-term reductions in the numbers of the
dominant in portions of study areas where animal numbers were
relatively high. Traps were checked hourly for 24 h, and some
individuals of the dominant species were removed for 1-3 h. In all
experiments removals reduced the local numbers of the dominant
species, on an hourly basis, to 0-90% of the numbers known to be
alive. The responses of the subordinate species in each community
were assessed simply by counting the numbers of subordinate in-
dividuals captured per hour throughout the 24-h experimental
periods. The results of long-term removal experiments have been
presented elsewhere, and are not further discussed (Dickman 1986a,
b, 1988a and unpublished).
To detect indirect olfactory interactions between species, I used
odour-impregnated traps. In each study area, in some months, I set
three aluminium traps at randomly selected trap stations. One was
~baited" with cotton wool containing the faeces, urine and other
bodily odours of one of the competing species in the community,
the second was baited similarly with cotton wool containing the
odours of the second species, the third contained odourless cotton
wool. Odour-impregnated cotton wool was obtained either from
captive animals or from individuals captured in traps off the grids,
and was used within 6-24 h of collection. All traps were cleaned
prior to use in deodorizing detergent and handled with gloves in the
field. The traps were set in "Y" - shaped clusters, with their entran-
ces facing inward, 10-15 cm apart, to ensure equal probability of
encounter for visiting mammals. Used traps were replaced im-
mediately; the impregnated cotton wool in unused traps was re-
plenished every 24 h. The responses of small mammals to inter-and
intraspecific odours are assessed by counting their frequencies of
capture in each trap, and testing for association by %~. Differences
in the level of association between species are detected by compar-
ing the deviation of expected and observed values in each cell of
the contingency tables, and expressing this as a percentage of the
expected values (Simpson et al. 1960).
Field observations
Animals were observed directly under natural conditions in the field
to assess whether encounter or territorial interactions occur.
Antechinus
species were observed in two study areas near nest sites.
466
Observations were made for 1-389 h a night, under red or white
torchlight, 1-3 times a month between October 1978 and November
1979. Observations began after dusk when
A. stuartii
becomes
active, and were discontinued by midnight. Shrews were observed
in four study areas between May and August 1983. Observations
were made 1-6 times a month in each area for periods of 89 to 289 h
early in the morning or around dusk. In each study area, observa-
tions were made at up to six sites that were known to be visited by
the two species. Shrews could be observed from distances of only
1 2 m without being obviously disturbed, but the dasyurids ap-
peared to be distracted by my presence at distances < 289 m. For
analysis, I divided observation periods for all species into 15-min
blocks, and counted the number of blocks in which one species,
both species or neither had been seen. Association was tested by ~2.
Observation periods that yielded no animals were omitted from the
analysis.
Enclosure observations
Temporary enclosures were constructed in the study areas to ob-
serve the effect of the dominant species of insectivore in each
community on the rate of capture of prey of the subordinate species.
All enclosures were circular and constructed from galvanized zinc
or aluminium; in Australia they were 2 m diameter and 0.5 m deep,
in England they were 1 m diameter by 0.3 m deep. One enclosure
was used in a single randomly chosen study area for each mammal
community. The enclosure was set on flat ground at sites where the
insectivores were known to forage.
Enclosure observations in Australia were carried out at night.
Shortly after dusk, one individual of each insectivore species was
live-trapped within 50 m of the enclosure. A light source (different
coloured beta light or ampulla containing cyalume; Buchler 1976)
was glued lightly to the fur between the scapulae of each animal;
they were then released into the enclosure. In England, enclosure
observations were made in the late afternoon and did not involve
attachment of a light source; procedures were otherwise the same.
Initial observations indicated that prey occurring naturally in the
enclosure were small and cryptic, and ! could not reliably determine
when an insectivore encountered one. Hence, for all observations,
I introduced known numbers of invertebrates into the enclosure
immediately after release of the insectivores. Ten individual in-
vertebrates were added initially to the 1 m diameter enclosure, and
12 to the 2 m one. Larvae of flour beetles
(Tenebrio molitor)
were
added in most experiments; however, I also used other species of
beetles (adults of
Anoplognathus
sp.). hemipteran bugs
(Euander
sp.), cockroaches
(Blatella germanica)
and spiders
( Araneus diade-
matus)
if these were known to be eaten by the insectivore and when
enough individuals could be collected locally. Only one species of
invertebrate was used in the enclosure observations at a time.
Observations began 15-30 rain after both insectivores and
invertebrates had been added to the enclosure, and continued until
the subordinate insectivore had encountered and killed > 3 of the
introduced invertebrates. The dominant species of insectivore was
then removed from the enclosure, and observations were continued
until _> 3 further invertebrates had been captured. The time taken
to capture prey by the subordinate insectivore was recorded by
stopwatch both in the presence and the absence of the dominant;
notes were also made on the pattern of movement and behaviour
of both species. Prey density was maintained at a constant level
throughout the enclosure observations by adding another in-
vertebrate each time one was eaten. Observations were discontinued
if no prey were encountered after 1 h. Enclosures were removed
after observations had finished on any night, and set up subsequent-
ly at another location on the same study area to ensure that no
individuals were observed on more than one occasion. Rates of prey
capture were calculated for the subordinate species of insectivores
before and after the removal of the dominant from the enclosure,
and differences detected by paired-sample t-tests.
Abundance of invertebrates
Invertebrates were censused in each study area to investigate their
speed of recovery after the dominant species of insectivore had been
removed. In Australia, 25 pitfall traps (diameter 70 mm with 5%
formalin as preservative) were set in each study area to sample
freely-moving fauna. Eight soil cores were also taken using a steel
corer (10 cm deep, 31.5 cm square), and sessile invertebrates extrac-
ted subsequently using Berlese funnels. Invertebrates were sampled
similarly in England, except that 32 pitfall traps were used in each
study area and the soil corer was 9 cm deep and 30 cm square.
Invertebrates were sampled at the beginning of or during mam-
mal trapping in each study area, immediately prior to experimental
removal of the dominant species of insectivore. In Australia,
A. swainsonii
was removed from 3 of the 7 study areas; in England,
S. araneus
was removed from 4 of the 8 study areas. Remaining
study areas were left unmanipulated as controls. Invertebrates were
re-sampled in all study areas about once a month for 3 months after
the dominant insectivore had been removed. In all sampling, in-
vertebrates were sorted, identified to order or below, measured and
then dried to constant weight. Invertebrates not eaten by the insec-
tivores (soil mites, springtails, ants, millipedes and slugs) were
omitted from the analyses. The effect of predation on invertebrates
is expressed as the percentage change in total dry invertebrate
biomass before and after species removal in the experimental study
areas compared with the percentage change in the controls.
Results
Do direct interactions occur between the species?
In 116 h of observation in both communities, I did not
see one direct encounter between a dominant and subor-
dinate insectivore. In contrast, single-species groups of
2-8 individuals were observed relatively frequently
(Fig. 1). Except for groups comprising juveniles with one
or more adults, in which interactions were usually amic-
r
._o
>
,.Q
O
"5
0
c
O"
14
12
8-
6-
4-
2-
0
16 -
14-
12-
10-
8-
6-
4-
2-
0
Antechinus stuartii
Sorex minutus
2 3 4 5 6
~,i Antechinus swainsonii
i
7 8
Group
size
~ Sorex araneus
2 3 4 5
Fig. 1. Frequency of observations of different sized groups of four
species of insectivorous mammals in two communities
Table 1. Co-occurrence of insectivorous mammal species at field
observation sites within the same 15-rain blocks of time
Antechinus swainsonii
§
Antechinus + 0 43
stuartii - 36 t 37
Z~ =9.29, P< 0.01
Sorex araneus
§
Sorex + 2 22
minutus - 56 128
Z~ =4.12, P<0.05
able, most encounters within groups @=29, 56%) ap-
peared to be agonistic. Audible vocalizations occurred in
all species, and chasing and fighting were observed in
A. stuartii and S. araneus. Other encounters elicited no
obvious intolerant behaviour, and involved play, sniff-
ing, courtship, huddling or simply moving apart.
In both communities, both species of insect+votes
were seen during the same observation period on 35
occasions (51%). However, only two joint occurrences of
dominant and subordinate species occurred within the
same 15-min time block, whereas occurrences of each
species alone were relatively frequent (Table 1). These
observations provide no evidence that competition be-
tween the species is mediated by direct encounters.
Do indirect (olfactory) interactions occur between the
species ?
In each community there was a strong tendency for the
subordinate species to avoid traps containing the odour
of the dominant, and to prefer traps containing con-
specific odour or no odour (Table 2). Captures of the
467
Table 3. Rates of prey capture by insectivorous mammals in field
enclosures, before and after removal of a dominant competitor
Insectivore n Prey No. prey captured/ t
species min. ~2 + S.D.
Before After
competitor competitor
removed removed
Antechinus 9
stuartii
8
6
Sorex 7
s
12
7
Tenebrio 0.21+0.13 0.28• 2.95*
molitor
larvae
Euandersp. 0.14• 0.19+0.08 2.25
Anoplognathussp. 0.40+0.39 0.49+0.23 0.44
Tenebrio 0.24+0.18 0.37J:0.27 2.66*
molitor
larvae
Araneus 0.59+0.38 0.63• 0.29
diadematus
Blatella 0.13:50.11 0.25+0.16 3.66*
9ermanica
*P < 0.05
dominant species also tended to be more frequent than
expected in traps containing conspecific odour. There
was a further tendency for reciprocal avoidance by the
dominant species of traps containing the odour of subor-
dinate species (Table 2).
These findings show that each species discriminates
and responds to trap odour, and that avoidance by the
subordinate species of the odour of the dominant is
marked.
Does the dominant species affect the rate of prey
capture of the subordinate species?
Rates of prey capture by the subordinate species gener-
ally increased after removal of the dominant species from
the field enclosures (Table 3). In A. stuart+i, increased
Table 2. Responses of insectivorous mammals to trap odours in two communities
Raw frequency data Deviation of expected
to observed values
A, stuartii A. swainsonii NO A. stuartii A. swainsonii No
odour odour odour odour odour odour
Antechinus stuartii 45 21 40 + 42.5 - 42.3 + 5.2
A. swainsonii 14 47 31 - 48.9 + 48.9 - 6.0
Z2 = 26.5, P< 0.001
S. minutus S. araneus No S. minutus S. araneus No
odour odour odour odour odour odour
Sorex minutus 15 4 17 + 43.7 -54.5 + 1.4
S. araneus 23 28 44 - 16.5 + 20.6 - 0.5
Z2 = 6.3, P< 0.05
468
=_
0
"6
0
..,Q
E
Z
12
10
8
6
4
2
0
[] Antechinus stuartii
nHiil,,iiiilili,ii i iL l lllll
[] Antechinus stuartii
[] Antechinus swainsonii
10
[] Sorex minutus
8 7
r] [7 Sorex araneus
,
ILl r I ktl ,,l f
10 -I []
Sorex minutus r'7
| 17f-] [] Sorex araneus I I
Time
Fig. 2. Hourly numbers of captures of dominant and subordinate
species of insectivorous mammals in two communities. Numbers of
the dominant species in each community (Anteehinus swainsonii,
Sorex araneus) were regulated by removal from the study areas
for 1-3 h
capture success appeared to be due to an increase in the
rate of movement, and hence rate of encounter with prey,
after A. swainsonii was removed. In the enclosure with its
larger congener, A. stuartii tended to remain still with its
eyes and ears alert; it moved rapidly away when A.
swainsonii approached to within ca. 50 cm. The foraging
movements of S. minutus were similar before and after
S. araneus was removed from the field enclosure, except
when the larger species approached to within ca. 25 cm.
At this close proximity S. minutus moved rapidly away
from its larger congener and resumed foraging again
after a flight of > 20 cm.
There was no evidence that the enclosures obviously
affected the behaviour of the insectivores, since the forag-
ing patterns of all species were qualitatively similar in the
enclosures and the natural environment. The results
therefore indicate that the rates of prey capture of the
subordinate species in each community are inhibited
simply by the presence of the dominant species.
oo
o3
m
E
0
35
E3
1.2
0.8
0.4
0
2.0
1.6
Antechinus
swainsonii
[
removed }
I
I0 0
t+7 T,
,'o .'Jo :+, i+
I I I I
1.2
0.8-
0.4
I
Sorex
arane
] .o
:
T'
I;
el 7
I
I [ I I I I )
0 2 4 6 8 10 12
Weeks
Fig. 3. Dry biomass of invertebrates in pitfall traps and litter sam-
ples before and after removal of a dominant species of insectivore.
Means are shown _ SD. Closed symbols, control study areas; open
symbols, removal areas. Circles, pitfall trap samples; diamonds,
litter samples
How quickly does the subordinate species respond to
changes in the numbers of the dominant ?
Hourly manipulations of the numbers of the dominant
species were carried out in the two communities over four
24-h periods. In three of these periods, the numbers of
dominant individuals at liberty in each trapping area
appeared to be negatively correlated with the numbers of
subordinate individuals captured (Fig. 2). Correlation
coefficients were not computed because of dependency in
the capture data from one hour to the next. These find-
ings suggest that the activity of the subordinate species
as shown by trapping responds on at least an hourly basis
to changes in the numbers of the dominant species.
How quickly does invertebrate abundance respond to
changes in the numbers of the dominant species?
Removal of the dominant species of insectivores
produced little consistent change in the biomass of pit-
fall-trapped invertebrates, but a tendency for a slow
recovery of the litter fauna (Fig. 3). Between the time of
removal of S. araneus and sampling 12 weeks later, the
biomass of litter invertebrates declined on average by
8.1% in the control study areas but increased by 15.9 %
in the removal areas (F= 6.51, P< 0.05); no other com-
parisons were statistically significant. Casual live-trap-
ping of insectivores during the period of invertebrate
sampling yielded very few S. araneus or A. swainsonii
469
(0-15% of the numbers originally removed), indicating
that the removal experiments had been effective. The
results suggest that the invertebrate populations re-
covered slowly after removal of the predators, and per-
haps required more than 3 months to show a clear re-
sponse.
Discussion
The results suggest clearly that competition between the
insectivores occurs by some form of interference. Al-
though no direct encounters were observed (see also
Dickman 1988b) individuals of the subordinate species
in each community actively avoided contact with the
dominant species on an instantaneous basis, and re-
sponded at least hourly to changes in the numbers of the
dominant. The rapidity of these responses contrasts
strongly with the slow and inconsistent recovery of in-
vertebrate biomass after removal of the dominant spe-
cies, and thus further supports the interpretation of inter-
ference.
Interference competition in other vertebrates is usu-
ally mediated by pheromones and other olfactory stimuli
(Jaeger and Gergits 1979; Roudebush and Taylor 1987)
and by growth inhibitors (Steinwascher 1978), but per-
haps most conspicuously by aggressive behaviour and
fighting. Aggression can be manifested by calls, displays
and other threatening behaviour (Morse 1971a, b); fight-
ing can involve chasing, mobbing and direct physical
contact, and may result in the death of subordinate
competitors (Barnett and Spencer 1951; Dow 1977). In-
terspecific aggression is likely to be advantageous to the
dominant species only if it can expect to acquire and
control a shared resource, and this expectation can de-
pend in turn on the structure of the habitat (Oksanen et
al. 1979). In open habitats, subordinates are easy to see
and evict, but in relatively complex habitats they may
escape aggression by hiding. In the present study, all
communities occupied sites with a high density of vegeta-
tion and uneven topography at ground level, where the
costs to one species of aggressively evicting the other
presumably outweigh the benefits gained by controlling
resources. Avoidance of encounters on a moment-to-
moment basis has been demonstrated among other com-
munities of small mammals inhabiting complex forest or
scrub environments (Crowcroft 1957; Hoffmeyer 1973;
Greenwood 1978; Hall and Lee 1982), but apparently
not among communities inhabiting more open areas.
The lack of direct encounters between the insectivore
species suggests that subordinate individuals detect the
presence of the dominant species at long range, and take
early evasive action. Detection may sometimes occur by
sight but, in the structurally complex habitats studied,
more probably occurs by sound and smell. In each of the
communities, I could often hear individuals of the domi-
nant species moving through vegetation or over dry leaf
litter, at distances up to 3 m. All the dominant species
appear to be capable of producing ultrasound (Sales and
Pye 1974; Poduschka 1977; D. P. Woodside, personal
communication), and all utter occasional squeaks while
foraging. In addition, A. swainsonii and S. araneus inter-
mittently produce soft whittering sounds while moving,
that may also alert the subordinate species to their
presence.
Discrimination of interspecific odours has been well
documented for species of shrews (Hawes 1976) and
rodents (Moore 1965; Randall 1981) as well as among
the present study species (Table 2). Although such dis-
crimination is usually assumed to facilitate reproductive
isolation, especially among congeners, it may also permit
avoidance of direct interspecific encounters. In the
present study, individuals of all species paused at ir-
regular intervals while moving, and sniffed the air. Al-
though the distance at which any individual of one spe-
cies can detect the presence of an individual of another
is unknown, it is likely to be much greater than 20-25 cm,
the distance at which I could distinguish the odours of
the dominant species.
It is also possible that subordinate individuals reduce
the probability of being in the same place as dominant
individuals by avoiding sites, such as runways and bur-
rows, where the activity and odours of the dominant are
likely to be most concentrated (e.g. Thomson and Pears
1962; Jaeger and Gergits 1979). However, there are two
reasons why such probabilistic avoidance is unlikely.
First, if each of the subordinate species avoided burrows
and other odour-impregnated sites, they would presum-
ably not respond to the removal of the dominant species
until all lingering odours had disappeared. However,
Fig. 2 shows that responses occurred within an hour. Sec-
ond, if avoidance was probabilistic, physical encoun-
ters would be expected to occur occasionally at least in
lightly scented areas, or in the open. Yet, no such encoun-
ters were seen, and none have been reported for the study
species elsewhere (Crowcroft 1957; Hall and Lee 1982).
Case and Gilpin (1974) argued that interference com-
petition between two species would be favoured if re-
source overlap was high and if the benefits to the interfer-
ing species in controlling that resource were high relative
to the cost. I have shown elsewhere that the present study
species overlap in diet and prefer the same sizes of prey,
and also that the effects of interference are relatively
great (Dickman 1988a). The costs of interference to the
dominant species also seem minimal. There is little short-
term or long-term effect on demography or resource use
of removal of A. stuartii on A. swainsonii (Dickman
1986a), and my observations in the present study suggest
that the dominant insectivores are almost "unaware" of
the presence of the subordinates. If interference is effec-
tive and relatively cost-free for the dominant species in
each community, it is relevant to ask whether the subor-
dinate species have counterbalancing advantages that
allow them to coexist.
In both communities, coexistence for the subordinate
species is possible because they exploit microhabitats
that are largely unavailable to the dominant species.
Both S. minutus and A. stuartii show some arboreal
activity (Holisova 1969; Dickman 1986a) and may es-
cape their larger congeners on a moment-to-moment
basis by climbing above ground. The smaller size of the
subordinate species may also allow them access to cracks
in logs or fissures in rocks and tree bark that are again
470
too small to permit access by dominants. However, de-
spite the apparent availability of such refugia in the study
areas, selection for microhabitat specialization is unlikely
for three reasons. First, the terrestrial microhabitats
occupied by the dominant species contain the largest and
most profitable prey for all the community members
(Dickman 1988a). Second, the refugium microhabitats
used by the subordinate species mostly contain epigeal
invertebrates that are not available in winter (Dickman
1986a); hence specialization on the prey associated with
these microhabitats may not be possible (Fretwell 1972).
Finally, all microhabitats exploited by the insectivores
are closely juxtaposed. Hence, if the subordinate species
can detect the approach of a dominant before an encoun-
ter occurs, the benefits of exploiting the resource-rich
terrestrial microhabitats presumably outweigh the small
costs of flight. A similar balance of advantages may
explain the coexistence of carnivorous weasels and
stoats. The smaller and more elongate weasels forage in
vole runways that are inaccessible to the stoat; their
access to larger prey elsewhere is limited by interference
from the stoat (King and Moors 1979; Simms 1979).
Although interference in the present communities has
a low cost and large effect, as predicted by Case and
Gilpin (1974), it is not clear why it should entirely pre-
clude direct interspecific contact. A possible explanation
is that, although the dominant species generally recog-
nize the odours of the subordinates, they may sometimes
mistake the smaller species for invertebrate prey. Ci-
cadas, mole crickets and locusts attain the sizes of the
small subordinate insectivores, and all are eaten by the
dominant species. The jerky movements of
A. stuartii
and
S. minutus
are also reminiscent of these insects. Since the
dominant insectivores in general are opportunistic
feeders, they presumably lack specific search images and
could approach and attack unwary subordinates by mis-
take (cf. Nishikawa 1987). Although it is unlikely that
the dominant insectivores would expend the energy re-
quired to subdue and eat subordinates, it should still be
advantageous for the smaller species to detect and evade
the larger species to avoid risk of injury. Strong avoid-
ance by subordinate species of salamanders of the odours
or physical presence of dominant species has also been
interpreted as an anti-predator response (Southerland
1986; Roudebush and Taylor 1987).
In summary, I suggest that interference is the pre-
dominant mechanism of competition among insec-
tivorous mammals. In most communities, dominant and
subordinate insectivores may maximize their foraging
returns by exploiting microhabitats containing the same
large, profitable prey (Dickman 1988a). Interference by
the dominant species is favoured because the cost is low
and the benefit of gaining exclusive access to such micro-
habitats is high. Conversely, interference is tolerated by
subordinate species because the benefit of obtaining even
temporary access to profitable microhabitats outweighs
the small costs of vigilance and flight to nearby refugia.
Despite the lack of direct physical interactions between
species of insectivores, the form of interference exhibited
accords most closely with Schoener's (1983) definition of
encounter competition.
Acknowledgements.
Research was carried out using the facilities of
the Zoology Departments of the Australian National University,
and the University of Oxford. I thank D.C.D. Happold and D.W.
Macdonald for much logistical support, D.L. Anderson, K. Dick-
man and C.A. McKechnie for helping to analyse samples of in-
vertebrates, and D.C.D. Happold, C.A. McKechnie and D.P.
Woodside for discussion and criticism. Many people assisted with
fieldwork, especially C.P. Doncaster and J. Winter. Funds were
provided by a Commonwealth Scholarship and Fellowship Plan
award (ANU), the Mammal Society and British Ecological Society
(Oxford).
References
Abramsky Z, Pinshow B (1989) Changes in foraging effort in two
gerbil species correlate with habitat type and intra-and inter-
specific activity. Oikos 56:43-53
Barnett SA, Spencer MM (1951) Feeding, social behaviour and
interspecific competition in wild rats. Behaviour 3:229-242
Brown JH (1971) Mechanisms of competitive exclusion between
two species of chipmunks. Ecology 52:305-311
Buchler ER (1976) A chemiluminescent tag for tracking bats and
other small nocturnal animals. J Mammal 57:173-176
Case TJ, Gilpin ME (1974) Interference competition and niche
theory. Proc Nat Acad Sci USA 71:3073-3077
Churchfield S (1982) Food availability and the diet of the common
shrew,
Sorex araneus,
in Britain. J Anim Ecol 51 : 15-28
Crowcroft P (1957) The life of the shrew. Max Reinhardt, London
Daly M, Wilson MI, Behrends P (1980) Factors affecting
rodents' responses to odours of strangers encountered in the
field: experiments with odour - baited traps. Behav Ecol Socio-
biol 6: 323-329
Dickman CR (1982) Some ecological aspects of seasonal breeding
in
Antechinus
(Dasyuridae, Marsupialia). In: Archer M (ed)
Carnivorous marsupials. Roy Zool Soc New South Wales, Syd-
ney, pp 139-150
Dickman CR (1986a) An experimental study of competition be-
tween two species of dasyurid marsupials. Ecol Monogr
56: 221-241
Dickman CR (1986b) An experimental manipulation of the inten-
sity of interspecific competition: effects on a small marsupial.
Oecologia 70:536-543
Dickman CR (1988a) Body size, prey size, and community structure
in insectivorous mammals. Ecology 69: 569-580
Dickman CR (1988b) Detection of physical contact interactions
among free-living mammals. J Mammal 69:865-868
Dickman CR (1989) Demographic responses of
An techinus stuartii
(Marsupialia) to supplementary food. Aust J Ecol 14:387-398
Dickman CR, Doncaster CP (1987) The ecology of small mammals
in urban habitats. I. Populations in a patchy environment.
J Anita Ecol 56:629-640
Dickman CR, Woodside DP (1983) A test of a competition model
with reference to three species of small mammals in south-
eastern Australia. Oecologia 60:127--134
Dow DD (1977) Indiscriminate interspecific aggression leading to
almost sole occupancy of space by a single species of bird. Emu
77:115-121
Florence RG (ed) (1973) A resource and management survey of the
Cotter River catchment. Dept Forestry, Australian National
University, Canberra
Fox B J, Pople AR (1984) Experimental confirmation of interspecific
competition between native and introduced mice. Aust J Ecol
9 : 323-334
Fretwell SD (1972) Populations in a seasonal environment. Prince-
ton University Press, New Jersey
Frye RJ (1983) Experimental field evidence of interspecific aggres-
sion between two species of kangaroo rat
(Dipodomys).
Oecolo-
gia 59: 74-78
Gliwicz J (1981) Competitive interactions within a forest rodent
community in central Poland. Oikos 37:353-362
471
Green RH (1972) The murids and small dasyurids in Tasmania.
Part 5,
Antechinus swainsonii
(Waterhouse 1840). Rec Queen
Vict Mus 46:1-13
Greenwood PJ (1978) Timing of activity of the bank vole
Cleth-
rionomys 91areolus
and the wood mouse
Apodemus sylvaticus
in
a deciduous woodland. Oikos 31:123-127
Hall S (1980) The diets of two coexisting species of
Antechinus
(Marsupialia: Dasyuridae). Aust Wildl Res 7:365-378
Hall S, Lee AK (1982) Habitat use by two species
of Antechinus
and
Rattus fuscipes
in tall open forest in southern Victoria. In:
Archer M (ed) Carnivorous marsupials. Roy Zool Soc New
South Wales, Sydney, pp 209-220
Hawes ML (1976) Odor as a possible isolating mechanism in
sympatric species of shrews
(Sorex vagrans
and
Sorex ob-
scurus).
J Mammal 57:404-406
Hoffmeyer I (1973) Interaction and habitat selection in the mice
Apodemus flavicollis
and
A. sylvaticus.
Oikos 24 : 108-116
Holisova V (1969) Vertical movements of some small mammals in
a forest. Zool Listy 18:121-141
Jaeger RG, Gergits WF (1979) Intra - and interspecific communica-
tion in salamanders through chemical signals on the substrate.
Anim Behav 27:150-156
King CM, Moors PJ (1979) On co-existence, foraging strategy and
the biogeography of weasels and stoats
(Mustela nivalis
and
M. erminea)
in Britain. Oecologia 39:129-150
Krebs CJ (1966) Demographic changes in fluctuating populations
of
Microtus californicus.
Ecol Monogr 36: 239-273
Lemen CA, Freeman PW (1983) Quantification of competition
among coexisting heteromyids in the southwest. Southwest Nat
28 : 41-46
Lemen CA, Freeman PW (1986) Interference competition in a
heteromyid community in the Great Basin of Nevada, USA.
Oikos 46: 390-396
Miller RS (1967) Pattern and process in competition. Adv Ecol Res
4: 1-74
Miller RS (1969) Competition and species diversity. Brookhaven
Symp Biol 22:63-70
Montgomery WI (1981) A removal experiment with sympatric
populations of
Apodemus sylvaticus
(L.) and
A. flavicollis
(Mel-
chior) (Rodentia: Muridae). Oecologia 51:123-132
Moore RE (1965) Olfactory discrimination as an isolating mecha-
nism between
Peromyscus maniculatus
and
Peromyscus polion-
otus.
Am Midl Nat 73:85-100
Morse DH (1971a) The foraging of warblers isolated on small
islands. Ecology 52:216-228
Morse DH (1971b) The insectivorous bird as an adaptive strategy.
Ann Rev Ecol Syst 2:177-200
Munger JC, Brown JH (1981) Competition in desert rodents: an
experiment with semipermeable exclosures. Science
211:510-512
Nishikawa KC (1987) Interspecific aggressive behaviour in sala-
manders: species - specific interference or misidentification?
Anim Behav 35:263-270
Oksanen L, Fretwell SD, Jarvinen O (1979) Interspecific aggression
and the limiting similarity of close competitors: the problem of
size gaps in some community arrays. Am Nat 114:117-129
Park T (1954) Experimental studies of interspecies competition. II.
Temperature, humidity, and competition in two species of
Tri-
bolium.
Physiol Zool 27:177-238
Pernetta JC (1976) Diets of the shrews
Sorex araneus
L. and
Sorex
minutus
L. in Wytham grass/and. J Anita Ecol 45:899-912
Poduschka W (1977) Insectivore communication. In: Seboek TA
(ed) How animals communicate. Indiana University Press,
Bloomington, pp 600-633
Randall JA (1981) Olfactory communication at sandbathing loci by
sympatric species of kangaroo rats. J Mammal 62: 12-19.
Roudebush RE, Taylor DH (1987) Chemical communication be-
tween two species of desmognathine salamanders. Copeia
1987:744-748
Sales G, Pye D (1974) Ultrasonic communication by animals. Chap-
man and Hall, London
Schoener TW (1983) Field experiments on interspecific competition.
Am Nat 122:240-285
Schroder GD (1987) Mechanisms for coexistence among three spe-
cies of
Dipodomys:
habitat selection and an alternative, Ecology
68:1071-1083
Simms DA (1979) North American weasels: resource utilization
and distribution. Can J Zool 57:504-520
Simpson GG, Roe A, Lewontin RC (1960). Quantitative zoology.
Harcourt, Brace & Co, New York
Southerland MT (1986) Behavioral interactions among four species
of the salamander genus
Desmognathus.
Ecology 67:175-
181
Steinwascher K (1978) Interference and exploitation competition
among tadpoles of
Rana utricularia.
Ecology 59:1039-1046
Stoddart DM (1980) The ecology of vertebrate olfaction. Chapman
and Hall, London
Thomson JA, Pears FN (1962) The functions of the anal glands of
the brushtail possum. Vict Nat 78:306-308
Tilman D (1987) The importance of the mechanisms of interspecific
competition. Am Nat 129:769-774
Wolff JO, Freeberg MH, Dueser RD (1983) Interspecific territorial-
ity in two sympatric species of
Peromyscus
(Rodentia : Criceti-
dae). Behav Ecol Sociobiol 12:237-242
