Oecologia (2003) 134:154–162
DOI 10.1007/s00442-002-1107-1
COMMUNITY ECOLOGY
M. Corrie Schoeman · David S. Jacobs
Support for the allotonic frequency hypothesis
in an insectivorous bat community
Received: 24 May 2002 / Accepted: 15 October 2002 / Published online: 9 November 2002
Springer-Verlag 2002
Abstract The allotonic frequency hypothesis proposes
that certain insectivorous bat species can prey upon moths
that can hear bat echolocation calls by using echolocation
frequencies outside the sensitivity range of moth ears. The
hypothesis predicts that the peak frequencies of bat
echolocation calls are correlated with the incidence of
moths in the diets of these bats. The aim of this study was
to test this prediction on a bat community dominated by
bats using low duty cycle echolocation calls, i.e. aerial
foraging, insectivorous species using frequency modulated calls. The community consisted of nine species, two
molossids, Sauromys petrophillus and Tadarida aegyptiaca, five vespertilionids, Eptesicus capensis, Eptesicus
hottentotus, Miniopteris schreibersii, Myotis tricolor, and
Myotis lesueuri, one rhinolophid, Rhinolophus clivosus,
and one nycterid, Nycteris thebaica. The insect fauna in
the habitat used by the bat community was suited to the
testing of the allotonic frequency hypothesis because
more than 90% of the moths comprising the insect fauna
were tympanate. These included Pyralidae (3.8%), Geometridae (44.9%), Notodontidae (3.8%), Arctiidae
(4.6%), Lymantriidae (0.8%) and Noctuidae (32.4%).
As predicted, peak echolocation frequency was correlated
with the incidence of moths in the diets of these nine
species (r=0.98, df=7, P<0.01). Furthermore, multivariate
analysis revealed that echolocation frequency (t=9.91,
n=129, P<0.001) was a better predictor of diet than
forearm length (t=5.51, n=129, P<0.001) or wing area
(t=–3.41, n=129, P<0.001). This suggests that the selection pressure exerted by moth hearing might have acted
directly on call frequency and secondarily on body size
and wing morphology, as part of the same adaptive
complex. It is unlikely that dietary differences were due to
temporal and spatial differences in the availability of prey
because the pattern of differences in skull morphology of
M.C. Schoeman · D.S. Jacobs ())
Department of Zoology, University of Cape Town, Private Bag,
Rondebosch 7701, South Africa
e-mail: [email protected]
Tel.: +27-21-6504011
Fax: +27-21-6503301
the nine species supported our dietary analyses. The skull
morphology of a bat represents a historical record of the
kind of diet it has become adapted to over its evolutionary
history. These results suggest that prey defences may
mediate other factors structuring bat communities, e.g.
competition. Competition may be reduced for those
species of bats that can circumvent prey defences.
Keywords Allotonic · Co-evolution · Diet ·
Echolocation · Morphology
Introduction
Aerially-foraging insectivorous bats use echolocation for
orientation and/or to locate their prey. When hunting,
these bats often emit intense pulses of sound (>110 dB)
and analyse the much fainter echoes reflected by objects
to detect, localise and possibly classify potential prey
(Schnitzler and Kalko 2001). However, echolocation has
the obvious drawback of potentially being used by prey as
an early warning system provided the prey could evolve
the necessary capacity to do so. Certain nocturnal insects,
including moths, lacewings, beetles and praying mantises
appear to have evolved tympanic ears as a defence against
bat predation (e.g. Roeder 1967; Miller 1983; Fullard
1987; Surlykke 1988).
With some unusual exceptions (Conner 1999), moth
ears appear to have no function other than to detect
approaching bats (Roeder 1975; Bailey 1991; Fullard and
Yack 1993). Firstly, tympanate moths are most sensitive
to frequencies between 20 and 60 kHz, coinciding with
the peak-frequency range of most echolocating bats
(Fullard 1987; Fenton et al. 1998). Secondly, moth
hearing sensitivity appears to be correlated with the
echolocation calls of the most common sympatric bats.
Moths sampled at sites with high bat diversity and density
have significantly higher auditory sensitivity, particularly
pronounced at both low-frequency (5–25 kHz) and highfrequency (80–110 kHz) ranges, than moths at sites with
low bat diversity and density (Fullard 1982, 1987). Lastly,
155
Fullard (1994), Fullard et al. (1997) and Surlykke, et al.
(1998) have shown that day-flying moths, no longer
subjected to bat-predation, display advanced auditory
degeneration.
Hearing combined with a complex suit of evasive
flight manoeuvres (Roeder 1967) allows tympanate moths
to be 40% more successful at evading bat predation than
non-tympanate moths (Roeder 1967; Rydell 1992;
Acharya and Fenton 1999). Thus, there is considerable
selective pressure on bats favouring adaptations to
overcome these prey defences.
Although tympanate ears are effective – some moths
can hear a bat echolocating up to 30 m away (Roeder
1967) – sensitivity falls off sharply to frequencies above
55 kHz (Fullard 1987; Surlykke 1988). One possible
counter-adaptation employed by bats may thus be to
exploit the frequencies above or below the moth’s
optimum hearing range (Novick 1977; Fenton and Fullard
1979). Fullard (1987) called these frequencies ‘allotonic
frequencies’. The allotonic frequency hypothesis (AFH)
predicts that the incidence of eared insects should be
highest in the diet of bats whose echolocation calls are
dominated by frequencies outside the 20–60 kHz range.
There is evidence supporting this (Fullard and Thomas
1981). For example, the Western Long-eared vespertilionid Euderma maculatum emits echolocation calls with
the most energy between 8 and 15 kHz and appears to
feed heavily on moths (Fullard and Dawson 1997). At the
other end of the spectrum, the hipposiderid Cloeotis
percivali echolocates as high as 212 kHz and feeds almost
exclusively on moths (Whitaker and Black 1976; Jacobs
2000).
Thus, across bat species with calls dominated by
frequencies >20 kHz the incidence of tympanate insects
in their diet should increase with echolocation frequency.
This seems to be the case whether the focus of the study is
global, incorporating a number of families of bats, or
regional focusing on a single community of bats. Jones
(1992) compared published echolocation and dietary data
from around the world on bats belonging to the families
Hipposideridae and Rhinolophidae. These families incorporate bats that use calls dominated by a single frequency
[constant frequency (CF) calls]. Jones found a positive
relationship between call frequency and the proportion of
moths in the diet of these bats and a negative relationship
between call frequency and the proportion of beetles (i.e.
non-tympanate insects). Bogdanowicz et al. (1999)
extended this study to include species that use frequency
modulated (FM) calls (species from the families Vespertilionidae and Molossidae). They found support for the
AFH in the form of a parabolic relationship between moth
consumption and echolocation call frequency for bats
whose echolocation calls are dominated by frequencies<100 kHz. However, this relationship was not significant for bat species using echolocation calls dominated
by sounds >100 kHz, suggesting that for these species
morphological characteristics rather than call frequency
may limit the range of potential prey items (Bogdanowicz
et al. 1999).
Jacobs (2000) argued that although the approach used
by Jones (1992) and Bogdanowicz et al. (1999) has strong
statistical validity, their data were collected in a number
of different ways and at different times. Their dietary and
echolocation data were not therefore collected at the same
time or in the same place for each species in the analyses.
Pavey and Burwell (1998) tested the predictions of the
AFH on three sympatric bat species using CF calls and
Jacobs (2000) tested them on a single bat community also
dominated by bats using CF calls. In both studies
echolocation and dietary data were collected at the same
time. As predicted by the AFH, echolocation frequency
was positively related to the proportion of moths in the
diets of the bats. In addition, echolocation frequency was
a better predictor of diet than wing loading, suggesting
that selection pressure exerted by moth hearing might
have acted directly on echolocation frequency and
secondarily on wing parameters (Jacobs 2000).
However, for the AFH to be generally applicable at the
level of the community, it must be tested on multiple
communities. Our aim was therefore to test the AFH on a
community that was different to that studied by Jacobs
(2000). The community we studied was temperate rather
than subtropical and consisted of at least nine bat species,
four more than that comprising the community studied by
Jacobs (2000). Our community was also dominated by
species using FM rather than CF calls.
Materials and methods
We did this study at Algeria Forestry Station (AFS, 3222'S;
1903'E) in the Cederberg, in the Western Cape Province of South
Africa during the summer and winter of 2001. Fieldwork was done
at the beginning of summer in September (21–26) in the middle of
summer in December (9–14) and towards the end of summer in
March (23–28) and April (20–26). One field trip was undertaken in
winter during July (13–18).
We caught bats using three or four mist nets at a time. We set
the nets, which ranged in length from 9 m to 12 m, a half hour
before dusk across forest flyways, along forest edges, across and
along a riverbed (the Rondegat River) and across suspected roost
entrances in toilets and office buildings of AFS. We monitored the
nets up to 0200 hours each night. In addition, at each trapping
location, we opened nets and monitored them for at least one full
night (from dusk to dawn) to ensure that we were catching species
that might be foraging after 0200 hours. We checked the nets
regularly, every 5–10 min during the first 2 h, and at least every
hour during the rest of the night to ensure that the bats were not
injured while caught in the nets. Only adult bats were included in
the analyses. We distinguished juvenile bats from adults by the
presence of cartilaginous epiphyseal plates in their finger bones
(Anthony 1988).
We measured body mass (to nearest 0.5 g) and forearm length
(to nearest 0.1 mm) of each captured bat. The extended right wing
of each bat (after Saunders and Barclay 1992) was photographed
with a Sony MVC-FD75 FD Mavica digital camera ensuring that
the camera was positioned at a 90 angle above the wing. We used
these images to measure wingspan, b (cm) and wing area, S (cm2),
including body area without the head, and the area of the
uropatagium; Norberg and Rayner 1987) using SigmaScan Pro 5
software (version 5.0.0; SPSS). These wing measurements were
used to calculate wing loading (WL=Mg/S where M is total body
mass in kg, g is gravitational acceleration in ms–2 and S is wing area
in m2), and aspect ratio (AR=b2/S; Norberg and Rayner 1987).
156
Diets of the species comprising the community were determined
by examining faecal pellets collected from captured bats placed in
cotton bags and kept for at least 1 h or overnight to ensure the
collection of enough faeces. We released bats unharmed after
clipping the fur on its back to avoid sampling the same bat twice.
Bats caught in September and December with identical forearm
measurements (see below) to individuals caught in March and April
were not considered for analyses because of the possibility of
clipped fur having re-grown. A minimum of five pellets from each
individual bat (Whitaker et al. 1996) and 20 pellets from each
species (Whitaker et al. 1999) were analysed. We teased apart each
pellet under 70% ethanol and identified the arthropod exoskeleton
fragments to order using a taxonomic text (Scholtz and Holm 1985)
and a reference collection of insects trapped at each site. We
calculated the percent volume of each insect order in each faecal
sample as described by Whitaker (1988).
Insects were trapped in two 22 W battery-operated black-light
traps (BioQuip Products, El Segundo, Calif.). We operated the traps
at the same time that we captured bats. The traps were placed at 1 m
and 4 m above the ground, about 100 m from the mist nets to ensure
that the light-traps did not affect bat activity at the nets. We
classified all insects to order. Moths were pinned and classified to
morpho-species (Oliver and Beattie 1993) for later identification to
family. Morpho-species have been shown to be useful as surrogates
for species, particularly in estimates of invertebrate biodiversity
and species turnover (Oliver and Beattie 1993, 1996).
We recorded echolocation calls of the bat species from handreleased bats. Bats were followed for as long as possible after
release to ensure that search phase calls were recorded (O’Farrell et
al. 1999). We released the bats just before dusk the day after they
were captured. This ensured that there was sufficient light for us to
follow them. We released each bat in the habitat in which we
caught it. The high frequency output of either a Pettersson D980
(Pettersson Electronik, Uppsala, Sweden) bat detector or a QMC
Model S200 (QMC Instruments, London, UK) bat detector, was
recorded on a Racal Store 4D tape recorder operated at 76 cm s–1. A
Tektronix oscilloscope (THS 710 A, 60 Mhz Scope/DMM Digital
Real-time 250 MS/s) was used to monitor the recording levels. The
slowed down recordings were analysed using BatSound Pro
software (version 3.20, Pettersson Elektronik, Uppsala, Sweden)
on a Compaq Presario 1400 notebook computer. We used a
sampling rate of 44,100 Hz (16 bits, mono) and a threshold of 15.
We measured peak frequency of the dominant harmonic from the
FFT power spectrum (size 512). A Hanning window was used to
eliminate effects of background noise.
We took skull measurements (see Jacobs 2000) from museum
specimens (Transvaal Museum and South African Museum) of all
species captured at AFS with dial callipers under a Leica Zoom
2000 dissecting microscope (15 magnification). Where available,
five male and five female skulls were measured for each species
from specimens that were collected from the Cederberg area (see
Appendix).
We regressed the arcsine of the mean percentage volume of
moths in the bats’ diet against the log of their mean peak call
frequency. To control for phylogeny, the relationship between call
frequency and diet was analysed by the method of independent
contrast (Felsenstein 1985) using CAIC software (version 2.0;
Purvis and Rambaut 1995). Felsenstein (1985) has shown that
species cannot be treated as independent data points in certain types
of analyses. Specifically, it is necessary to take into account
phylogenetic relationships to examine whether the evolution of two
characters is correlated. The phylogeny for the species comprising
the community was derived from Simmons and Geisler (1998)
(Fig. 1). We lacked branch-length information on actual times of
divergence for standardising contrasts therefore all branch lengths
were set equal. Doing so defines branch length as the number of
steps along a branch as indicated by the cladistic analysis (Garland
et al. 1992) and hypothesises a punctuational mode of evolution
(Grafen 1989). Previous work has indicated that the use of branch
lengths so defined is preferable to obtaining correlations with no
correction for phylogeny (Martins and Garland 1991). In some
Fig. 1 Phylogeny of species (after Simmons and Geisler 1998)
making up the bat community at Algeria Forestry Station, South
Africa
cases, a diversity of different branch length definitions will yield
essentially the same results (Garland et al. 1992).
We did a backward stepwise regression (Zar 1999) to determine
whether call frequency or morphology is the better predictor of diet.
Percentage moths in the diet was taken as the dependent variable,
and flight parameters (see above, including FA) and peak call
frequency as the independent variables. The percentage volume of
moths in the diets was arcsine transformed and the flight parameters
were log transformed. Bats for which we had dietary, wing, and
peak call frequency data were included in the analysis. In all
regressions, we plotted the residuals against their normal scores to
test the normality of the residuals. The residuals were normally
distributed in all regression models. We also did principal
components analysis on the skull morphology data to investigate
the correspondence, if any, between skull morphology and diet. We
used STATISTICA (version 5.5; StatSoft) for all statistical
analyses.
Results
Diet, echolocation and morphology
We did not catch any bats during the winter month of
July. It rained four out of the five sampling nights and bat
activity was low. We only detected one rhinolophid pass
on three out of five nights with the bat detector. In
contrast, we caught nine species of bats belonging to four
families over 23 nights of trapping (n=129 bats) during
the summer months. There were two molossids, Sauromys
petrophilus and Tadarida aegyptiaca, five vespertilionids,
Eptesicus capensis, Eptesicus hottentotus, Miniopteris
schreibersii, Myotis tricolor, and Myotis lesueuri, one
rhinolophid, Rinolophus clivosus, and one nycterid,
Nycteris thebaica (Table 1). Only one individual for both
Myotis tricolor (in September) and Miniopteris schreibersii (in March) were caught. Myotis lesueuri was not
caught during March, and N. thebaica was not caught
during April. For all other species, we caught at least one
individual every month (i.e. in March, April, September
and December). None of the bats called above 100 kHz
with R. clivosus calling at the highest peak frequency
157
Table 1 Mean € SD echolocation and wing parameters of the nine
species of bats caught at Algeria Forestry Station: Eptesicus
capensis (E.c.), Eptesicus hottentotus (E.h.), Miniopterus schreiberBat species
No. of bats
Echolocation parameters
Peak frequency (kHz)
Call type
Intensity
Wing parameters
Forearm length (cm)
Wing area (cm2)
Wing span (cm)
Wing loading (Nm–2)
Aspect ratio
sii (M.s.), Myotis tricolor (M.t.), Myotis lesueuri (M.l.), Sauromys
petrophillus (S.p.), Tadarida aegyptiaca (T.a.), Rhinolophus clivosus (R.c.), and Nycteris thebaica (N.t.)
E.c.
10
E.h.
6
M.s.
1
M.l.
4
M.t.
1
S.p.
51
T.a.
43
R.c.
5
N.t.
8
41.9€1.4
LDFMa
High
31.8€1.0
LDFM
High
53.6
LDFM
High
36.8€0.9
LDFM
High
50.0
LDFM
High
29.2€1.8
LDFM
High
22.8€1.9
LDFM
High
92.7€0.5
HDCFb
High
77.4€2.7
LDFM
Low
35.9€1.2
87.6€7.4
24.0€0.8
7.3€1.0
6.6€0.4
48.9€1.4
173.7€14.1
33.0€0.5
10.3€1.6
6.3€0.6
46.6
137.7
30.9
7.8
6.9
36.0€0.4
93.3€1.6
23.8€1.1
7.5€1.4
6.1€0.7
46.9
156.6
30.6
6.9
6.0
39.0€1.0
93.1€11
26.3€1.9
11.2€1.1
7.4€0.6
45.1€1.2
115.2€8.4
30.8€1.0
13.1€1.6
8.3€0.6
53.4€1.5
175.9€5.8
33.8€1.1
9.2€0.8
6.5€0.3
48.2€0.8
155.4€7.4
28.7€1.1
8.3€0.9
5.3€0.3
a
Low
b
duty echolocation dominated by frequency modulated calls
High duty echolocation dominated by constant frequency calls
Table 2 Mean € SD percent volume of prey categories in the diets of the nine bat species caught at Algeria Forestry Station: abbreviations
of names as in Table 1
Bat species
No. of bats
No. faecal pellets
E.c.
10
89
E.h.
6
67
M.s.
1
20
M.l.
4
39
M.t.
1
11
S.p.
51
464
T.a.
43
314
R.c.
5
46
N.t.
8
51
Prey category
Lepidoptera
Coleoptera
Hemiptera
Diptera
Hymenoptera
Trichoptera
Orthoptera
Neuroptera
Ephemeroptera
Arachnida
Unknown
0
12.5€21.5
46.6€33.7
37.6€28.4
0
0
0
0
2.3€5.5
0
1€3.2
0
88.3€9.0
1.5€3.0
4.5€9.0
4.8€6.2
0
0
0
0
0
1€2
6
6
33
49
0
6
0
0
0
0
6.3€4.9
32.5€22.8
47.3€37.3
6.8€13.5
2.5€5.0
0
0
1.5€3.0
0
3.3€6.5
0
50
25
0
25
0
0
0
0
0
0
0.3€2.2
19.5€23.1
32.9€23.7
37.8€31.2
7.9€16.2
0.4€2.9
0
0.6€3.0
0
0
1.0€3.3
5.4€13.4
25.5€30.7
19.2€25.6
34.6€33.4
8.4€22.2
0
0.6€3.4
1.5€6.5
2.7€7.7
0
1.4€7.1
62.6€30.8
29.4€24.8
2.0€4.5
0
1.4€3.1
0
0
0
0
0
4.6€6.4
34.9€27.3
21.3€17.6
1.5€4.5
3.8€10.6
1.6€4.6
0
23.5€19.6
0
0
13.5€27.4
0
0
(92.7 kHz), and T. aegyptiaca at the lowest (22.8 kHz).
The peak frequency for Myotis tricolor was obtained from
Taylor (1999), since we were unable to obtain clear calls
for this species. The peak frequencies recorded for these
species correspond with published data (Taylor 1999;
Jacobs 2000).
On average 8.2 faecal pellets per individual bat were
analysed but<20 pellets for Myotis tricolor could be
obtained from the single individual caught (Table 2). Five
species took moths, but only two species, N. thebaica and
R. clivosus, had more than 10% of moths in their diets
(Table 2). The percentage moth in the diets of the nine bat
species did not fluctuate between the four months
(Kruskal Wallis, all Ps>0.3). There was a positive
relationship between peak call frequency of the bats and
the proportion of moths in their diet (Fig. 2). When we
controlled for phylogeny using independent contrasts, the
relationship remained significant (r=0.85, n=8, P<0.01).
Stepwise backward regression of moth incidence
(dependent variable) against peak echolocation frequency
and wing parameters, including forearm length (Table 1)
yielded a model (r=0.77, F(3,125)=60.1, P<0.001) which
eliminated all parameters except peak echolocation
call frequency (t-test: t=9.91, n=129, P<0.001), forearm
length (t=5.51, n=129, P<0.001) and wing area (t=–3.41,
n=129, P<0.001). Both peak echolocation call frequency
and forearm length were positively associated with moth
incidence while wing area was negatively associated.
Peak echolocation call frequency was a better predictor of
moth incidence than forearm length or wing area by virtue
of its higher t-value.
The first two unrotated principal components (PCs)
accounted for 81% of the total variance in the skull
morphology of the nine species and separated the bats
into nine groups along species divisions (Fig. 3A).
Plotting the factor loadings resulted in the clear separation
of condyle height (g) from the rest of the variables
(Fig. 3B). On the opposite side, skull length (a), length of
maxillary tooth row (b) and length of dentary (d)
clustered together (Fig. 3B). Varimax rotation did not
clarify or alter these patterns appreciably. These patterns
were interpreted as follows. Firstly, PC1 was a measure of
skull size and bats that loaded high on PC1 had relatively
larger skulls than bats that loaded low on PC1 (Table 3).
158
Fig. 2 Relationship between the peak echolocation frequency and
the mean percentage by volume of moths in the diets of nine
species of bats caught at Algeria Forestry Station, South Africa:
Rhinolophus clivosus (Rc), Nycteris thebaica (Nt), Miniopterus
schreibersii (Ms), Myotis tricolor (Mt), Myotis lesueuri (Ml),
Eptesicus capensis (Ec), Eptesicus hottentotus (Eh), Sauromys
petrophillus (Sp), and Tadarida aegyptiaca (Ta). The curve of best
fit (not shown) y=5.8x2–18.1x+14.2, r2=0.97, P<0.01 where y=arcsine (% moth in the diet) and x=Log(peak echolocation frequency
in kHz)
Therefore, Myotis lesueuri and E. capensis have smaller
sized skulls than either E. hottentotus or R. clivosus, while
the remaining bats have intermediate sized skulls.
Secondly, PC2 was a measure of skull robustness. There
was a significant difference in all skull parameters
between the bats (Table 3). A post hoc Tukey-test
revealed that E. hottentotus, which loaded highest on
PC2, had a significantly longer condyle than any of the
other bat species (all Ps<0.01), except for Myotis tricolor
(P=0.9) and Miniopteris schreibersii (P=0.1). Compared
to the other species, E. hottentotus had a longer coronoid
process (Tukey test; all Ps<0.01) and dentary thickness
(Tukey test; all Ps<0.01, except for Myotis tricolor, P=1.0
and R. clivosus, P=0.8). These characteristics are indicative of a robust skull (Freeman 1979, 1981). Hence, bats
that load high on PC2 have skulls that are more robust and
eat more hard-bodied prey, i.e. non-tympanate prey such
as beetles. Bats that load low on PC2 have more gracile
skulls, indicative of a diet comprised of soft-bodied prey,
such as moths. This corresponds with the dietary analysis
of the nine species of bats. N. thebaica and R. clivosus
which had the highest incidence of moths in their diets,
loaded lowest on PC2. E. hottentotus took no moths, but
had the highest incidence of beetles in its diet and loaded
the highest on PC2 (Table 2).
Insect and moth abundance
Flies and moths dominated the nocturnal insect fauna at
AFS. This was the case throughout the data collecting
periods. These two insect categories represented, respectively, 47% and 35% of the number of items recorded
Fig. 3A, B Principal component analysis of skull parameters taken
from Rhinolophus clivosus (Rc), Nycteris thebaica (Nt), Miniopterus schreibersii (Ms), Myotis tricolor (Mt), Myotis lesueuri
(Ml), Eptesicus capensis (Ec), Eptesicus hottentotus (Eh), Sauromys petrophillus (Sp), and Tadarida aegyptiaca (Ta). A Plot of
component scores for individual bats on the first two principal
components. B Plot of component weights for each skull measurement on the first two principal components. Abbreviations are: a
length of skull (from the occipital to the alveolus of the canine); b
length of the left maxillary tooth row (from the front of the left
fourth premolar to the back of the left third molar); c distance from
the anterior surface of the mandibular fossa to the origin of the
masseter muscle (bottom of the left angular process); d length of
the dentary (from the back of the left condyle to the epiphysis of the
dentary); e length of the masseter muscle scar (from the back of the
left condyle to the furthest extend of the scar on the left dentary); f
distance from the top of the left condyle to the insertion of the
masseter muscle (bottom of the left angular process); g height of
the condyle (top of the left condyle to the plane of the alveoli of the
left first and second molar); h height of the coronoid process (top of
the left coronoid process to the plane of the alveoli of the left first
and second molar); i dentary thickness (from the plane of the
alveoli of the left first and second molar to the bottom of the left
dentary). (After Jacobs 2000)
(n=8,557). However, ca. 90% of the Diptera were small
(<7 mm body length). Moths were by far the most
abundant medium and large sized insect prey found in the
light traps. There was no significant difference amongst
the 4 months in the number of individuals caught
weighted by the number of trapping hours for each of
the insect orders. The only exception was that significantly more Diptera were caught in September (F=5.4,
df=3, 9, P<0.05).
159
Table 3 Mean € SD skull parameters (mm) of the nine bat species caught at Algeria Forestry Station: abbreviations of names as in Table 1.
Skull parameter abbreviations as in Fig. 3. There is a significant difference between the species in all skull parameters
Bat Species
No. of bats
E.c.
105
Skull parameters
a
14.6€0.6
b
3.8€0.2
c
6.2€0.1
d
10.4€0.5
e
3.2€0.2
f
2.0€0.1
g
2.2€0.4
h
3.2€0.1
i
1.3€0.1
a
E.h.
10
M.s.
8
M.l.
3
M.t.
10
S.p.
10
T.a.
7
R.c.
10
N.t.
F-valuea
19.8€0.6
5.2€0.2
8.5€0.3
14.8€0.4
4.8€0.3
2.9€0.2
3.0€0.6
5.0€0.4
1.7€0.1
15.2€0.2
4.6€0.2
5.4€0.2
11.3€0.2
3.1€0.2
2.2€0.3
2.5€0.3
2.4€0.2
1.1€0.1
13.7€0.4
3.8€0.2
5.9€0.1
10.2€0.3
2.8€0.1
2.0€0.1
2.3€0.3
2.9€0.2
1.1€0.1
17.6€0.7
5.4€0.1
7.8€0.1
13.1€0.3
4.0€0.2
2.5€0.3
2.7€0.3
3.5€0.4
1.6€0.2
17.1€0.3
4.8€0.2
5.8€0.3
12.1€0.5
3.7€0.2
2.9€0.2
2.1€0.3
2.3€0.2
1.4€0.1
17.3€0.3
5.1€0.1
7.7€0.2
12.4€0.2
3.4€0.1
2.6€0.1
2.3€0.3
3.3€0.3
1.1€0.1
22.0€0.9
6.4€0.2
7.1€0.2
15.2€0.4
4.3€0.3
3.0€0.3
1.6€0.1
2.9€0.2
1.6€0.2
19.5€0.4
5.5€0.2
6.9€0.3
12.9€0.2
4.0€0.2
2.5€0.1
1.4€0.1
2.7€0.1
1.2€0.1
241.9
186.8
164.7
186.4
66.2
49.0
22.0
82.4
26.7
One-way ANOVA-tests; all P<0.01
The 2,956 moths that were trapped were classified to
278 morpho-species belonging to 15 families. Not all
families of moths have been tested for hearing, hence for
the purposes of this study we followed Scoble’s (1992)
classification of families composed mostly of species
which can hear ultrasound. These families and the
proportion they made up in the AFS moth fauna were
Pyralidae (3.8%), Geometridae (44.9%), Notodontidae
(3.8%), Arctiidae (4.6%), Lymantriidae (0.8%) and
Noctuidae (32.4%). Thus, most of the morpho-species
belonged to the Geometridae and the Noctuidae. The nontympanate families captured included Chrysopolomidae
(0.4%), Crambidae (0.8%), Ethmiidae (0.4%), Eupterotidae (2.7%), Limacodidae (1.1%), Psychidae (0.4%),
Saturniidae (1.4%), Thyretidae (1.9%) and Tortricidae
(0.8%). Tympanate moths, in terms of morphospecies
richness, predominated in all trapping months, and their
overall proportion was 90% of the total moth population.
This dominance was also evident in terms of number of
moths. The overall proportion of tympanate moths was
83% of the total number of moths.
Discussion
The significant relationship between peak call frequency
and the incidence of moths in the diet of bats (Fig. 2)
supports our interpretation of the allotonic frequency
hypothesis. A minor caveat is the small sample sizes for
Myotis tricolor and Miniopteris schreibersii. However,
the diet of Myotis tricolor sampled at De Hoop Nature
Reserve, Western Cape Province of South Africa (n=5
bats) contained no moths, but was dominated by
coleopterans and hemipterans (unpublished data). Similarly, dietary analyses for Miniopteris schreibersii based
on larger sample sizes indicate that moths may comprise
up to 25% of its diet (Jacobs 2000). However, this does
not alter the significance of the relationship between call
frequency and diet (r=0.78, n=8, P<0.01).
The large-eared N. thebaica is a gleaner that homes in
on sounds emanating from its prey (Fenton et al. 1983). It
could thus be argued that N. thebaica should not be
included in the analysis because the AFH deals with aerial
hawking bats that use echolocation calls more or less
continuously to catch insects. Gleaning bats that rely on
prey-generated sound do not give moths a signal on which
to base their evasive manoeuvres (Faure et al. 1993).
However, Fenton et al. (1983) found that echolocation
was still important for N. thebaica to track and catch prey.
We therefore felt justified in including N. thebaica in the
analysis (see also Fullard and Thomas 1981).
However, frequency might not be the only factor in the
echolocation calls of bats that influences whether or not
they can prey on moths (Fullard and Thomas 1981). For
example, although the peak frequency used by N.
thebaica is high, the calls of this species are also of low
intensity and very short. Both the latter characteristics are
likely to make its calls less detectable by moths (Jones
and Waters 2000). Furthermore, the high proportion of
moths in the diet of R. clivosus might be due to its use of
long CF calls adapted for detection and classification of
prey rather than its high echolocation frequency (Schnitzler and Kalko 2001). Large-winged insects such as moths
may be more easily detected by R. clivosus because they
are likely to produce greater amplitude modulations than
smaller-winged insects. The difference between the
proportions of moths in the diet of R. clivosus compared
to other bats in the community using FM calls, might
simply be because the latter group of bats have greater
difficulty detecting moths.
If most moths at AFS were non-tympanate, then the
degree of difference in the proportion of moths in the
diets of these bats could not be explained by the allotonic
frequency hypothesis. Instead it would suggest that moths
are using defences other than hearing to evade bat
predation and that some bats are using mechanisms other
than allotonic frequency echolocation calls to circumvent
these defences. However, although the faecal analysis
does not allow the identification of moths eaten down to
species level, 90% of the moth morpho-species that we
caught in light traps at the foraging sites of the bats at
AFS were tympanate. This is similar to the proportions
reported by other studies. Based on the families of moths
known to be present in southern Africa, Fenton and
160
Fullard (1979) estimated that 85% of moths in southern
Africa are tympanate. This predominance of tympanate
moths is not unique to southern Africa. Pavey and
Burwell (1998) sampled a number of tropical and
subtropical sites in Australia and found that at least
80% of moths at these sites were tympanate while Fullard
(1990) reported the prevalence of tympanate moths at
temperate and tropical localities to be as high as 95%.
Analysis of the skull morphology (Fig. 3) suggests that
the diets of these bats are unlikely to vary, in terms of the
amount of soft-bodied tympanate prey eaten, to an extent
where the relationship between diet and call frequency
disappears. There is a relationship between skull morphology and diet in insectivorous bats (Freeman 1979,
1981; Jacobs 1996; Bogdanowicz et al. 1999). Species
with more robust skulls, i.e. with characteristics such as
larger masseter muscles, a greater sagital crest, longer
condyle lengths and coronoid processes, usually take
more hard-bodied prey such as beetles, while species with
more gracile skulls usually take more soft-bodied prey
such as moths. This pattern can also occur within the
same species. For example, Jacobs (1996) found differences in the jaw morphology of two populations of
Lasiurus cinereus (Vespertilionidae), corresponding to
differences in the amount of soft-bodied versus hardbodied prey eaten. The skull morphology is thus an
indirect record of the type of prey (i.e. soft bodied or hard
bodied) that a bat species has become adapted to over its
evolutionary history. Analysis of the skull morphology of
the nine species making up the AFS community (Fig. 3)
shows a strong correspondence between the general diet
and the robustness of the skull. The more robust skull of
E. hottentotus is indicative of the more hard-bodied
insects such as beetles that it eats. Moreover, the similarly
sized but more gracile skulls of N. thebaica and R.
clivosus suggests that they have become adapted to feed
on more soft-bodied prey such as moths. Thus, although
E. hottentotus has a skull that could process moths, the
fact that it does not eat moths suggests that other factors,
e.g. echolocation, might influence its diet.
Contrary to Bogdanowicz et al. (1999), who found a
negative relationship, and Jacobs (2000), who did not find
any relationship, we found a positive relationship between
forearm length and the proportion of moths in the diet of
the bats comprising the AFS community. This is
supported by the results of the principal components
analysis (Fig. 3). Species that ate a large percentage of
moths, had relatively larger, albeit more gracile, skulls.
Body size (as measured by forearm length) rather than
call frequency may therefore be responsible for differences in the incidence of moths in the diet of the species
considered here. This is unlikely, however, because call
frequency was a better predictor of diet than forearm
length and despite E. hottentotus being one of the largest
bats in the community, its diet was dominated by beetles
rather than by moths. Instead, the relationship between
forearm length and diet is probably related to the size of
moth prey. Larger bats, i.e. N. thebaica and R. clivosus,
echolocate at higher frequencies and ate a larger propor-
tion of moths than smaller bats. Moths were by far the
most abundant medium- to large-sized insect prey found
in the light traps, and the large size of these bats probably
facilitated their exploitation of these larger prey items.
The negative relationship between wing area and the
incidence of moths in the diet is more difficult to explain.
However, the decrease in wing area with an increase in
body size could be related to higher flight speeds needed
to catch larger, and therefore probably faster flying,
moths. Thus, selection probably acted primarily on
echolocation call frequency and secondarily on body size
and wing morphology permitting the capture of large
tympanate insects.
High frequency sound is attenuated by the atmosphere
to a greater degree than low frequency sound (Lawrence
and Simmons 1982). The detection range of bats using
high call frequencies is therefore shorter than that of bats
using low call frequencies (Barclay and Brigham 1991).
Bats using high frequencies therefore have to be smaller
to increase their flight manoeuvrability so that they can
react over short detection distances. Body size is therefore
a good negative predictor of call frequency (von Helversen 1989; Barclay and Brigham 1991; Jones 1996;
Bogdanowicz et al. 1999). However, at call frequencies
<100 kHz atmospheric attenuation gets progressively less
severe and does not diminish detection range as much as
it does at call frequencies >100 kHz (Lawrence and
Simmons 1982). Atmospheric attenuation therefore constrains body size to a lesser extent in species echolocating
at <100 kHz than in species echolocating at >100 kHz.
This explains why Bogdanowicz et al. (1999) found a
relationship between call frequency and forearm length
only for bats calling at >100 kHz. It also explains why
Jacobs (2000) found no relationship. Three of the five
species in his analysis called at frequencies below
100 kHz. A consideration of our results and those of
Bogdonawicz et al. (1999), Bowie et al. (1999) and
Jacobs (2000) suggests that echolocation is a major
constraint on bat foraging strategies. The constraining
influence of morphology (Aldridge and Rautenbach 1987;
Norberg and Rayner 1987) thus appears to be mediated by
the constraints of echolocation strategies.
In conclusion, the AFH appears to be valid in both the
community on which we worked and the very different
community studied by Jacobs (2000). It might therefore
be valid for a wide range of bat communities. However, it
is crucial that the AFS be tested on a number of
communities and in opposition to other factors, e.g.
competition, that influence community structure to
determine the extent of its contribution to such structure.
Acknowledgements We thank Daniel De Lemos Ribeiro, Geeta
Eick, Mark Kratzman, Rory Millam, Lee Potter, Aiden RossGillespie, Natalie Schoeman and Samantha Stoffberg for their
assistance in the field. We thank Robert Barclay and James Fullard
for reading an earlier version of this manuscript as well as two
anonymous referees for their valuable comments. We also thank the
Algeria Forestry Station and the Cape Nature Conservation Board
for permission to work at Algeria Forestry Station, and the
Transvaal and South African Museums for the loan of their
161
specimens. This study was funded by a grant to D.S.J. from the
University Research Committee of the University of Cape Town
and grants to D.S.J. and C.M.S. from the National Research
Foundation of South Africa.
Appendix
Transvaal Museum (TM) and South African Museum (ZM) codes
for the bat skulls that were measured for the principal component
analysis.
Eptesicus capensis – TM35095, TM35105, TM38434, TM38435,
TM39389, TM39396, TM39402, TM40525, TM40559, TM40632
Eptesicus hottentotus:
TM40630, TM40631
–
TM35150,
TM38411,
TM38412,
Miniopteris schreibersii: – TM35149, TM38421, TM40671TM40676, TM41521, TM42459
Myotis leseuri: – TM29511, TM35121, TM35152, TM35172,
TM38422- TM38424, TM38428
Myotis tricolor: – TM40577, TM40578, TM41690
Nycteris thebaica: – ZM5578, ZM5580, ZM5582, ZM5584,
ZM5586, ZM5591, ZM5595, ZM5598, ZM16407, ZM16412
Rinolophus clivosus: – ZM10132, ZM10306, ZM13185, ZM33493,
ZM33496, ZM35606, ZM35666
Sauromys petrophillus: – TM24081, TM24103, TM24277,
TM24278, TM24279, TM24280, TM25409, TM25607, TM25608,
TM25560
Tadarida aegyptiaca – TM40605, TM40618, TM38431, TM35086,
TM35122, TM38429, TM40617, TM38432, TM35153, TM38430
References
Acharya L, Fenton MB (1999) Bat attacks and moth defensive
behaviour around streetlights. Can J Zool 77:27–33
Aldridge HDJN, Rautenbach IL (1987) Morphology, echolocation
and resource partitioning in insectivorous bats. J Anim Ecol
56:763-778
Anthony ELP (1988) Age determination in bats. In: Kunz TH (ed)
Ecological and behavioural methods in the study of bats.
Smithsonian Institution Press, Washington, D.C. pp 47–58
Bailey WJ (1991) Acoustic behaviour of insects. An evolutionary
perspective. Chapman and Hall, London
Barclay RMR, Brigham RM (1991) Prey detection, dietary niche
breadth, and body size in bats: why are aerial insectivorous bats
so small? Am Nat 137:693–703
Bogdanowicz W, Fenton MB, Daleszczyk K (1999) The relationships between echolocation calls, morphology and diet in
insectivorous bats. J Zool 247:381–393
Bowie RCK, Jacobs DS, Taylor PJ (1999) Resource use by two
morphologically similar insectivorous bats (Nycteris thebaica
and Hipposideros caffer). S Afr J Zool 34:27–33
Conner WE (1999) “Un chant d’appel amoreau”: acoustic
communication in moths. J Exp Biol 202:1711–1723
Faure PA, Fullard JH, Dawson JW (1993) The gleaning attacks of
the northern long-eared bat, Myotis septentrionalis, are relatively inaudible to moths. J Exp Biol 178:173–189
Felsenstein J (1985) Confidence limits on phylogenies: an approach
using the bootstrap. Evolution 39:783–791
Fenton MB, Fullard JH (1979) The influence of moth hearing on
bat echolocation strategies. J Comp Physiol 132:77–86
Fenton MB, Gaudet CL, Leonard ML (1983) Feeding behaviour of
the bats Nycteris grandis and Nycteris thebaica (Nycteridae) in
captivity. J Zool 200:347–354
Fenton MB, Portfors CV, Rautenbach IL, Waterman JM (1998)
Compromises: sound frequencies used in echolocation by
aerial-feeding bats. Can J Zool 70:1174–1182
Freeman PW (1979) Specialised insectivorous beetle-eating and
moth-eating molossid bats. J Mammal 60:467–479
Freeman PW (1981) Correspondence of foraging habits and
morphology in insectivorous bats. J Mammal 62:166–173
Fullard JH (1982) Echolocating assemblages and the effects on
moth auditory systems. Can J Zool 60:2572–2576
Fullard JH (1987) Sensory ecology and neuroethology of bats and
moths: interactions on a global perspective. In: Fenton MB,
Racey P, Rayner JMV (eds) Recent advances in the study of
bats. Cambridge University Press, Cambridge, pp 244–272
Fullard JH (1990) The sensory ecology of moths and bats: global
lessons in staying alive. In: Evans DL, Schmidt JO (eds) Insect
defences: adaptive mechanisms and strategies of prey and
predators. State University of New York Press, Albany, N.Y.
pp 203–226
Fullard JH (1994) Auditory changes in noctuid moths endemic to a
bat-free environment. J Evol Biol 7:435–445
Fullard JH, Dawson JD (1997) The echolocating calls of the spotted
bat, Euderma maculatum, are relatively inaudible to moths.
J Exp Biol 200:129–137
Fullard JH, Thomas DW (1981) Detection of certain African,
insectivorous bats by sympatric tympanate moths. J Comp
Physiol 143:363–368
Fullard JH, Yack JE (1993) The evolutionary biology of insect
hearing. Trends Ecol Evol 8:248–252
Fullard JH, Dawson JF, Otero LD, Surlykke A (1997) Bat-deafness
in day-flying moths (Lepidoptera, Notodontidae, Dioptinae).
J Comp Physiol 181:477–483
Garland TP, Harvey PH, Ives AR (1992) Procedures for the
analysis of comparative data using phylogenetically independent contrasts. Syst Biol 41: 8–32
Grafen A (1989) The phylogenetic regression. Philos Trans R Soc
Lond B 326:119–157
Helversen O von (1989) Bestimmungsschlssel fr die europischen Fledermuse nach ußeren Merkmalen. Myotis 27:41–60
Jacobs DS (1996) Morphological divergence in an insular bat,
Lasiurus cinereus semotus. Funct Ecol 10:622–630
Jacobs DS (2000). Community level support for the allotonic
frequency hypothesis. Acta Chiropter 2:197–207
Jones G (1992) Bats vs. moths: studies on the diets of rhinolophid
and hipposiderid bats support the allotonic frequency hypothesis. In: Horček I, Vohralik V (eds) Prague studies in
mammology. Charles University Press, Prague, Czech Republic, pp 87–92
Jones G (1996) Does echolocation constrain the evolution of body
size in bats? Symp Zool Soc Lond 69:111–128
Jones G, Waters DA (2000) Moth hearing in response to bat
echolocation calls manipulated independently in time and
frequency. Proc R Soc Lond B 267: 1627–1632
Lawrence BD, Simmons JA (1982) Measurements of atmospheric
attenuation at ultrasonic frequencies and the significance for
echolocation by bats. J Acoust Soc Am 71:585–590
Martin EP, Garland T (1991) Phylogenetic analyses of the
correlated evolution of continuous characters: a simulation
study. Evolution 45:534–557
Miller LA (1983) How insects detect and avoid bats. In: Huber F,
Markl M (eds) Neuroethology and behavioural physiology.
Springer, Berlin Heidelberg New York, pp 251–266
Norberg UM, Rayner JMV (1987) Ecological morphology and
flight in bats (Mammalia; Chiroptera): wing adaptations, flight
performance, foraging strategy and echolocation. Philos Trans
R Soc Lond B 316:335–427
Novick A (1977) Acoustic orientation. In: Wimsatt WA (ed)
Biology of bats, vol III. Academic Press, New York, pp 73–287
O’Farrel MJ, Miller BW, Gannon WL (1999) Qualitative identification of free-flying bats using the Anabat detector. J Mammal
80:11–23
Oliver I, Beattie AJ (1993) A possible method for the rapid
assessment of biodiversity. Conserv Biol 7:562–568
162
Oliver I, Beattie AJ (1996) Invertebrate morphospecies as surrogates for species: a case study. Conserv Biol 10:99–109
Pavey CR, Burwell CJ (1998) Bat predation on eared moths: a test
of the allotonic frequency hypothesis. Oikos 81:143–151
Purvis A, Rambaut A (1995) Comparative analysis by independent
contrasts (CAIC): an Apple Macintosh application for analysing comparative data. Comput Appl Biosci 11:247–251
Roeder KD (1967) Nerve cells and insect behaviour. Harvard
University Press, Cambridge, Mass.
Roeder KD (1975) Neural factors and evitability in insect behavior.
J Exp Zool 194:75–88
Rydell J (1992) Exploitation of insects around streetlamps in
Sweden. Funct Ecol 9:744–750
Saunders MB, Barclay RMR (1992) Ecomorphology of insectivorous bats: a test of predictions using two morphologically
similar species. Ecology 73:1335–1345
Schnitzler H-U, Kalko EKV (2001) Echolocation by insect eating
bats. BioScience 51:557–569
Scholtz CH, Holm E (1985) Insects of Southern Africa. University
of Pretoria, Pretoria, South Africa
Scoble MJ (1992) The Lepidoptera. Form, function and diversity.
Oxford University Press, Oxford
Simmons NB, Geisler JH (1998) Phylogenetic relationships of
Icaronycteris, Archaeonycteris, Hassianycteris, and Paleochiropteryx to extant bat lineages, with comments on the evolution
of echolocation and foraging strategies in Microchiroptera. Bull
Am Mus Nat Hist 235:1–182
Surlykke A (1988) Interaction between echolocating bats and their
prey. In: Nachtigall PE, Moore PWB (eds) Animal sonar:
processes and performance. Plenum Press, New York, pp 551–
566
Surlykke A, Skals N, Rydell J & Svensson M. (1998) Sonic hearing
in a diurnal geometrid moth, Archiearis parthenias, temporally
isolated from bats. Naturwissenschaften 85:36–37
Taylor PJ (1999) Echolocation calls of twenty South African bat
species. S Afr J Zool 33:114–124
Whitaker JO (1988) Food habits analysis of insectivorous bats. In:
Kunz TH (ed) Food habits analysis of insectivorous bats.
Smithsonian Institution Press, Washington, D.C. pp 171–189
Whitaker JO, Black H (1976) Food habits of cave bats in Zambia,
Africa. J Mammal 57:199–204
Whitaker JO, Neefus C, Kunz TH (1996) Dietary variation in the
Mexican free-tailed bat (Tadarida brasiliensis mexicana).
J Mammal 77:716–724
Whitaker JO, Suthakar S, Marimuthu G, Kunz TH (1999) Seasonal
variation in the diet of the Indian pygmy bat, Pipistrellus
mimus, in southern India. J Mammal 80:60–70
Zar JH (1999) Biostatistical analysis 4th edn. Prentice-Hall,
Englewood Cliffs, N.J.