Bol. Mus. Mun. Funchal, 55 (312): 5-15, 2004 ISSN 0870-3876
DIFFERENCES IN SEARCH BEHAVIOUR IN LARVAE OF THE
TWO MADEIRAN SPECKLED WOOD BUTTERFLIES,
PARARGE AEGERIA AND PARARGE XIPHIA
IMPLICATIONS FOR INTERSPECIFIC COMPETITION?
By M. GIBBS
, L. A. LACE
, M. J. JONES
& A. J. MOORE
With 1 figure and 1 table
ABSTRACT. Two species of speckled wood butterfly on Madeira, Pararge
aegeria (L., 1758), a recent colonist, and the endemic P. xiphia (F., 1775), may
compete for resources during their larval stages. This competition may induce P.
aegeria and P. xiphia larvae to disperse to new host-plants. Laboratory
experiments were performed to determine how successful the larvae of both
species were at locating new hosts. Pararge xiphia larvae spent significantly
longer searching for plants, and were also significantly less likely to locate plants
than P. aegeria larvae. This may indicate that P. aegeria larvae have a
competitive advantage when locating new host-plants.
A detailed analysis of P. aegeria larval search behaviour suggested that they
are adapted for local searching. Age was not found to be a significant factor in
search time, but starvation significantly decreased the time taken to find host-
plants. These results suggest that P. aegeria are adapted for dispersal and can
forage efficiently in areas where their host-plant (Brachypodium sylvaticum) is
Department of Biological Sciences, Manchester Metropolitan University, Chester Street, Manchester
M1 5GD, U. K.
The University of Manchester, School of Biological Sciences, 3614 Stopford Building, Oxford
Road, Manchester M13 9PT, U. K. E-mail: firstname.lastname@example.org
6 Boletim do Museu Municipal do Funchal (História Natural) No. LV, Art. 312
RESUMO. Durante a fase larvar, duas espécies de borboletas da Madeira,
Pararge aegeria, uma recente colonização, e P. xiphia, endémica, podem
competir pelo alimento. Esta competição pode levar a que ambas as espécies
procurem novas plantas-hospedeiro. Foram realizadas experiências em laboratório
com objectivo de determinar quão bem sucedidas são as larvas destas espécies
na localização de novas plantas-hospediro. As larvas de P. xiphia demoraram
significativamente mais tempo e foram significativamente menos bem sucedidas
na procura de plantas do que as larvas de P. aegeria. Estes resultados parecem
indicar que as larvas de P. aegeria possuem uma vantagem competitiva na
localização de novas plantas-hospedeiro. Uma análise detalhada do
comportamento de busca das larvas de P. aegeria parece indicar que elas estão
adaptadas para a busca local. A idade parece não constituir um factor significativo
no tempo de busca, ao contrário do jejum, o qual diminuiu significativamente o
tempo para encontrar as plantas-hospedeiro. Estes resultados parecem sugerir
que as larvas de P. aegeria estão adaptadas para a dispersão e podem alimentar-
se eficientemente em áreas onde a sua planta-hospedeiro (Brachypodium
sylvaticum) é localmente densa.
Madeira is the only place in the world where two species of speckled wood
butterfly (Lepidoptera: Satyrinae; Pararge) are sympatric (O
WEN et al., 1986). One
of these, Pararge xiphia (F., 1775), is endemic to Madeira, the other P. aegeria (L.,
1758), is a European mainland species and was first recorded on the island in 1976
IGGINS, 1977). Both species are now widespread and common on Madeira (SHREEVE
& SMITH, 1992; SHREEVE et al., 1992; GOTTHARD et al., 1999), and have overlapping
ALMON & WAKEHAM-DAWSON, 1999; SWASH & ASKEW, 1981). Both
species use the same host-plants and have overlapping life histories (F
UBIO & GARCÍA-BARROS, 1995; SALMON & WAKEHAM-DAWSON, 1999; JONES
et al., 1998; OWEN et al., 1986), and it is possible that Pararge aegeria and P. xiphia
compete for access to resources.
Competitive interactions can play an important part in the dynamics of insect
herbivore communities (F
ISHER et al., 2000), and recent sympatry of the two Pararge
species offers an almost unique opportunity to critically examine the importance of
interspecific competition. The larvae of both species are known to feed on the same
food plants in Madeira (F
ERNÁNDEZ-RUBIO & GARCÍA-BARROS, 1995), and
2004 Gibbs et al., Larval search behaviour in P. aegeria and P. xiphia 7
females of both species have been observed using the same oviposition sites in some
ALMON & WAKEHAM-DAWSON, 1999; JONES et al., 1998; OWEN et al.,
ONES et al. (1998) observed that although the favoured food plant
(Brachypodium sylvaticum, Beauv., 1762) was very widespread and abundant, the
eggs laid by the two species were not uniformly distributed; P.aegeria eggs greatly
out-numbered those of P.xiphia and eggs of both species were seen on the same
blade of grass. Upon hatching, P.xiphia larvae are almost twice the size of P.aegeria but
the development time to pupation occurs at a much slower rate for P.xiphia (G
Dispersal during the larval stage, when larvae are crowded, may help avoid
periods of food shortages (E
RELLI & ELKINTON, 2000). It is possible therefore,
that competition during larval development may stimulate P. aegeria and P. xiphia
larvae to disperse to new host-plants. The timing of this dispersal may prove critical
in ensuring their survival. C
AIN et al. (1985) found that early instar larvae of Pieris
rapae (L., 1758) were less mobile than later instar larvae, and in the three butterfly
species that were examined by J
ONES (1977), larger larvae travelled faster than small
larvae. The success rate of plant location was also greater for larger larvae, particularly
if the larvae were starved before searching commenced (J
ONES, 1977). Therefore,
during dispersal to a new host-plant, early instar larvae may be at a disadvantage and
have increased vulnerability to ground-dwelling predators (D
ICHOLLS & JAMES, 1996).
The dispersal behaviour of P. xiphia and P. aegeria will be examined
experimentally, and the results obtained will be used to predict the outcome of any
larval competition that may occur between these two species.
To monitor how successful each species was at locating food, we adapted a
method described by C
AIN et al. (1985) carried out under laboratory conditions. On
the laboratory floor 25 potted plants of Brachypodium sylvaticum were uniformly
arranged in five rows to create 16 cells of 0.5 x 0.5 m (Fig. 1). Temperature has been
shown to affect the searching behaviour of larvae (J
ONES, 1977). Therefore a
temperature of 21 +/- 3° C and a humidity of 55 +/- 10% was maintained throughout
the experiment. The light intensity was maintained at 700 Lux, and the laboratory floor
provided a uniformly smooth, even surface for the searching larvae. One larva was placed
at the centre of each cell so that it was equidistant from each plant. A 1 hour time limit
was imposed on searching and larvae that failed to find a plant, or left the area were
denoted as failures (as in C
AIN et al., 1985).
8 Boletim do Museu Municipal do Funchal (História Natural) No. LV, Art. 312
Comparing the search success of P. aegeria and P. xiphia larvae
A total of 95 P. aegeria larvae, from instars II-IV were observed, and the time
taken for each larva to find a plant was recorded. These data were used to determine the
effect of age (and hence size) on P. aegeria search success. Larvae from instars II and
III were classed as young (and small, with weights between 0.03 g and 0.07 g) larvae,
and those from instars IV and V were classed as old (large, with weights above 0.07 g)
larvae. Additionally, these data were also used to estimate the appropriate numbers of
sampling units (i. e. number of larvae) required to detect an observable difference in
larval search time at a 5% confidence level (for details of this analysis see J
al., 1998). This analysis was performed to reduce the number of animals used in the
following experimental procedures, and to avoid stressing animals unnecessarily. An
estimate of 20 larvae was determined as appropriate.
To compare the search success of P. aegeria and P. xiphia larvae the experiment
described previously was repeated for 21 P. xiphia larvae. The P. xiphia larvae were
weighed and those with weights between 0.04 g and 0.1 g were determined as being
from instars II and III. The search success of these larvae was compared with that of P.
aegeria larvae also from instars II and III (e. g. with weights between 0.03 g and 0.07 g).
Pararge xiphia larvae proved to be difficult to rear under laboratory conditions
and only 21 larvae were available at the time of these experiments. Thus, further
experiments with P. xiphia were not possible and the following procedures were
performed with only P. aegeria larvae.
Detailed analysis of the search behaviour of P. aegeria larvae
Using a method similar to that outlined above, a detailed examination of the
searching behaviour of an additional 25 unstarved P. aegeria larvae was recorded. An
experimental plot was set-up as described previously but no plants were placed into the
area. The larvae were positioned into the sub-plots as described above. The positions
of the larvae were recorded at 5 minute intervals using numbered discs. The markers
were placed at the end of the larva’s anal claspers, and the larvae were followed around
the area for 1 hour. The markers were used to map the movements of the larvae and to
record the distance and direction travelled during each 5 minute time interval.
Effect of starvation on P. aegeria larval searching success
An experimental plot was set-up as described previously (see Experimental
design). Prior to their release into the search arena, 20 P. aegeria larvae were starved
overnight for a period of approximately 12 hours. The larvae were then released into
the experimental plot and allowed to search for a host-plant. The time taken for each
2004 Gibbs et al., Larval search behaviour in P. aegeria and P. xiphia 9
larva to find a plant was recorded. Additionally, whether the larva started feeding
immediately on reaching a plant was also recorded.
Two-way analysis of variance was used to investigate the effects of starvation,
age and species on larval search time. Data on search time were log transformed to
meet assumptions of normality. A Pearson chi-square test was used to compare the
proportion of larvae that found a host-plant, and to determine whether there were species
differences, or any effect of starvation, on host-plant location. All tests were two-
tailed. The statistical procedures were performed using Systat 9.0.
Comparing the search success of P. aegeria and P. xiphia larvae
Pararge xiphia larvae (
= 30.1 ± 1.02 min, n = 8) spent significantly longer
time searching for plants than P. aegeria larvae ( = 17.0 ± 1.15 min, n = 16; F =
5.368, d. f. = 1, P = 0.030). There were also significant species differences in plant
finding ability, with a significantly larger proportion of P. xiphia larvae (9/20) failing
to locate a plant compared to P. aegeria (16/99; =
8.090, d. f. = 1, P = 0.004).
Fig. 1 - Mean distance travelled (cm) per five minute time interval by P. aegeria larvae.
10 Boletim do Museu Municipal do Funchal (História Natural) No. LV, Art. 312
Detailed analysis of the search behaviour of P. aegeria larvae
The larvae turned frequently, with individuals having a turn bias (i. e. with a
larger proportion of their total turns) towards either the left or the right, resulting in
larvae searching in a circular motion. Overall, larvae did not show a turn bias towards
any one direction, as 42% of larvae had a turn bias towards the right and 54% towards
the left (n = 24).
As the recording period progressed larvae tended to turn less and their
movements became straighter and faster. The mean distance travelled per 5 minute time
increment increased in a linear fashion over time, with the largest increase in distance
travelled occurring after 35 minutes of searching. Thus, as larvae travelled, their rate
of movement increased. Only one larva failed to move for the entirety of the recording
period. Scanning for a resource was observed, but was not quantified, scanning took the
form of head waving (as described by J
ONES, 1977), and only occurred when the larvae
were stationary (saltatory searching, B
TABLE 1 - The effect of age and starvation on P. aegeria larval search time
Category Mean time (± SE)/ mins n
Old and Starved 9.87 (1.26) 6
Old and unstarved 18.14 (1.12) 26
Young and starved 9.88 (1.23) 8
Young and unstarved 27.19 (1.21) 9
Effect of larval age and starvation on P. aegeria search success
Starved larvae had a mean search time of 9.88 ± 1.15 min (n = 14) and unstarved
larvae had a mean search time of 19.11 ± 1.06 min (n = 81), and this difference was
highly significant (ANOVA, F = 18.055, d. f. = 1, P = < 0.001). Food deprivation, however
did not increase the likelihood of a larvae finding a plant, 14/17 of the starved larvae
and 83/99 of the unstarved larvae were successful at finding a new plant (χ
= 0.001, d.
f. = 1, P = 0.976). Of the larvae that began to feed immediately on reaching a host-
plant, 3/43 were from the unstarved group and 2/13 were from the starved group.
Age was not a significant factor in search time (Table 1). ANOVA of
search time indicated that older larvae did not reach plants quicker than young
2004 Gibbs et al., Larval search behaviour in P. aegeria and P. xiphia 11
larvae (F = 1.13, d. f. = 1, P = 0.293). There was not a significant age x starvation
interaction (F = 1.12, d. f. = 1, P = 0.296).
ELL et al. (1985) suggested that food deprivation increases an animal’s
responsiveness to resource-related cues. Our results support this hypothesis in part.
Pararge aegeria larvae that were deprived of food found host-plants significantly faster
than non-starved larvae, but equivalent numbers of larvae from each group were
successful at locating food. Therefore, starvation would appear to increase the
responsiveness of P. aegeria larvae without improving their chance of host location.
Starved larvae were twice as likely to begin feeding immediately on reaching a host-
plant than non-starved larvae. Overall, however, small percentages of larvae actually
began to feed on reaching a plant. This may be due to the need for a rest period to allow
recovery after searching, or may indicate that a larva’s need for camouflage is a large
motivational force behind host-plant location.
As a result of competitive interactions it is possible that larvae may experience
food shortage/deprivation prior to their dispersal. Pararge aegeria larvae, however,
will not be disadvantaged if such a situation were to arise because starvation appears to
increase their motivation for host location. Age-related differences in searching success
were not observed for P. aegeria larvae, suggesting that old and young larvae are equally
successful at locating new plants, this may be expected, however, as there is the potential
for food to become depleted at any time during larval development. These results would
suggest that P. aegeria larvae are adapted for dispersal. The high success rate of P.
aegeria larval dispersal may partly be explained by adult ecology. Pararge aegeria
females oviposit on small Brachypodium sylvaticum plants in areas where the host-
plant is locally dense (G
IBBS, M., unpublished data). Given that P. aegeria larvae
search intensively, in a manner characteristic of that described by B
ELL (1990) and
TANTON (1986) for animals that restrict their foraging to locally dense areas,
oviposition on small plants in dense areas will increase the amount of food available to
larvae by facilitating dispersal. L
AWRENCE (1990) suggested that local movement of
larvae reduced predation and resulted in higher survival rates because moving groups had
fewer predators. It is possible, therefore, that the oviposition behaviour of females has
evolved to ‘encourage’ local larval dispersal to increase the survival of P. aegeria larvae.
Larger larvae have a longer stride than small larvae and are therefore able to
move faster (R
EAVEY, 1993). Pararge xiphia larvae have a larger body size (at all
instars) than P. aegeria, it could be expected, therefore, that they would move quicker.
However, P. xiphia larvae both took significantly longer to locate a new host-plant, and
were less successful at finding plants than P. aegeria larvae. It is possible that P. xiphia
larvae have different periods of inactivity than P. aegeria larvae, and they may not forage
12 Boletim do Museu Municipal do Funchal (História Natural) No. LV, Art. 312
during these times even if they perceive resource-specific cues. They may also have
different search tactics. Some large caterpillars do not move if it increases their
vulnerability to predators (R
EAVEY, 1993). Differences in the visual and olfactory
sensitivity of P. xiphia and P. aegeria larvae could account for the differences in their
searching success. Pararge xiphia larvae may be more dependent on environmental
cues (e. g. orientation of the sun) than P. aegeria larvae, and the artificial conditions of
the laboratory may have hindered their searching.
Pararge xiphia larvae are larger than P. aegeria larvae (F
& GARCÍA-BARROS, 1995), and there may be less urgency for these larvae to find a
new plant since their large size decreases their physiological vulnerability (e. g. they
are less vulnerable to starvation because they have larger energy reserves and lower
metabolic rates (A
LJETLAWI & LEONARDSSON, 2003; REAVEY, 1993). However, the
risks from predation will still exist for P. xiphia larvae. Longer search times increases
the vulnerability of larvae to predation (D
ETHIER, 1959; NICHOLLS & JAMES, 1996),
indicating that P. xiphia larvae may be more vulnerable than P. aegeria larvae.
OKOLOWSKI (1985) found that for Drosophila melanogaster (MEIGEN, 1830)
intraspecific variation in larval foraging behaviour was under genetic control. Genetic
differences may, therefore, also be responsible for the intra- and interspecific variation
in foraging behaviour observed for P. aegeria and P. xiphia larvae.
Adult ecology may also help to explain why P. xiphia are less successful at
locating new food plants. Female P. xiphia lay fewer eggs (than P. aegeria) and
distribute them more widely over host-plants (M. J
ONES, pers. com.), suggesting that
larval competition may have occurred less frequently for this species (before the
colonisation of P.aegeria) and therefore, the larvae may not be adapted for dispersal.
This suggests that larval displacement due to interspecific competition for food may
be more detrimental to P. xiphia larvae than P. aegeria larvae.
Differences in the larval searching success of P. aegeria and P. xiphia larvae
may indicate that the recent colonists (P. aegeria) are at a competitive advantage during
host-location. Such a situation would be predicted by the Taxon cycle hypothesis
ILSON, 1959, 1961; RICKLEFS & COX, 1972, 1978), which suggests that recent
colonists are at a competitive advantage (J
ONES et al., 1998). Therefore, further work
would be valuable to assess the impact of interspecific competition on the P. aegeria
and P. xiphia ecology.
The Department of Biological Sciences at Manchester Metropolitan University
provided the funding for this study. We acknowledge Paulo and Luisa Oliveira and their
families for their help and support during our collecting trip on Madeira. We would
also like to acknowledge Richard Preziosi for his helpful comments on an earlier
version of this manuscript.
2004 Gibbs et al., Larval search behaviour in P. aegeria and P. xiphia 13
ALJETLAWI, A. A. & K. LEONARDSSON:
2003. Survival during adverse seasons reveals size-dependent competitive ability in a
deposit-feeding amphipod. Monoporeia affinis. Oikos, 101: 164-170.
BELL, W. J.:
1990. Searching behaviour patterns in insects. Annual Review of Entomology, 35:
BELL, W. J., C. TORTORICI, R. J. ROGGERO, L. R. KIPP & T. R. TOBIN:
1985. Sucrose-stimulated searching behaviour in Drosophilia melanogaster in a uniform
habitat: modulation by period of deprivation. Animal Behaviour, 33: 436-48.
CAIN, M. L., J. ECCLESTON & P. M. KAREIVA:
1985. The influence of food-plant dispersion on caterpillar searching success. Ecological
Entomology, 10: 1-7.
DETHIER, V. G.:
1959. Food-plant distribution and density and larval dispersal as factors affecting insect
populations. Canadian Entomologist, 88: 581-596.
ERELLI, M. C. & J. S. ELKINTON:
2000. Factors influencing dispersal in neonate gypsy moths (Lepidoptera: Lymantriidae).
Population Ecology, 29: 509-513.
FERNÁNDEZ-RUBIO, F. & E. GARCÍA-BARROS:
1995. Taxonomic situation of the species of the genus Pararge Hübner, 1819 – P.
aegeria, P. xiphia and P. xiphioides – (Lepidoptera: Satyridae) in the
Macaronesian Islands. Boletim do Museu Municipal do Funchal, 47 (260):
FISHER, A. E. I., S. E. HARTLEY & M. YOUNG:
2000. Direct and indirect competitive effects of foliage feeding guilds on the performance
of the birch leaf-miner Eriocrania. Journal of Animal Ecology, 69: 165-176.
GOTTHARD, K., S. NYLIN & C. WIKLUND:
1999. Mating system evolution in response to search costs in the speckled wood butterfly,
Pararge aegeria. Behavioural Ecology and Sociobiology, 45: 424-429.
14 Boletim do Museu Municipal do Funchal (História Natural) No. LV, Art. 312
HIGGINS, L. G.:
1977. The Speckled Wood (Pararge aegeria L.) in Madeira. Entomologist’s Record
and Journal of Variation, 89: 22-23.
JONES, A., R. REED & J. WEYERS:
1998a. Practical skills in biology. Addison Wesley Longman Limited.
JONES, M. L., L. A. LACE, E. C. HARRISON & B. STEVENS-WOOD:
1998b. Territorial behaviour in the speckled wood butterflies Pararge xiphia and P.
aegeria (Satyridae) of Madeira: a mechanism for interspecific competition.
Ecography, 21: 297-305.
JONES, R. E.:
1977. Search behaviour: A study of three caterpillar species. Behaviour, 60: 237-259.
LAWRENCE, W. S.:
1990. The effects of group size and host species on development and survivorship of a
gregarious caterpillar Halisidota caryae (Lepidoptera: Arctiidae). Ecological
Entomology, 15: 53-62.
NICHOLLS, C. N. & R. JAMES:
1996. Host selection and juvenile survivorship in the swallowtail butterfly Papilio
machaon britannicus Seitz (Lepidoptera: Papilionidae). Entomologist’s Gazette,
OWEN, D. F., T. G. SHREEVE & A. G. SMITH:
1986. Colonisation of Madeira by the speckled wood butterfly, Pararge aegeria and its
impact on the endemic Pararge xiphia. Ecological Entomology, 11: 349-352.
1993. Why body size matters to caterpillars. In: Ecological and evolutionary
constraints on foraging, (eds.: N. E. Stamp & T. M. Casey), pp. 248-279.
Chapman & Hall, New York.
RICKLEFS, R. E. & G. W. COX:
1972. Taxon cycles in the West Indian avifauna. American Naturalist, 106: 195-219.
1978. Stage of the taxon cycle, habitat distribution and population density in the avifauna
of the West Indies. American Naturalist, 112: 875-895.
2004 Gibbs et al., Larval search behaviour in P. aegeria and P. xiphia 15
SALMON, M. A. & A. WAKEHAM-DAWSON:
1999. Thomas Vernon Wollaston and the Madeiran butterfly fauna – A re-appraisal.
British Journal of Entomology and Natural History, 12: 69-88.
SHREEVE, T. G. & A. G. SMITH:
1992. The role of weather related habitat use on the impact of the European speckled
wood Pararge aegeria on the endemic Pararge xiphia on the island of Madeira.
Biological Journal of the Linnean Society, 46: 59-75.
SHREEVE, T. G., A. G. SMITH & M. BÁEZ:
1992. Egg-laying sites, distributions and host-plants of members of the genus Pararge
(Lep: Satyrinae). Entomologist’s Record and Journal of Variation, 104:
SOKOLOWSKI, M. M.:
1985. Genetics and ecology of Drosophila melanogaster larval foraging and pupation
behaviour. Journal of Insect Physiology, 31: 857-864.
STANTON, M. L.:
1986. Spatial patterns in the plant community and their effects upon insect search. In:
Herbivorous insects: Host seeking behaviour and mechanisms (ed.: S. Ahmad),
pp. 125-152. Academic Press, New York.
SWASH, A. R. & R. R. ASKEW:
1981. A survey of Madeiran butterflies. Boletim do Museu Municipal do Funchal,
WILSON, E. O.:
1959. Adaptive shift and dispersal in tropical ant fauna. Evolution, 13: 122-144.
1961. The nature of the taxon cycle in the Melanesian ant fauna. American Naturalist,
Date received: 17-05-2004.