dual mimicry in the dimorphic eusocial wasp …so356/pdfmmdual.pdf · dual mimicry in the dimorphic...

Biological Journal of the Linnean Society (1999), 66: 501–514. With 5 figures

Article ID: bijl.1998.0273, available online at http://www.idealibrary.com on

Dual mimicry in the dimorphic eusocial waspMischocyttarus mastigophorus Richards(Hymenoptera: Vespidae)

SEAN O’DONNELL∗

Department of Psychology 351525, University of Washington, Seattle, WA 98195, U.S.A.

FRANK J. JOYCE

Apartado 32–5655, Monteverde, Provincia Puntarenas, Costa Rica

Received 3 May 1998; accepted for publication 14 August 1998

The eusocial vespid wasp Mischocyttarus mastigophorus exhibits two colour morphs, with malesand females of each morph co-occurring at Monteverde, Costa Rica. Each morph closelyresembles a different sympatric species of swarm-founding wasp in the genus Agelaia. Wepropose that the Agelaia species are models for a dual mimicry system. The Agelaia species(A. yepocapa, mimicked by the M. mastigophorus pale morph, and A. xanthopus, mimicked by theM. mastigophorus dark morph) are locally abundant wasps with large, aggressively defendedcolonies. The mimic and models are restricted to high-elevation habitat in the Monteverdearea, and the elevational ranges of both Agelaia species partially overlap the elevational rangeof M. mastigophorus. Relative frequencies of the M. mastigophorus colour morphs vary withelevation, with the pale morph predominating at lower elevations. Elevational differences inthe relative abundances of the Agelaia species suggest that the models act as a selectiveforce maintaining the M. mastigophorus colour polymorphism at Monteverde. Mischocyttarusmastigophorus overlaps only A. xanthopus in the northern part of its range (S. Mexico), andoverlaps only A. yepocapa in the southern part of its range (Ecuador). We hypothesize thatthe M. mastigophorus colour morphs evolved in allopatry and later came into contact in CentralAmerica. Appropriate high-elevation habitat for cloud forest species is distributed as discretepatches in Central America and Northern South America. The island-like nature of suitablehabitat may favour the isolation and rapid evolutionary diversification of vespid species thatare restricted to high elevations in the Neotropics.

1999 The Linnean Society of London

ADDITIONAL KEY WORDS:—Agelaia – biogeography – Batesian mimicry – cloud forest– elevational gradient – Mullerian mimicry – necrophagy.

CONTENTS

Introduction . . . . . . . . . . . . . . . . . . . . . . . 502Material and methods . . . . . . . . . . . . . . . . . . . 503

∗Corresponding author. Email: [email protected]

5010024–4066/99/0040501+14 $30.00/0 1999 The Linnean Society of London

S. O’DONNELL AND F. J. JOYCE502

Study site . . . . . . . . . . . . . . . . . . . . . . 503Monteverde elevational distribution of M. mastigophorus colonies . . . . 503Meat baiting at Monteverde to sample Agelaia elevational distributions . . 503Monteverde elevational distribution of Agelaia colonies . . . . . . . 504Determination of geographical and elevational ranges at other sites . . . 504

Results . . . . . . . . . . . . . . . . . . . . . . . . 505Resemblance of M. mastigophorus to proposed models . . . . . . . . 505Monteverde elevational distribution of M. mastigophorus colour morphs . . 505Monteverde elevational distribution of Agelaia foragers at meat baits . . . 505Monteverde elevational distribution of Agelaia colonies . . . . . . . 507Geographic ranges of subject species . . . . . . . . . . . . . 508

Discussion . . . . . . . . . . . . . . . . . . . . . . . 509Evolution of mimicry in the genus Mischocyttarus . . . . . . . . . 509Evolutionary maintenance of M. mastigophorus colour polymorphism at

Monteverde . . . . . . . . . . . . . . . . . . . . 510Evolutionary origin of M. mastigophorus dual mimicry . . . . . . . . 512

Acknowledgements . . . . . . . . . . . . . . . . . . . . 512References . . . . . . . . . . . . . . . . . . . . . . . 512Appendix . . . . . . . . . . . . . . . . . . . . . . . . 514

INTRODUCTION

Tropical invertebrate mimicry complexes often involve eusocial wasps of thefamily Vespidae as models (Wickler, 1968). Potent stings used in defence againstvertebrates and possibly invertebrates combine with aposematic colour patterns invespids, making them effective model species for Batesian mimicry by more palatableinsects (O’Donnell & Jeanne, 1990; Starr, 1990; O’Donnell, 1996). Mimicry com-plexes among eusocial vespid wasps have also been described (Richards, 1978;West-Eberhard & Carpenter, 1995). Swarm-founding Neotropical Vespidae (tribeEpiponini) often have large, aggressively defended colonies (hundreds to ten thousandsof adults), and are often mimicked by other eusocial wasps. In contrast, the largelyNeotropical genus Mischocyttarus is characterized by species with independently-founded small colonies (several dozen adults) of relatively non-aggressive wasps.Many Mischocyttarus species apparently mimic other vespid wasps (Richards, 1978).At least some Mischocyttarus species wasps are capable of stinging humans. However,Mischocyttarus wasps are often reluctant to sting humans even in nest defense, andindividuals of most species we have observed remain immobile or even flee theirnest when disturbed (S.O’D. and F. J. J., pers. obs.). Because they possess stings butare relatively docile, Mischocyttarus species may mimic other Vespidae through acombination of Batesian and Mullerian processes.

Mischocyttarus mastigophorus Richards exhibits two colour morphs which co-occurin the population at Monteverde, Costa Rica (O’Donnell & Joyce in press). Herewe present evidence that M. mastigophorus exhibits dual mimicry, with two vespidwasp species from the swarm-founding genus Agelaia (A. xanthopus (de Saussure) andA. yepocapa Richards) serving as models for the colour morphs. To our knowledge,this is the first reported case of a eusocial wasp species with discrete colour morphscoexisting within a population (other than colour differences between female castes,reviewed by O’Donnell [1998]. See Wenzel [1992] for an example from independent-founding wasps). Other Mischocyttarus species are known to vary in body colour overtheir geographic ranges ( James Carpenter, pers. comm.).

The aim of this study was to determine whether the presence of two colour

DUAL MIMICRY IN THE DIMORPHIC EUSOCIAL WASP 503

morphs within the M. mastigophorus population could be maintained by differencesin the relative abundances of the putative model species across an elevationalgradient. We asked whether the frequencies of the M. mastigophorus colour morphscorresponded to local patterns of abundance of the Agelaia species models at differentelevations. If presence of the model species affected the fitness of the mimic morphs,we predicted that the relative abundances of the M. mastigophorus morphs wouldcorrespond to the abundances of the model species. We also determined the extentof overlap in elevational and geographic range of the three species throughoutCentral and South America by making additional collections and by consultingpublished reports of collection localities. We used information on geographic rangesto estimate the extent of sympatry among species in the mimicry complex. Weconclude by inferring possible evolutionary origins of the mimicry complex fromcurrent patterns of geographic overlap.

MATERIAL AND METHODS

Study site

Data on the elevational distributions of M. mastigophorus colour morphs and Agelaiaspecies models were collected in Monteverde, Prov. Puntarenas, Costa Rica (10°18′N,84°49′W). The Monteverde area includes a well-studied tropical cloud forest, andis characterized by varying topography and extensive forest cover at higher elevations(Hartshorn, 1983). Along with several other eusocial wasp species, the subject speciesare largely restricted to higher elevations in the tropical lower montane wet forestlife zone (Holdridge, 1967; O’Donnell, in press). Although seasonality of climate(e.g. amount of rainfall) and vegetation cover change dramatically with elevation atMonteverde, the study area is characterized by extensive cloud cover and precipitation(either rain or wind-driven mist) for much of the year (Clark, Lawton & Butler, inpress).

Monteverde elevational distribution of M. mastigophorus colonies

We recorded the elevations of all M. mastigophorus colonies located from 1993 to1997. Elevations of colonies were determined using a hand-held altimeter calibratedto a known-elevation location (in Monteverde, the entrance to the MonteverdeCloud Forest Reserve at 1540 m). The altimeter was precise to within 10 m elevationwhen measurements were repeated on different days. Colonies were located bysearching the eaves of all buildings to which we had access. In addition, ad libitumsearches for colonies were performed by scanning roadside vegetation. For eachcolony we noted whether all dark morph, all pale morph, or a mixture of morphswas present. Colonies were examined at night whenever possible to ensure that alladults would be present on the nest. All adults present were collected, or the adultwasps were counted at night, for a subset of the M. mastigophorus colonies.

Meat baiting at Monteverde to sample Agelaia elevational distributions

All Agelaia species studied to date exhibit necrophagy (collection of flesh fromvertebrate and large invertebrate carcasses; O’Donnell, 1995). Agelaia foragers visit


and collect flesh from meat baits placed on the ground in forest and along forestedges. To sample the elevational distributions of A. xanthopus and A. yepocapa foragers,we placed meat baits at varying elevations in the Monteverde area between 18 Julyand 6 August 1997. Baits consisted of a mixture of uncooked chicken flesh and beef,and weighed approximately 0.5 kg. Baits were refrigerated at approximately 4°Cbefore and between trails, and no bait was used for more than three days of baitingtrials (Agelaia foragers are more readily attracted to fresh meat than to spoiled meat;F. Joyce unpublished data; O’Donnell, 1995). For each baiting trial, baits wereplaced in pairs at approximately the same elevation but separated by at least 50 mdistance. The elevation of each trial location was determined as above. Baits withinpairs were monitored sequentially for 10 min intervals at each bait, for a total trialduration of 100 min of observation at each bait pair. During each observationinterval we recorded the total number of foragers of each Agelaia species that visitedeach bait. Because individual foragers may have returned to a bait several timesduring a trial, we also recorded the largest number of A. xanthopus and A. yepocapaforagers observed simultaneously on each bait. All trials were conducted between0830 h and 1450 h local time, and baiting was not performed during periods ofheavy rain. Baiting trials were conducted along a roughly linear East-West transectfrom the Monteverde Cloud Forest Reserve down the main Monteverde road towardthe town of Santa Elena.

Monteverde elevational distribution of Agelaia colonies

As is typical of all but one species in the genus Agelaia ( Jeanne, 1975; Richards,1978), A. xanthopus and A. yepocapa nest in cavities. We noted the location andestimated the elevation of all colonies of A. xanthopus and A. yepocapa that weencountered in the Monteverde area from 1988 to 1997. Colonies were locatedthrough our own searching efforts, surveys of bird nesting boxes, and when residentsof the area drew them to our attention. We confirmed the identity of the Agelaiaspecies in all cases. Elevations of Agelaia species nesting sites were determined asabove.

Determination of geographical and elevational ranges at other sites

Additional searches for colonies and foragers of the three subject species wereconducted in Costa Rica in the Central Volcanic Plateau (vicinity of Heredia,10°10′N, 84°25′W, 24 August 1997; San Antonio de Escazu, 9°55′N, 84°08′W,24–25 August 1997) and the Northern Talamanca Range (in the vicinity of SanGerardo de Dota, 9°32′N, 83°52′W, 26–27 August 1997). Elevations at which nestsand foragers were located were determined as above. We surveyed publishedcollection locations of the three subject species from the literature and from collectingreports by colleagues. For each collection locality we determined the latitude/longitude coordinates to the nearest second by examination of maps of the area(1:500 000 to 1:1 000 000 scale). Collections reported in the literature typically notedthe nearest town to the collection site. When only the nearest town was reported,we assigned the location of collection to the town, and estimated the elevation inmeters for that location from topographic maps (elevation isoline intervals 300 m to


500 m). It is possible that the actual elevations of collection differed from elevationof the nearest town in some cases.

RESULTS

Resemblance of M. mastigophorus to proposed models

Mischocyttarus mastigophorus morphs differ dramatically in colour hues and patterns(Fig. 1). No intergradation between the morphs was evident; all individuals withyellow on the sides of the thorax were assigned to the pale morph. Resemblance ofthe M. mastigophorus morphs with their putative models is detailed (Fig. 1). The palemorph and A. yepocapa are extensively yellow when viewed laterally, and yellow withblack markings on the dorsum. Agelaia yepocapa and pale morph M. mastigophorusshare similar patterns of black markings, particularly on the dorsum of the gaster(distal segments of the abdomen). The coloration of the dark morph is almostidentical to that of its putative model, A. xanthopus. The dark morph’s body is darkbrown, with yellow on the legs and yellow mandibles. Mischocyttarus mastigophorusmales occur in both morphs, and are similar to females in colour and in pattern.Males are recognized by the possession of elongate, filamentous antennae, and aresmaller in body size than females.

Monteverde elevational distribution of M. mastigophorus colour morphs

We collected morph composition data on 78 M. mastigophorus colonies. Wefound M. mastigophorus colonies from approximately 1450 m to 1600 m in elevation.Mischocyttarus mastigophorus nests sites were restricted to sheltered locations in oradjacent to cloud forest. Most nests that we found were constructed on the wallsand eaves of buildings, although nests were also found suspended under leaves andon exposed rootlets in roadcuts and landslides. Colonies comprising only dark morphadults were too rare to allow statistical tests of their abundance (n=3), so wecompared the relative abundance of colonies with dark morph adults (sum of alldark and mixed colonies) with the abundance of colonies with only pale morphadults. The proportion of colonies with dark morph adults increased with elevationat Monteverde (Fig. 2; Wilcoxon 2-sample test, Z=4.15, P<0.001). Colonies withdark morph adults comprised approximately 50% of the Monteverde populationabove 1530 m elevation, but were rare at and below 1500 m elevation. We countedadults present or collected adults from 63 of the 78 colonies. The proportion ofdark morph individuals in these colonies increased with elevation (Spearman rankcorrelation r=0.37, P<0.05).

Monteverde elevational distribution of Agelaia foragers at meat baits

Foragers of the Agelaia model species overlapped extensively in elevational rangeat Monteverde, and foragers of both species frequently visited the same meat baits.However, only A. xanthopus forgers were recorded at baits placed at elevations above1550 m in our transect (Fig. 3). Maximum numbers of foragers at baits differed


Figure 1. Photographs of the dorsum of (clockwise, from upper left): Mischocyttarus mastigophorus (darkmorph), Agelaia yepocapa, A. xanthopus and M. mastigophorus (pale morph).

significantly between the Agelaia species with elevation (Wilcoxon 2-sample test, Z=−5.27, P<0.001); maximum numbers of A. xanthopus foragers were greater at higherelevation baits, and A. xanthopus foragers were rare below 1450 m elevation. Similarly,total numbers of A. xanthopus foragers were greater at higher elevations (Wilcoxon2-sample test, Z=−7.55, P<0.001). We did not observe foragers of either Agelaia


1.00

0.00

Above1550

Elevation (nearest 10 m)

Pr

(Col

onie

s)

Below1500

1530 1500

0.75

0.50

0.25

1540

11(0)

9(9)

4(2)

35(12)

19(7)

Figure 2. Elevational distributions of Mischocyttarus mastigophorus colonies with (Ε) and without (Φ)dark morph adults, surveyed at Monteverde. Number of colonies in the sample for each elevationinterval is given at the top of the bar (number of pre-emergence colonies in parenthesis).

4

0

1600(8)

Elevation to nearest 50 m;(number of baits placed at elevation)

Mea

n (

max

. nu

mbe

r of

for

ager

s)

1250(6)

1450(0)

1350(2)

2

1550(10)

3

1

1500(2)

1400(4)

1300(0)

Figure 3. Mean maximum numbers of Agelaia xanthopus (Ε) and A. yepocapa (Φ) foragers arriving atmeat baits placed at different elevations along a linear transect at Monteverde. Numbers of baits foreach elevation interval are given in parenthesis beneath the X-axis.

model species at baits placed below 1230 m (n=4 baits), although A. panamensisforagers visited these baits.

Monteverde elevational distribution of Agelaia colonies

Most A. xanthopus colonies were located at or below ground level in fallen treetrunks, while all A. yepocapa nests were built above the ground in cavities in standing(often living) tree trunks or in bird nesting boxes. Nesting elevations of the Agelaiaspecies overlapped with each other and with M. mastigophorus. However, A. xanthopuscolonies’ nests were found at significantly higher elevations than A. yepocapa colonies


7

0

Above1550

Elevation to nearest 50 m

Nu

mbe

r of

col

onie

s

12501450 1350

5

4

2

1550

6

3

1

1500 1400 1300

Figure 4. Elevational distributions of Agelaia xanthopus (Ε) and A. yepocapa (Φ) colonies located atMonteverde.

(Fig. 4; Wilcoxon 2-sample test, Z=3.27, P<0.01). Most A. xanthopus colonies (7/8)were found at or above 1550 m elevation, while most A yepocapa colonies (20/24)were found at or below 1450 m elevation. We did not observe nests of either Agelaiamodel species below approximately 1250 m elevation.

Geographic ranges of subject species

Most collections of all three subject species have been made at or above 1300 melevation throughout their geographic ranges (Appendix). Because collection localitiesin the literature were often assigned to a nearby town but did not provide preciseelevational data, collections reported from low elevation locations (e.g. 450 m inEcuador for A. yepocapa; 700 m for A. xanthopus in Costa Rica) may have been madeat higher elevations. Mischocyttarus mastigophorus has been recorded from SouthernMexico south to Southern Ecuador, and the entire range of M. mastigophorus likelyoverlaps one or both of the model Agelaia species (Fig. 5; A. yepocapa has not beencollected from as far south in Ecuador as M. mastigophorus). The southern limit of A.xanthopus’ range is in Costa Rica, while the northern limit of A. yepocapa occurs inSouthern Guatemala. Therefore, the range of M. mastigophorus coincides only withA. xanthopus in Southern Mexico, and only with A. yepocapa in South America.Collection data do not permit assessment of whether M. mastigophorus and its modelsare syntopic outside of Monteverde, because they have not always been collectedat identical localities. However, extensive geographic and elevational overlap suggestthat these species are sympatric over much or all of their ranges.

Elevational distributions of the subject species differed among collection locations.For example, we collected A. yepocapa foragers at over 2000 m elevation in theNorthern Talamanca Range of Costa Rica, well above the apparent upper elevationallimit of this species at Monteverde (approximately 1550 m for foragers and 1570 mfor nests; highest elevations of peaks surveyed in Monteverde were approximately1800 m). However, A. xanthopus foragers were collected at higher maximum elevations


MmAx

Ax

AxAy

Mm, Ax, Ay

Mm

Ax

Ay

Ay

Mm

North

Figure 5. Map of Central and Northern South America showing locations of collections of Mischocyttarusmastigophorus (Mm), Agelaia xanthopus (Ax) and A. yepocapa (Ay). Arrows point to approximate locationsof collections. Elevations are indicated by shading as follows: light grey – Sea level; white – 0–1299 m;dark grey – 1300 m and above.

than A. yepocapa foragers in the Talamanca range, suggesting that the elevationalsegregation of these species persists when their absolute elevation ranges shift.

DISCUSSION

Evolution of mimicry in the genus Mischocyttarus

Mischocyttarus is the largest genus of eusocial wasps in the Western Hemisphere,with approximately 200 described species (Richards, 1978). The facts that Mis-chocyttarus species are often involved in mimicry complexes, and that many (perhapsmost) Mischocyttarus species are mimics of other eusocial Vespidae, suggest that thegenetic and/or physiological mechanisms of colour determination are evolutionarilylabile in this genus. We have not determined the developmental mechanism(s) ofdifferentiation among the M. mastigophorus colour morphs. An individual’s colour


morph does not correspond consistently with body size (pers. obs.), and we haveobserved nestmates of both morphs eclosing from their pupal cases within one hourof each other. Dramatic differences in environmental experience for simultaneouslydeveloping larvae are unlikely. Therefore, we suspect that the polymorphism has agenetic component. However, environmental factors such as differential feeding(O’Donnell, 1998), or genotype X environment interactions, may play roles inmorph determination.

Many of the recognized mimicry complexes among eusocial wasps (Richards,1978; West-Eberhard & Carpenter, 1995) include ecologically dominant species asmodels, with dominance indicated by relative abundance of individuals and colonies( Jeanne, 1991). According to Jeanne’s (1991) review of collections of eusocialVespidae from eight Neotropical localities, Agelaia ranks third among 16 genera inabundance of individuals and sixth in abundance of colonies. However, all of thelocations surveyed by Jeanne were lowland sites (below 300 m elevation). Agelaiaspecies appear to be by far the most abundant eusocial wasps at higher elevations(above 1300 m) in the Monteverde area (S.O’D. & F. J. J., pers. obs.), and thisgenus may be particularly successful at high elevations in other locations as well.Mischocyttarus colonies are often abundant in Neotropical habitats (colony abundanceranked second among 16 genera), but due to their small colony populations,their individual abundance is generally much lower than that of swarm-foundingEpiponines (individual abundances of Mischocyttarus wasps ranked fourteenth amongthe 16 genera that were surveyed).

Polistes wasps also participate in mimicry complexes among Vespidae, but ap-parently not nearly as often as Mischocyttarus. Polistes species have similar socialstructure, and similar individual and colony abundance, to Mischocyttarus species inthe Neotropics ( Jeanne, 1991). The relative scarcity of epiponine mimicry by Polistesspecies may be partly explained by higher levels of aggression (e.g. greater willingnessto sting) by Polistes wasps than Mischocyttarus wasps, but this possibility remains to betested.

Evolutionary maintenance of M. mastigophorus colour polymorphism at Monteverde

The dark morph of M. mastigophorus almost certainly mimics A. xanthopus. Weknow of no other eusocial wasp that shares their colour pattern. It is possible thatthe pale morph of M. mastigophorus represents a more generalized vespid aposematiccoloration pattern (yellow body with dark striped markings). Another yellow andblack species of swarm-founding vespid, A. areata, is common at lower elevations inthe Monteverde area and probably overlaps the lower elevational range of M.mastigophorus. However, A. areata differs from the pale morph of M. mastigophorus andA. yepocapa in having a lighter yellow background and less extensive dark markings.Close resemblance and extensive elevational overlap suggest that M. mastigophorusmimicry of A. yepocapa is specific.

Although vertebrates are often implicated as predators driving the evolution ofvisual mimicry among insects (Cuthill & Bennett, 1993; Dittrich et al., 1993; Mappes& Alatalo, 1997), visually-hunting invertebrates may also act as important selectiveagents favouring resemblance. Hunting dragonflies have been observed rejectingeusocial wasps as prey in Monteverde (O’Donnell, 1996). Because M. mastigophorus


nests are often located in sheltered, dark sites, mimetic protection may accrue mostlyto foragers when away from their nests, rather than to individuals at their nests.

Elevational segregation of the Agelaia models was reflected in both nest and foragerabundances, although our data suggest that foragers range to elevations beyondtheir nesting limits. Coevolving species, including Batesian and Mullerian mimics,rarely have perfectly overlapping geographic ranges. Furthermore, selection forBatesian mimicry does not require complete geographic or phenological overlap ofmimic species with their models (Waldbauer, 1988; Thompson, 1994). The abund-ance of M. mastigophorus shows greater phenological variation at Monteverde thanthe models’ abundance. Swarm-founding wasp colonies, including Agelaia species,remain large throughout the colony cycle and do not pass through a solitary phase.Even newly-arrived swarms of the putative model species comprise thousands ofadults (S. O’D. & F.J.J., pers. obs.). In contrast, the proportion of newly-foundedM. mastigophorus colonies varies among seasons at Monteverde, with greater than75% of colonies we sampled from August to January (late wet season to dry season)comprising small groups of foundresses (one to six individuals) without workerspresent. Peak population density of M. mastigophorus apparently occurs in the earlywet season (colonies sampled May to July), and corresponds to nesting and fledgingperiods of many bird species at Monteverde (unpublished data).

The degree to which M. mastigophorus mimicry evolved through Batesian orMullerian processes is unclear. This distinction may be unimportant from a co-evolutionary perspective. Because M. mastigophorus is rare relative to its models, evenat peak population densities, it is unlikely to affect selection on the models’ colourpatterns even if M. mastigophorus is relatively defenceless against predators and istherefore a Batesian colour mimic (Gilbert, 1983; Alatalo & Mappes, 1996). Thevespid sting apparatus is exocrine and can deliver venom outside the wasp’s body.Since wasps can sting potential predators and survive, self-sacrifice of modelindividuals may not be necessary to induce avoidance learning by predators. Exocrinedefense may enhance the effectiveness of Vespidae as Mullerian mimics.

Elevational variation in climate, and its effects on insect thermoregulation, is animportant alternative to mimicry as a possible selective factor favoring the increasedabundance of dark morph M. mastigophorus at higher elevations. Persistent cloudcover and low temperatures may make efficient thermoregulation difficult for insectsat higher elevations in Monteverde (Hartshorn, 1983; Clark et al., in press). Theincreased abundance of dark morph M. mastigophorus individuals with elevation ispredicted on the basis of climatic effects on thermoregulation. Many of the vespidwasp species that nest in Monteverde are dark-bodied, and some species are darkerthan congeners or even conspecific populations at lower elevations on the Pacificslope (O’Donnell, in press). Body colour effects on foraging efficiency at low ambienttemperatures or in overcast conditions may also contribute to the partial elevationalsegregation of A. xanthopus and A. yepocapa at Monteverde and in the Talamancarange. Agelaia yepocapa foragers collected in the Northern Talamanca range of CostaRica (at elevations higher than at Monteverde) had more extensive dark markingsthan A. yepocapa from the Monteverde population (S.O’D., pers. obs.), furthersuggesting a role of climate as a selective factor in establishing elevational patternsof wasp body coloration. However, climatic gradients across the Monteverdeelevational range of M. mastigophorus do not correspond to dramatic colour differenceswithin other species of Polistinae (O’Donnell, in press), and are unlikely to be strongenough to maintain discrete colour polymorphism in the absence of other factors.


Evolutionary origin of M. mastigophorus dual mimicry

Mischocyttarus mastigophorus appears to exhibit a wide latitudinal range, but isrestricted to high elevation habitats along with its putative model species. Mischocyttarusmastigophorus has been collected at high elevations in Mexico, Costa Rica, andEcuador (Appendix). The fragmentary distribution of high-elevation habitat in theNorthern Neotropics (Fig. 5) may isolate populations of species restricted to montanehabitats, thereby favouring rapid evolutionary diversification. Similar processes havebeen invoked to explain the evolution of polymorphic mimicry complexes in lowlandbutterfly species in response to fragmentation of forest habitat following climatechange (Turner & Mallet, 1996). Recent evidence suggests that changes in mimeticmorph frequencies can change rapidly (on the order of several decades) followinghuman-induced forest disturbance (Linares, 1997).

Mischocyttarus mastigophorus overlaps only one of the putative model species at eitherend (N and S) of its geographic range (Fig. 5). We hypothesize that the M. mastigophoruspale morph evolved in the south in sympatry with A. yepocapa, later coming intocontact in the center of its range (e.g. Costa Rica) with the dark morph, which hadevolved to the north in sympatry with A. xanthopus. We predict that distributions ofM. mastigophorus morphs will be shown to correspond with geographic segregationof their model species.

ACKNOWLEDGEMENTS

Special thanks to Eva Chun, who first noted and drew our attention to colourdimorphism in M. mastigophorus. Susan Bulova, Sara Ranger, and Nicolas Spangassisted with field work. Thanks to Paul Hanson and Humberto Lezama for helpin transporting specimens. James Carpenter identified the wasps. Scott Mooreassisted with implementation of GIS software. Thanks to Karen London, EmıliaMartins, Fred Nijhout, Mary Jane West-Eberhard, and two anonymous reviewersfor helpful comments and discussion. The Monteverde Cloud Forest Reserve, theMonteverde Conservation League, the Hotel Fonda Vela, and members of theMonteverde community gave us permission to conduct research on their land andprovided logistical support. Research and collecting permits were obtained by S.O’D. and F.J.J. from the Costa Rican Ministry of Natural History. Financial supportwas provided by an NSF Postdoctoral Fellowship and the University of WashingtonRoyalty Research Fund (grants to S.O’D.)

REFERENCES

Alatalo RV, Mappes J. 1996. Tracking the evolution of warning signals. Nature 382: 708–710.Clark KL, Lawton RO, Butler P. In press. The physical environment of Monteverde. In: Nadkarni

NM, Wheelright NT, eds. The Natural History and Ecology of Monteverde, Costa Rica. New York: OxfordUniversity Press.

Cuthill IC, Bennett ATD. 1993. Mimicry and the eye of the beholder. Proceedings of the Royal Societyof London B253: 203–204.

Dittrich W, Gilbert F, Green P, McGregor P, Grewcock D. 1993. Imperfect mimicry: a pigeon’sperspective. Proceedings of the Royal Society of London B251: 195–200.


Gilbert LE. 1983. Coevolution and mimicry. In: Futuyma DE, Slatkin M, eds. Coevolution. Sunderland:Sinauer Associates, 263–281.

Hartshorn G. 1983. Plants: Introduction. In: Janzen DH, ed. Costa Rican Natural History. Chicago:University of Chicago Press, 118–157.

Holdridge LR. 1967. Life Zone Ecology. San Jose: Tropical Science Center.Jeanne RL. 1975. The adaptiveness of social wasp nest architecture. Quarterly Review of Biology 50:

267–287.Jeanne RL. 1991. The swarm-founding Polistinae. In: Ross KG, Matthews RW, eds. The Social Biology

of Wasps. Ithaca: Cornell University Press, 191–231.Linares M. 1997. The ghost of mimicry past: laboratory reconstitution of an extinct butterfly ‘race’.

Heredity 78: 628–635.Mappes J, Alatalo RV. 1997. Batesian mimicry and signal accuracy. Evolution 51: 2050–2053.O’Donnell S. 1995. Necrophagy by Neotropical swarm-founding wasps (Hymenoptera: Vespidae,

Epiponini) Biotropica 27: 133–136.O’Donnell S. 1996. Dragonflies (Gynacantha nervosa Rambur) avoid wasps (Polybia aequatorialis Zavattari

and Mischocyttarus sp.) as prey. Journal of Insect Behavior 9: 159–162.O’Donnell S. 1998. Reproductive caste determination in eusocial wasps (Hymenoptera: Vespidae).

Annual Review of Entomology 43: 323–346.O’Donnell S. In press. Eusocial wasps of Monteverde (Hymenoptera: Vespidae). In: Nadkarni NM,

Wheelright NT, eds. The Natural History and Ecology of Monteverde, Costa Rica. New York: OxfordUniversity Press.

O’Donnell S, Jeanne RL. 1990. Notes on an army ant (Eciton burchelli) raid on a social wasp colony(Agelaia yepocapa) in Costa Rica. Journal of Tropical Ecology 6: 507–509.

O’Donnell S, Joyce FJ. In press. A dual mimicry complex involving three eusocial wasp species atMonteverde (Hymenoptera: Vespidae). In: Nadkarni NM, Wheelright NT, eds. The Natural Historyand Ecology of Monteverde, Costa Rica. New York: Oxford University Press.

Richards OW. 1978. The Social Wasps of the Americas. London: British Museum, Natural History.Starr CK. 1990. Holding the fort: colony defense in some primitively social wasps. In: Evans DL,

Schmidt JO, eds. Insect Defenses. Stonybrook: SUNY Press, 421–463.Thompson JN. 1994. The Coevolutionary Process. Chicago: University of Chicago Press.Turner JRG, Mallet JLB. 1996. Did forest islands drive the diversity of warningly coloured butterflies?

Philosophical Transactions of the Royal Society of London B351: 835–845.Waldbauer GP. 1988. Asynchrony between Batesian mimics and their models. The American Naturalist

Supplemental Issue 131: 103–121.Wenzel JW. 1992. Extreme queen-worker dimorphism in Ropalidia ignobilis, a small-colony wasp

(Hymenoptera: Vespidae). Insectes Sociaux 39: 31–43.West-Eberhard MJ, Carpenter JM. 1995. In: Hanson PE, Gauld ID, eds. The Hymenoptera of Costa

Rica. Oxford: Oxford Science Publications, 561–587.Wickler W. 1968. Mimicry in Plants and Animals. New York: McGraw-Hill Book Co.


APPENDIX

Collection locations and elevations of Mischocyttarus mastigophorus, Agelaia xanthopus and A. yepocapa.Collections where elevation was reported are given in bold type.

Department/ Approx.State/ Nearest Elev. Reference/

Lat/Long Country Province town (meters) Source

Mischocyttarus mastigophorus19° 09′N, 96° 57′W Mexico Vera Cruz Huatusco 1300 1

10° 18′N, 84° 48′W Costa Rica Puntarenas Monteverde 1460–1700 2

9° 45′N, 84° 18′W Costa Rica San Jose San Antonio de 1500 3Escazu

3° 45′S, 79° 45′W Ecuador El Oro Moromoro 1500 1

Agelaia yepocapa14° 34′N, 90° 44′W Guatemala Sacatepequez Antigua 1500 1

14° 30′N, 90° 57′W Guatemala Chimaltenango Yepocapa 1400 1


9° 31′N, 83° 40′W Costa Rica Puntarenas Division 2360 2

9° 32′N, 83° 45′W Costa Rica Puntarenas San Gerardo 2200 2de Dota

0° 15′S, 79° 10′W Ecuador Pichincha Santo Domingo 450 1

Agelaia xanthopus18° 51′N, 97° 06′W Mexico Vera Cruz Orizaba 1284 1

17° 04′N, 96° 43′W Mexico Oaxaca Oaxaca 1550 1

15° 16′N, 90° 12′W Guatemala Baja Verapaz Purulha 1500 1

15° 10′N, 90° 20′W Guatemala Alto Verapaz Head Rio 1300 1Polochic

15° 03′N, 90° 12′W Guatemala San Jeronimo Santa Cruz 1800 1


9° 55′N, 83° 50′W Costa Rica Cartago Volcan Irazu 1829–2133 1

9° 54′N, 83° 41′W Costa Rica Cartago Turriabla 700 1

9° 55′N, 84° 10′W Costa Rica San Jose San Antonio 2000 2de Escazu

9° 32′N, 83° 54′W Costa Rica Puntarenas San Gerardo 2200–2780 2de Dota

References/sources for collections: 1. Richards (1978). 2. This study. 3. Martin Cooper, pers, comm.

dual mimicry in the dimorphic eusocial wasp …so356/pdfmmdual.pdf · dual mimicry in the dimorphic...

Documents