Ecological Characterization of a Costa Rican Dry Forest Caterpillar Fauna

Author(s): Daniel H. Janzen

Source: Biotropica, Vol. 20, No. 2 (Jun., 1988), pp. 120-135
Published by: The Association for Tropical Biology and Conservation
Stable URL: http://www.jstor.org/stable/2388184 .
Accessed: 11/06/2014 18:51

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

The Association for Tropical Biology and Conservation is collaborating with JSTOR to digitize, preserve and
extend access to Biotropica.

http://www.jstor.org

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions
Ecological Characterizationof a Costa Rican Dry
Forest CaterpillarFauna1

Daniel H. Janzen
of Pennsylvania,Philadelphia,Pennsylvania19104, U.S.A.
Departmentof Biology,University

ABSTRACT
The dryforests of Costa Rica's SantaRosa NationalPark (11?N lat., 0-350 m elevation,6 mo withoutrain,900-2300 mm rain/
yr) rangefromdeciduous and 2 m tall to nearlyevergreenand 30 m tall. This vegetationand a complexmosaic of secondary
successionin old fieldsand pasturessupportsat least 3 140 speciesof caterpillars. Butterfly caterpillarsconstituteonly 11 percent
of the speciesand less than 1 percentof the biomass.Althoughthe mothcaterpillars are veryspecies-rich, each of the 725 species
of vascularplantsis attackedby less than20 speciesof caterpillars. Fromthe viewpointof a speciesof plant,the caterpillar species
richnessin Santa Rosa is therefore less thanfromthe viewpointof anyof a numberof extratropical
substantially treespecies;they
are fed on by manymorespeciesof caterpillars at a singlelocation.AlthoughSanta Rosa does have some speciesof caterpillars
thatfeedon seeds,dung,detritus,bark,etc.,morethan95 percentof thespecieseat greenleaves.Of these,about 37 percentfeed
exposedon the leafsurface.
A Santa Rosa woodyplantcommonlyloses 1-20 percentof its leafarea to defoliators duringthe firsthalfof the rainyseason
and verylittleafterthat,but on rareoccasionsplantsare totallydefoliated.These defoliationsare almostalwayscommittedby a
singlegeneration feedingon their1 or 2 sole hostplants.Nearlyall the speciesof SantaRosa plants
of 1 or 2 speciesof caterpillars
are not fed on by any given speciesof caterpillareitherbecause the plantscontainchemicalsthat renderthem unacceptableor
becausethe caterpillar refusesto eat foliagethatlacks a feedingstimulant.Simultaneously, it is the negativeimpactof inclement
weatherand carnivores on Santa Rosa caterpillarpopulationsthatpreventsanygivenspeciesof caterpillar fromusuallydefoliating
the fewspeciesof plantsthatit can eat.
The SantaRosa carnivore faunarangesfromabsolutelymonophagousto highlypolyphagous.Itscaterpillar foodsupplyfluctuates
enormously withyear,season,positionin a season,and microsite.This favorsgeneralizedand flexiblecarnivores on the one hand,
and strongly specializedcarnivoreson the otherhand. Many caterpillars are not available to the carnivoresbecause theyare too
crypticto be found,personallywell-defendedby urticatingspines or chemicals(and are aposematic),mimicsof distastefulor
dangerouscaterpillars, or participatingin predator-satiating synchronizationof life-stages.Probablyno carnivorerecognizesmore
thanabout threekindsofpreyor host:thosethatareperceivedand rejected,thosethatare eatenor used directly whenencountered,
and thosethatrequiresome specialtreatment. No two speciesof carnivores divide the manyspeciesof caterpillars equallyamong
thesethreekindsof preyor hosts.Certainlyno carnivoreviewsthe Santa Rosa caterpillar faunaas consistingof 3140 species.
Virtuallyall Santa Rosa caterpillarspeciessharetheirhost plantswithless than 20 othercaterpillar species(the averagemay
be less than 5). This means thatfromthe directperspectiveof any one speciesof caterpillar, the Santa Rosa caterpillarfauna is
quite small and potentialcompetitiveinteractions are verylimited in ecologicaltime. Furthermore, actual directcompetitive
interactionsare rare(if theyoccurat all) owingto the generallylow densityof caterpillars at any givenpoint in space and time.
However,thereare many potentialopportunitiesfor indirectintra-and interspecific caterpillarinteractions that are mediated
throughsharedcarnivores (such as parasitoidHymenopteraand Diptera,diseases,vertebrates, spiders,etc.).

CATERPILLARS ARE THE LARVAE of Lepidoptera,and in most vegetationrangedfrom2-m-talltotallydeciduousforest

forestsof theworldtheyconsumemorelivingleavesthan on serpentine(peridotite)ridgesoverlookingthe ocean,
all otheranimals combined.The understanding of an to 30-m-tallnearlyevergreen foreston mesas,hills,and
herbivorefauna thus becomesin greatpart the under- swalesat 300 m elevation.This vegetationhas been se-
standingofthecaterpillar fauna.My goal is to understand verelyperturbedby ranching,burning,agriculture, log-
theherbivore faunaof SantaRosa NationalPark,a-semi- ging,and huntingsincethe late 1500s; priorto that,at
forested10,800-ha preserveon the Pacificcoastal plain leastsomeof it was subjectto indigenousagriculture.The
(0-350 m elevation,11?N lat.) of northwestern Guana- rainyseasonis approximately6 mo long(May-December)
caste Province,Costa Rica (Fig. 1). This park contains and receives900-2300 mm of rain; the dry season is
some of the last (and tiny)remnantsof originallowland rainfree.Additionalcharacterization
essentially of thesite
dryforest(also calleddeciduousforest)on thePacificside can be foundinJanzen(198 la, 1984a, b, 1985b, 1986a,
ofCentralAmerica.Overall,thevegetationof SantaRosa b, d, 1987a-c) and Hartshorn(1983).
is a mosaicofabandonedpasturesand fields,and of ? 400- HereI offeran approximate ecologicalcharacterization
yr-oldsecondarysuccessionalforestrangingfromalmost faunaof Santa Rosa. It is also the first
of the caterpillar
The original
totallydeciduous(Fig. 2) to nearlyevergreen. attemptat an ecologicalcharacterizationof thecaterpillar
faunaof any tropicalhabitat.The studybegan in 1977
' Received 12 December 1985, revisionaccepted30 December and will continue.Upcomingfindings will undoubtedly
1986. modifyand enrichthe patterns,but this approximate

120 BIOTROPICA 20(2): 120-135 1988

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions
 treeon the other,and the otheris the response(s)to the
two trees'proximity by some thirdorganism.All other
measuresof proximity are anthropocentric conveniences,
usefulto humansbuta restrictive waytoportray biological
The samephilosophy
reality. appliesto thecharacterization
fauna.Thereis,at theleast,thetaxonomist's
ofa caterpillar
viewpoint,thehostplant'sviewpoint, thecarnivore'sview-
point, and the caterpillar'sviewpoint.Here I offeran
approximation different
of theserelatively viewpoints,the
amalgamationof whichmightbe termedthe ecologist's
viewpoint.
COSTA RICA .

METHODS
The factsand opinionsexpressedhereare based on 10 yr
and
of intenseobservationof the Santa Rosa caterpillars
adult Lepidopterafaunain the wet and dryseasons,be-
ginningin 1977. Moths have been intensively collected
at 15-W blacklights and fluorescentwhitelightsat fixed
FIGURE 1. LocationofSantaRosa NationalParkin north- positionsoverlookingthe forestcanopyand throughthe
westernCostaRica.Thedottedlineis theInteramericanHigh-
way.Thefivemostnorthernvolcanos deGuana- forest
oftheCordillera At foursitestheselightsweremaintained
understory.
by theirspecific almostcontinually
de Tilaranare indicated
casteand Cordillera for3 yrand for2-4 mo in each ofthe
names. otheryears.Over 50,000 mothshave been collectedand
preparedas museumspecimens,but this is less than 1
percentof thoseexamined.The rateof capture(at lights)
image is intendedto encourageothersto exploreentire of macromothsand largemicrolepidoptera thatare new
faunasin the habitatstheyknowwell.
caterpillar to theSantaRosa collectionhas declinedto an averageof
Thereare onlytwo biologicallymeaningful measures only about one per month,and I am certainthat the
oftheproximity of two trees.One is theinfluenceof each figuresforSanta Rosa in Table 1 are greaterthan 90

TABLE 1. Comparative ofthecaterpillar

speciesrichness siteswithdiversesuccessional
faunas ofthreelocal continental and
histories
approximately
covering 100 km2.

Santa Rosa Ithaca, Kevo Station,

National Park New York northernFinlandb
Taxon (110 N lat.) (420 N lat.)a (700 N lat.)

(incl. Hesperiidae,Lycaenidae,Riodinidae)
Butterflies 345C 105 21
Arctiidaesensulato (incl. Ctenuchidae,Pericopidae) 90 41 1
Lymantriidae 5 7 0
Limacodidae 20 13 0
Cossidae 25 3 0
Sphingidae 84 34 2
Saturniidae 35 11 0
Mimallonidae 8 2 0
Apatelodidae 10 2 0
Lasiocampidae 5 5 2
Geometridae 400 206 46
Notodontidae 150 50 2
Noctuidae 800 487 22
Microlepidoptera (incl. Pyralidae) 1130 600 168
Miscellaneous 35 11 0
Totald 3142 1577 264

and J.
a Based on the synopticcollectionof Lepidopteraat CornellUniversity, pers.comm.
Franclemont,
b Based on extensiveLepidopterasurveys(Koponen et al. 1982).
otherthanHesperidae,Lycaenidae,and Riodinidaeare derivedfromDeVries (1983).
c Butterflies
species).
d Totals should be discountedby about 3 percentto accountforwaifsor strays(nonbreeding

Ecology of Costa Rican CaterpillarFauna 121

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions
 A~~~~~~~~~~~~~~~~~~~~A

4~~~12~~

V~

FIGURE 2. (A) 60- to 80-yr-oldsecondarysuccessionaldryforestmnthe firthalfof the rainyseason(June). (B) Same view in
the last thirdof the dryseason (April).__________________________________

percentof therealtotalforthesegroups.Some speciesof of leafminersin the dryforestsof the COMELCO area

microlepidoptera have been as thoroughly collected(e.g., nearBagaces,about 70 km southof Santa Rosa.
Cossidae,Pyralidae,Oecophoridae,Tortricidae)as have I and numerousassistants havespentmorethan5000
been the macromoths. person-hours observingand rearingmore than 10,000
Approximately 400 speciesofmothshavebeenreared accessionsof wild-caughtcaterpillars (and theirparasit-
from more than 5000 wild-caughtcaterpillars,and oids) in Santa Rosa's dryforest.Free-livingcaterpillars
onlyone of thesespeciesalso has not been takenat the have beenobtainedforrearingby virtually everyconceiv-
lights.Herbs,vines,shrubs,and treesin all habitatsare able kind of collecting,indudingcuttinglargebranches
searchedforlarvae.The completefaunaof saturniids and out of treecrownsand inspectingtheirleaves,climbing
almostall sphingidshas beencollectedand reared(Janzen treeswithflashlights at night,and searchingforthe cat-
havebeenintensively
1984b, 1986d). Butterflies collected erpillars
thathavedroppedspecific kindsoffrasson sheets
by DeVries(1983), and further collectingand rearingof spreadbelow trees(foggingwithpesticideshas not been
thehesperiids,lycaenids,and riodinidsin thepresentproj- used because the goal is to rearthe caterpillars encoun-
ect leads me to condude that the butterflies have been tered).In the approximation of caterpillarecologypre-
relatively
thoroughly collected.Leaf-mining and othermi- sentedin thisessay,I dwell onlyon thosepatternsthat
crolepidoptera thatdo not come to lightshave been es- appearso emergentand/orrepeatedlythattheyare un-
timated(Table 1) throughconversations withtaxonomists likelyto be significantly alteredby further study.
workinginthesegroups,myobservations ofleafand other of the plantsand otherorganismsal-
Identifications
kindsof mines,and rearingof microlepidoptera; P. A. luded to hereare based eitheron mypersonalconfidence
Opler (pers. comm.) feelsthatthisfigureis as much as in the names of organismswell known to those of us
30 percenttoo low, based on hisrearingand observations workingin thisforest (e.g.,Janzen& Liesner1980, Janzen
122 Janzen

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions
 by specialists(see
1983a) or on specificdeterminations forIthaca,New York,and Kevo, Finland,are 59 and 36
acknowledgments). All photographsweretakenin Santa percent,respectively. All threeof thesepercentages were
Rosa National Park. determined by characterizing thegroups,or subgroupsof
them,in Table 1 as livingeitherinsideor on leafsurfaces.
In short,a substantial fractionofthespeciesin a caterpillar
THE TAXONOMIST'S VIEWPOINT faunalivesinsideof a plantpartand is thusnotgenerally
Sinceeveryspeciesof mothand butterfly has a caterpillar availableto foliagegleanersthatdo not penetrateor rip
(technically a "larva"), Santa Rosa is presently knownto open plantparts.
be occupiedby about 3140 speciesof caterpillars (Table Adult butterflies (includingHesperiidae)are a very
1). Howwever,thisfigureshouldprobablybe discounted visiblepartofthelepidopterous faunaofanytropicalsite.
by about 3 percentto account for nonbreedingstrays However, they constituteonly a small fractionof the
(Janzen1986c, 1986d). Such straysappearat about this speciesof Lepidoptera.At SantaRosa, Ithaca,and Kevo,
frequency inlargegroupsofSantaRosa Lepidoptera whose theyare 11, 7, and 8 percent,respectively, of the fauna
residency statusis well known(e.g., butterflies, Saturni- (Table 1). Theirecologicalinsignificance is emphasizedby
idae, Sphingidae,Noctuidae, Notodontidae,Pyralidae). the factthat theyconstitute well less than 1 percentof
As indicatedearlier,the figureof 3140 is a compositeof thebiomassof caterpillars or adultsat Santa Rosa. How-
data fromwild-caughtadultsand larvae,and estimations ever,wildland butterfly biologyhas receivedfar more
forsomemicrolepidoptera families.The microlepidoptera attention (e.g.,Owen 1971, Gilbert& Singer1975, Vane-
maybe underestimated by severalhundredspecies(P. A. Wright& Ackery1984) thanhas thatof moths.
Opler,pers.comm.),but thatunderestimate will notsig- Perhapsthe most glaringtraitto the taxonomistis
nificantly affectthe conclusionspresentedhere. that the Santa Rosa Lepidopterafauna is not local or
The Santa Rosa mothfaunais probablyat its peak "endemic."Amongthe35 speciesofsaturniids (30 breed-
in speciesrichness withrespectto thenearpastand future. ing residentsand 5 strays)and 84 speciesof sphingids
The progressive insularizationofthispieceofforest during (74 usually-to-occasionally breedingresidentsand 10
the past 400 yrhas probablynot yeteliminatedall the strays),onlyone speciesis restricted ("endemic") to the
species that cannot persistindefinitely in the scattered area (Janzen 1984b, 1986d). This is Schausiellasanta-
fragments of thediversehabitatsoccurring in a dryforest rosensis, a medium-large ceratocampine saturniid (Lemaire
area of 108 km2 on diversetopography.Because Santa 1982, 1987) that seems to occur only in eastern Santa
Rosa was an operatingcattleranchand a mosaicof rice Rosa and thenceto thelowerfoothills ofthevolcanicrange
fields,cornfields,selectively loggedforest, and variously- 10-15 km to the northeastof the park (Fig. 1). This
aged old pasturesuccessionwhen it became a national oddlyrestricted geographicdistribution ofabout 500 km2
parkin 1972, it nowcontainsjustabouteveryimaginable occursdespitethefactthatthecaterpillar is a monophagous
kindof habitatand successionalstagethatcan be derived feederon the leaves of a tree(Hymenaeacourbaril,Le-
fromdryforest.As thishabitatdiversity is graduallyre- guminosae)that rangesfromMexico to Brazil. The re-
duced throughsuccessionto a relatively continuous(and maining118 speciesof Santa Rosa saturniidsand sphin-
eventually pristine)dryforestcover,some speciesof Lep- gids have geographicdistributions thatextendover 10-
idopterawill undoubtedlydisappear.This should result 60 degreesof latitudeand occupyverydiversehabitats
in a net loss of species,because the maturingforestscan (Janzen 1986d). Among the 345 speciesof butterflies,
no longeracquirea fullpristineforestLepidopterafauna no local speciesor evendistinctive populationshave been
fromareasadjacentto SantaRosa. The pristinedryforest encountered. Among the 150 species of Notodontidae
areasoutsideof the park weredestroyed long ago. knownfromSantaRosa (morespeciesthanin all ofNorth
Based on information in theliterature, lepidopterist's Americanorthof Mexico;J. Franclemont, pers.comm.),
folklore, and therearingprogramin SantaRosa, it is clear none appearendemicto Santa Rosa.
thatits Lepidopterafaunacontainsa smattering of seed AnotherconspicuoustaxonomictraitoftheSantaRosa
predators (mostlyphycitine and chrysaugine pyralids)and macrolepidoptera is that less than 20 percentwere un-
fruitminers(mostlypyralids;e.g., Janzen1983c), stem describedat thetimethestudybegan.Therewereno new
borers,gall borersor makers,detritivores, dung eaters, sphingids,one new saturniid,fivenew notodontids, and
and carnivores. It evenhas a tineidmicrolepwhoseadult approximately15 new pyraustinepyralids.The reason
femalesride on the backs of Liomyssalvini mice from why the Santa Rosa moth fauna consistsof described
nestto nest (Davis et al. 1985). However,at least 95 speciesis not because collectionsweremade in thispart
percentof thespeciesof Santa Rosa Lepidopteraare con- ofCostaRica byearlycollectors. Rather,eachofthespecies
sumersof greenleaves. Only approximately 37 percent in Santa Rosa was collectedelsewherein its largerange.
feedwhileexposedon In short,the caterpillar
of all of the speciesof caterpillars phenotypesthatsurvivein Santa
theleafsurface,whilethe remainder are leafminers,leaf Rosa also surviveelsewhere; biogeographically, SantaRosa
rollers,orsomekindofcase-bearer. The comparablefigures has beena veryordinary place (Janzen1986d). However,
Fauna
EcologyofCosta RicanCaterpillar 123

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions
Santa Rosa is now a forestislandfloatingin an ocean of will view the forestas being quite species-rich in cater-
pastureand cropland.This insularization willundoubtedly pillars.
take its toll of widespreadspecies. Eventheaveragefigures are a misleadingindicatorof
How species-rich is the Santa Rosa caterpillarfauna what is experienced by a plant. The Ithaca,New York,
in comparisonwiththatof othersmalland well-surveyed area has about 1580 species of caterpillars(Table 1).
mainlandareas?Two aretabulatedin Table 1. The overall TompkinsCounty,New York, whichroughlyapproxi-
increasein Lepidopteraspeciesrichness thatis encountered matestheareasurveyedfortheIthacaLepidopteracensus,
in movingfromhigherto lowerlatitudesis not equally containsabout 1800 speciesof nativevascularplants(R.
intensefor all familiesof Lepidoptera,but thereis an Wesley,pers. comm.). Kevo, in northernFinland,has
approximate doublingin movingfromIthaca,New York, about260 speciesofcaterpillars (Table 1) and 248 species
to Santa Rosa. of nativevascularplants(Makinen& Kallio 1979). On
Five yearsof intensivecollectingof mothsat about averageeach of the two extratropical siteshas about one
20 rainforest sites in Costa Rica (a country-wide moth speciesof caterpillar per plantspecies,as comparedwith
survey,D. H. Janzen& W. Hallwachs,pers.obs.) have a littlemore than fourper host speciesat Santa Rosa.
leftthe impressionthat Santa Rosa is about as richin However,the directionand intensity of thisgradientin
speciesof mothsas are nearbylowland rainforest sites. averagescertainly does not reflectthe speciesrichnessof
This equalityamong dissimilarhabitatsis due in partto theinteractions betweena treeand itscaterpillars at each
certainfamiliesbeingsomewhatmorespecies-rich in Costa of the sites.
Rican rainforests than at Santa Rosa (e.g., Saturniidae, Anotherquestionto ask a plantis how muchdamage
Geometridae),whereasothersare less so (e.g., Pyralidae, it sustainsfromcaterpillars:
Noctuidae,Ethmiidae).
REDUCTION IN LEAF AREA.-On an annual basis, the her-
bivoreloads in Santa Rosa remove1-10 percentof the
expandedleavesretainedbythetree.Another' 10 percent
THE PLANT'S VIEWPOINT of the leaf areas is removedin the formof leaves eaten
Thereare approximately 725 speciesofvascularplantsin entirely or leaves thatare sufficiently damaged thatthey
SantaRosa (Janzen& Liesner1980, pers.obs.). The 3140 are shed by the tree.Certainspeciesof Santa Rosa trees
speciesofcaterpillars aredistributedquiteunequallyamong [e.g., Simaroubaglauca (Simaroubaceae),Tabebuia rosea
them(numerousspeciesof plantshave onlyone species (Bignoniaceae),Guarea glabra and Trichilia americana
of caterpillar recordedfromthemto date). However,no (Meliaceae),Allophyllus occidentalis (Sapindaceae),Hemi-
speciesofplant,evenlargeand commontrees,is threatened angiumexcelsum (Hippocrateaceae)lexperience almostno
bymorethanabout 20 speciesofcaterpillars in ecological leafherbivory. At the otherextremeare treeswhoseleaf
time. For example,after6 yr of intensiverearingwild- cropis oftenseverelydamagedby theend of itslife[e.g.,
caughtcaterpillars in Santa Rosa, therecordlepidopteran C. candidissimum, Alibertiaedulis,and Genipaamericana
herbivore load is the 17 speciesof caterpillarsthateat the (Rubiaceae),Manikara chicle(Sapotaceae),Casearia cor-
leavesofManilkarachicle,a commonand largeevergreen ymbosa(Flacourtiaceae),Quercusoleoides(Fagaceae), Li-
treeof pristineand old secondarysuccessionalforest(3 cania arborea(Chrysobalanaceae), Spondiasmombin (An-
saturniids, 1 sphingid,2 lycaenids, 2 notodontids, 1 mim- acardiaceae),Guazuma ulmifolia (Ulmaceae), Luehea
allonid, 1 limacodid,1 arctiid,1 pyralid,2 noctuids,2 speciosa(Tiliaceae)}.
geometrids,and no leaf miners).Anothertreewith an The small annual amountsof leaf damage are ex-
exceptionally largefauna (15 spp. of caterpillars) is Ca- tremely difficultto measuredirectly, and theirpertinence
lycophyllum candidissimum (Rubiaceae). Thereis no hint to any directmeasureof plant fitnesscan onlybe deter-
thattheherbaceousplantsin theparkaresupporting more minedby species-specific and habitat-specific experiments
mothspeciesliststhatareon averageanylongerthanthose (e.g., Marquis 1984). How theselossescomparewiththe
of the woodyplants.At thisdate in the sampling,it is cost of the standingdefensesthatkeep the other3120-
too earlyto calculateeithera reasonableoverallaverage 3140 speciesof caterpillars (and otherherbivores)from
or standarddeviation,but judgingby therateof addition eatingthe foliageof a treeis unknowableat present,but
of new hostrecordsforparticular plantspecies,I estimate standingdefensesare undoubtedlymuchmoreexpensive
thatthe averagewill be between4 and 8 speciesof cat- thantheusual annualdamage to thatplant.Incidentally,
erpillarper plantspecies. the proportionand absoluteamountof the leaf damage
A plant at Santa Rosa therefore differsgreatlyfrom done by insectsotherthancaterpillars varieswidelyfrom
plantsofsimilarindividualsizesinmorenorthern latitudes, yeartoyear,conspecific to conspecific,and speciesto species.
wherea treespeciesis commonlyfedon by manytensof
speciesof caterpillars (e.g., Niemela & Neuvonen 1983). TOTAL DEFOLIATION.-At long multiyear intervals,one of
A SantaRosa treewillviewa tropicalforest as beingquite thespeciesofmonophagousornearlymonophagous(feed-
species-poor in caterpillars,whereasan IthacaorKevo tree ing on 2-3 closelyrelatedplants)caterpillars defoliatesa

124 Janzen

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions
Santa Rosa treeor severelydamages the leaf crop (e.g., case (Perigonialusca,Sphingidae,in Janzen1984b), and
Fig. 3). More than 40 such eventshave occurredin the I suspectinmanyothers,theypass one ormoregenerations
past 9 yr at Santa Rosa (e.g., Janzen 198 la, 1984a, in the wetterforest(wherethe habitatis probablyless
1985a). Such a defoliationeventnormallylasts foronly carnivore-rich thanSantaRosa becomesafteritsfirst gen-
one generation, are on theplant(s)for erationofcaterpillars
and thecaterpillars at thebeginning oftherainyseason).
2-3 wk. Usuallythe caterpillars thatdo the defoliation In othercasesthereis no suggestion thattheadultsmigrate
produceenormousnumbersof healthypupae, almostall out of Santa Rosa. Some of thesespeciesof caterpillars
adultseclosewithina fewweeks,and theythendisappear spendtheremainder of theyearas pupae, and otherspass
(in many cases, apparentlymigratingaway fromSanta a subsequentinconspicuous generation in thesecondhalf
Rosa, Janzen1986d, 1987b). Thereis normallyno hint oftherainyseason(Janzen1987b); a thirdgroupofspecies
ofeventhebeginning ofsubsequentdefoliating generation. haspopulationsthatcontainindividualsdoingboththings.
Even thoseindividualhostplantsthatare stillleafyafter Whateverthe behaviorof the mothsor caterpillars,
thedefoliation event(normallybecausetheywerecleaned defoliation eventsare clearlynot normallydirectly extin-
of caterpillarsby carnivoressuch as ants, spiders,true guishedby carnivores thatbuild up on successiveconspe-
bugs,and vertebrates; e.g.,Janzen1985a) are notheavily cificgenerations of caterpillars(Hylesia lineata has been
fedon further duringthatseason. In the one case where the onlyexception).Caterpillaroutbreaksat Santa Rosa
a seconddefoliating generationwas initiatedby the rel- generally sustainlessthan5 percentparasitization bypara-
ativelypolyphagoussaturniidcaterpillars of Hylesia li- sitoids(e.g., Janzen1985a). By definition the vertebrate
neata,thesecondgeneration was conspicuouslyterminated carnivores musthave been satiatedby the defoliators (or
beforepupationbya virusdisease(Janzen198 la, 1984a, a defoliation eventwouldnothaveoccurred).On theother
b). hand,migration or pupal diapause duringthe rainy(and
Such defoliatingeventsoccurat such long intervals dry)seasoncan be viewedas the termination of a defoli-
that to date only one repeatperformance has been ob- ationeventby carnivores in evolutionary ratherthaneco-
served.Randia subcordataand R. aculeata (Rubiaceae) logicaltime.Were a mothto beginevolvingin waysthat
weredefoliatedby a singlegenerationof Aellopostitan led to successivehigh-density generations, it is easy to
(Sphingidae)caterpillarsin theearlyrainyseasonof 1978 imagineitsdemiseat SantaRosa throughthecombination
and 1983 (Janzen1985a; one hostwas incorrectly iden- of carnivores and depletionof the foliageof its particular
tifiedas Randia karstenii).However,otherrepeatshave hostspecies.
occurred inwhichthecaterpillars wereextremely common, Clearly,10 yrof rearingrecordsand observations at
but not abundantenough to have caused conspicuous Santa Rosa are inadequatefordiscerning inter-or intra-
defoliationof theirhosts. specificmultiyearpatternsto defoliationevents.For ex-
Defoliationsof trees(e.g., T. rosea,Tabebuia impe- ample, the decade beginningin 1977 has had onlyone
tiginosa)by theleaf-feeding larvaeof chrysomelid beetles yearwithoutanyspeciessuffering a majordefoliation and
at Santa Rosa followthe same patternas thatdescribed one year of defoliationof many species. However,the
above forcaterpillars.However,monophagousspeciesof absenceof repeatdefoliations strongly suggeststhata tree
adult meloidbeetles(blisterbeetles)aggregateon a very speciesthatis defoliatedis likelyto growformanyyears
fewindividualsof theirhostplants[e.g.,Pisonia macran- withoutbeingdefoliatedagain(thoughduringthatperiod
thocarpa(Nyctaginaceae),Diphysa robinioides (Legumi- it is used as a host plant by its herbivoreload at low
nosae)l and maystripoffan entireleafcropin one night. density).Also, the impact (if any) of the Santa Rosa
This occurson an annual basis at the beginningof the caterpillar faunaon thestructure of arraysofplantspecies
rainyseason. is throughthe subsequentgrowthand reproductive be-
Leaf replacement ratesvarystronglyamong species. haviorofa defoliatedspeciesofplant,ratherthanthrough
Leaves may be replacedalmostimmediately followinga themoreindirect pathwayofa speciesofcaterpillar build-
defoliationevent[e.g., S. mombin(Anacardiaceae)- defo- ing up to outbreakdensitieson one speciesof plant and
liatedbyeuteliinenoctuidsin 19841, afterseveralmonths thenmovingon to damage otherspeciesof plantsin the
(e.g.,Randia spp. defoliatedbyA. titan),or nearlya year same or subsequentgenerations.
laterat the timeof population-wide leafturnover by the
host(e.g., M. chicleafterdefoliationby manycaterpillars HOST-HERBIVORE SPECIFICITY.-The species-poorcaterpillar
of Othorene purpurascens thatwereexperimentally main- faunaofeach SantaRosa treespeciesis usuallynotwidely
tainedat a highdensity). sharedwithotherspeciesofhostplants.Forexample,only
Reasonsforthelack ofa secondsuccessivedefoliating one(Erinnyis ello,Sphingidae)ofthe 17 speciesmentioned
generationvaryamong plant and caterpillarspecies.In earlieras eatingM. chicleleaves also eats the leaves of
manycases,theadultsmigrateout ofthepark,apparently threeor moreothertreespecies(and even in thiscase it
towetter forestinotherpartsofCostaRica (Janzen1984b, is restricted to a fewlatex-rich species;e.g., Dillon et al.
1987c) wheretheypass theremainder of the Santa Rosa 1983; Janzen1984b, 1985a). The arctiidalso eatsMas-
wetseasonand theSantaRosa dryseason.In at leastone tichodendron capiri (Sapotaceae). The other 15 species

Ecology of Costa Rican CaterpillarFauna 125

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions
126 Janzen

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions
 XI,X',"{

FIGURE 4. Ventralfrontalview of a livingspinypocketmouse (Liomyssalvini) thatwas carrying a prepupaof Cautethiaspuria

(Sphingidae)in one of itspouches.The head of theprepupahas been bittenoff,and the entireinsectwould have been eatenlater
in the burrow.

appearto be monophagous.The hostrecordsaccumulated the remainderoligophagousto only a few species,the

to date suggestthatwhen the finalcountsare done, at intensityof caterpillar
challengeto a givenplant species
leasthalfof the Santa Rosa caterpillarswill be foundto is not directlyrelatedeitherto the numberof speciesof
have onlyone host plant speciesat Santa Rosa, and at in thehabitator to thedetailsof the traitsof
caterpillars
least 80 percentof the remainderwill have just a few nearlyall of the otherspeciesof plantsin thehabitat.In
chemically or taxonomically relatedspeciesof hosts(e.g., otherwords,H. courbarildoes not in any directway
see thehostlistsforSphingidaeinJanzen1984a, 1985a). experiencethe presenceof the herbivoreload thatfeeds
The highlypolyphagousspecieswith20+ speciesof host on T. rosea.
plantsin numerousplantfamilies[e.g.,H. lineata (Janzen Threeobservations suggestthatplantdefensivechem-
1984a); Hypercompe spp.,Arctiidae)aredearlya miniscule istryhas littleor nothingto do directlywith the daily
proportion of the caterpillar
fauna. challengeby a plant'sherbivore faunain SantaRosa: the
In sum, each of the Santa Rosa plants "sees" (in membersof thatfauna obviouslycan eat the foliageof
ecologicaltime) its habitat as containingonly a small thatplantspecies,thereis no suggestion
thatleafchemistry
numberof speciesof caterpillars.Becauseat leasthalfthe changesseasonally(eitherin the rearingrecordsor in
speciesof Santa Rosa caterpillarsare monophagous,and analyses;e.g.,Jansen& Waterman1984), and duringthe

FIGURE 3. (A) Crownof largeadult Acosmium panamensis(Leguminosae)defoliatedby a singlespeciesof noctuidcaterpillar

(June 1984). (B) CrownofAnnonapurpurea(Annonaceae)defoliatedbycaterpillars of Gonodontapyrgo(Noctuidae); an undamaged
vineof Cissusrhombifolia
(Vitaceae) formsa roughX in the centerof the Annonacrown(June 1983). (C) Crownof Enterolobium
cyclocarpum(Leguminosae)defoliatedby a speciesof megalopygidcaterpillar
(cf Fig. 9b) (October 1982).

Ecology of Costa Rican CaterpillarFauna 127

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions
FIGURE 5. Last instarlarvaof Manducalefeburei(Sphingidae)witha sib groupof parasiticHymenoptera(Braconidae)newly
emergedfromits body cavityand spinningcocoons.

secondhalfoftherainyseasontheplantsarein fullfoliage
yetsupportalmostno caterpillars. However,it is equally
dear thatplantdefensive chemistry keepsthevastmajority j
ofcaterpillar
species(as well as fungi,microbes,and other
insects)fromeatingthe foliageof any particularplant
species.
However,the intensity of caterpillarchallengeto a
SantaRosa plantis directly
relatedto climateand carnivore
components of thehabitat.And thecamivores,excepting
the monophagousspecies,thatfeedon thecaterpillars of
any givenplant speciesare at least in part sustainedby
thatfeedon some otherspeciesof plant.
the caterpillars

THE CARNIVORE'S VIEWPOINT

Santa Rosa has an abundantand species-rich camivore
fauna.However,whethera Santa Rosa caterpillar leads a
riskier
lifethandoes an extratropicalcaterpillar
(e.g.,Jan-
zen 198 lb) has not been studied.Santa Rosa has many
highlypolyphagousvertebrate and invertebratecamivores:
coatis,Nasua narica; white-faced monkeys,Cebuscapu-
cinus;spidermonkeys,Atelesgeoffroyi; collaredpeccanes,
Dicotylestajacu; spinypocketmice,L. salvini (Fig. 4); Ji 4%
grayfoxes,Urocyon cinereoargentus;
9-bandedarmadillos,
Dasypus novemcinctus; squirrelcuckoos,Piaya cayana;
motmots,Momotusmomota,Eumomotasuperciliosa;tro-
gons,Trogonspp.; rufous-naped wrens,Campylorhyncbus

FIGURE 6. AdultEnicospilus parasiticwasp (Ichneumonidae)

lebeau(Satumiidae)
ofa speciesthatparasitizesonlyRothschildia
in Santa Rosa National Park.

128 Janzen

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions
FIGURE 7. Last instargeometridmoth caterpillarin restingpositionon its sole species of host plant, Hymenaeacourbaril
(Leguminosae).The caterpillar is thesame shadeofyellowish-greenas arethestemsand newleaveson whichit feeds.The caterpillar
is the terminalthirdof the "stem" and is holdingtightlyto the stemwithits posteriorprolegsat the pointwherethe petioleof
the secondpair of leafletsleavesthe stem.The caterpillar's
shadow is markedsharplyon the leafletbelow it.

rufinucha; at least another50 speciesof small birdsthat caterpillar escapetraits,anygivenmemberofthecarnivore

gleanand otherwise harvestcaterpillars;
scorpions;spiders; arraywill make use of (is a threatto) onlya verysmall
ants;socialwasps; carabidbeetles,neuropterans; crickets; fractionof the speciesof Santa Rosa caterpillars. To a
pentatomidbugs; reduviidbugs. Thereare also well over trogon,forexample,thereare threebasic "species" of
300 speciesof caterpillar-eating and pupa-eatingtachinid caterpillars: thoseseen and rejected,thoseseen and eaten
fly,ichneumonidwasp, and braconidwasp larvae (e.g., directly, and thoseseen and eaten but requiringspecial
Fig. 5). The latterthreegroupsof insectparasitoidsrange processing.There is a fourth"species" as well, the cat-
fromabsolutelymonophagous[e.g., each of the three erpillarsneverseen at all (e.g., Fig. 7); thisis probably
speciesof Enicospilusichneumonidwasps (Fig. 6) that the most species-rich categoryforall camivores.Within
attackSaturniidaehave onlyone knownspeciesof larval the firstcategoryof caterpillars are coralsnake-patterned
hostat SantaRosa) to attackinga dusterofrelatedspecies caterpillars (Fig. 8) thatare presumablyrejectedbecause
(e.g.,one speciesof BelvosiatachinidflyparasitizesRoths- of genetically based fearor repulsion(e.g., Smith 1975,
childia,Eacles,Citheronia,andhemileucine one 1977). For example,thebrightly
saturniids; ringedlargecaterpillars
speciesofBelvosiaparasitizes therelativelysmallsphingids of Rothschildiaerycina(Satumiidae)(e.g., coverphoto-
thatpupatein thelitter;and one speciesof Belvosiapar- graphof Janzen1983a) and Arsenuraarmida (Satumi-
asitizesthe largersphingidsthatpupate deep below the idae) (Janzen1982) are passed overby squirrelcuckoos,
soil surface). trogons, andothermedium-sized caterpillar-hunting birds,
Most membersof theSantaRosa carnivore faunacan butreadilyeatenbylargelizards,Ctenosaura similis(pers.
potentiallyencounter nearlyall 3140 speciesofcaterpillars obs.). The latterexamplereflects thegeneralphenomenon
on an ecologicaltimescale. However,owingto selective thata caterpillar thatis rejectedbyone speciesofcarnivore
foragingbehaviorby the carnivores,physiologicaland maybe readilyeatenbyanother;forexample,white-faced
morphological incompatibility withpreyand hosts,and monkeysrejecturticating (Fig. 9), but squirrel
caterpillars
EcologyofCosta RicanCaterpillar
Fauna 129

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions
 V

FIGURE 8. (A) Last instarcaterpillar of Diphtherafestiva(Noctuidae); the body is black and white,witha red head capsule
and redposteriorportionof thelastabdominalsegment,and an orangelowerlateralareajustabove theabdominallegs(hostplant:
Waltheriaspp., Sterculiaceae).(B) Last instarcaterpillar
ofNeophaenisrespondens(Noctuidae); it is blackand yellowish-white,
with
a red head and red-orangedorsalbump at the posteriorend (host plant: Petraea volubilis,Verbenaceae).(C) Penultimateinstar
of Lirimirisguatemalensis
caterpillar (Notodontidae);it is black and white,witha red head capsuleand red dorsalbump at the
posteriorend (host plant: Guazuma ulmifolia,Sterculiaceae).All threeof thesespeciesare Mullerianor Batesianmimicsof the
"coral-snakeimage" in a bird's mind,or cdassicalmodels.

cuckooseat themregularily (afterfirst

smashingthespines a Wisconsinwoodloton the same basis as would a tax-
on a treebranch). onomist.A foragingbird in Santa Rosa and Wisconsin
It is clearthatthe camivoreswould not comparethe mightwellcondudethatthe"speciesrichness"ofthetwo
caterpillar faunaof Santa Rosa withthatof,forexample, sitesis the same (four"species" in each), althoughthe
130 Janzen

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions
relativeproportionsof each are different.Likewise,the indirectinteractionin thispart of the ecosystem.A con-
monophagouscarnivores in bothhabitatswould conclude spicuousoutbreakof a speciesof highlyediblecaterpillar
thatthediversitywas identicalin thetwoforests, sinceall may satiatequite generalistcarnivores (birds,monkeys).
the nonpreyand nonhostsdo not directly affectthem. These carnivores will thensearchless forthosespeciesof
The demographicand microgeographic propertiesof (or otherinsects)thatare at low densitiesor
caterpillars
the Santa Rosa caterpillar
fauna can affectcarnivores in on fewspeciesof plantsin the habitat.For example,in
at least threemajor ways. Like the traitsof individual yearsofhighcaterpillar density,thereis conspicuously
less
larvaetouchedon above,theseproperties arenotreflected damage by birdsto the wingsof large saturniidmoths
in thelengthsof faunallistsforthe site. thatrest(hide) in the foliagein the daytimethan is the
case in yearsof low caterpillar
density.
CATERPILLARDENSITY.-The overalldensityand biomass
of caterpillars
at Santa Rosa fluctuates enormously from
THE CATERPILLAR'S VIEWPOINT
yearto year[forexample,thecaterpillar densityin 1977
U(anzen1980) was farhigherthan it has been in the 9 Most Santa Rosa caterpillars occurat such low density
yrsincel.However,overallmeasuresofcaterpillar biomass, relativeto theirfood that thereappearsto be littleop-
speciesrichness,caterpillardensity,etc.,probablymean portunity fordirect,one-on-oneinteractions with other
littlebiologically(i.e., no carnivoreis as polyphagousas overfood. However,caterpillars
caterpillars can and will
are the generalizedsamplingtechniquesassociatedwith in a numberof waysthroughthe me-
interactindirectly
thesemeasures).The degreeto whicha particular carnivore diumoftheresource budgetofa hostplant(Janzen1973)
individualorspeciesis influenced bysuchfluctuations will and throughcarnivores. It is throughsuch processesthat
dependlargelyon whetheritsparticular preyor hostsare perceivesthecaterpillar
a caterpillar faunaof Santa Rosa.
contributors to the fluctuation and whetherothercarni-
voresshareitspreyor hosts.This appliedto polyphagous CARNIVORE SHARING.-As mentionedearlier,an outbreak
armadillosas wellas to monophagousichneumonid wasps. of one or a fewspeciesof caterpillars may draw the at-
tentionofpolyphagouspredators awayfromotherspecies.
SEASONAL CHANGE.-Duringthe year,overallcaterpillar However,if some different and edible species occur in
biomassand densityis closeto zero duringthe 6-mo dry high densitiesforseveralsuccessiveyears,the predator
season (late Decemberto earlyMay), builds up rapidly densitymay then be at its maximumwhen a year of
duringthe first2 mo of the rainyseason (late May to densityoccurs.This will createex-
overalllow caterpillar
mid-July), declinesprecipitously in lateJulyand August, ceptionallyintensehuntingpressureon the already(or
and thenremainsverylow throughtheremaining halfof of otherspecies.On a narrower
usually)scarcecaterpillars
therainyseason(e.g.,Janzen1980, 1987b). This general timescale,the same processoccurswithina singlerainy
patternis mostcloselyfollowedby the leaf-eating cater- season.
pillars.Heinrichand Collins(1983) have notedthatfor- In thesame vein,an outbeakof caterpillars maywell
agingbirdsmaycue on leafdamage as an aid in locating satiatetheparasitoidsthatcan use thatcaterpillarspecies,
caterpillarsin extratropicalforests.In Santa Rosa, the but an outbreakalso generatesa largeabsolutenumber
accumulationof damageto theleavesis veryrapidin the of parasitoids.These parasitoidswill thenmoreseverely
firstmonthof therainyseason,and thecaterpillars move challengesubsequentlow-density caterpillargenerations
a greatdeal fromday to day. During laterpartsof the thanwould be thecase iftheparasitoiddensitywereonly
rainyseason the damage persists,but no caterpillars ac- thatwhichis supportedby a low densityof caterpillars
companyit. The consequenceis thata damaged leafis a (see Janzen 1976, 1983b foranalogousexampleswith
verypoorindicator ofthepresenceofa caterpillar in Santa If an outbreakis quelled by a polyphagous
vertebrates).
Rosa's forests. disease(suchas occurredwithH. lineata in 1979; Janzen
The caterpillars
ofnumerousspeciesofseed-and fruit- 198 la, 1984a), an enormousamountof inoculantis left
miningmicrolepidoptera are most numerousduringthe on foliage,branches,and the ground;redistributed later
dryseason,andwood-boring caterpillars(Cossidae)appear in the year and in subsequentyearsthroughrainflow,
to be presentin moderatenumbersthroughout the year. splash,and wind,the infective stagesmay attenuatethe
However,thesecaterpillars are largelyinaccessibleto car- multispecificeffect
of an outbreakformanygenerations.
nivoresand so do not compensatethe predatorsforthe
absenceof foliage-eaters. MIMICRY.-Modeland mimicry systemsamong caterpil-
larsdependon the componentcaterpillar species,as well
PREDATORSATIATION.-Since different speciesof caterpil- as certaintraitsof the predators.For example,the many
larsappearinlargenumbersin different years,a carnivore's species of spinyyet harmlesscaterpillars(Fig. 9c) are
year-to-year perceptionof caterpillar densitywill depend dependenton a certainbackgrounddensityof urticating
on how polyphagousit is. Thereis, however,a distinctive caterpillars(Fig. 9a) for effectiveness of theirmimetic
EcologyofCosta RicanCaterpillar
Fauna 131

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions
 AA

132JanzenW

132 Janzen

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions
defenses.However,this may not be true (in ecological that a particularenvironmental eventwill favorone or
time)forthe caterpillars thatdepend on theirpossession more of them. In this sense,it mattersto a caterpillar
of a colorpattern(Fig. 8) thatis genetically frightening specieshow manyotherspeciesof caterpillars are in the
to thepredator(e.g., Smith 1975, 1977). habitat,even if theyfeedon otherspeciesof plants.
On the otherhand, a habitatthatis species-poorin
RESOURCE AND DEFENSE BUDGET.-Each time a plant is caterpillarsmay averagea highercaterpillar biomassper
defoliated, theallocationschedules(and quantitiesofhar- speciesthandoesa habitatthatisspecies-rich incaterpillars.
vestableresource)of the plant are changed.On the one Bothhabitatsmaytherefore sustainaboutthesamespecies
hand,suchchangeswill affectotherspeciesof caterpillars richness ofcaterpillar
parasitoids;indeed,parasitoidspecies
thatare usingsubsequentproductssuch as leaves,shoot richnessdoes not attainits highestlevelsin the species-
tips,flowers,fruits,and seeds. Defoliatedtreesin Santa richlowland tropicalforests(Janzen 1981b). The rela-
Rosa commonlyabortor otherwisefail to producetheir tivelylow speciesrichnessof Santa Rosa tachinidsand
floweror fruitcrop in the subsequentsexual cycle;for ichneumonids (M. Wood, J.Gauld, and H. Townes,pers.
example,theRandia spp. defoliatedby A. titan (Janzen comm.,afterexaminingmanyyearsof rearingsand mal-
1985a) bore no fruitsin the rainyseason followingthe aise trapmaterial)suggestthatthe increasein caterpillar
defoliation. The factthatmanySanta Rosa treesdo not speciesrichnessfrom,forexample,Ithaca to Santa Rosa
producea new leaf crop until monthsafterdefoliation (Table 1) hasnotled toa concomitant increaseinparasitoid
mustalso affectthe othermembersof a tree'sherbivore speciesrichness;thatis, whetherit mattersto a particular
load. Forexample,whentheSantaRosa Tabebuiaochracea speciesof sphingidcaterpillar in Santa Rosa to co-occur
lost theirentireleaf crop to the leaf-mining larvae of a with3, 5, or 20 otherspeciesof sphingidcaterpillars will
chrysomelid beetle at the beginningof the 1984 rainy dependprimarily on howthecarnivores viewthisgradient.
season,T. ochraceatreesof all sizesdid not producenew
leavesuntilthe beginningof the 1985 rainyseason-a
yearlater. CONCLUSIONS
On the otherhand, the interaction of Santa Rosa's The emergingimage of the Santa Rosa caterpillar fauna
plantsand caterpillars maywell lack theinduceddefenses differs fromthat of extratropical caterpillarfaunasin a
thatappear to be a major aspectof caterpillar-plant in- numberof waysthatwill probablybe importantin un-
teractions in extratropical I have rearedbatchesof dertanding
forests. how theSanta Rosa caterpillar faunainteracts
monophagouslarvaemanytimeson enclosedor exposed withits hostsand its carnivores. On the otherhand, to
foliagethathas been subjectto repeateddefoliation;in viewthe Santa Rosa caterpillar faunaas representative of
such rearingsthereis no suggestionthat the regrowth the "tropics"would be a seriouserror:forexample,the
foliageis any less suitableas food thanwas the original Santa Rosa caterpillar faunadiffers in manymajorways
cropof leaves.In fact,chemicalchangesfollowingdefol- not hereindiscussedfromthat of otherwetterand/or
iationare not expectedin a systemin which,because of colderhabitatsonlya fewtensof kilometers fromSanta
carnivores, seasons,migratory behavior,etc.,Lepidoptera Rosa.
populationsdo not build up on a particulartreeduring It willbe moreyearsbeforea thoroughimageofSanta
subsequentgenerations. Rosa's caterpillarfauna is available. In the mean-
time,however,a simultaneouslong-termintensivelook
PARASITOID SPECIES RICHNESS.-Becausethereare many at other faunasin thesameand different
tropicalcaterpillar
speciesof caterpillars at Santa Rosa, it is temptingto habitatscould be conductedprofitably by otherpeople.
suggestthatat any one season or year,at least one host It is onlythroughan intensive understanding of a variety
speciesmaybe sufficiently abundantto sustaina somewhat oftropicalhabitatsthatwe willbe able bothto manipulate
polyphagousparasitoid;thatis, the morelifeformsand themand to displaythemin such a manneras to attract
speciesofcaterpillars ina habitat,thegreaterthelikelihood public interestfortheirsurvival.There is a lot more to

FIGURE 9. (A) Penultimateinstarcaterpillar of Automeriszugana (Saturniidae),a severelyurticating that is part of

caterpillar
the model complexfora species-richarrayof mimicsof urticatingcaterpillars. It is greenwith a red head capsule and red and
whitemarkingslaterallyand ventrally(host plant: Annonapurpurea,Annonaceae).(B) Last instarcaterpillarof a megalopygid
moth,a strongly thatis partof the model complexof urticating
caterpillar
urticating It is black withwhitespotsand
caterpillars.
denselycoveredwith long silkyorangehairs (host plant: Enterolobium cyclocarpum, Leguminosae;see Fig. 3C). (C) Last instar
of Emesismandora(Riodinidae). This butterfly
caterpillar caterpillaris brilliantred with black spines and a black head capsule
withwhitemarkings.It is a harmlessmimicof urticating caterpillars.

Ecology of Costa Rican CaterpillarFauna 133

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions
the tropicsthanmacaws in rain foresttreesand lions in de ParquesNacionalesde Costa Rica. The symposiumforwhich
game parks. If we expectpeople to see the beautyand the manuscriptwas preparedwas conceivedby J. Schultzand
supportedby fundsfromtheEntomologicalSocietyofAmerica.
thenwe have
intriguein thesmallerworldof caterpillars, I. Gauld, W. Hallwachs,and A. Pescadorcommentedon the
to learnenoughabout themto tell theirstories. manuscript.Taxonomic assistancewas providedby C. Porter,
H. Townes, I. Gauld, N. Woodley, M. Wood, C. Lemaire,
J.-M. Cadiou, E. Munroe,J. Franclemont, R. Poole,P. DeVries,
ACKNOWLEDGMENTS R. Robbins, D. Ferguson,R. Hodges, J. F. Gates-Clarke,A.
Hayes,W. Haber, P. Thiaucourt,E. Todd, J. Rawlins,A. Wat-
This studywas supportedby NSF BSR 83-07887, BSR 84- son,J. Holloway,J. Miller,S. Miller,D. Wagner,J. Donahue,
00351, BSR 83-08388, and DEB 80-11558, and bytheServicio P. Whalley,J. Heppner,J. Powell,and D. Davis.

LITERATURECITED
DAVIS, D. R., D. H. CLAYTON, D. H. JANZEN, AND A. BROOKE. 1985. NeotropicalTineidae,II. Biologicalnotesand descriptions
of two new mothsphoreticon spinypocketmice in Costa Rica (Lepidoptera:Tineoidea). Proc. Entomol.Soc. Wash. 88:
98-109.
DEVRIES, P. J. 1983. Checklistof butterflies. In D. H. Janzen(Ed.). Costa Rican naturalhistory,pp. 654-678. University of
Chicago Press,Chicago.
DILLON, P. M., S. LOWRIE, AND D. McKEY. 1983. Disarmingthe "evil woman"; petiole constriction by a sphingidlarva
circumvents mechanicaldefensesof its hostplant,Cnidoscolus urens(Euphorbiaceae).Biotropica15: 112-116.
GILBERT, L. E., AND M. C. SINGER. 1975. Butterfly ecology.Annu. Rev. Ecol. Syst.6: 365-397.
HARTSHORN, G. S. 1983. Plants.In D. H. Janzen(Ed.). Costa Rican naturalhistory, pp. 118-157. University of Chicago Press,
Chicago.
HEINRICH, B., AND S. L. COLLINS. 1983. Caterpillarleafdamage, and the game of hide-and-seekwithbirds.Ecology64: 592-
602.
JANZEN, D. H. 1973. Hosts plantsas islands. II. Competitionin evolutionary and contemporary time. Am. Nat. 107: 786-
790.
1976. The depressionof reptilebiomassby largeherbivores. Am. Nat. 110: 371-400.
1980. Heterogeneity of potentialfood abundancefortropicalsmall land birds. In A. Keast and E. S. Morton(Eds.).
Migrantbirdsin the neotropics,pp. 545-552. SmithsonianInstitutionPress,Washington,D.C.
1981a. Patternsof herbivory in a tropicaldeciduousforest.Biotropica13: 271-282.
1981b. The peak in NorthAmericanichneumonidspeciesrichnesslies between380 and 42?N. Ecology62: 532-537.
1982. Guia para la identificacion de mariposasnocturnasde la familiaSaturniidaedel Parque Nacional Santa Rosa,
Guanacaste,Costa Rica. Brenesia19/20: 255-299.
(Ed.). 1983a. Costa Rican naturalhistory.University of Chicago Press,Chicago,816 pp.
1983b. The Pleistocenehuntershad help. Am. Nat. 121: 598-599.
1983c. Larvalbiologyof Ectomyelois muriscis (Pyralidae:Phycitinae),a Costa Rican fruitparasiteof Hymenaeacourbaril
(Leguminosae:Caesalpinioideae).Brenesia21: 387-393.
1984a. Natural historyof Hylesia lineata (Saturniidae:Hemileucinae)in Santa Rosa National Park, Costa Rica. J.
Kans. Entomol.Soc. 57: 490-514.
1984b. Two ways to be a tropicalbig moth: Santa Rosa saturniidsand sphingids.OxfordSurveysin Evolutionary
Biology 1: 85-140.
1985a. A hostplantis morethanits chemistry. Ill. Nat. Hist. Surv. Bull. 33: 141-174.
1985b. Spondiasmombin deprivedin megafauna-free
is culturally forest.J. Trop. Ecol. 1: 131-155.
1986a. GuanacasteNational Park: ecologicaland culturalrestoration. UNED, San Jose,Costa Rica.
1986b. Mice, big mammalsand seeds: it matterswho defecateswhatwhere.In A. Estradaand T. H. Fleming(Eds.).
Frugivoresand seed dispersal,pp. 251-271. Dr. W. JunkPublishers,Dordrecht.
1986c. Lost plants.Oikos 46: 129-131.
1986d. Biogeography of an unexceptional place: whatdetermines thesaturniidand sphingidmothfaunaof SantaRosa
National Park,Costa Rica, and what does it mean to conservation biology?Brenesia25/26: 51-87.
1987a. Managementof habitatfragments in a tropicaldryforest:growth.Ann. Mo. Bot. Gard. In press.
1987b. How mothspass the dryseasonin a Costa Rican dryforest.InsectSci. Appl. In press.
1987c. Insectdiversityof a Costa Rican dryforest:whykeep it and how? Biol. J. Linn. Soc. 30: 343-356.
AND R. LIESNER. 1980. Annotatedcheck-list of plantsof lowlandGuanacasteProvince,Costa Rica, exclusiveof grasses
and non-vascular cryptogams. Brenesia18: 15-90.
, AND P. G. WATERMAN. 1984. A seasonalcensusof phenolics,fibreand alkaloidsin foliageof foresttreesin Costa Rica:
some factorsinfluencing theirdistribution and relationto hostselectionby Sphingidaeand Saturniidae.Biol. J. Linn. Soc.
21: 439-454.
KOPONEN, S., E. M. LAASONEN, AND E. T. LINNALUOTO. 1982. Lepidopteraof InariLapland,Finland.Kevo Notes 6: 1-36.
LEMAIRE, C. 1982. Trois Saturniidaeineditsdu Costa Rica et du Perou (Lepidoptera).Rev. Fr. Entomol.(N.S.) 4: 79-85.
1987. Les SaturniidaeAmericains.3. Ceratocampinae.Museo Nacional de Costa Rica, San Jose,Costa Rica. In press.
MAKINEN, Y., AND P. KALLIO. 1979. The vascularplantsof InariLapland, Finland.Kevo Notes 4: 1-45.
MARQUIS, R. J. 1984. Leaf herbivoresdecreasefitnessof a tropicalplant. Science226: 537-539.

134 Janzen

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions
NiFMELA, P., AND S. NEUVONEN. 1983, Speciesrichness of herbivores
on hosts:how robustarepatterns revealedby analysing
publishedhostplantlists?Ann.Entomol.Fenn.49: 95-99.
OWEN, D. F. 1971. Tropicalbutterflies:theecologyand behaviour ofbutterflies
in thetropicswithspecialreference
to African
species.ClarendonPress,Oxford.
SMITH, S. M. 1975. Innaterecognition of coralsnakepatternbya possibleavianpredator.Science187: 759-760.
1977. Coral-snakepatternrecognitionandstimulus bynaivegreatkiskadees
generalization (Aves:Tyrannidae). Nature
265: 535-536.
VANE-WRIGHT, T. I., AND P. R. ACKERY (Eds.). 1984. The biologyofbutterflies.
AcademicPress,London.

ERRATA
L. P. Brower,B. E. Horner,M. A. Marty,C. M. Moffitt, P. spicilegus,
maniculatus,
andB. Villa-R,"Mice(Peromyscus
and Microtus mexicanus) of overwintering
as predators monarch (Danaus plexippus)in Mexico,"
butterflies
Biotropica17(2): 89-99. Specimens maniculatus
as Peromyscus
identified havebeenreidentified
andP. spicilegus
as Peromyscus J.A. AllenandChapmanandNeotomodon
melanotis aistoniMerriam, Forre-checking
respectively.
we areindebted
theseidentifications to Drs. M. D. Carleton,G. G. Musser,and M. Engstrom.

EcologyofCosta RicanCaterpillar
Fauna 135

This content downloaded from 132.248.172.184 on Wed, 11 Jun 2014 18:51:39 PM

All use subject to JSTOR Terms and Conditions