Behavior and Mimicry in Heliconius Butterflies

	The tropics of South America have lavish exhibits of sympatric butterfly mimicry rings.  Under 
strict coevolutionary (Mullerian) mimicry, this feature is of an enigma: all unpalatable species should 
converge as co-mimics to the same pattern.  If mimicry has evolved in unpalatable species by one-sided 
(Batesian) evolution, however, mimicry rings based on different models could remain distinct.  If mimicry 
rings were also segregated by habitat, a diversity of mimicry rings could be stabilized (Mallet and Gilbert, 
1995). This paper reports correlation between behavior and mimicry of nine unpalatable Heliconius 
species.  It is reported, that co-mimics fly in similar habitats, and non-mimics fly in different habitats with 
much overlap of habitats of other non-mimics (Mallet and Gilbert, 1995).  This is contrary to other reports 
that indicate that there is little difference in flight heights of heliconiine mimicry rings; all species fly from 
ground level to the canopy.  However, co-mimics roost at night in similar habitats and at similar heights 
above the ground, but in different habitats and at different heights from species in other mimicry rings 
(Mallet and Gilbert, 1995).  Heliconius (especially in erato taxonomic groups) are renowned for roosting 
gregariously, and co-mimics roost gregariously with each other more often than with non-mimics (Mallet 
and Gilbert, 1995).  Gregarious roosting is therefore common between species, as well as within species.  
Thus, there are strong links between mimicry and behavioral ecology in Heliconius (Mallet and Gilbert, 
1995).  The paradoxical correlation between nocturnal roosting and visual mimicry is presumably 
explained by bird predation at dusk when roosts are forming, or at dawn before they have disbanded 
(Mallet and Gilbert,1995).  Direct evidence of predation is lacking, but there are high rates of disturbance 
by birds at these times.  These results, together with knowledge of the phylogeny of Heliconius, suggest 
that species from the melpomene-group of Heliconius have radiated to occupy mimetic niches protected by 
model species in the Ithomiinae and the erato-group of Heliconius (Mallet and Gilbert, 1995).  Sympatric 
mimicry rings are maintained because key models fail to converge, while rapidly-evolving, unpalatable 
mimics evolve towards the color pattern of the models.  The maintenance of mimetic diversity would be 
aided by the habitat and behavioral differences between mimicry rings, provided that different predators 
are found in different habitats (Mallet and Gilbert,1995).  

	Tropical butterflies (Lepidotera) that are unpalatable to predators advertise this ability with warning 
colors (Mallet and Gilbert, 1995).  These butterflies frequently belong to mimicry rings (groups of 
unpalatable species, together with some palatable species, that have converged on the same warning color 
pattern) (Mallet and Gilbert, 1995).  Mimicry rings where first explained by Bates (1862), who found that 
presumably unpalatable Ithomiinae, Danainae and Heliconiinae (danaoid and acraeoid Heliconiidae) and 
presumably palatable Dismorphiinae often converged in color pattern (Bates, 1862).  Unprotected species 
would be selected to resemble more common species which are protected from predators by nauseous 
smell or taste.  This is known as Batesian mimicry (Bates, 1862).  Batesian mimicry was used to explain 
mimicry between pairs of species, both of which were presumed unpalatable (Mallet and Gilbert, 1995).  
Bates argued that rare, unpalatable species, such as the silvaniform known as Heliconius numata 
(Heliconiinae), should converge on the patterns of more common or more highly protected model species, 
such as Melinaea spp. (Ithomiinae): a rare, protected species would otherwise be vulnerable to predators 
that had not experienced its color pattern (Mallet and Gilbert, 1995).  Mimicry which involves unilateral 
convergence; as Batesian mimicry does, is usually only applicable to palatable species.  Bates (1862) 
applied unilateral convergence to Lamarckian explanation of mimicry between pairs of unpalatable 
ithomiines where both species are common.  Muller (1879) showed that both unpalatable species of a 
mimetic pair of arbitrary related density could benefit, allowing the potential for coevolutionary mimicry in 
which both species approach an intermediate pattern.  This bi-directional coevolutionary interpretation 
became known as Mullerian mimicry.  Batesian vs. Mullerian mimicry is talked about in terms of benefits 
and costs.  If both mimic and model benefit from mimicry, then the mimicry is Mullerian; if only the mimic 
benefits by deceitful parasitism of the models signal to predators, then the mimicry is Batesian (Wicker, 
1968; Edmunds, 1974; Vane-Wright, 1976; Gilbert, 1983; Turner, 1984).  The exact model of evolution 
is never certain, it is sensible to use definitions of Mullerian vs. Batesian mimicry based on cost, even 
though many so-called Mullerian mimics may have done what Bates originally proposed--to converge 
unidirectionaly on a model (Mallet and Gilbert, 1995).  The existence of bi-directional Mullerian 
convergence is still in doubt because more common species would probably lose protection by mimicking 
rarer species (Turner, 1977,1984).  There is only slender evidence for any coevolution in mimicry 
(Gilbert, 1983).  

	An obvious corollary of Mullerian mimicry (both coevolutionary and unilateral) is that all 
unpalatable species in an area should converge on a single abundant and effective warning pattern.  
Panglossian hypothesis for mimicry can immediately be rejected (Mallet and Gilbert, 1995).  Unpalatable 
butterflies have massively diversified in color pattern, as though adaptive radiation has occurred.  This 
reaches a peak in the Heliconiiae, where mimetic radiation has occurred again and again in different 
lineages (Turner, 1976).  Within a number of mimicry rings in any one locality , the habitats or 
microhabitat of the species within a mimicry ring are expected to be similar for three reasons: (1) A mimic 
should show itself preferentially to predators that have encountered the model.  It should avoid predators 
that have not encountered the model,  (2) Alternatively, if the mimic adapts to a habitat or microhabitat 
already occupied by a common aposematic model, the mimic should converge on that species. (These 
arguments are phrased in unilateral, Batesian terms, but similar coevolutionary arguments can be made.), 
and (3) If color patterns have important effects on thermal ecology or background matching, model and 
mimic could share habitat requirements.  Different mimicry rings are therefore expected to be separated by 
habitat, while co-mimics are expected to share habitats (Malley and Gilbert, 1995).	

	Butterflies of different mimicry rings fly at different heights in the rainforest, but different mimetic 
groups do not fly in different habitats.  The same birds are found at all levels in the forest, and  butterflies 
are unlikely to minimize overlap with predators trained to avoid other mimicry rings (Papageorgis, 1975).  
Mimicry rings are distributed in the canopy in the reverse of the order expected based on thermal grounds: 
dark-colored mimicry rings, which absorb heat faster, are often found higher in the canopy than pale 
mimicry rings, suggesting that thermal ecology is unimportant (Papageorgis, 1975).  An angument is that 
each mimicry ring was camouflaged against its background, and that disruptive pattern was a microhabitat-
limited constraint on the evolution of mimetic patterns (Papageorgis, 1975).  This idea suggests that color 
patterns are under dual selection for camouflage and warning coloration.  Papageorgis rejected purely 
mimetic and thermal ecology hypotheses and ignored the possibility that mate choice might explain the 
origin and maintenance of mimicry rings (Mallet and Gilbert, 1995).  A feature absent from Papageorgis 
work is any discussion of the function of flight which must constrain the butterflies behavior (Malley and 
Gilbert, 1995).  Heliconius search in leaned home ranges for larval food plants for oviposition (Passiflora 
spp., Passifloraceae), for coevolved nectar and pollen sources (Psiguria and Gurania spp., 
Cucurbitaceae), for mates, and for roosting sites (Gilbert, 1975; Tuner, 1983; Mallet, 1986; Murawski and 
Gilbert, 1986).  Heliconius butterflies fly in major habitat types that are similar to those of co-mimics, but 
different from those of non-mimics (Smiley, 1978).  There is often a mimicry ring of small transparent 
ithomiines that fly near the ground in deep forest.  These are not conspicuous, and could possibly have 
evolved transparency for camouflage as Brown (1973) and Papageorgis (1975) suggest.  Casual 
observations suggest that the four mimicry rings involving heliconiines (the tiger mimicry ring, including 
Ithomiinae and Heliconiinaee; the red and blue mimicry rings of the genus Heliconius; and the orange 
mimicry ring of other heliconiines) overlap in flight height far more than reported by Papageorgis (Mallet 
and Gilbert, 1995).  By studying the behavioral ecology of mimetic butterflies using ecological segregated 
by habitat or height the question why there are so many mimicry rings in an area.

	Mimicry and flight height   are an important pat of the is research relative to the forest canopy, and 
methods given by Papageorgis, (1974) recorded the height at which mimetic butterflies entered the forest 
canopy from logged areas.  The flight heights of butterflies entering the forest was impossible to measure 
due to the irregular canopy of the primary forest (Mallet and Gilber, 1995).  

	Mimicry in daytime flight habitat separation has been found for Heliconius mimicry rings in 
Corcovado and elsewhere in Costa Rica (Smileey, 1978). The virtually uniform presence of the yellow 
mimicry ring on its own in the shady understory of the primary forest in Corcovado (Longino, 1984) is 
particularly striking.  The second growth and light gap areas, however, where Heliconius are most 
abundant, are sites of great overlap for all Heliconius mimicry rings, including the yellow ring (Mallet and 
Gilbert, 1995).

	In relation to nocturnal roosting and mimicry, Heliconius are renowned for faithfully returning to 
gregarious roost sites night after night (Tuner, 1971, 1975, 1983).  There is a correlation between 
mimicry, roosting habitat, roosting height, and gregarious roosting with members of other species (Mallet 
and Gilbert, 1995).  The roosting habitat of all Heliconius is on dead twigs of tendrils suspended from 
mats of dense vegetation or from branches of tall trees.  They roost away from living plant tissue, and are 
commonly found in primary forest or in second growth which is very shady.  There are differences 
between mimicry rings, but also differences between species within rings.  H. erato and H. melpomene 
tend to roost in low second growth and secondary forest, whereas H. hecale and H. ismenius prefer to 
roost in taller, secondary forest.  The yellow ring is alone in their roosting in mature forest, paralleling 
their diurnal habits, but they also roost in other habitat (Mallet and Gilbert, 1995).  

	There is a great deal of overlap between the daytime flight heights of Heliconius and other 
heliconiines with different color patterns.  However, mimicry rings do show habitat differences (Smiley 
1978), and this study exhibits strong habitat and height differences between mimicry rings in their 
nocturnal roosting habitats (Mallet and Gilbert, 1995).  Correlations between mimicry, habitat, and 
behavior cannot be dismissed on the grounds of phylogenetic inertia because members of the two major 
taxonomic groups of Heliconius, the erato-group and the melpomene-group (Eltringham, 1916; Brown, 
1981; Brower,1994a), are represented in two and three mimicry rings, respectively (Mallet and Gilbert, 
1995).  This can be compared to the results of the Papageorgis study (1975), which found no difference in 
habitat between Heliconius mimicry rings , but strong segregation in flight height.  These results contrast 
with the results of Mallet and Gilbert (1995) study.  These differences are hard to explain, although some 
discrepancy may be due to differences in the species composition between Peru and Costa Rica.  This does 
not agree with the impressions; Heliconius overlap in flight heights though out the neotropics (Mallet and 
Gilbert, 1995).  Brown (1988), found much more overlap in flight height between Heliconius mimicry 
rings than Papageorgis, but Brown (1988) did not have quantitative information.  All of the groups were 
overlapping between 2-5 m, and had overlapping in the canopy.  The red ring, which flies low (1-5m), 
was an exception.  Differences in flight height exist between non-heliconiine mimicry rings and ithomiine 
mimicry rings are stratified in flight heights (Mallet and Gilbert, 1995).  

	The nocturnal roostmates, roosting habitats and roosting heights of Heliconius seem from Mallet 
and Gilbert (1995) data to be more strongly influenced by diurnally-visible mimicry than daytime 
behavior.  Why should a visual pattern that is probably ignored by nocturnal predators affect nocturnal 
roosting?  A possibility is that during the middle of the day mimicry is unimportant, because poikilthermic 
butterflies can easly dodge predators (Mallet and Gilbert, 1995).  At sunrise or sunset Heliconius fly slow 
due to the fact that the butterflies are not operating at optimal temperatures for flight.  In the early morning 
H. erato leave their roosts and spead their wings and bask in the sun nearby; also late in the day, or after a 
cloudy spell this behavior is exhibited (Mallet and Gilbert, 1995).  Heliconius may be vulnerable to birds 
that search foliage and chase insects that are disturbed.  Butterflies are less quick and less likely to escape 
from gleaners at this time of day (Chai, 1996).  Bird predation at dusk and dawn is high, because roosting 
groups are forming or disbanding at is time.  Attacks by sallying birds such as Jacamars (Galbula spp.) 
and tyrannid flycatchers may be more common because of slower escape in cooler conditions (Mallet and 
Gilbert, 1995).  Neotropical birds forage at different heights in the canopy and in different habitats (Chai 
1996).  Heliconius reduces the diversity of predators by roosting in narrow habitats and height range, 
which may provide selection for habitat and microhabitat divergence between Heliconius in different 
mimicry rings, or may even reduce convergence between rings of different habitats (Mallet and Gilbert, 
1995).  

	Behavioral mimicry hypotheses would be more strongly supported if there were evidence for 
predation on Heliconius, especially near roosts.  Birds frequently flew near roosts of H. erato when 
butterflies were still present in the morning.  On average, each H. erato was disturbed by potential predator 
every third morning (Mallet, 1986).  While birds never seen to attack the butterflies, frugivores or gleaners 
mainly specialize on slow moving insects living in rolled leaves; the high rate of disturbance suggests that 
there is at least a high potential for predation (Mallet and Gilbert, 1995).  Evidence for birds attacks shows 
up in the form of Jacamar and other beakmarks on the wings of Heliconius, and from various estimates of 
selection on mimetic pattern (Benson, 1972).

	An explanation for differences in roosting height and habitat is that mimetic patterns differ in their 
ability to absorb solar radiation. This is also found in non-mimetic butterflies (Douglas and Grula, 1978).  
The arrangement of mimicry rings in roosting height seems to support this hypothesis; species possessing 
the darker color patterns (red and yellow mimicry rings) roost lower, than paler tiger species.  This is 
contrasts with Papageorgis (1975), which found the darker blue [similar to Corcovado yellow ring] and 
red mimicry rings flew higher than the tiger rings.  Heliconius roost in shade and must fly to sunny areas 
to bask in the sun.  The thermal ecology of these species is not effected by roost heights (Mallet and 
Gilbert, 1995).  

	Disruptive coloration may have nothing to do with daytime flight environment of Heliconius 
(Papageorgis, 1975).  Predators may find it hard to follow Heliconius flying against dappled background; 
however, this would be true of most objects moving against a contrasting background (Mallet and Gilbert, 
1995).  Heliconius are among the most visible butterflies in flight.  The color pattern minimizes any 
disruptive effect the pattern may have.  Colored spots or bands of red, yellow, or orange are neatly 
separated from each other and from the edge of the wing by borders of black pigment, and some are 
overlaid with iridescent blue (Mallet and Gilbert, 1995).  Patterns that break up outlines are termed 
disruptive patterns, but these bright markings of Heliconius are almost always clearly demarcated by dark 
edges (Mallet and Gilbert 1995).  Red rings are low and tiger rings are high, which falsifies the notion that 
fleck size and mimetic pattern are in some way correlated (Papageorgis, 1975).  Camouflage is also 
unlikely attribute of Heliconius, as it would be hard to design a more conspicuous patterns.  These 
comments also apply to heliconiines in the orange mimicry rings.  Orange heliconiines in flight are visible 
from far away, and their colors stand out from the rainforest background (Mallet and Gilbert, 1995).  

	Evolution of mimicry with mate choice suggests that mimicry may diversify because there is an 
upper limit to the sexual confusion caused by many species within a mimicry ring, and female-limitation in 
Batesian mimicry has almost certainly evolved because of sexually selected constraints on male patterns 
(Turner, 1978).  Heliconius are found in Mullerian rings (Turner, 1976), and speciation is often 
accompanied by mimetic switch.  Supporting this idea Brown and Benson (1974) found that orange and 
red rags attract the orange tiger pattern H. numata.

	If mimicry between unpalatable species evolves unilaterally, the development of mimicry rings 
would depend only on a few key species, or the model.  Except for H. doris which is based, the 
Heliconius in this study can be divided into two taxonomic groups: melpomene-group and erato-group 
(Brown, 1981).  The members of erato-group are closely related, but the eggs, larvae and pupae are 
diverse and can be gregarious or solitary (Brown,1981).  Hybrids between sympatric species have never 
been found in nature.  The roosts of these species are usually gregarious, and individuals rarely roost next 
to co-mimics (Mallet and Gilbert 1995).

	Mallet and Gilbert (1995) uncovered behaviors that reduce overlap between mimicry rings and 
improve the probability that mimics are found at the same heights and microhabitats.  Co-mimics are likely 
to have similar habits because the diversifying melpomene-group species evolved unilaterally towards 
ithomiines or erate-group Heliconius.  Evolution occurred first in habitat, followed by alterations in color 
pattern and behavior (Mallet and Gilbert, 1995).


References

Bates HW. 1862. Contributions to an insect fauna of the Amazon valley. Lepidoptera: Heliconidae. 
Transactions of the Linnean Society of London 23: 495-566.

Benson WW. 1972. Nature Selection for Mullerian mimicry in Heliconius erato in Costa Rica. Science 
176: 936-939.

Benson WW. 1978. Resource partitioning in passion vine butterflies. Evolution 32: 493-518.

Brower AVZ. 1994a. Phylogeny of Heliconius butterflies inferred from mitochondrial DNA sequences 
(Lepidoptera: Nymphalinae). Molecular Phylogenetics and Evolution 3: 159-174.

Brown KS. 1973. A Portfolio of Neotropical Lepidopterology. Rio de Janeiro, Brazil: privately 
published.

Brown KS. 1981. The biology of Heliconius and related genera. Annual Review of Entomology 26: 427-
456.

Brown KS. 1988. Mimicry, aposematism and crypsis in neotropical Lepidoptera: the importance of dual 
signals. Bulletin de la Societe Zoologique de France 113: 83-101.

Chai P. 1996. Butterfly visual characteristics and ontogeny of responses to butterflies by a specialized 
tropical bird. Biological Journal of the Linnean Society 59: 37-67.

Douglas MM, Grula JW. 1978. Thermoregulatory adaptations allowing ecological range expansion by the 
Pierid butterfly Nathalis iole Boisduval. Evolution 32: 776-783.

Edmunds M. 1974. Defense in Animals. Harlow, Essex: Longmans.

Eltringham H. 1916. On specific and mimetic relationships in the genus Heliconius. Transactions of the 
Entomological Society of London 1916: 101-148.

Gilbert LE. 1972. Pollen feeding and reproductive biology of Heliconius butterflies. Proceedings of the 
National Academy of Sciences, USA 69: 1403-1407.

Gilbert LE. 1983. Coevolution and mimicry. In: Futuyma DJ, Slatkin M, eds. Coevolution. Sunderland, 
Mass.: Sinauer Associates, 263-281.

Mallet J., Gilbert LE. 1995. Why are there so many mimicry rings? correlations between habitat, 
behaviour and mimicry in Heliconius butterflies. Biological Journal of the Linnean Society 55: 159-180.

Muller F. 1879. Ituna and Thyridia; a remarkable case of mimicry in butterflies. Transactions of the 
Entomological Society of London 1879: xx-xxix.

Papageorgis C. 1975. Mimicry in neotropical butterflies. American Scientist 63: 522-532.

Smiley JT. 1978. The host plant ecology of Heliconius butterflies in Northeastern Costa Rica. Ph.D. 
Dissertation, University of Texas at Austin.

Turner JRG. 1976. Adaptive radiation and convergence in subdivisions of the butterfly genus Heliconius 
(Lepidoptera: Nymphalidae). Zoological Journal of the Linnean Society 58: 297-308.

Turner JRG. 1977. Butterfly mimicry-genetical evolution of an adaptation. Evolutionary Biology 10: 163-
206.

Turner JRG. 1978. Why male butterflies are non-mimetic: natural selection, sexual selection, group 
selection, modification and sieving. Biological Journal of the Linnean Society 10: 385-432.

Turner JRG. 1984. Mimicry: the palatability spectrum and its consequences. In: Vane-Wright RI, Ackery 
PR, eds. The Biology of Butterflies (Symposia of the Royal Entomological Society of London, 11). 
London: Academic Press, 141-161.

Vane-Wright RI. 1976. A unified classification of mimetic resemblances. Biological Journal of the 
Linnean Society 8: 25-56.

Wickler W. 1968. Mimicry in Plants and Animals. New York: McGraw Hill.