Annals of the Entomological Society of America, XX(X), 2020, 1–9

doi: 10.1093/aesa/saz073
Research Research

Wing Morphology and Body Design in Opsiphanes and

Caligo Butterflies Match the Demands of Male Mating

Downloaded from https://academic.oup.com/aesa/advance-article-abstract/doi/10.1093/aesa/saz073/5698063 by guest on 14 February 2020

Displays (Lepidoptera: Nymphalidae)
Carla M. Penz1 and Susan F. Williams
Department of Biological Sciences, University of New Orleans, 2000 Lakeshore Drive, New Orleans, LA 70148 and 1Corresponding
author, e-mail: [email protected]

Subject Editor: Donald Thomas

Received 24 July 2019; Editorial decision 19 December 2019

Abstract
There is noticeable variation in male mate-seeking behavior among species of Brassolini butterflies. Males of
Opsiphanes Doubleday and Caligo Hübner (Lepidoptera, Nymphalidae) species perform crepuscular displays
along forest edges and in light gaps. While male Opsiphanes perform aerial displays, Caligo males perch and
wait for receptive females. A comparison of five species of each genus suggests that male display behavior is
associated with, and has likely influenced the evolution of their wing attributes and body design. Opsiphanes
males have higher wing aspect ratio and more distal centroid position than congeneric females, suggesting
that the energetic demands of aerial displays led to sexual dimorphism in wing morphology. In contrast, male
and female Caligo generally showed similar wing morphology, which possibly results from the lower energy
expenditure of perching behavior when compared with active flight. Likely due to a genetic correlation be-
tween sexes, female Opsiphanes and Caligo are more similar in wing morphology and body design to their
congeneric males than to each other. Based on our analyses, we make predictions about reproductive behavior
for four species in which male mating displays are unknown.

Key words:  aspect ratio, centroid, flight, lekking, patrolling

The ability to fly is regarded as an influential factor for insect taxo- survival-related activities like finding food and evading predators,
nomic diversification (Dudley 2000, Grimaldi and Engel 2005). Due selection on flight performance is expected to be similar between the
to its evident role in locomotion, flight is integral to a broad range sexes. In contrast, reproductive activities require sex-specific flight
of activities such as finding nutritional resources, escaping from behaviors, and selection could potentially lead to wing shape di-
predators, and locating potential mates. Insects underwent remark- morphism by acting differently on males and females (Srygley 2001,
able morphological diversification in evolutionary time, including DeVries et al. 2010). Regardless of sex, flight behavior is influenced
a multitude of wing adaptations. Familiar examples include elytra by multiple attributes such as wing aerodynamic properties, body
in beetles, halteres in flies, and scaled wings in moths and butter- design (i.e., allometric proportions such as relative thorax size and
flies. The latter illustrates particularly well how insect wings have abdomen length), and the way these structures function together
acquired various functions beyond flight. Butterfly wing colors play (Marden and Chai 1991). For example, flight speed appears to be
a role in thermoregulation (e.g., Kingsolver 1987) and are used as positively correlated with wing loading (body mass/total wing area;
signals for male–female communication (e.g., Silberglied 1988) and Dudley 2000). This leads to the prediction that fast flying butterflies
predator–prey interactions (e.g., Bates 1862). Finally, various wing would have large thoraces and small wings, and from a functional
attributes are known to influence flight performance, and the evo- standpoint, a high wingbeat frequency (Bartholomew and Casey
lution of butterfly flight behavior and associated wing morphology 1978, Dudley 2000).
has also been the focus of various investigations (e.g., DeVries et al. Morphological properties of the wings also influence butterfly
2010). flight performance. For instance, distally produced wings have a
Like other flying insects, butterflies employ flight to engage in higher aspect ratio than their shorter counterparts. Given that the
activities related to survivorship and reproduction. During dispersal distal portion of the wing generates greater force per unit of area, a
flights or while searching for nutritional resources, butterflies can high aspect ratio therefore increases lift while promoting energy effi-
be exposed to predator attacks. Therefore, within the context of ciency (Dudley 2000; see also Marden 1987, Marden and Chai 1991,

© The Author(s) 2020. Published by Oxford University Press on behalf of Entomological Society of America. 1
All rights reserved. For permissions, please e-mail: [email protected].
2 Annals of the Entomological Society of America, 2020, Vol. XX, No. XX

DeVries et al. 2010). The first moment of wing area, or centroid, also and also that members of these genera will differ in body design
varies with wing shape and is correlated with aspect ratio (Dudley (Opsiphanes having smaller forewings than Caligo proportionately
2000). Experimental studies show that a distally placed wing centroid to the size of their thoraces). As our analyses uncovered significant
yields greater acceleration at takeoff (Berwaerts et al. 2002, 2006). differences in wing morphology within Caligo, we also explore and
There is considerable variation in male butterfly mating behavior, propose hypotheses that might explain such variation. To integrate
ranging from active search to displaying in aggregations in specific morphology and natural history, we compiled published information
locations (see Wiklund 2003 for a review). Two behaviors are of on Opsiphanes and Caligo mating behavior and also report unpub-
interest here: 1) males that engage in active flight to search for, or lished personal observations.
become visible to, females and 2) males that perch alone or in leks.

Downloaded from https://academic.oup.com/aesa/advance-article-abstract/doi/10.1093/aesa/saz073/5698063 by guest on 14 February 2020

It is reasonable to infer that these behaviors incur different energetic
costs, and that selection operating on male performance favors dif- Material and Methods
ferent attributes of wing and body morphology to match each type Species and Specimens Sampled
of behavior. Wickman (1992) suggested that males that use active
We examined 10 species of the focal genera (a sample is illustrated in
flight for patrolling are selected for endurance, while perching males
Fig. 1): Opsiphanes boisduvallii Doubleday, 1849; Opsiphanes cassina
benefit from high takeoff speed and prompt approach. If patrolling
C.  Felder and R.  Felder, 1862; Opsiphanes invirae (Hübner, 1808);
males spend significant time flying, they should possess the neces-
Opsiphanes quiteria (Stoll, 1780); and Opsiphanes sallei Doubleday,
sary muscle mass and evolve energy-saving wing adaptations. In
1849 and Caligo atreus (Kollar, 1850); Caligo beltrao (Illiger, 1801);
males that perch in groups, however, takeoff speed might determine
Caligo idomeneus (Linnaeus, 1758); Caligo illioneus (Cramer, 1775);
whether a particular male gains (or loses) a courtship opportunity.
and Caligo martia (Godart, 1824). For comparison, we also examined
As such, increased lift force that expedites the response to an ap-
Brassolis astyra Godart, 1824, Brassolis sophorae (Linnaeus, 1758),
proaching female would be advantageous.
Mielkella singularis (Weymer, 1907), and Narope cyllastros Doubleday,
Brassolini butterflies are members of a fruit-feeding guild that
1849 (Fig. 1). Specimens were obtained from the American Museum
forms an important component of Neotropical butterfly communi-
of Natural History, Florida Museum of Natural History (University of
ties (DeVries et al. 2012, Fordyce and DeVries 2016). Adults of most
Florida), Milwaukee Public Museum, Phil DeVries Personal Collection,
Brassolini are crepuscular and, depending on the species, mating and
and the United States National Museum (Smithsonian Institution).
oviposition occur either at dawn, dusk, or both (Fruhstorfer 1910,
Collection localities are given in Supp Table S1.
DeVries 1987). As ambient temperature is lower during crepuscular
hours, Opsiphanes and Caligo butterflies shiver to elevate the thor-
acic temperature in preparation for reproductive behaviors (Srygley Images and Variables
1994), and other crepuscular species likely do so as well. Mating Digital images of specimens were taken in dorsal and ventral views
displays vary across the tribe; Opsiphanes males use fast flight to alongside a paper grid divided in 5 mm squares. We used specimens
perform aerial displays (Anton Fassl in Fruhstorfer 1910, Srygley with at least one intact forewing and hind wing, except for cases of
1994), whereas Caligo males perch in leks, occasionally engaging in minor damage where the wing outline could be easily determined.
male–male interactions (Freitas et al. 1997, Srygley and Penz 1999). Calculations of aspect ratio and wing centroid position were based on
Although most Brassolini species are found in the forest under- composite images of the coupled wing pair. Because it is wide and flex-
story (e.g., Caligo), adults of some taxa inhabit the canopy (e.g., ible, the hind wing anal margin can dry in different positions during
Opsiphanes; Fordyce and DeVries 2016 and references therein). specimen preparation. Prior to photographing, by eye we selected spe-
A study of fruit-feeding nymphalid butterflies by Graça et al. (2017) cimens that had the anal margin extended in a similar way.
suggested that stout-bodied species tend to associate with forest To compare aspect ratio and centroid, we followed the same
canopy, whereas those that inhabit the understory are generally methods as in DeVries et al. (2010). Composite images were assem-
more slender. Of particular interest here is that these authors exam- bled in Adobe Photoshop by overlapping the wing pair and aligning
ined three species of Opsiphanes that were similarly stout but dif- the hind wing costal margin to forewing vein CuA2. This was done
fered in vertical habitat preference (two were associated with the to emulate natural wing overlap of the aerodynamically active wing
canopy, one with the understory). Their study did not sample Caligo, surface (Betts and Wootton 1988, DeVries et al. 2010) and to stand-
the largest-bodied Brassolini, and known to inhabit the forest under- ardize the wing coupling position across all taxa. Adobe Illustrator
story (Fordyce and DeVries 2016). was used to overlay 14 concentric lines radially equidistant from one
Male mate-seeking behavior and flight demands differ between another onto the composite image of the wing pair, resulting in 15
male Opsiphanes and Caligo, but are expected to be comparable concentric sectors, which were cut and weighed with a Mettler Toledo
in females of both genera. Here, we investigate potential effects of analytical scale (d = 0.001 g). ImageJ software (http://rsb.info.nih.gov/
male mating behavior on the evolution of wing morphology and nih-imagej) was used to measure forewing length (base to apex) and
body design by addressing the following questions: 1) Is there sexual area of the wing pair. These measurements were used to calculate as-
dimorphism in wing aspect ratio and centroid in Opsiphanes and pect ratio and wing centroid with equations in Ellington (1984).
Caligo? 2) Do male Opsiphanes and Caligo differ in wing aspect ratio Fresh specimens were not available to calculate wing loading
and centroid? 3) Do female Opsiphanes and Caligo differ in wing (body mass/total wing area), so we used ‘thorax width3/total fore-
aspect ratio and centroid? 4)  Accounting for sex, do Opsiphanes wing area’ as a proxy—pWL hereafter. The area of the left forewing
and Caligo differ in body design? The foundation for these questions was measured (unless it was damaged) and this value was doubled
and inferences comes from research on butterfly mating behavior to yield total forewing area. Using forewing area only allowed us
(Wickman 1992, Freitas et al. 1997, Srygley and Penz 1999), flight to measure specimens with damaged hind wings, thus increasing
kinematics (Dudley 2000 and references therein), and flight perform- sample size. It also eliminated error due to differential folding of the
ance (e.g., Berwaerts et al. 2002). Given their differences in mating hind wing anal margin. We are aware, however, that the hind wings
display, we predict that male Opsiphanes will have higher wing as- of Opsiphanes are proportionately smaller than those of Caligo (es-
pect ratio and more proximally placed centroid than male Caligo, pecially in males, see Fig.  1; CMP, unpublished data), and that in
Annals of the Entomological Society of America, 2020, Vol. XX, No. XX 3

Downloaded from https://academic.oup.com/aesa/advance-article-abstract/doi/10.1093/aesa/saz073/5698063 by guest on 14 February 2020

Fig. 1.  Sample of male (M) and female (F) butterflies examined. (a) Opsiphanes cassina (both specimens from Ecuador, Sucumbios, Garza Cocha); (b) Opsiphanes
invirae (both specimens from Ecuador, Sucumbios, Garza Cocha); (c) Caligo illioneus (male from Ecuador, Sucumbios, Garza Cocha; female from Costa Rica,
Heredia, Finca La Selva); (d) Caligo idomeneus (male from Ecuador, Sucumbios, Garza Cocha; female from Peru); (e) Brassolis astyra (both specimens from
Brazil, Rio de Janeiro); (f) Mielkella singularis (both specimens from Mexico, Chiapas); (g) Narope cyllastros (male: no data, female: Brazil, Santa Catarina, Nova
Teutônia).

both genera males and females of the same species seem to differ in Thorax size is a good predictor of flight muscle mass (Dudley 2000),
relative hind wing area [also the case in Pierella helvina (Hewitson, and although thorax width3 underestimates total thoracic volume, it
1860) (Nymphalidae, Satyrinae); Stylman et al. 2019]. An explora- is likely associated with the volume of flight muscle. As our measure-
tory test with total wing area of one male and one female of each ments account for an important component of body mass (thoracic
genus produced comparable results as those reported below. The muscle) and the area of the wing that more directly powers flight
forewing/hind wing allometry between sexes and genera will be ex- (Dudley 2000, Jantzen and Eisner 2008), we are confident that there
plored in a future study. is a correspondence between pWL and wing loading.
Thorax width was measured with calipers between the forewing
bases of pinned specimens because the scale vestiture did not allow
for precise measurements from photographs. Although we used Analyses
specimens where thoraces did not seem compressed, deformation Although phylogenetic analyses by Penz et  al. (2013) showed that
that might have occurred during storage and preparation possibly Opsiphanes and Caligo are monophyletic, their species-level re-
explains the broad range of values among conspecific specimens. lationships were poorly resolved and we could not account for
4 Annals of the Entomological Society of America, 2020, Vol. XX, No. XX

phylogeny. To increase sample size for statistical comparisons, we and centroid. Male Opsiphanes were predicted to have higher wing
pooled species data and performed the analyses at the genus level. aspect ratio and more proximally placed centroid than male Caligo due
As an inspection of species means indicated that one Caligo spe- to differences in mate seeking behavior, and our analyses supported this
cies noticeably differed from others (Table 1), we compared it with prediction [Table  2 (E)]. Moreover, female Opsiphanes also showed
the pooled data for the four species that fit the same general trend. significantly higher aspect ratio and a more proximally positioned cen-
We used two-tailed t tests for unequal variances implemented in troid than female Caligo [Table 2 (F)]. Female Opsiphanes are more
Microsoft Excel to compare wing aspect ratio, centroid position, similar to their conspecific males than to female Caligo (Fig. 2).
and pWL (our proxy for wing loading) between sexes and genera.
Body Design

Downloaded from https://academic.oup.com/aesa/advance-article-abstract/doi/10.1093/aesa/saz073/5698063 by guest on 14 February 2020

Male and female mean pWL for Opsiphanes (19 males, 7 females)
Results
and Caligo (9 males, 6 females) are shown in Fig. 3a and Table 1.
Wing Aspect Ratio and Centroid Position We found no significant difference in body design between sexes for
We first asked whether species of Opsiphanes and Caligo show either genera [Table 2 (A, B)], but detected opposite trends between
sexual dimorphism in wing aspect ratio and centroid. When species genera. Opsiphanes males showed a higher pWL average than con-
were pooled, we found that male Opsiphanes had higher aspect ratio generic females (Fig.  3a), suggesting that males have proportion-
than congeneric females [Table 2 (A), see also Fig. 2], consistent with ately smaller forewings than females (in our sample, thorax width
their intense flight activity during mating displays. Centroid position did not differ between sexes, results not shown). In contrast, Caligo
also differed between sexes [Table  2 (A)], being more proximally males had larger forewings than females (Fig.  3a; in our sample,
positioned in male than female Opsiphanes (Table  1). Given the thorax width did not differ between sexes, results not shown). We
wing morphology variation among Caligo species (Fig. 2), our ana- then asked whether Opsiphanes and Caligo differ in body design,
lyses of this genus were done in two parts. We first pooled all species accounting for sex. Male Opsiphanes had significantly higher pWL
and found that Caligo males and females differed in aspect ratio and values than male Caligo, and the same was found for the female sex
wing centroid [Table  2 (B)]. A  second set of analyses showed that [Table 2 (E, F)]. To provide a visual example of the difference in male
C. idomeneus was driving this pattern. Male and female C. idome- body design between genera, we scaled the outlines of two individ-
neus differed significantly in both aspect ratio and wing centroid uals to identical thorax size. Male Opsiphanes have proportionately
[Table 2 (C), see also Figs. 1 and 2]. Nonetheless, when C. atreus, smaller forewing area than male Caligo (Fig. 3b).
C.  beltrao, C.  illioneus, and C.  martia were pooled, males and fe-
males did not differ in aspect ratio nor wing centroid [Table 2 (D)],
Discussion
confirming that C.  idomeneus was the only Caligo species in our
data set showing sexual dimorphism in wing morphology. Male Mating Displays in Opsiphanes and Caligo
Our second and third questions focus on whether male and female Species of Opsiphanes typically perform aerial displays, and
Opsiphanes differ from male and female Caligo in wing aspect ratio some form mating aggregations (Srygley 1994, C.P., personal

Table 1.   Mean values for wing aspect ratio and centroid position for species of Opsiphanes, Caligo, and three additional genera

Aspect ratio Centroid position

Species (n) Male Female Male Female

Focal taxa
  Opsiphanes cassina (7M, 4F) 2.648 2.445 0.514 0.535
  Opsiphanes sallei (3M, 1F) 2.645 2.563 0.513 0.527
  Opsiphanes quiteria (5M, 4F) 2.609 2.419 0.507 0.537
  Opsiphanes invirae (5M, 4F) 2.595 2.501 0.522 0.537
  Opsiphanes boisduvallii (2M, 2F) 2.555 2.424 0.526 0.552
  Opsiphanes—all species pooled
means ± SD (22M, 15F) 2.618 ± 0.07 2.458 ± 0.08 0.515 ± 0.01 0.538 ± 0.01
  Caligo idomeneus (5M, 5F) 2.624 2.434 0.521 0.535
  Caligo atreus (5M, 5F) 2.499 2.359 0.534 0.543
  Caligo illioneus (5M, 5F) 2.371 2.356 0.542 0.548
  Caligo martia (4M, 3F) 2.366 2.323 0.542 0.553
  Caligo beltrao (4M, 5F) 2.176 2.195 0.557 0.553
  Caligo—all species pooled
means ± SD (23M, 23F) 2.419 ± 0.17 2.332 ± 0.11 0.538 ± 0.01 0.546 ± 0.01
  Caligo (atreus, illioneus, martia, beltrao)
means ± SD (18M, 18F) 2.362 ± 0.14 2.304 ± 0.1 0.543 ± 0.01 0.549 ± 0.01
Additional taxa
  Brassolis sophorae (1M, 1F) 3.022 2.932 0.506 0.518
  Brassolis astyra (2M, 2F) 2.924 2.81 0.576 0.538
  Mielkella singularis (2M, 2F) 2.703 2.534 0.523 0.55
  Narope cyllastros (2M, 2F) 2.541 2.335 0.512 0.544

Species are listed within genera based on highest to lowest male aspect ratio. Sample sizes are in parentheses after the species names and headings of pooled
species means. When data were pooled for multiple but not all species in a genus, their names appear in parentheses. Means and Standard Deviations (SD) are
given for pooled data.
Annals of the Entomological Society of America, 2020, Vol. XX, No. XX 5

Table 2.  Results of t-test comparisons in the same site as C.  illioneus and C.  oileus (C Penz pers.) and
displayed both at dawn and dusk in an insectary (Srygley 1994).
t-test comparisons P-value
Fruhstorfer (1910) described aggregations of male C. beltrao (Illiger,
(A) Opsiphanes male vs. female 1801) perching sequentially along forest edges in Brazil, noting that
  Aspect Ratio <0.001 it was active at both dawn and dusk. Finally, again in Brazil, males of
  Wing Centroid <0.001 C. idomeneus (Fig. 1d) were observed to patrol back-and-forth over
 pWL 0.124 n.s. a stretch of 10 to 50 m before perching in their lek arena, but some
(B) Caligo male vs. female males did not settle and apparently displayed exclusively on the wing
  Aspect Ratio 0.042 (Freitas et al. 1997).

Downloaded from https://academic.oup.com/aesa/advance-article-abstract/doi/10.1093/aesa/saz073/5698063 by guest on 14 February 2020

  Wing Centroid 0.044
As they are active during cooler hours of the day, crepuscular
 pWL 0.107 n.s.
Opsiphanes and Caligo shiver to increase thoracic temperature, and
(C) C. idomeneus male vs. female
  Aspect Ratio <0.001 during displays male thoracic temperature can increase 15.4–17.5°C
  Wing Centroid 0.02 above ambient in Opsiphanes and 3.8–10.7°C in Caligo (Srygley
(D) C. (atreus, beltrao, illioneus, martia) male vs. female 1994). As body temperature affects flight performance (Chai and
  Aspect Ratio 0.161 n.s. Srygley 1990), the inability to bask in the sun implies that crepus-
  Wing Centroid 0.098 n.s. cular butterflies allocate a considerable amount of energy to repro-
(E) male Opsiphanes vs. male Caligo ductive activities. The high thermal excess in Opsiphanes suggests
  Aspect Ratio <0.001 that their intense flight activity during displays generates consider-
  Wing centroid <0.001 able metabolic heat, which is not the case with perching Caligo.
 pWL <0.001
(F) female Opsiphanes vs. female Caligo
  Aspect Ratio <0.001 Male Mating Displays and Sexual Dimorphism
  Wing centroid 0.019 Butterfly flight behavior includes morphological, behavioral, and
 pWL 0.007 physiological components. Wing aspect ratio and centroid position,
muscle mass, wing loading, wingbeat frequency and amplitude, the
When data were pooled for multiple species, their names appear in paren- proportion of time spent flapping or gliding, and body temperature all
theses. The abbreviation pWL corresponds to our proxy for wing loading (see contribute to the particular way an insect flies (Dudley 1990, Marden
Methods). and Chai 1991, Srygley 1994, Almbro and Kullberg 2008, Cespedes
et al. 2015). This study assessed some of these variables and suggest an
observations). In Panama, Opsiphanes cassina males (Fig. 1a) were evolutionary association between male mating behavior, wing morph-
observed at dusk displaying in groups of four to six individuals. ology, and body design in Opsiphanes and Caligo butterflies.
They flew in broad circles ca. 1.5–2 m above the ground, landing for Our analyses point to a consistent association between male
only a few seconds between bouts of fast flight, and did not appear mating behavior and sexual dimorphism in all five species of
antagonistic towards conspecifics. Much like skippers (Hesperiidae), Opsiphanes, but not Caligo. Male Opsiphanes reproductive activ-
they flapped their wings at a high and seemingly regular wingbeat ities likely incur higher energy expenditure than do female’s, which
frequency (C.P., personal observations). Unfortunately courtship could explain their divergence in wing morphology. As it could be
and mating were not observed. These field observations agree with expected, males have higher aspect ratio and more proximally posi-
accounts by Anton Fassl (in Fruhstorfer 1910), who noted that tioned centroid than females (Fig.  2). Males seem to have propor-
O. camena Staudinger, 1886 males flew together at dusk, performing tionately more flight muscle per unit of forewing area than females
short bouts of fast flight. Finally, in the Brazilian cerrado (Brasilia, (Fig. 3a), which corresponds to the demands of their aerial displays.
DF), a single Opsiphanes sp. male (likely O.  invirae, Fig.  1b) was This allows us to predict that male Opsiphanes might employ a
observed at dusk patrolling along a road, flying ca. 2 m above the higher wingbeat frequency than congeneric females, which could be
ground at an estimated land speed of 15 km/h as measured by fol- tested in future studies. As male Caligo spend more time perching
lowing alongside it with a motor vehicle (C.P. and P. DeVries, per- than flying during reproductive displays, finding that Caligo males
sonal observations). and females were generally similar in wing aspect ratio and cen-
Male Caligo are known to form mating aggregations, but un- troid was not surprising. Caligo beltrao and C.  illioneus males
like the fast-flying Opsiphanes they perch in leks (Fruhstorfer 1910, are known to favor perching (Fruhstorfer 1910, Penz and Srygley
Freitas et al. 1997, Srygley and Penz 1999). At dusk in Panama, male 1999), and the latter patrols only when population density is high
C. illioneus (Fig. 1c) and C. oileus C. Felder and R. Felder, 1861 ar- and perching spots seem limited at the lek site (C.P., personal ob-
rived simultaneously at a specific lek site and perched ca. 1 m from servations). Caligo idomeneus was exceptional by showing sexual
the ground. Such males did not fly unless disturbed and remained differences in aspect ratio and centroid that were even higher than
perched in a position that made them visible to potential mates those of Opsiphanes (Fig.  2), and the males of this species have
(Srygley and Penz 1999). Perching C. illioneus were nonetheless terri- been observed to use both perching and active flight at a display
torial and responded to visual stimuli, such as conspecific or congen- site (Freitas et al. 1997). Finding wing morphology dimorphism in
eric males and conspecific females, when the approaching butterfly some Caligo species but not others indicates that flight demands of
reached a distance of ca. 1 m (Srygley and Penz 1999). Upon arrival reproductive activities likely vary among species, suggesting that
at the lek site, receptive females flew towards a male, which would speciation and behavioral diversification are interconnected in this
immediately take flight and begin courtship display. Sometimes fe- large genus of brassoline butterflies. Mating behavior is known to
males were pursued by more than one male (C.P., personal observa- show plasticity in Lepidoptera, with males changing their behavior
tions). Individual males often returned to the same perching location in response to density of conspecifics (C. illioneus, C.P., personal ob-
each day, and some remained on the lek site for 3 mo (Srygley and servations), ambient temperature (Wickman 1985a), or age (Turner
Penz 1999). Caligo eurilochus (Cramer, 1775)  perched at dawn 2015). Further field observations of male mating behavior in species
6 Annals of the Entomological Society of America, 2020, Vol. XX, No. XX

Downloaded from https://academic.oup.com/aesa/advance-article-abstract/doi/10.1093/aesa/saz073/5698063 by guest on 14 February 2020

Fig. 2.  Plots of aspect ratio and wing centroid for Opsiphanes and Caligo species (range, vertical line; mean, horizontal bar). Abbreviations: Opsiphanes, cas—
cassina, sal—sallei, qui—quiteria, inv—invirae, boi—boisduvallii; Caligo, ido—idomeneus, atr—atreus, ill—illioneus, mar—martia, and bel—beltrao.

of Caligo coupled with an investigation of wing morphology and aerial displays in Opsiphanes and perching in Caligo. To enhance
body design might also shed light on geographical variation within performance and promote energy conservation during flapping
broadly distributed species in this genus. flight, selection likely favors wing attributes that increase lift force
Although reproductive behaviors for most brassoline species have and reduce drag (Betts and Wootton 1988, Wickman 1992). Male
not been documented, our findings can be used to make some predic- Opsiphanes had higher wing aspect ratio than male Caligo [Table 2
tions. In Brassolis astyra Godart, 1824, Brassolis sophorae, Mielkella (E)], fitting this expectation and providing support for our initial
singularis (Weymer, 1907), and Narope cyllastros Doubleday, 1849 prediction (see Introduction). Flight speed seems to be an important
(Fig. 1), differences in wing aspect ratio and centroid between sexes component of Opsiphanes male display behavior, but a correlation
are comparable to what we found in Opsiphanes (Fig. 4), suggesting between speed and aspect ratio has not been found (Dudley 1990,
that males of these species use active flight to find receptive females. 2000). Instead, flight speed is known to be associated with thoracic
Cleare and Squire (1934) noted that male and female B. sophorae mass and wingbeat frequency (Bartholomew and Casey 1978, Betts
rest during the day and start activity at sunset (about 17:30 h) when and Wootton 1988) and high wing loading (Chai and Srygley 1990).
mating occurs, and that oviposition takes place after dark. Field ob- Our results are consistent with these findings inasmuch as male
servations suggest that female B.  sophorae moved very little from Opsiphanes have proportionately smaller wings compared with
where they pupated, and that individuals have ca. 100 to 300 mature perching Caligo (Fig.  3b). Although analyses of European butter-
eggs at eclosion, which are laid as a cluster (Cleare and Squire 1934, flies by Wickman (1992) suggested that species with patrolling males
Carvalho et al. 1998). Although these observations on Brassolis con- had smaller thorax/body mass ratios, lower wing loading, and lower
form to our morphology-based prediction of males employing active wing aspect ratio than species with perching males, the taxa included
flight to find receptive females, nothing is known about the mating in Wickman’s study varied in a continuum and not all fit the pro-
behaviors of Narope or Mielkella singularis. Lastly, differences in posed general trend—something we found in Opsiphanes.
wing aspect ratio and centroid between sexes of C. atreus approach In contrast, if perching males benefit from a prompt response
those of C. idomeneus (Fig. 4), leading to the prediction that male to an approaching female (Wickman 1992), selection should favor
C. atreus employ patrolling behavior. increased muscle mass (Marden 1987) and a distally placed wing
centroid (Berwaerts et  al. 2002) to yield greater acceleration at
takeoff. Indeed, in Caligo the wing centroid is generally positioned
Wing Morphology and Body Design more distally than in Opsiphanes (except for C. idomeneus; Fig. 2,
Our analyses uncovered wing and body morphology attributes Table  1). Moreover, Caligo males seem to have proportionately
that seem consonant with male mate-seeking behaviors: active larger wings than females [Fig. 3a, Table 2 (E)], which would also
Annals of the Entomological Society of America, 2020, Vol. XX, No. XX 7

Downloaded from https://academic.oup.com/aesa/advance-article-abstract/doi/10.1093/aesa/saz073/5698063 by guest on 14 February 2020

Fig. 3. (  a) Comparisons of thoracic width3/total forewing area (pWL) as proxy for wing loading (species pooled). Range (vertical line) and mean (horizontal
bar) for male (M) and female (F) Opsiphanes and Caligo. Note that high numbers in the Y-axis indicate small forewing size relative to thorax size, whereas low
numbers indicate large forewing size relative to thorax size. (b) Outline of two selected males scaled for thorax width to highlight genus-level differences in
body design.

allow for more lift at takeoff. Thus, Caligo males are equipped for a in Caligo. Observations on live butterflies will be required to com-
fast response to approaching females. Similar behaviors are known pare thoracic and abdominal mass between sexes and taxa.
in other butterfly groups (Baker 1972, Dowes 1975, Lederhouse
1982, Wickman 1985b) and are consistent with the demands to Body Design and Habitat Use
reach high speed upon takeoff. To confirm our interpretations of Reproductive behavior is one of the multiple life history components
wing morphology attributes and body design in male Opsiphanes that could affect the evolution of wing morphology and body design.
and Caligo, field studies are required to compare wingbeat frequen- Based on canopy and understory samples of nymphalids in four sub-
cies, flight and takeoff speeds, and wing loading for species in both families, Graça et al. (2017) found a positive interaction between the
genera. ratio wing length2/thoracic volume and vertical stratification; stout-
bodied species showed high canopy fidelity, and slender-bodied taxa
Flight Demands in Females were more prevalent in the understory. They found no interaction
The flight demands of female reproductive activities differ from between thoracic volume and canopy fidelity, indicating that both
males. After mating, females search for larval hosts, and in the case strata included species with small to large body mass. Three species
of Opsiphanes and Caligo, they use a locally comparable range of of Opsiphanes were sampled by Graça et al. (2017), and although the
plant species for oviposition (Arecaceae, Maranthaceae, Musaceae; wing length2/thoracic volume was similar among them, O. cassina and
see Penz et al. 1999 for comprehensive host plant records) and man- O. invirae were sampled mostly in the canopy, and O. quiteria in the
euver around large plants to select suitable leaves. Abdominal mass understory. Caligo was not represented in their study, but Fordyce and
varies during an individual female’s lifetime due to egg maturation DeVries (2016) found Caligo species to be confined to the understory.
and oviposition, and they must compensate for this variation in mass Measurements of thoracic width in our samples showed an
during flight (see Srygley 2001). Female Opsiphanes lay eggs singly overlapping range for male Opsiphanes (4.9–7.4  mm) and Caligo
while Caligo generally oviposits in multiples or in clusters (DeVries (6.2–7.5  mm), but wing sizes markedly differed between genera
1987), thus making it likely that the egg loads and amount of time (Figs.  1 and 3b). This indicates a strong difference in body design
searching for oviposition sites differ between genera. Probably due to between genera [see pWL comparisons in Fig.  3a and Table  2 (E,
a genetic correlation between sexes, Opsiphanes and Caligo females F)], with Opsiphanes having a putatively higher wing loading than
were more similar to congeneric males than to each other (Table 1, Caligo (by proxy). It is possible that life in the canopy and aerial
Fig. 2, various comparisons in Table 2). Nonetheless, having propor- male mating displays jointly contributed to the evolution of body
tionately larger forewings compared with thorax size (Fig. 3a) could design in Opsiphanes butterflies. A  similar reasoning could be ap-
be interpreted as an advantage when carrying large egg loads such as plied to understory Caligo, where males of some species are known
8 Annals of the Entomological Society of America, 2020, Vol. XX, No. XX

Joseph Beasley for help with data collecting. For comments on the manu-
script, we thank Phil DeVries, Logan Crees, and two anonymous reviewers.
S.F.W.  was funded by the College of Sciences Undergraduate Research
Program, University of New Orleans.

References Cited
Almbro, M., and C. Kullberg. 2008. Impaired escape flight ability in butter-
flies due to low flight muscle ratio prior to hibernation. J. Exp. Biol. 211:

Downloaded from https://academic.oup.com/aesa/advance-article-abstract/doi/10.1093/aesa/saz073/5698063 by guest on 14 February 2020

24–28.
Baker, R. R. 1972. Territorial behaviour of the nymphalid butterflies, Aglais
urticae (L.) and Inachis io (L.). J. Anim. Ecol. 41: 453–469.
Bartholomew, G. A., and T. M. Casey. 1978. Oxygen consumption of moths
during rest, pre-flight warm-up, and flight in relation to body size and
wing morphology. J. Exp. Biol. 76: 11–25.
Bates,  H.  W. 1862. Contributions to an insect fauna of the amazon valley.
Lepidoptera: Heliconidae. T. Linn. Soc. London. 23: 495–566.
Berwaerts, K., H. Van Dyck, and P. Aerts. 2002. Does flight morphology re-
late to flight performance? An experimental test with the butterfly Pararge
aegeria. Funct. Ecol. 16: 484–491.
Berwaerts,  K., P.  Aerts, and H.  Van  Dyck. 2006. On the sex-specific mech-
anisms of butterfly flight: flight performance relative to flight morph-
ology, wing kinematics, and sex in Pararge aegeria. Biol. J. Linn. Soc. 89:
675–687.
Betts,  C.  R., and R.  J.  Wootton. 1988. Wing shape and flight behaviour in
butterflies (Lepidoptera: Papilionoidea and Hesperioidea): a preliminary
Fig. 4. Plot of the difference in mean male and female aspect ratio (AR).
Higher values indicate greater degree of sexual dimorphism (see Table  1 analysis. J. Exp. Biol. 138: 271–288.
for mean values). Vertical lines connect species of the same genus. Carvalho,  M.  C., P.  C.  Queiroz, and A.  Ruszczyk. 1998. Protandry and fe-
Abbreviations: Opsiphanes, cas—cassina, qui—quiteria, boi—boisduvallii, male size-fecundity variation in the tropical butterfly Brassolis sophorae.
inv—invirae, sal—sallei; Caligo, ido—idomeneus, atr—atreus, mar—martia, Oecologia. 116: 98–102.
ill—illioneus, bel—beltrao; other taxa, cyl—Narope cyllastros, sin—Mielkella Cespedes,  A., C.  M.  Penz, and P.  J.  DeVries. 2015. Cruising the rain forest
singularis, ast—Brassolis astyra, and sop—Brassolis sophorae. floor: butterfly wing shape evolution and gliding in ground effect. J. Anim.
Ecol. 84: 808–816.
Chai, P., and R. B. Srygley. 1990. Predation and the flight, morphology, and
to perch and wait for approaching females. Nonetheless, we cannot
temperature of neotropical rain-forest butterflies. Am. Nat. 135: 748–765.
separate the effects of habitat use and male mating behavior on Cleare,  L.  D., and F.  A.  Squire. 1934. The coconut caterpillar, Brassolis
wing morphology and body design. To do so, future studies should sophorae L. (Lep., Brassolidae) in British Guyana. Agricult. J. Br. Guyana.
examine natural history, behavior, and morphology within a phylo- 5: 166–199.
genetic context and include a larger sample of taxa within Brassolini. DeVries,  P.  J. 1987. The butterflies of Costa Rica and their natural his-
tory. Papilionidae, Pieridae, Nymphalidae. Princeton University Press,
Princeton.
Concluding Remarks DeVries, P. J., C. M. Penz, and R. I. Hill. 2010. Vertical distribution, flight be-
haviour and evolution of wing morphology in Morpho butterflies. J. Anim.
We propose that male mating behavior influenced the evolution of
Ecol. 79: 1077–1085.
wing and body morphology in Opsiphanes and Caligo butterflies.
DeVries,  P.  J., L.  G.  Alexander, I.  A.  Chacon, and J.  A.  Fordyce. 2012.
The evolution of sexual dimorphism in wing morphology and body Similarity and difference among rainforest fruit-feeding butterfly commu-
design in Opsiphanes likely resulted from the demands of male aerial nities in Central and South America. J. Anim. Ecol. 81: 472–482.
mating displays. Although high wing loading seems associated with Dowes,  P. 1975. Territorial behaviour in Heodes virgaureae L.  (Lep.,
canopy fidelity in Opsiphanes, it is also a requirement for fast flight Lycaenidae) with particular reference to visual stimuli. Norw. J. Entomol.
used during male displays. In contrast, in four out of five Caligo 22: 143–154.
species examined here, males and females have similar performance- Dudley, R. 1990. Biomechanics of flight in neotropical butterflies: morphomet-
related wing morphology attributes, which allow males to execute rics and kinematics. J. Exp. Biol. 150: 37–53.
Dudley,  R. 2000. The Biomechanics of Insect Flight: Form, Function,
a fast response to an approaching female. Field observations and
Evolution. Princeton University Press, Princeton.
morphological studies on other Brassolini taxa can contribute to the
Ellington,  C.  P. 1984. The aerodynamics of hovering insect flight. II.
understanding of the evolution and diversification of mating behav-
Morphological parameters. Philos. T. Roy. Soc. B. 305: 17–40.
iors in this group of butterflies. Fordyce, J. A., and P. J. DeVries. 2016. A tale of two communities: Neotropical
butterfly assemblages show higher beta diversity in the canopy compared
to the understory. Oecologia. 181: 235–243.
Supplementary Data Freitas, A. V. L., W. W. Benson, O. J. Marini–Filho, and R. M. Carvalho. 1997.
Supplementary data are available at Annals of the Entomological Society of America Territoriality by the dawn’s early light: the Neotropical butterfly Caligo
online. idomeneus (Nymphalidae: Brassolinae). J. Res. Lep. 34: 14–20.
Fruhstorfer,  H. 1910. Family: Brassolidae, pp 285–332. In A.  Seitz ed., Die
Gross-schmetterlinge der erde, Volume 5. Alfred Kernen, Stuttgard.
Acknowledgments Graça,  M.  B., P.  A.  C.  L.  Pequeno, E.  Franklin, and J.  W.  Morais. 2017.
For specimen loans, we thank Phil DeVries (personal collection), Milwaukee Coevolution between flight morphology, vertical stratification and sexual
Public Museum, Florida Museum of Natural History, The United States dimorphism: what can we learn from tropical butterflies? J. Evol. Biol. 30:
National Museum, and The American Museum of Natural History. We thank 1862–1871.
Annals of the Entomological Society of America, 2020, Vol. XX, No. XX 9

Grimaldi,  D., and M.  S.  Engel. 2005. Evolution of the Insects. Cambridge Srygley,  R.  B. 1994. Shivering and its cost during reproductive behavior in
University Press, Cambridge. Neotropical owl butterflies, Caligo and Opsiphanes (Nymphalidae:
Jantzen, B., and T. Eisner. 2008. Hindwings are unnecessary for flight but es- Brassolinae). Anim. Behav. 47: 23–32.
sential for execution of normal evasive flight in Lepidoptera. Proc. Natl. Srygley,  R.  B. 2001. Sexual differences in tailwind drift compensation in
Acad. Sci. U. S. A. 105: 16636–16640. Phoebis sennae butterflies (Lepidoptera: Pieridae) migrating over seas.
Kingsolver, J. G. 1987. Evolution and coadaptation of thermoregulatory be- Behav. Ecol. 12: 607–611.
havior and wing pigmentation pattern in pierid butterflies. Evolution. 41: Srygley, R. B., and C. M. Penz. 1999. The lek mating system in Neotropical
472–490. owl butterflies: Caligo illioneus and C. oileus (Lepidoptera, Brassolinae).
Lederhouse,  R.  C. 1982. Territorial defense and lek behavior of the black J. Insect Behav. 12: 81–103.

Downloaded from https://academic.oup.com/aesa/advance-article-abstract/doi/10.1093/aesa/saz073/5698063 by guest on 14 February 2020

swallowtail butterfly, Papilio polyxenes. Behav. Ecol. Sociobiol. 10: Stylman, M.,  Penz, C. & DeVries P.. 2019. Large hind wings enhance gliding
109–118. performance in ground effect in a Neotropical butterfly (Lepidoptera:
Marden, J. H. 1987. Maximum lift production during takeoff in flying ani- Nymphalidae). Annals of the Entomological Society of America113:
mals. J. Exp. Biol. 130: 235–258. 15-22.
Marden, J. H., and P. Chai. 1991. Aerial predation and butterfly design: how Turner, J. R. 2015. The flexible lek: Phymatopus hecta, the gold swift, dem-
palatability, mimicry, and the need for evasive flight constrain mass alloca- onstrates the evolution of leking and male swarming via a hotspot
tion. Am. Nat. 138: 15–36. (Lepidoptera: Hepialidae). Biol. J. Linn. Soc. 114: 184–201.
Penz, C. M., A. Aiello, and R. B. Srygley. 1999. Descriptions of Early Stages of Wickman, P. C. 1985a. The influence of temperature on the territorial and mate
Caligo illioneus oberon Butler, and C. idomeneus (L.) From the Lowland locating behaviour of the small heath butterfly, Coenonympha pamphilus
Rainforest of Panama (Nymphalidae, Brassolinae), With Remarks on (L.) (Lepidoptera: Satyridae). Behav. Ecol. Sociobiol. 16: 233–238.
Larval Food Plants for the Subfamily. J. Lep. Soc. 53: 142–152. Wickman,  P.  O. 1985b. Territorial defence and mating success in males of
Penz,  C.  M., A.  V.  L.  Freitas, L.  A.  Kaminski, M.  M.  Casagrande, and the small heath butterfly, Coenonympha pamphilus (L.) (Lepidoptera:
P. J. Devries. 2013. Adult and early‐stage characters of Brassolini contain Satyridae). Anim. Behav. 33: 1162–1168.
conflicting phylogenetic signal (Lepidoptera, Nymphalidae). Syst. Ent. 38: Wickman,  P.  O. 1992. Sexual selection and butterfly design-a comparative
316–333. study. Evolution. 46: 1525–1536.
Silberglied,  R. 1988. Visual communication and sexual selection in butter- Wiklund, C. 2003. Sexual selection and the evolution of butterfly mating sys-
flies, pp 207–223. In R. I. Wright-Vane, P. R. Ackery eds., The Biology of tems, pp 67–90. In C. L. Boggs, W. B. Watt, P. R. Ehrlich eds., Butterflies,
Butterflies. Princeton University Press, Princeton. Ecology and Evolution Taking Flight. University of Chicago Press, Chicago.