H

|
1
l‘
1

i 1

i

i
$

H11!

Ii
‘
!

WIN

1‘

l
I

 

TH _

This is to certify that the
thesis entitled

THE INFLUENCE OF FEMALE WING COLOR
VARIATION ON MALE MATE CHOICE IN TIGER
SWALLOWTAIL BUTTERFLIES (PAPILIO SPP.)

presented by

Matthew L. Aardema

has been accepted towards fulﬁllment
of the requirements for the

degree in Entomology

 

 

lﬂ/M %%

’ Majormofessor’sSi , ature
7 glee; 5% gm)

Date

MSU is an Afﬁrmative Action/Equal Opportunity Employer

 

LIBRARY
Michigan State
University

 

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KilProlecc8-Pres/ClRC/DateDueiindd

THE INFLUENCE OF FEMALE WING COLOR VARIATION ON MALE MATE
CHOICE IN TIGER SWALLOWTAIL BUTTERFLIES (PAPILIO SPP.)

By

Matthew L. Aardema

A THESIS

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
MASTER OF SCIENCE
Entomology

Ecology, Evolutionary Biology and Behavior

2010

ABSTRACT

THE INFLUENCE OF FEMALE WING COLOR VARIATION ON MALE MATE
CHOICE IN TIGER SWALLOWTAIL BUTI'ERFLIES (PAP/LIO SPP.)

By
Matthew L. Aardema

Variation both within and between populations is an important component
of the speciation process. This variation can be found in all biological arenas and
includes ecological variation, developmental variation and variation in sexual
characteristics. In Lepidoptera, this last source of variation has had an important
role in producing divergence between populations. It has been shown that males
of the eastern tiger swallowtail, Papilio glaucus, and the Canadian tiger
swallowtail, F’. canadensis, both prefer to mate with female P. glaucus in two-
choice ﬁeld experiments. This suggests that there is a species difference
between females of the two species that is recognizable to males. One possible
difference is the amount of blue scaling on the dorsal wing surface of females. P.
glaucus females have relatively more area of their wing surface covered by this
blue coloration than do P. canadensis females. However, my experiments to
determine the importance this blue coloration has for males selecting a mate
suggests that its inﬂuence is minimal and hence the difference in blue coloration
seen between P. glaucus and P. canadensis is unlikely to explain differences in
male mate choice. Other explanations for asymmetrical male mate choice may
be differences in body size and/or yellow coloration in females of these two
species. Although quantiﬁed, these were not experimentally tested to determine

their importance in male mate selection.

To G-ma

ACKNOWLEDGEMENTS

I would like to thank Rodrigo Mercader and Gabe Ording for discussions
and overall good advice, Shane Empie and John Acorn for advice on UV
photography, Jay Sobel for assistance with spectrophotometery, Ange Kosnick
and Brittany Sonke for assistance rearing Ohio females for tethering
experiments, Howard Romack and Bob Howtz for rearing Vermont and
Pennsylvania females for tethering experiments, Sean Ryan for ﬁeld assistance,
Matthew Douglas for original mentorship, and Amanda Lorenz for being there.

I would also like to thank my guidance committee, Anthony Cognato, Doug
Schemske, and Jeff Conner.

Lastly, and most importantly, I would like to thank my advisor Mark Scriber.

This work was funded by MAES project #01644, NSF DEB# 0716683 and
NSF DEB# 0918879, a Scriber Scholar award, a Hutson research grant, the
Jeffery Boettcher Travel Fund, and MSU College of Natural Science

undergraduate research support.

TABLE OF CONTENTS

List of Tables ......................................................................................... vi
List of Figures ..................................................................................... viii
Chapter 1:

Introduction ......................................................................................... 01
The Tiger Swallowtail System .................................................................. 02
Male Mate Choice in Tiger Swallowtails ..................................................... 11
Research Objectives ............................................................................. 13
Chapter 2:

Variation in wing coloration between Papilio canadensis and P. glaucus
populations

Introduction ......................................................................................... 14
Materials and Methods .......................................................................... 19
Results ............................................................................................... 24
Discussion .......................................................................................... 29
CHAPTER 3:

Courtship and mating response of male Papilio spp. to variation in female-limited
wing coloration

Introduction ......................................................................................... 39
Materials and Methods .......................................................................... 40
Resutts ............................................................................................... 47
Discussion .......................................................................................... 51
CHAPTER 4,

Possible alternative explanations for asymmetrical male mating preferences in
Papilio spp.

Introduction ......................................................................................... 56
Materials and Methods .......................................................................... 59
Results ............................................................................................... 61
Discussion .......................................................................................... 69
CHAPTER 5;

Summary and Conclusions ..................................................................... 71
Appendix 1: Record of Deposition of Voucher Specimens ............................. 74
Appendix 1.1: Voucher Specimen Data ................................................... 75
LITERATURE CITED ............................................................................ 76

LIST OF TABLES

Table 1.1. Summary of morphological, ecological and genetic differences
between P. glaucus and P. canadensis .................................................... 05

Table 2.1. Results from analysis of covariance (ANCOVA) for a comparison
between female P. canadensis and yellow morph P. glaucus for the total area of
the wings and the area of blue coloration with species as a cofactor. The data
was |0910 transformed to homogenize variance between the two species. The
additive model appears to have the greatest predictive value and suggests that
the two species increase in blue area in similar fashion with increased overall
size (statistically equivalent slopes), but that P. glaucus females have more blue
on an absolute scale. Bolded numbers represent the values for the preferred
model in the three different analyses (Adj. R2, AIC and BIC). See text for more
details ................................................................................................ 31

Table 3.1. Number of two—choice pairings for each population examined. When
male preferences were examined over muftiple years for the same population,
results are combined. No difference was signiﬁcant (see text for values) ......... 49

Table 3.2. Number of one-choice pairings for each population examined. When
male preferences were examined over multiple years for the same population,
results are combined. No difference was signiﬁcant (see text for values) ......... 51

Table 4.1. Summary statistics for the relative relationship between forewing
length (mm) and abdomen weight (g) in female P. glaucus and P. canadensis...62

Table 4.2. Results from analysis of covariance (ANCOVA) for a comparison
between P. canadensis and P glaucus females for wing length and abdomen
weight with species as a cofactor. The simple model which does not take into
account species appears to have the greatest predictive value and suggests that
the two butterﬂies do not differ signiﬁcantly in their basic allometric relationship
between wing length and abdomen weight. Bolded numbers represent the
values for the preferred model in the three different analyses (Adj. R2, AIC and
BIC). See text for more details ................................................................ 63

vi

Table 4.3a. Summary statistics for hue value in female P. glaucus and P.
canadensis ......................................................................................... 66

Table 4.3b. Summary statistics for chroma value in female P glaucus and P
canadensis ......................................................................................... 66

Table 4.4a. Summary statistics for hue value in populations of female P
canadensis (Alaska and Michigan) and P. glaucus (Ohio and Florida) .............. 67

Table 4.4b. Summary statistics for chroma value in populations of female P.
canadensis (Alaska and Michigan) and P. glaucus (Ohio and Florida) .............. 67

Table 4.5. Signiﬁcance (p) values from Tukey’s HSD for all pairwise comparisons
between populations in the yellow color analysis between P. canadensis (Alaska
and Michigan), and P. glaucus (Ohio and Florida). Values on the lower left side
of the ﬁgure represent chroma comparisons and values on the upper right side of
the ﬁgure represent hue comparisons. Bolded values in the dark gray boxes are
signiﬁcant (p50.05). See text for more details ............................................ 68

vii

LIST OF FIGURES

Figure 1.1. Distribution map of P. canadensis, P glaucus, and areas of
hybridization between them in North America ............................................. 03

Figure 1.2. a.) Dark morph, P. glaucus female; believed to mimic the distasteful
pipe-vine swallowtail, Battus philenor b.) Battus philenor male 0.) Yellow morph,
P. glaucus female d.) P. glaucus male ....................................................... 06

Figure 2.1. Locations of sample populations for blue color analysis. Gray circles
represent P. canadensis populations and black circles represent P. glaucus
populations. a.) Fairbanks-North Star Co., Alaska b.) Charlevoix 00., Michigan
c.) Bennigton Co., Vermont d.) Jackson and Gallia 005., Ohio e.) Habersham
and Rabun Cos., Georgia f.) Levy Co., Florida ........................................... 21

Figure 2.2. Process of wing analysis to determine total wing area and area of
blue coloration. a.) First, one forewing and one hindwing were photographed.
This was done for both dorsal and ventral sides b.) The image was made binary
and the total area was determined. A ruler in the photograph was used to
determine scale. c.) Lastly, all coloration except the blue was removed and the
area of this remaining coloration was determined. See text for more detail. . . ...22

Figure 2.3. Mean relative blue area 2!: SD for all species/sex/color morph
combinations. Pc : P. canadensis, Pg : P. glaucus, YW : yellow morph females
and DK : dark morph females. For each combination, the light grey bar shows
the mean blue coloration on the dorsal wing surface and the dark grey bar shows
the mean blue coloration on the ventral wing surface. Bars with the same letter
are not signiﬁcantly different at the p<0.05 level (pairwise t test with Holm’s
correction) .......................................................................................... 26

Figure 2.4. Mean relative blue area 1 SD for all female P. canadensis (Po) and
yellow morph P glaucus (Pg). For each combination, the light grey bar shows
the mean blue coloration on the dorsal wing surface and the dark grey bar shows
the mean blue coloration on the ventral wing surface. Bars with the same letter
are not signiﬁcantly different at the p<0.05 level (painrvise t test with Holm’s
correction) .......................................................................................... 27

Figure 2.5. Mean relative blue area 1: SD for all male P. canadensis (Pc) and P.
glaucus (Pg). For each combination, the light grey bar shows the mean blue
coloration on the dorsal wing surface and the dark grey bar shows the mean blue
coloration on the ventral wing surface. Bars with the same letter are not
signiﬁcantly different at the p<0.05 level (pairwise t test with Holm’s correction).
......................................................................................................... 28

viii

Figure 2.6. Log transformed comparison of absolute total wing area and area of
blue coloration for P. canadensis and yellow morph, P. glaucus females. Area
values are in cmz. Regression lines are drawn using the additive model (See
table 2.10). The slopes are not signiﬁcantly different. See text for more details.
........................... 32

Figure 2.7. An example of my photographic examination of ultraviolet coloration
in tiger swallowtail butterﬂies. a.) The visual image of the dorsal side of a yellow
morph (left) and dark morph (right), female P. glaucus. b.) This image was taking
using a UV bypass ﬁlter. The wings displayed are the same as those in ‘a.’ Areas
that reﬂected UV light are white and areas that absorbed UV light are black. See
text for more details .............................................................................. 34

Figure 2.8. An example of the reﬂectance spectrum patterns of blue, yellow and
black coloration from one yellow morph, P glaucus female. The other three
individuals assessed in this manner had similar spectral patterns. This ﬁgure
shows a clear UV signature (300-400 nm) for blue coloration, and virtually no UV
reﬂectance for yellow and black coloration. See text for more details .............. 35

Figure 3.1. Three examples of female P. glaucus dorsal colors showing the
”spectral reﬂectance signature of blue coloration, black coloration and blue
coloration covered by black ink from a sharpie brand marker. UV coloration is
that which occurs between 300 and 400 nm. Black ink effectively conceals both
the UV and visual elements of the blue coloration ....................................... 45

Figure 4.1. Comparison of wing length (mm) and abdominal weight (g) in P.
canadensis and yellow morph, P. glaucus females. Regression lines are drawn
using the simple model (See table 4.2). The slopes are not signiﬁcantly different.
See text for more details ........................................................................ 64

CHAPTER 1:
INTRODUCTION

It has been argued that variation within and between populations is one of
the most interesting, yet overlooked aspects of biological diversity (Mayr, 2005).
Subtle differences among individuals within and between populations may lead to
remarkable divergences and biological radiations (Schluter, 2000). For example,
it has been suggested that differences in beak width and depth helped to
generate the diversity of Darwin’s ﬁnches on the Galapagos Islands (Grant,
1986), morphological variation and the inﬂuence this has on pollinator complexes
lead to the diversiﬁcation of orchids (T remblay et al., 2005), and variation in teeth
shape lead to ungulate mammalian diversiﬁcation (Jemvall et al., 1996).

Among most lepidopteran groups important ecological (Ayres and Scriber,
1994; Bossart and Scriber, 1995, Mallet and Joron, 1999), developmental
(Beldade and Brakeﬁeld, 2002), and sexual variation (Ellers and Boggs, 2003) is
common within and between extant populations and such variation likely played
a large role in the historical diversiﬁcation of this clade (Boggs et al., 2003 and
chapters within). Several factors contribute to the high degree of population
variation observed in butterﬂies including a complex life history (Beck and
Fiedler, 2009) and rapid generation times (Thomas et al., 2010).

Variation in traits related to mating and reproduction has been especially
important in butterﬂy speciation (Chamberlin et al., 2009). In Lepidoptera, the
male is the homogametic sex and the sex chromosomes are designated 2 [=X]

and W [=Y] respectively. This sex determination system allows for the generation

of a relatively large amount of sexual dimorphism in many species, possibly
because genes that generate ‘preference’ for particular sexual traits can be
maintained on the W chromosome in females, without the disadvantages of also
being found in males that have the potentially costly trait (Hastings, 1994). The
presence of such sexual dimorphism (speciﬁcally variation in secondary sexual
characteristics and preferences for these characteristics) can lead to differential
selection and assortative mating (Naisbit et al., 2001) when subsets of males and
females preferentially mate with one another at higher frequencies than they
mate with other individuals of the population. This process can lead to the

formation of new species.

The Tiger Swallowtail System

Within the swallowtail family, Papilionidae, there are between 475 to 563 species
worldwide (Scriber, 1995). Of these, the tiger swallowtail clade comprises
between 5 and 8 species in North America. In the eastern United States and
Canada, two species occur with a parapatric distribution (Scriber, 1996a). The
Canadian tiger swallowtail, Papilio canadensis, is found from central Alaska,
throughout much of Canada, south into northern Minnesota, VVrsconsin and
Michigan, and east into the New England States (Fig. 1; Scriber, 1996a). The
range of F’. canadensis overlaps with that of the eastern tiger swallowtail, P.
glaucus, in a narrow zone of hybridization through central Wisconsin, Michigan

and New York, and along the northern Appalachian mountain range extending

I P. canadensis

I Hybrid Zone

 

P. glaucus

Figure 1.1. Distribution map of P. canadensis, P. glaucus, and areas of

hybridization between them in North America.

south. From this hybrid zone, the range of P. glaucus reaches south to central
Florida and west to Texas, Arkansas, Missouri and Illinois. Despite continued
gene ﬂow between these two species, they exhibit many important
morphological, ecological and genetic differences (summarized in Table 1).

P. glaucus individuals are larger and have a relatively narrower black band
on the anal cell of their hind wings (Luebke et al., 1988). Females have more
blue swling on the dorsal side of their hindwings (see Ch. 2) which extends
slightly onto the forewing. Additionally, female P. glaucus occur as two color
morphs, a yellow morph which is similar in coloration to males, or a dark morph
which is believed to mimic the black and blue coloration of the male pipevine
swallowtail, Battus philenor (Fig. 1.2; Brower, 1958). The frequency of these two
morphs within a population appears to be correlated with the density of B.
philenor occurring congruently (Brower and Brower, 1962).

In contrast to P. glaucus, individual P. canadensis are generally smaller
and have a relatively wider black band. Females have less overall dorsal blue
coloration which does not extend onto the upper wing. Female dark morphs do
not exist in pure P. canadensis populations due to the absence of a melanism
gene (b+, found on the W- chromosome and maternally inherited) and the
additional presence of a suppresser gene (found on the Z- chromosome and
paternally inherited in females; Scriber et al. 1996).

The two most important ecological differences occur in diapause induction

(and corresponding voltinism patterns; Rockey et al., 1987), and larval host-use

Table 1.1. Summary of morphological, ecological and genetic differences

between P. glaucus and P. canadensis

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[ TRAIT | P. CANADENSIS I P. caucus | REFERENCE
Morphological
Smaller (35 - 50 Larger (50 - 70 Luebke et al.,
Size mm uppenrving mm uppenrving 1988; Aardema
length) length) Pers. Obs.
Wider (50 - 90% Narrower (10 - Luebke et al.,
Blackband Width of anal cell 50% of anal cell 1988; Aardema
width) width) Pers. Obs.
Female Blue
Coloration Extends No Yes Aardema, Pers.
- Obs.
onto Upperwrngs
Hagen and
Female Coloration OngoYell‘ow Yellilir‘voothark Scriber, 1989;
'p ’9 Scriber et al., 1996
Ecological
Diapause Induction Obligate Facultative $5);er et al.,
Very
Very
Larval Host Use polyphagous, but 332323983156: :1: Scriber, 1988
tulip tree IS toxrc toxic
Genetic
. . Andolfatto et al.,
hgzﬁfggﬁnaelnigl 19 Fixed Differences 2003; Putnam et
9 al, 2007
Absence of Dark Presence of
Morph enabler Dark Morph
gene on W enabler gene on
. . chromosome W chromosome Hagen and
Mrmrcry Genes and Presence of and absence of Scriber; 1989
Dark Morph Dark Morph
Suppresor on Z Suppresor on 2
chromosome chromosome
Z—Iinked Allozyme _ _ _ _ Hagen and
Pgd -125/ 80/ 150 100 or 50 Scriber, 1991
Z-linked Allozyme Hagen and
tan ‘80 °' 4° 40° Scriber; 1991
Autosomal Allozyme Hagen and
Hk '110 '90 °' ‘100 Scriber, 1991
AFLP Analysis Many Species Level Differences \zllggtger and Porter,

 

 

 

 

 

 

Figure 1.2. a.) Dark morph, P glaucus female; believed to mimic the distasteful
pipe-vine swallowtail, Battus philenor b.) Battus philenor male c.) Yellow morph,

P. glaucUs female d.) P glaucus male

abilities (Scriber, 1988). P. canadensis are obligatory diapausing and only
complete only one generation per year (univoltine).

They ovenrvinter in the pupal stage and must experience a prolonged
period of cold temperatures before they will break diapause. Under natural
conditions, diapause is entered regardless of the photoperiod experienced by the
larvae (Rockey et al., 1987). In contrast, P. glaucus is facultatively diapausing
and can either enter a diapause stage like P. canadensis, or else directly develop
into an adult within the same year that larval development is completed. Which
of these two strategies is initiated in the pupal stage is based primarily on light
cues (day length or ‘photoperiod’), although temperature and host quality may
also play a role (Rockey at al., 1987). All populations of P. glaucus complete at
least two generations per season (bivoltine), and in some locations (e.g. southern
Georgia and Florida), three or more generations may be possible (multivoltine;
Rockey et al., 1987, Scriber at al., 2008).

Tiger swallowtail butterﬂies are among the most polyphagous of all
butterﬂies with regards to larval host use abilities. Larvae can utilize a large
number of deciduous tree species from 10 disparate families (Scriber et al.,
1982; Scriber, 1988). These include black cherry (Pmnus semtina), ash
(Fraxinus spp.) and hop tree (Ptelea tn'foliata). However, adult female
ovipositional preferences tend to be narrower. One of the most preferred
ovipositional hosts of P glaucus is tulip tree (Lin'odendnon tulipifera). Larvae of P
glaucus can utilize tulip tree effectively and grow and survive on it similar to other

good host species (Mercader et al., 2009). In contrast, tulip tree is toxic to P.

canadensis larvae and causes complete mortality in larval families fed this
species (Scriber et al., 1982). Conversely, one of the preferred ovipositional
hosts of P. canadensis is quaking aspen (Populus tremuloides) which is toxic to
larvae of P. glaucus (Scriber et al., 1989). Interestingly, in ovipositional assays P.
canadensis females will often lay a high proportion of their eggs on tulip tree
(Mercader and Scriber, 2007; 2008a). This maladaptive tendency is unlikely to
have any negative impact on natural populations as the range of tulip tree does
not overlap with P. canadensis (Mercader and Scriber, 2007). In ovipositional
assays, P. glaucus females lay very few eggs on quaking aspen (Mercader and
Scriber, 2007; 2008a).

Lastly, there are many genetic differences between the two species.
These include ﬁxed nucleotide differences in the COI and COII mitochondrial
genes (Andoffatto et al., 2003) and in several genes found on the Z—chromosome
(Putnam et al., 2007). Additionally, AF LP markers exhibit clear species
diagnostic differences (Winter and Porter, 2009). There are ﬁxed species
differences in the Z-linked allozymes Pgd and th, as well as the autosomal Hk
(Hagen and Scriber, 1989; 1991). Finally, the aforementioned
enabler/suppressor complex of genes which facilitates the presence of dark
morph females in P. glaucus populations may be considered an important genetic
difference (Scriber et al., 1996a).

Local adaptation between populations of each species has been observed
(Ayres and Scriber, 1994; Bossart and Scriber, 1995; Scriber 1996b; 2002),

indicating that in some circumstances selection pressures can be strong enough

to overcome the homogenizing tendencies of migration and gene ﬂow. However,
between P glaucus and P. canadensis, historical hybridization has clearly
occurred, facilitating the exchange of genomic regions and producing hybrid
populations (Luebke and Scriber, 1988; Scriber, 1990a; Stump et al., 2003;
Scriber and Ording, 2005, Kunte et al., In Prep). This is an interesting
observation as hybridization between P glaucus and P. canadensis results in
both direct ﬁtness reductions (in part due to a Haldane effect; Hagen and Scriber,
1995) and extrinsic ﬁtness reductions that include failure to enter diapause in
areas at northern latitudes where it is impossible to complete two generations in
one season (Ritland and Scriber, 1985; Hagen and Scriber, 1989), and the
initiation of diapause after one generation (obligate diapause) in more southerly
regions where the time spent in this stage during the warmest parts of the year
incurs metabolic costs which reduce survival and ﬁtness (e.g. fecundity, etc.;
Scriber et al., 2002; Mercader and Scriber, 2008b).

There are several explanations to account for incomplete reproductive
isolation and continued gene ﬂow between P. glaucus and P. canadensis. First,
hybridization occurring directly between pure parental genotypes is likely rare
(Donovan and Scriber, 2003). Gene ﬂow more likely occurs along a gradient of
backcrossed individuals throughout the hybrid zone from pure P glaucus south of
the zone to pure P. canadensis north of the zone (Hagen, 1990). This means
that direct costs to hybridization (e.g. reduced viability) are diminished and may
not be strong enough to generate signiﬁcant reinforcement to prevent

hybridization. Furthermore, extrinsic costs to hybridization are further reduced as

speciﬁc genetic regions that cause strong ﬁtness reductions are quickly purged
from populations (e.g. facultative diapause in the northern parts of the hybrid
zone is ‘suicidal’; Martinsen et al., 2001; Scriber et al, 2008).

Another possible explanation for the presence of continued gene ﬂow is
an asymmetrical mating preference observed in male tiger swallowtails. In two-
choice mate preference assays, Deering and Scriber (2002) observed that males
of both P. glaucus and P. canadensis strongly preferred to mate with P. glaucus
females. This suggests that within and close to the hybrid zone, P. glaucus
males may exhibit reduced inclinations to mate with P. canadensis and possibly
hybrid females. However, it is also probable that male P. canadensis would
readily mate with any P. glaucus females they encounter and likely hybrid
females as well. There are two mveats to this suggestion. First, Deering and
Scriber did not examine male mating preferences near or within the hybrid zone
and as seen in other hybrid zone systems (eg. Peterson et al., 2005) mating
preferences may be weaker or stronger depending on proximity to areas where
hybridization occurs. Secondly, as the cues that resulted in males of both
species preferring P. glaucus females are unknown (see below), we have no
knowledge of how these cues may be inherited and thus how males of either
species would respond to hybrids of various genetic backgrounds.

Differential preference for speciﬁc phenotypes could manifest itself as
complete refusal to engage a female or relatively less time/energy spent in
pursuit of a female who initially rejects a male’s advances (Krebs, 1988;

Aardema, pers. obs.). The interspeciﬁc dynamic in male mate preference may, in

10

part, explain continued hybridization between the two species, as well as the
speciﬁc ecological and genetic characteristics of several unique hybrid
populations that have been observed (Scriber and Ording, 2005, Mercader et al.,

2008; Kunte et al, In Prep.).

Male Mate Choice in Tiger Swallowtails

Male mate choice is deﬁned as a differential response of individual males
to conspeciﬁc, reproductively mature females which vary in some quality
recognizable to the male (Bonduriansky, 2001). Male mate choice is expected to
arise in systems where the male contributes resources to the formation of
offspring aside from sperm (Bonduriansky, 2001). Additionally, when there is a
large amount of variance in female quality, males are also predicted to exhibit
some degree of discrimination.

When male tiger swallowtails mate, they pass a large packet of nutrients
and sperm (known as a sperrnatophore) to the female. To produce this packet,
males engage in a behavior called puddling, which involves the uptake of
moisture and nutrients from wet substrate, dung, carrion, and other similar
resources (Arms et al., 1974). Male tiger swallowtail butterﬂies are frequent
puddlers, regularly accumulating in large numbers at high quality puddling
locations (Arms et al., 1974; Scriber and Ayres, 1988; Otis et al., 2006). Male
Papilio are severely limited in their reproductive ability if they are unable to intake
nutrients (Lederhouse et al., 1990). The need of males to puddle in order to

acquire sufficient resources to form a sperrnatophore likely limits the number of

11

successful copulations males can engage in during their lifetime (Lederhouse et
al., 1990; Bissoondath and Wiklund, 1996; Lewis and Wedell, 2007).
Furthermore, puddling puts males in danger from predation (Morris, 1953;
Rawson and Bellinger, 1953). Finally, the ﬁrst sperrnatophore of a male is
signiﬁcantly larger than subsequent sperrnatophores (Lederhouse et al., 1990).

The sperrnatophore represents a signiﬁcant contribution of time, energy,
and resources (Arms of al., 1974). As encountered females will vary in the
number of eggs they have available (both absolutely and due to previous
oviposition), it is perhaps not surprising that males exhibit mating preferences.
Examinations of spennatophore number have revealed that in most
circumstances a single mating is likely sufﬁcient to fertilize the entirety of a
female’s eggs (Levin, 1973; Lederhouse and Scriber, 1987). However, when
initial spennatophores are small, subsequent matings may be necessary to
ensure all eggs are fertilized (Lederhouse and Scriber, 1987). Additional matings
may also provide important nutrients for female somatic maintenance.
Accordingly, females frequently solicit male matings (Krebs, 1988). .

The high cost of a spennatophore creates a strong impetus for males to
exhibit some degree of female mate choice. Furthermore, mating studies also
lend credence to the assertion that males exhibit mate choice and preferences
(Brower, 1959; Deering and Scriber, 2002). The female characteristics that
Papilio males use to select conspeciﬁc mates are presently unknown. Nor is it
known whether traits being assessed in intraspeciﬁc encounters are the same

responsible for the interspeciﬁc mating preferences exhibited by P. canadensis

12

males for P. glaucus females. However, it is reasonable to suspect that the trait
(or traits) used by males to judge females in intraspeciﬁc encounters are the
same that generate asymmetrical, interspecies preferences and that this trait

generally exists in a more ‘preferred’ state in P. glaucus females.

Research Objectives

Given the asymmetrical mating preferences exhibited by male tiger
swallowtail butterﬂies, l postulate that a single morphological trait or small
number of traits differs between P. glaucus and P. canadensis and is
recognizable and distinguishable by male tiger swallowtails. This trait is likely
female limited. I also predict that variation in some quality of this trait will result in
a greater or less likelihood of males attempting to copulate. This trait may either
be directly correlated with ﬁtness, or may simply be a mate recognition cue that
allows males to recognize females of their own species (or a species that is very
similar in appearance). In this case, interspeciﬁc preferences may be the result

of this trait eliciting a stronger mate recognition response.

13

CHAPTER 2:

VARIATION IN WING COLORATION BETWEEN PAPILIO CANADENSIS AND
P. GLAUCUS POPULATIONS

Introduction

In the context of evolutionary processes, large amounts of variation in
morphological characteristics exist and persist for several different reasons. One
explanation for the maintenance of such variation is an absence of consistent,
strong stabilizing selection on a trait (Heegaard, 1997; Kotanen and Bergelson,
2000). Selection that acts in a relatively regular fashion will generally narrow a
trait’s phenotypic variance by reducing genetic variation contained within
individuals that are less ﬁt than population conspeciﬁcs (Kaufman et al., 1977).
Conversely, an absence of some regular selection permits a trait to vary to the
extent genetic variability allows (Hunt at al., 2007). Environmental inﬂuences,
polygenic effects, and differences in gene regulation can all increase phenotypic
variation in a trait (T urelli and Barton, 2004).

Another cause of extensive morphological variation is the presence of
multiple, divergent selection pressures that act on a trait, preventing stabilization
and the reduction of variance (Grant, 1986). This may happen within a
population, causing the variance of a trait to remain broad and preventing the
narrowing of allelic diversity (Saint-Laurent et al., 2003, Yeaman and Jarvis,
2006), or it may be that speciﬁc pressures differ in nature and intensity between
populations, creating local adaptation with limited variation within speciﬁc areas,

but wide variation within the species as a whole (Linhart and Grant, 1996).

14

 

Depending on their extent, migration and gene ﬂow will impact tendencies for
local adaptation substantially (Yeaman and Jarvis, 2006).

Tiger swallowtails in the eastern USA are known to vary in many
ecological and morphological traits within populations (Luebke et al., 1988;
Hagen, 1990; Mercader and Scriber, 2007; Mercader et al., 2009), between
diﬂ‘erent populations of the same species (Hagen, 1990; Scriber, 1996b; 2002;
Kunte et al., In Prep.), and between different species (Scriber at al., 1989; Stump
et al., 2003; Kunte et al., In Prep.). Much of this variation is linked to speciﬁc
climatic conditions in the respective geographical ranges of these species and
populations. Climate and temperature inﬂuences variation either directly due to
selection created by dilferent thermal tolerances (Mercader and Scriber, 2008b),
or else indirectly as mediated through host-plant ranges and phenologies
(Scriber and Ording, 2005). This selection is particularly important for
maintaining the species boundary between P. canadensis and P. glaucus, mostly
through a post-zygotic reduction in survival of hybrid offspring outside the hybrid
zone.

Male and female mating preferences may also have importance for
maintaining these species’ identities in addition to inﬂuencing hybrid zone
introgression patterns (Stump et al., 2003; Scriber and Ording, 2005; Kunte et al.,
In Prep). Variable traits that are important for either mate recognition and/or
mate preference either within or between these species remain unknown,
although evidence suggests this may be another important and possibly non-

thennally linked area of variation (Deering and Scriber, 2002). Male mating

15

 

preferences are particularly intriguing as tiger swallowtail males have been
shown to exhibit choice for both conspeciﬁcs (Brower, 1959; Deering and Scriber,
2002) and heterospeciﬁcs (Deering and Scriber, 2002), and also for females that
appear ‘young’ versus those that appear ‘old’ (as indicated by scale loss; Scriber
and Aardema, unpublished data). Furthermore, males will reject potential
conspeciﬁc mates outright (Krebs, 1988; Aardema, pers. obs.), possible because
they appear older and are likely to have less eggs or possibly because of some
perception of genetic inferiority. This further indicates that male mate
preferences may be a fruitful area of investigation for understanding potential
pre-zygotic isolating mechanisms between tiger swallowtail butterﬂies.

The color vision of butterﬂies in the genus Papilio has a wide spectral
range that extends down into the UV (300 - 400 nm) and up into the infrared
(700 — 800 nm) (Briscoe, 2000). Additionally, these butterﬂies may be among the
best of any organism at discriminating between similar shades of color
(Koshitaka et al., 2008). Visual assessment of females by males is common in
many Lepidopteran species (Bonduriansky, 2001). These facts suggest that a
visual cue or cues may be important in tiger swallowtail mate recognition and/or
mate assessment. In particular, a consistent difference in some component of
coloration between P. canadensis and P. glaucus females may explain previously
observed asymmetrical male mating preferences (Deering and Scriber, 2002).

The extent of dorsal blue coloration has been noted to appear qualitatively
different in the two species, with P. glaucus females seeming to have

substantially more of their wing surface covered by blue scaling (Deering and

16

Scriber, 2002). Additionally, although both males and females have some dorsal
blue scaling, extensive blue coloration appears limited to females (Scriber,
1990b; Aardema, pers. obs.). Both female color morphs of P. glaucus have this
coloration, making it a potential 'common recognition characteristic' that may aid
males in recognizing both color morphs as conspeciﬁc females (Platt et al., 1984;
Hereau and Scriber, 2002). In populations where both color morphs are present,
individual males do recognize females of each morph as potential mates and will
attempt to copulate with individuals of both colors (Aardema, pers. obs.).

To locate females, males generally patrol forest edges along streams,
roadways, power cuts and other, similar linear paths. They ﬂy above nectar
sources along these paths looking for conspeciﬁcs. It is believed that yellow
coloration and movement provide initial recognition cues, as males will engage
both males and females when ﬁrst encountered, generally approaching from
above and behind (Aardema, pers. obs.). After initial contact, males usually
recognize other males and break off contact quickly. Conversely, if a female is
encountered the male will pursue her until she either accepts his attempts to
copulate or else loses him via rapid ﬂight through the undergrowth (Deering and
Scriber, 2002). Although some subtly differences in this behavior between P.
canadensis and P. glaucus males have been noted (Deering and Scriber, 2002),
general patterns of mate location are similar between the two species.

This type of mate lomtion strategy suggests that a visually-based, dorsally
located, gender-speciﬁc cue such as blue coloration is highly likely to aid males

in locating and recognizing females. Afemale with reduced blue coloration may

17

appear ‘male-like’ and receive reduced copulation attempts (and a potentially
corresponding reduced ﬁtness) relative to females with more extensive blue.
Additionally, as the sperrnatophore is a valuable resource and males seemingly
exhibit some degree of mate selection, it is possible that this blue coloration may
be linked to female ﬁtness in some presently unknown way.

To determine whether the relative amount of blue coloration on the dorsal
wing surface differs between female P. glaucus and P. canadensis (independent
of overall size), I quantiﬁed the relative extent of blue coloration in females and
males of both species on both the ventral and dorsal side of the wings. Based on
my qualitative observations, I predicted that the difference between males and
females of both species would be signiﬁcant on the dorsal side of the wings, but
not the ventral side. I also postulated that female P. glaucus and P canadensis
will be signiﬁcantly different in dorsal coloration. I predicted that in no possible
comparison (between sexes or species), would ventral blue coloration differ
signiﬁcantly.

Additionally, because ultraviolet color is important in many lepidopteran
systems for mate recognition and selection (Brunton and Majerus, 1995), I
measured wing regions of blue, yellow and black coloration for the presence or
absence of UV reﬂectance. Given that these Papilio butterﬂies can see well into
the UV spectrum, I hypothesized that the blue coloration would reﬂect a relatively
large amount of W light, the yellow would reﬂect a relatively small amount of UV
light and the black would reﬂect no UV light. A ﬁnding of UV reﬂectance would

suggest potentially fruitful avenues of future inquiry.

18

 

Materials and Methods
Samples

To analyze differences in blue coloration between populations and species
of tiger swallowtail butterﬂies I collected pristine or near pristine (little or no wing
damage and scale loss) wild females from populations in Fairbanks-North Star
Co., Alaska (n=11), Charlevoix Co., Michigan (n=12), Jackson and Gallia Cos.,
Ohio (yellow morph: n=8, dark morph: n=9) and Levy Co., Florida (yellow morph:
n=15, dark morph: n=10). I also obtained females from two additional
populations, one in Bennigton 00., Vermont (n=7) and one in Habersham and
Rabun Cos., Georgia (yellow morph: n=4, dark morphs: n=2). Collectively these
six populations span the latitudinal range of both P glaucus and P. canadensis
(Fig. 2.1). Males from each of these populations were also collected for
female/male comparisons (Alaska= 10; Michigan=12; Ohio= 12; Florida= 11;
Vennont= 9; Georgia= 3).
Wing Preparation and Photography

Using a ﬁne bladed scissors, l carefully removed all four wings from each
specimen at the wing base where it attaches to the thorax. After removal, wings
were kept in glassine envelopes until photographed. I used an Olympus FE-310
(8.0 megapixels) digital camera to photograph both sides of minimally one hind
wing and one forewing from each specimen (Fig. 2.2a). In some instances all
four wings were in pristine condition and l was able to take two photographs, one

of each wing set. I generally tried to match wings from the same side of the

19

body, but occasionally this was not possible because of damage or scale loss to
two separate wings.
Analysis of Relative Blue Area

To analyze these photographs, I used the imaging program ImageJ
(Abramoff et al., 2004) with standard tools to measure the total area of the two
wings (Figure 2.2b). I then used the Threshold_Colour plugin (Landini, 2010) to
remove all coloration in the image except blue and then the standard ImageJ
tools to measure the area of this coloration (Figure 2.2c). For both
measurements, area was calculated as the number of pixels per unit
measurement. With the measurements of total wing area and blue area, I was
able to calculate a relative proportion of blue coloration. This allowed me to
make comparisons of blue area that were independent of the general size
differences between P. glaucus and P. canadensis. When two images were
taken for the same butterﬂy, the mean of their relative blues areas was used in
analysis.
Ultraviolet Examination
I ﬁrst assessed the presence or absence of ultraviolet coloration qualitatively by
using a UV bypass ﬁlter attached to a Nikon Coolpix 990 (3.34 megapixels)
digital camera. With this combination of ﬁler and camera, I photographed six
fore- and hind wing pairs from P. glaucus females in a darkened room (to reduce
full spectrum light) with a circular black light for illumination (Acorn, 2002). Using
this technique I could visualize areas that were W reﬂective and those that were

UV absorbent.

20

PM?“

 

Figure 2.1. Locations of sample populations for blue color analysis. Gray circles
represent P canadensis populations and black circles represent P glaucus
populations. a.) Fairbanks-North Star Co., Alaska b.) Charlevoix Co., Michigan
c.) Bennigton Co., Vermont d.) Jackson and Gallia 003., Ohio e.) Habersham

and Rabun Cos., Georgia f.) Levy Co., Florida

21

 

 

Figure 2.2. Process of wing analysis to determine total wing area and area of
blue coloration. a.) First, one forewing and one hindwing were photographed.
This was done for both dorsal and ventral sides b.) The image was made binary
and the total area was determined. A ruler in the photograph was used to
determine scale. c.) Lastly, all coloration except the blue was removed and the

area of this remaining coloration was determined. See text for more details.

22

I also measured the spectral reﬂectance signature of blue, yellow and
black wing areas using an Ocean Optics $2000 miniature ﬁber optic
spectrophotometer. I out three roughly 1 cm2 squares of wing that had a single
solid color (yellow, black or blue) from the pristine wings of four Levy Co., Florida
yellow females. Reﬂectance of color was measured in comparison to a white
standard.

Data Analysis

All statistical analysis was performed using the statistical package R
(Abramoff et al., 2004). I made overall mean comparisons between male P
canadensis, female P canadensis, male P glaucus, and yellow- and dark morph
female P glaucus for the relative blue area of both the dorsal and ventral sides of
the wing. To do this I used either a standard one-way analysis of variance
(ANOVA) when variances were homogenous (as determined by Levene’s test;
Levene, 1960), or, when variances were unequal, I used Welch’s ANOVA (Welch,
1951). I also performed ANOVAs comparing males and females of individual
populations, again for both dorsal and ventral relative blue wing area. I
performed comparisons for each of the populations from the above analyses
using pairwise t tests with adjusted p values using the Bonferroni step-down
correction method develop by Holm (Holm, 1979). This correction gives control
over the family wise error, but is less conservative than the simple Bonferroni
correction.

Finding a statistical difference for relative blue area in species

comparisons, I performed a one-way analysis of covariance (ANCOVA)

23

comparing the total wing area and the area of blue with species as a co-factor to

determine if observed differences in blue were an artifact of size diﬂ‘erences

between P canadensis and P glaucus or rather were indicative of a fundamental

allometric difference between the two species. This would be indicated by

statistically signiﬁcant differences in either the slope and/or y-intercept. I

compared three models: a simple model without species as a cofactor, an

additive model with species as a cofactor and a multiplicative model, again with F
species as a cofactor. Models were compared using adjusted coefﬁcient of I
determination (Adj. R2), the Akaike information criterion (AIC; Akaike, 1978) and
the Schwarz Bayesian Information Criterion (BIC; Schwarz, 1978). For all
statistical tests, an alpha value of 0.05 was considered signiﬁcant. No statistical
tests were performed for ultraviolet coloration, but rather a qualitative

assessment was made.

Results

For all comparisons, variances were homogenous except for the dorsal
comparison of relative blue area between species/sex results. There was a
signiﬁcant difference in dorsal blue area [F(4,57.11)=32.05, p <0.001].
Speciﬁcally, both yellow-morph and dark-morph P glaucus females differed
signiﬁcantly from female P canadensis, male P canadensis and male P glaucus
(Fig. 2.3). No other sex or species comparison differed signiﬁcantly. There were
no signiﬁcant differences in ventral relative blue area [F(4,128)=1.71, p=0.151];

(Fig. 2.3).

24

Between populations of female P canadensis and female yellow-morph P
glaucus there were signiﬁcant differences between the mean relative dorsal blue
area [F(4,48)=23.235, p<0.001]. In concordance with the species/sex result,
both P glaucus populations differed from the three P canadensis populations
(Fig. 2.4). No female P canadensis populations differed signiﬁcantly from one
another, nor did the three P glaucus populations differ from one another. There
were no signiﬁcant differences in ventral blue coloration between any population
of female P canadensis and female yellow-morph P glaucus [F (4,48)=1 .724,
p=0.160]; (Fig. 2.4)

Variance between the three populations of dark-morph female P glaucus
was homogenous for dorsal blue, and the means did not differ signiﬁcantly
between the populations [F(2,18)=0.735,p=0.494]. Variance between the two
populations for ventral blue was homogeneous and an ANOVA revealed no
difference in means as well [F(2,18)=0.276,p=0.762).

Surprisingly, there were differences in mean male dorsal blue between
populations [F(4,49)=7.590, p<0.001]. However, unlike in females this variation
did not occur in relation to species (Fig. 2.5). Males from the Vermont population
(P canadensis) differed from the Alaska and Michigan populations (also P
canadensis), and Ohio (P glaucus) likewise differed from Alaska and Michigan.
Also surprisingly, male ventral blue coloration differed signiﬁcantly between some

populations

25

9.0 -

 

 

 

 

 

   

 

 

 

 

 

     

 

 

 

 

 

 

 

 

 

g e
C 8.0 1 r
.9
E 7.0 a
% 6.0 —
O
3 5.0 -
.0
‘8 4.0 ‘ b b d “ b b b
E 3.0 _ a. a. ,C i a.c a. ,0 ac a, ,C C
re 2 0 l
o . - -
.5, l j
19 1.0 « I 1
0 " "' I
m 0.0 I'I' n" n n
Pc Female Pg Female Pg Female Pc Male Pg Male
(YW) (DK)

Figure 2.3. Mean relative blue area i SD for all species/sex/color morph
combinations. Pc : P canadensis, Pg : P glaucus, YW : yellow morph females
and DK : dark morph females. For each combination, the light grey bar shows
the mean blue coloration on the dorsal wing surface and the dark grey bar shows
the mean blue coloration on the ventral wing surface. Bars with the same letter
are not signiﬁcantly different at the p<0.05 level (pairwise t test with Holm’s

correction).

26

 

4.0

 

 

 

 

 

 

 

 

 

 

 

SQ? c.e e

v c,e
c:

s I l
E 3'0 ' a.c a.e.g l a.e,g

'3 b,c,d.g,f a,f
0 .
3 d

_ , a.

a 2.0 ab a'd ,

“5 ‘ a

as I I

9 I I

to

0 1.0

g .

.9

0

CE 00 n= n= n3

 

 

 

 

 

 

 

 

 

 

 

 

 

Alaska Vermont Michigan Ohio Georgia Florida
PC PC Pc Pg Pg Pg

Figure 2.4. Mean relative blue area :I: SD for all female P canadensis (Pc) and
yellow morph P glaucus (Pg). For each combination, the light grey bar shows
the mean blue coloration on the dorsal wing surface and the dark grey bar shows
the mean blue coloration on the ventral wing surface. Bars with the same letter
are not signiﬁcantly different at the p<0.05 level (painrvise t test with Holm’s

correction).

27

3.0 ‘

(1.9

b,e.f,g.h b,e,f,g.i

b.c.d a,c,g a,d,h,i

2.0 '

 

a.f

 

 

 

 

 

 

1.0 ‘

 

 

 

 

 

Relative area of blue coloration (%)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘0

I'Ii D II-

0.0

 

Alaska Vermont Michigan Ohio Georgia Florida
Pc P0 P0 Pg Pg Pg

Figure 2.5. Mean relative blue area 1 SD for all male P canadensis (Pc) and P
glaucus (Pg). For each combination, the light grey bar shows the mean blue
coloration on the dorsal wing surface and the dark grey bar shows the mean blue
coloration on the ventral wing surface. Bars with the same letter are not

signiﬁcantly different at the p<0.05 level (pairwise t test with Holm’s correction).

28

[F(4,49)=3.840, p=0.009]. Again, this cannot be generalized to species
differences (Fig. 2.5).

I performed an analysis of covariance between total wing area and area of
blue coloration for female P canadensis and yellow-morph P glaucus dorsal wing
blue as species-level differences were signiﬁcantly different (Table 2.7). An F test
showed that variance within the two species was not homogenous, so both blue
and total area were log transformed to normalize the data. I compared the
regression slopes for a simple model without species as a cofactor, an additive
model with species as a cofactor and a multiplicative model, again with species
as a cofactor using R2, AIC, and BIC (Table 2.10). The model with the greatest
predictive value appears to be the additive model that incorporated population
adjustments to the intercept. This suggests that within both female P glaucus
and P canadensis, absolute blue area increases in a similar fashion (statistically
equivalent slopes) from smaller to larger individuals, but that between females of
the same size, P glaucus females have fundamentally more absolute dorsal blue
coloration (Fig. 2.6).

Ultraviolet photography of dorsal wing coloration revealed clear areas that
reﬂect UV and areas that appear to be entirely UV absorbent (Fig 2.7).
Reﬂectance appeared to occur primarily in the blue regions of the wing. The
spectral analysis revealed a strong UV signature for blue coloration (Fig 2.8), but

virtually no UV reﬂectance for yellow or black coloration.

Discussion

29

Yarn _ V .. .

I examined differences in blue coloration between male and female tiger
swallowtail butterﬂies and between the sister species P glaucus and P
canadensis. As predicted, I found that males and females of P glaucus differed
signiﬁcantly in the relative amount of blue on their dorsal wing surface, but not
their ventral. However, I also predicted that female and male P canadensis
would differ signiﬁcantly in their dorsal wing coloration, but this proved to be
incorrect. Neither the dorsal nor the ventral amount of relative blue coloration
differed signiﬁcantly between the sexes in P canadensis.

Females of each species differed from one another in their relative amount
of dorsal blue coloration, but not their ventral. P glaucus females had
signiﬁcantly more relative blue coloration than did P canadensis females.
Furthermore, regression analysis suggests that the area of blue coloration
increases in a similar manner between P canadensis and P glaucus, but that P
glaucus has more blue on an absolute scale, irrespective of overall wing size.
Unexpectedly, there were signiﬁcant differences between male populations for
both dorsal and ventral relative blue coloration, but these were not species
speciﬁc. Finally, I found that blue coloration in general is UV reﬂective, whereas
other coloration is not.

These observations lend support to my hypothesis that female-limited,
dorsal coloration has importance for male mate recognition and possibly mate
preferences. If males use the extent of blue coloration as a cue to differentiate
between males and females than more blue coloration may result in a greater

‘preference’ for a particular female when multiple females are present. With

30

Table 2.1. Results from analysis of covariance (ANCOVA) for a comparison

between female P canadensis and yellow morph P glaucus for the total area of

the wings and the area of blue coloration with species as a cofactor. The data

was l091o transformed to homogenize variance between the two species. The

additive model appears to have the greatest predictive value and suggests that

the two species increase in blue area in similar fashion with increased overall

size (statistically equivalent slopes), but that P glaucus females have more blue

on an absolute scale. Bolded numbers represent the values for the preferred

model in the three different analyses (Adj. R2, AIC and BIC). See text for more

 

 

 

 

 

 

 

details.
. . Ad'. .
Model Specres Model Equation R! AIC BIC
. (log1oblue area) = 2.104 *
S I NA - .7 8 -210.9 -211.4
Impe (log1ototal_area) - 3.123 O 6
P. (Iog1oblue_area) = 1.043 *
canadensis (10910 total_area) - 1.951
Additive 0.836 -229.2 -182.6
log1oblue area) = 1.043 *
P. I I —
g aucus (log1o total_area) - 1 .636
P. (log1oblue_area) = 1.289 *
canadensis (10910 total_area) - 2.231
Multiplicitive 0.835 -228.0 -159.9
P. glaucus (Iog1oblue_area) = 0.842

 

 

(log1o total_area) - 1.362

 

 

 

 

31

 

 

 

 

A P glaucus

Log10 of blue area

 

 

e P. canadensis

 

 

 

 

10 1'5 20 25 30
Log10 of total wing area

Figure 2.6. Log transformed comparison of absolute total wing area and area of
blue coloration for P canadensis and yellow morph, P glaucus females. Area
values are in cmz. Regression lines are drawn using the additive model (See

table 2.10). The slopes are not signiﬁcantly different. See text for more details.

32

regard to previous interspec'rﬁc preference comparisons (Deering and Scriber,
2002), my results suggest that more relative blue coloration may explain why
males of both P glaucus and P canadensis preferred to mate with female P
glaucus over similar sized P canadensis females. Although it is common to ﬁnd
multiple females feeding in close proximity to one another (Aardema, pers. obs.),
most encounters likely involve single males and females. In these instances, the
amount of dorsal blue may inﬂuence a male's initial inclinations to engage the
female or the amount of effort given to pursuit if it occurs. Differences in blue
coloration may also inﬂuence the amount of nutrients and sperm a male passes
to the female in his sperrnatophore (Bonduriansky, 2001; Xu and Wang, 2009). It
is not presently known whether male tiger swallowtails vary the size of their
spennatophore based on a 'quality' evaluation of their mate. Although some
unusual and extreme patterns in male blue coloration have been observed
(Scriber, 1990b), more general variation in male blue coloration was unexpected.
Perhaps even more surprisingly, this variation does not correspond to the two
different species. It seems that deviation in blue coloration within the Michigan
population explains most observations of difference. Michigan had a relatively
small amount of blue both dorsally and ventrally compared to the majority of the
other populations. It is presently unknown what factors might result in this
population having less blue (or conversely, other populations having more blue).
Given the observed patterns of difference in blue coloration between the
species and the sexes, mate recognition and/or preferences remain very likely

candidates for selection pressures inﬂuencing the amount of blue on the wings,

33

 

Figure 2.7. An example of my photographic examination of ultraviolet coloration

in tiger swallowtail butterﬂies. a.) The visual image of the dorsal side of a yellow
morph (left) and dark morph (right), female P glaucus. b.) This image was taking
using a UV bypass ﬁlter. The wings displayed are the same as those in ‘a.’ Areas
that reﬂected UV light are white and areas that absorbed UV light are black. See

text for more details.

35%

30%
8 o
: 25/o
g 20%
.43 15%
(D
a: 10%
°\°

5%

0%

30 400 500 600

Yellow
.5... Blue

— Black

 

q

 

 

Wavelength (nm)

Figure 2.8. An example of the reﬂectance spectrum patterns of blue, yellow and

black coloration from one yellow morph, P glaucus female. The other three

individuals assessed in this manner had similar spectral patterns. This ﬁgure

shows a clear UV signature (300-400 nm) for blue coloration, and virtually no UV

reﬂectance for yellow and black coloration. See text for more details.

35

especially for females. However, it is easy to speculate that amount of blue
coloration may have multiple selection factors acting on it, some of which may be
antagonistic to one another.

One pressure that may work to reduce blue coloration is avian predation.
Birds have visual capabilities that extend into the UV spectrum and it has been
shown that Lepidoptera with extensive ultraviolet coloration may be at a higher
risk of attack relative to Lepidoptera that have little UV coloration (Church at al.,
1998; Majerus et al., 2000, Lyytinen at al., 2004). It is presently unknown how
predation pressure would act to reduce blue coloration in males and females,
and given differences in daily behavioral activities (eg. puddling in males versus
ovipositing in females), it is possible that selection on the two sexes by avian
predators differs.

Another inﬂuence on blue coloration may be temperature and a need for
therrnoregulation. Most ultraviolet coloration in Iepidopterans is structurally
based (Ghiradella, 1991; Prum at al., 2006) and such structural color has been
shown to signiﬁcantly increase absorbance of solar radiation (Bosi et al., 2008).
However, two lines of evidence suggest this is unlikely to be a large selective
force on blue coloration. First, the populations that would be predicted to have
the greatest need for increased solar absorbance would be those located in more
northerly localities. However, these populations have the least amount of blue
coloration. Secondly, it is only the basal one-third of the wing that strongly

inﬂuences thoracic and abdominal temperature (Douglas, 1981). This means

36

that most blue coloration in both males and females is unlikely to strongly affect
an individual's body temperature.

Yet another inﬂuence on the extent of blue coloration may be sexual
harassment by other butterﬂies either of the same species or different species.
Although polyandrous, female tiger swallowtails will in some circumstances
refuse male mating attempts, likely based on whether they have recently mated
and possibly due to perceived quality of the male. While females feed, they are
often subjected to harassment by males (Lederhouse, 1995; Deering and
Scriber, 2002; Aardema, pers. obs.). This type of intraspeciﬁc interaction has
been shown to have negative ﬁtness consequences for other insect species
(Svensson et al., 2005; Takahashi and Watanabe, 2010). If males utilize blue as
a mate recognition cue, than male harassment would likely oppose pressures
acting to increase blue coloration due to a disproportionally greater amount of
harassment directed towards females that have more extensive blue.
Additionally, in some populations of tiger swallowtail, interspeciﬁc harassment
occurs from males of other butterﬂy species (Deering and Scriber, 1998;
Aardema, pers. obs.). Like intraspeciﬁc harassment, interspeciﬁc harassment
likely has negative ﬁtness consequences on females (McLain and Pratt, 1999;
Nielsen and Watt, 2000). However, this pressure could act to either increase or
decrease blue coloration depending on what recognition cues the males of other
species react to. Furthermore, interspeciﬁc sexual harassment could have

negative impacts on males in addition to females (Groning and Hochkirch, 2008).

37

Although I have shown that dorsal blue coloration is signiﬁcantly different
between P glaucus males and females, and between female P canadensis and
P glaucus, its utility for mate recognition and/or mate selection is by no means a
certainty. Only further examinations will elucidate whether blue coloration is
actually utilized by males for the purpose of recognizing potential mates and/or
judging these mate's reproductive quality. An investigation of these potential
inﬂuences on tiger swallowtail blue, combined with studies of other selection
factors, will help to reveal why variation in blue coloration exists and possibly
what inﬂuences this variation could have on the dynamics of gene ﬂow in the
hybrid zone. Of most interest may be how pre-zygotic mating barriers are
inﬂuenced by this seeming sexual characteristic, and how it contributes to
divergence, reproductive isolation and adaptive radiation in the tiger swallowtail

species complex.

38

CHAPTER 3:

COURTSHIP AND MATING RESPONSES OF MALE PAPILIO SPP. TO
VARIATION IN FEMALE-LIMITED WING COLORATION

Introduction

Morphological traits that are limited to one sex and appear to have no
direct utility for either basic survival or fostering offspring may be suspected to
have some function in attracting and obtaining a mate (Darwin, 1879). These
“secondary sexual characteristics’ may function either through intra-sexual
conﬂict or intersexual mating preferences. General and consistent differences in
these traits between closely related species may be further suspected of having
importance in conspeciﬁc mate recognition and preference (Panhuis et al., 2001),
which may facilitate continued reproductive isolation.

I have shown that the relative amount of blue coloration on the dorsal side
of the wings in female P glaucus butterﬂies is signiﬁcantly greater than in males
(Chap. 2). Additionally, this blue coloration is relatively more extensive in female
P glaucus than in the sister species P canadensis. Dorsal color between male
and female P canadensis does not differ signiﬁcantly, nor does ventral blue color
between males and females of either species. Lastly, females of the two species
do not differ from one another in the relative amount of ventral coloration.

Males of both P canadensis and P glaucus prefer to mate with female P
glaucus in two-choice assays (Deering and Scriber, 2002). Given the observed
interspeciﬁc variation in relative blue coloration, it stands to reason that this trait

may be important for male mate recognition. Furthermore, although it is

39

presently unknown how the extent of blue correlates with ﬁtness, it is possible
that this trait may be used in intraspeciﬁc mate selection. This would be
especially true if there is a high degree of variation in the genetic quality of
females (something presently unknown). In addition to the attributes of blue
coloration listed above (interspeciﬁc and male/female variation), this
morphological feature is W reﬂective (Figs. 3.1&2) and found in both color
morphs of female P glaucus. It may thus act as a shared recognition
characteristic between the color morphs that enables individual males to identify
both colors as potential mates.

- If the presence of blue coloration acts as an important mate recognition
cue, and/or a potential mate selection cue, then I postulate that male tiger
swallowtail butterﬂies of both P canadensis and P glaucus will strongly prefer to
mate with females that exhibit a greater extent of blue coloration relative to
females that exhibit less blue coloration. In an extreme case, I predict that when
presented with females that either retain or lack dorsal blue coloration, males will

predominantly choose females that have their blue preserved.

Materials and Methods
Butterﬂies

My studies were conducted during the summer from 2008 to 2010. For all
years, I utilized adult, virgin females, captive reared from eggs laid by wild-caught
mothers collected from Bennigton Co. Vermont (VT; a P canadensis population)

and Lancaster Co. Pennsylvania (PA; a P glaucus population). Although ideally I

40

would have used females reared from the same location that the studies were
conducted in, the large number of females needed made this impractical, and
hence it was necessary to obtain females from sources that had the capacity to
rear large numbers of individuals. As my comparisons were between females
reared from the same populations, I feel that the origin of my butterﬂies did not
inﬂuence the outcome of my results. This notwithstanding, in 2008 l was able to
rear virgin females from wild mothers that I collected in Jackson Co. Ohio (OH; a
P glaucus population). For VT and PA females, l utilized individuals in the year
following their rearing (i.e. they overwintered in diapause). The OH females were
utilized in the same year in which I originally collected their mothers.
Experimental Regions

I conducted two-choice mate preference experiments in two P glaucus
populations, one during the ﬁrst seasonal ﬂight in Levy Co. Florida (in 2009 and
2010) and during both the ﬁrst (2010) and second (2008) seasonal ﬂights in
Jackson Co. Ohio. For the Florida experiments I used all females reared from PA
mothers. For the 2008 Ohio experiments (conducted during the second ﬂight), I
used females reared from mothers collected during the ﬁrst ﬂight of this
population. In 2010, I used females reared from PA mothers. I also conducted
two-choice experiments in a single P canadensis population in Charlevoix Co.,
Michigan (in 2008). For this trial, I used all females reared from VT collected
mothers.

All three regions were the same as those where females were collected

for my blue coloration analysis (Chap. 2). In some cases, collection of females

41

and the male mate—choice experiments occurred concurrently, but there is no
reason to believe that this affected my results in either case.

Single choice experiments were conducted in 2009 in a P canadensis
population in Charlevoix Co., Michigan using VT P canadensis and in 2010 in a
P glaucus population in Jackson Co, Ohio using PA P glaucus. These were the
same populations used in my two-choice studies. Based on the outcome of my
two-choice assays and the single-choice experiments (see results), I deemed it
unnecessary to conduct single-choice assays in Florida.

General Mate Choice Experimental Methods (Used in both 2- & single—choice
assays)

To examine how male mating preferences were inﬂuenced by the extent of
blue coloration on females, I conducted choice experiments in wild populations
with known high concentrations of males. To do this, I manipulated females so
that they either retained their dorsal blue coloration or had it entirely concealed. I
choose to do these extremes, as it was my hope that this would result in the
strongest differences and I could then look at more subtle variation in future
work. In the areas described above, I presented these manipulated females to
wild males. My methods are modiﬁed from Brower (1959) and Deering and
Scriber (2002).

I ﬁrst tied a ﬁne nylon thread carefully around the thorax of fresh (emerged
from the pupal stage within 10 days prior), virgin females between the front and
hind wings. The other ends of these threads were tied to metal alligator clips

which allowed me to attach the females to dowels that I inserted through holes in

42

wooden stakes driven into the ground. My stakes were each approximately 2
meters long. The length of the thread varied between 40 and 60 cm. I attempted
to have the females hang at a height similar to that of the nectaring plants that
were being utilized by the study population. Typically, this was 1 — 1.5 m above
the ground. To prevent excessive tangling, a small lead weight was attached to
the thread halfway between the butterﬂy and the alligator clip. This prevented
most tangling, although females would still occasionally ﬂy on the thread in a
manner which caused the line to kink. When this occurred, I manually untangled
these females.

All females in this study received one of two treatments. For the
experimental treatment (blue-concealed), l masked all dorsal blue coloration
using ink from a ﬁne tipped black sharpie brand marker. For the control treatment
(blue-retained) I covered up black areas of the wing with this same black ink, but
left all dorsal blue coloration visible. The area of coverage in both treatments
was approximately equal. This black ink effectively concealed both the visual
and UV spectral components of the blue coloration and appeared very similar
spectrally to natural areas of black coloration (Fig. 3.1). l assigned females to
one of these two treatments haphazardly without ﬁrst looking at the extent of her
dorsal blue coloration.

In most cases, I engaged in two tethering sessions each day, one in the
morning and one in late afternoon. My previous observations indicated that
these were the times when males would be most likely to initiate copulation.

Tethering sessions typically lasted two hours. No female was ever left out for

43

longer than three hours. In between tethering sessions, females were kept in a
cooler with ice.
Two—choice Assays

Two—choice male mate preference trials consisted of preparing pairs of ,
females in the manner described above. Within each pairing, the females were
of approximately the same size as measured by forewing length from base to tip
(1 1 mm). For the two—choice assay the wooden dowel was placed through the
top of the stake so that equal halves stuck out horizontally in opposite directions.
This allowed me to tether females approximately one meter apart. The stakes
were placed a minimum of eight meters from one another either along roads that
were being used as ﬂyways (Ohio and Florida) or else scattered about a meadow
where there were abundant nectar resources (Michigan).

Males ﬂying by these females would engage them if interested. Typically
this involved repeatedly circling of the female on her tether. Upon contact with a
male the female would generally become enlivened and ﬂutter rapidly at the end
of the string, spiraling in tight circles. At this point the male would either land on
the female and initiate copulation or in some cases ﬂy off. In many instances the
male would engage both females of a pair before copulation was initiated.

While tethered, I checked each pair at least once every ﬁve minutes. After
I observed a mating being initiated or else found a copulation in progress, I would
gently remove the male from the female. In most cases, I retained males for

future analysis as part of a separate study. Othenrvise, I released males back into

Blue Coloration

  

 

 

 

 

 
 

20%
8 __ Blue Coloration Covered
g by Black Ink
3 10% — Black Coloration
0)
c:
°\° ‘KIALIWNJVWM #1!
0% 2. ~ +95 .2 - - ~—~
300 350 400 450 500 550 600 650 700 750 800
180/ Wavelength (nm)
0
0)
U
c
(0
1'3
g 9%
(U
o:
°\°
0%. :‘ , 2,- Mm, ___r
300 350 400 450 500 550 600 650 700 750 800
Wavelength (nm)
25%
0)
U
c
to
4.:
u
.3).
H.—
0)
r:
o\° “WW
0% . ~

 

300 350 400 450 500 550 600 650 700 750 800
Wavelength (nm)

Figure 3.1. Three examples of female P glaucus dorsal colors showing the
spectral reﬂectance signature of blue coloration, black coloration and blue
coloration covered by black ink from a sharpie brand marker. UV coloration is
that which occurs between 300 and 400 nm. Black ink effectively conceals both

the UV and visual elements of the blue coloration.

45

the population after they had been marked with a unique number using a red
sharpie and the 1,2,4,7 numbering system developed by Ehrlich and Davidson
(1960). Most marked males were not seen again, but in a few cases the marked
males would return and attempt to mate with a different female in another pair or
occasionally the same female. I did not include any pairing with the same male
and female twice. I did, however, include multiple mate choices by the same
male for different females. These second choices account for four of my total
number of pairings. After each copulation, I switched the positions of the two
females on the tethering pole. There were no more than three total copulations
with any collective pair during a single tethering session.
Single-choice Assays

It has been suggested that choice experiments such as my two-choice
assay may lead to erroneous conclusions as an element of one of the females
may elicit courtship behavior in males, and after this‘occurs no further preference
is exhibited (Trnbergen, 1948). It is possible that the female that retained her
blue coloration compelled males to engage in mating behaviors and the female
he inevitable ended up with was not indicative of a choice. To account for this
possibility, I also carried out single-choice trials where treated females would be
available to males without a female exhibiting the opposite treatment in close
proximity.

For the single-choice assay, l tethered females singularly to dowels in a

fashion similar to that used in the two-choice assay except that only one female

46

was present rather than two. At all times during the trials there were equal
numbers of blue-concealed and blue-retained females available in the study
area. These females were not sized matched between the treatments. The total
number of individuals I tethered out during a given session varied from 6 to 14,
depending on availability. I kept single females at a minimum distance of eight
meters apart. Treated females were assigned to poles haphazardly and when a
male initiated copulation with a female, I exchanged her for a female of the
opposite treatment.
Statistical Analysis

I performed all analysis with the statistical package R (Abramoff et al.,
2004). To assess whether the number of males that preferred females with blue
retained was signiﬁcantly greater than the number of males which preferred
females with blue concealed, I used chi-square tests to compare the number of
matings for each treatment to the expected null-distribution of 0.5, which would
indicate random mating. I assessed each population in this manner individually,
with dark and yellow morphs separate and combined. Additionally, I combined all
observations for each treatment together. When the same experiment was
conducted in multiple years, I combined observations for my analysis. I did not
analyze years separately as samples sizes were generally small. Finally, in
cases where my sample sizes were too low (expected values <5.0), I did not

assess the data statistically.

Results

47

Two-Choice Trials

When all two-choice pairings were totaled, there were 34 male copulations
with females of both species and color morphs that had their blue-retained, and
44 copulations with blue-concealed (T able 3.1). This overall result is not
signiﬁcant, x2(1. N = 78) = 0.26, p =0.61.

In the Florida population (P glaucus), I observed 5 copulations with blue-
retained, yellow-morph females, and 7 pairings with blue-concealed, yellow-
morph females. This difference was not signiﬁcant, x2(1, N = 12) = 0.56, p =0.45.
For dark-morph females, I observed 5 pairings with blue-retained and 8 with blue-
reduced. This difference was also not signiﬁcant, x2(1. N = 13) = 0.41, p =0.52.
When I compare all the pairings of both color-morphs (10 blue-retained, 15 blue-
concealed), the result was not signiﬁcant, x2(1, N = 25) = 0.32, p =0.57.

Within the Ohio population (P glaucus), there were 6 mating observations
with yellow-morph females with blue-retained and 5 with blue-reduced. This
difference was not signiﬁcant, x2(1, N = 11) = 0.76, p =0.38. For dark-morph
females I observed 7 pairings for blue-retained, and 8 for blue-reduced. Again,
this difference was not signiﬁcant, x2(1, N = 15) = 0.80, p =0.37. A statistical
comparison of all the pairings of both color-morphs (13 blue-retained, 13 blue-
concealed) was unnecessary as the number of pairings for each treatment was
identical.

In the Michigan population (P canadensis), I obtained 11 pairings with

females that retained their blue and 16 pairings with females that had their blue

48

Table 3.1. Number of two-choice pairings for each population examined. When

male preferences were examined over multiple years for the same population,

results are combined. No difference was signiﬁcant (see text for values)

 

 

 

 

 

 

 

 

 

 

 

TOTAL

 

BOTH

 

ALL

 

78

 

 

Female Total Number of Number of
Color Number Pairings Pairings
Population Species Morph of with Blue- with Blue-
(P glaucus Pairin s Retained Concealed
only) 9 Female Female
. Yellow 12
Florida P glaucus
Dark 13
Yellow 11 6 5
Ohio P glaucus
Dark 15 7 8
Michigan R NA 27 11 16
canadensis

 

 

49

 

concealed. This was not a signiﬁcant difference, x2(1, N = 27) = 0.34, p =0.56.
Within the two P glaucus populations, I compared the observations for all dark-
morph and yellow-morph females. For the yellow, there were 11 observations of
pairing with blue-retained females, and 12 observations with blue-reduced
females. This was not a signiﬁcant difference, x2(1, N = 23) = 0.83, p =0.36. I
obtained 12 total pairings with blue-retained, dark-morph females, and 16
pairings with blue-reduced, dark-morph females. Again, this was not a signiﬁcant
result, x20. N = 28) = 0.45, p =0.50.
One-Choice Trials

Combining all one-choice pairing observations from both Michigan and
Ohio, I observed 16 pairings with females that had their blue retained and 21
pairings with females that had their blue reduced (Table 3.2). This difference was
not signiﬁcant, x2(1, N = 37) = 0.41, p =0.52. My Ohio samples for both yellow
and dark morph females were too small to compare using the chi-square test, but
for yellow-morphs, there were 2 pairings with blue-retained females, and 2
pairings with blue-reduced females. For dark-morph females, there were again 2
pairings with blue-retained, and 3 with blue reduced. Examining just the
Michigan population, I observed 11 pairings with females that retained their blue
and 17 pairings with females that had their blue reduced. This result was not
signiﬁcant, x20. N = 28) = 0.26, p =0.61.

50

Table 3.2. Number of one-choice pairings for each population examined. When

male preferences were examined over multiple years for the same population,

results are combined. No difference was signiﬁcant (see text for values).

 

 

 

 

 

 

 

Number of Number of
:30; "I???" Pairings Pairings
Population Species (P Iariicus of with Blue- with Blue-
' g . . Retained Concealed
only) Pairings Female Female
, Yellow 2
Ohio P glaucus
Dark 5
Michigan R NA 28 11 17
canadensis

 

 

 

 

 

 

TOTAL

 

BOTH

 

ALL

 

37

 

16

 

21

 

51

 

 

Discussion

I examined male mate recognition and preference for females with either
their dorsal blue coloration retained or concealed in one- and two-choice trails.
My results suggest that an absence or reduction of dorsal blue coloration on a
female does not reduce a male’s willingness to engage in copulation with her.
Males readily mated with females that had their entire dorsal blue coloration
concealed. This result was surprising as several lines of indirect evidence
suggest that blue coloration may be important for male mate recognition and
possibly mate preference.

Although not signiﬁcant, there may actually have been a slight trend for
males to prefer females that had their blue concealed. In all but two comparisons
(both in Ohio, 'all pairings' and 'yellow-morph, P glaucus females'), the number of
pairings was always greater for females with blue-concealed than for females
with blue-retained. I attribute this to my low sample sizes and believe that an
increase in the number of pairings would eliminate this slight trend. Nonetheless,
it is possible that a signiﬁcant preference for blue-concealed females would be
seen with additional pairings. However, given the original secondary evidence
that initially lead to this investigation, it is unlikely that this would be the case as
this would predict selection pressure for a reduction in the extent of blue
coloration. The continued presence of a large amount of female-limited blue
would, in turn, suggest a strong counter agent selecting for the retention of blue

coloration. Given my results, no such agent is presently known.

52

 

One problem with focusing on a single trait (such as blue coloration) is
that in most systems male mate recognition and preferences are likely to be
based on several concurrent and possibly interacting traits (Bonduriansky, 2001).
The most important factor inﬂuencing insect male mate choice is generally the
fecundity of the female or the number of eggs she has available. Genetic quality
is almost certainly second in terms of selection importance (Bonduriansky, 2001).
If blue were an important characteristic used in mate preference, than it would
likely advertise this second component of fitness, and thus be less'important than
indicators of fecundity such as body size. In the context of my experiments, the
absence of blue coloration (if it had any inﬂuence), may have been
overshadowed by the unmated state of my treated females and their
correspondingly large abdomens.

Likewise, pheromonal cues could also have played a role in male mate
selection. Olfactory cues have an important role in a vast number of
Lepidopteran systems where male mate selection is observed (Bonduriansky,
2001), and may be the most common form of mate assessment. Although
previous pheromonal analysis conducted within the tiger swallowtail system
suggests that pheromones have limited inﬂuence on males (Deering, 1998), this
research did not rule out close-range or contact olfactory inﬂuences on mate
selection.

In general, sexual characteristics which would advertise fecundity are
likely to be redundant and the resources spent on them better applied to

increasing the number or success of offspring (Fitzpatrick et al., 1995). This

53

notwithstanding, when the number of matings a male can achieve is limited
(independent of access to mates) such as is the case with tiger swallowtails, than
female genetic quality can signiﬁcantly impact a male's lifetime ﬁtness
(Bonduriansky, 2001). For instance, individuals may differ in their ability to utilize
locally common hosts, and this may be reﬂected in morphological characteristics.
If a male can recognize this, than it becomes beneﬁcial to mate with females that
exhibit traits which indicate they use hosts well.

Most likely, a combination of visual, olfactory, and tactile cues combine in
some sequence to create an overall impression of a female to males
(Bonduriansky, 2001). If one of these elements is removed, the high favorability
of others may be enough to overcome this deﬁciency. In the end, the presence
of such multiple traits may have lead to an apparent lack of importance for blue
coloration in this system.

Male ‘choice’ may also be expressed cryptically as variation in the amount
of sperm and/or nutrients a male passes on to a particular female (Bonduriansky,
2001). It is possible that male tiger swallowtails may have exhibited this type of
mate choice in response to variation in the presence of blue coloration.

Rather than classiﬁcation as a secondary-sexual characteristic, female
limited blue coloration may be better explained in the context of the Battus
philenor mimicry complex, of which dark-morph females are considered a part of
(Brower, 1958). B. philenor males have extensive blue coloration on their dorsal
hind wings, and are a model for several mimetic species. Mimicking this

coloration likely provides some level of protection from predators, and

accordingly many edible species whose range is congruent have very similar
coloration (Brower and Brower, 1962; Mullen et al., 2008). The underlying
genetics which result in differential dorsal blue coloration are unknown although it
would seem that the genetic regions that contribute to the generation of
extensive female blue coloration are separate from the regions that confer dark
or yellow morph coloration respectively, as both morphs clearly have extensive
blue independent of their base coloration. It is possible that there is strong
selection on dark-morph females to have large amounts of blue coloration and
that this carries over into yellow-morph females. Likewise, this trait may also
either be retained in P canadensis populations from its progenitor or else
maintained by contemporary or historically recent introgression. Evidence from
microsatellites and other traits that do not have strong divergent selection on
them indicate that gene ﬂow on such regions across the hybrid zone is high
(Nowak et al., 2009). Selection for a large amount of blue coloration in dark-
morph females may hence cany over into both yellow morph females and also
into P canadensis populations.

This alternative hypothesis is not without problems. The blue coloration in
tiger swallowtails is strongly W reﬂective whereas the blue coloration of B.
philenor is UV absorbent (Figure 3.2). Most birds have visual acuity which
extends well into the ultraviolet range (Chen et al., 1984). This suggests that
differences in UV reﬂectance between B. philenor and dark-morph, female P
glaucus may be distinguishable to birds and reduce the effectiveness of this

mimicry coloration. Furthermore, ultraviolet coloration in diurnal lepidoptera has

55

been shown to increase predation risk (Cuthill and Bennett, 1993; Majerus et al.,
2000; Lyytinen et al., 2004). Hence this UV coloration may make yellow-morph
females more conspicuous to predators relative to the blue-reduced coloration of
males.

It is clear that additional experimentation is needed to determine both why
female tiger swallowtails retain a greater amount of dorsal blue coloration than
males, as well as what traits males use in mate selection (other than blue) that

resulted in the asymmetrical male matings preferences for P glaucus females.

56

CHAPTER 4:

POSSIBLE ALTERNATIVE EXPLANATIONS FOR ASYMMETRICAL MALE
MATING PREFERENCES IN PAPILIO SPP.

Introduction

In tiger swallowtail butterﬂies, a large amount of secondary evidence
suggests that dorsal blue coloration is an important cue for mate recognition and
possibly mate selection. This evidence includes three observations: 1) dorsal
blue coloration is signiﬁcantly most extensive in female P glaucus relative to both
male P glaucus and all P canadensis (see chapter 2), 2) blue is likely highly
visible to these butterﬂies due to substantial ultraviolet reﬂectance (Briscoe,
2000; chapter 2) and 3) this blue coloration may be a shared characteristic
between the otherwise visually very divergent dark and yellow morph females
that occur in some P glaucus populations (see chapter 2). Combined with
repeated demonstrations that males exhibit mate preferences (Bower, 1959;
Deering and Scriber, 2002, Aardema and Scriber, unpublished data), this
evidence suggests that blue coloration may be important in mate recognition and
possibly mate choice in these butterﬂies.

However, as I have shown (Chapter 3), blue coloration in females is
unlikely to have a signiﬁcant on male mate choice in these tiger swallowtail
populations. Experimental alteration of this trait did not elicit any differential
mating response from wild male populations of P canadensis in Michigan or P

glaucus in Ohio and Florida.

57

 

Two questions arise from this outcome. First, if blue coloration is not
important in mating preferences, does it have other biological signiﬁcance? l
addressed this question in chapter three, postulating that the one explanation for
patterns of variation in blue coloration may be its utility to dark morph P glaucus
females for mimicking the toxic, pipevine swallowtail, Battus philenor. However,
this use may be conﬂicted by the UV component of the blue scaling and the
affect this has on avian predation frequencies. Needless to say, uncertainties
remain.

The second question concerns the traits males are using to choose mates.
Put another way, what else varies between P canadensis and P glaucus females
that may explain why P canadensis males strongly prefer P glaucus females?
(Deering and Scriber, 2002). The list of possible traits may be long, but the most
likely candidates are differences in abdomen size and variation in yellow
coloration. Larger abdomens are a very common trait used by males to select
females in other insect systems due to their direct correlation with fecundity
(Bonduriansky, 2001 ). Female yellow variation is an important part of mate
recognition and preferences in other butterﬂy systems (Lederhouse and Scriber,
1996; Ellers and Boggs, 2003), and would likely be a very conspicuous cue if
utilized by males for the purpose of recognizing and selecting a potential mate.

I postulate that P glaucus females have a fatter (heavier) abdomen
compared to P canadensis females relative to overall size as indicated by wing
length. With regards to yellow coloration, it is hard to conjecture what

characteristics male Papilio spp. prefer, but I hypothesized that some

58

measureable value of yellow coloration will signiﬁcantly differ between P glaucus
and P canadensis females, and that this difference may have inﬂuence on male

preferences.

Materials and Methods
Abdomen Size

To investigate differences in abdomen size I made use of 24 captive-
reared Vermont (Bennington Co.) P canadensis and 24 captive reared
Pennsylvania (Lancaster 00., n=21) or Georgia (Oglethorphe Co., n=3) P
glaucus females. These butterﬂies were all initially reared in 2009. I kept them in
diapause in 4°C over winter (2009/2010) for approximately eight months and then
placed them in 23°C to emerge as adults. After emergence, these butterﬂies
were kept refrigerated until I measured and weighed their abdomens.

I ﬁrst measured the length of the forewings to the nearest 1 mm from base
to tip with a standard metric ruler. This is a common approximate measurement
for butterﬂy size (see chapter 1 for more details). Next, I removed the head
followed by the abdomen at the ‘waist’ where it connects to the thorax. After
removal the abdomen was quickly weighed to the nearest 0.001 gram.

Yellow Coloration

I examined yellow coloration by measuring the spectral reﬂectance
signature of yellow dorsal areas using an Ocean Optics s2000 miniature ﬁber
optic spectrophotometer. I took spectral reﬂectance measurements from 2 -3

separate yellow areas of the forewing from 17 P canadensis females and 17 P

59

 

glaucus females. The minimum area for spectral measurements is approximately
0.75 cm, and hence the areas selected for measurement came from wing regions
that had a solid yellow patch at least this size without any other interfering colors.
Females were collected from Fairbanks-North Star 00., Alaska (n=9, P
canadensis), Charlevoix Co., Michigan (n=8, P canadensis), Jackson and Gallia
Cos., Ohio (n=6, P glaucus) and Levy Co., Florida (n=11, P glaucus).
Reﬂectance of color was measured in comparison to a white standard. Using the
reﬂectance values, I was able to calculate hue and chroma for each observation.
Hue is essentially the color we recognize and consider as distinctive (e.g.
“yellow”, "blue", etc.), and chroma is a measure of the saturation or amount of
diffusion of this color. These values were calculated based on Endler (1990),
with modiﬁcations from Sobel et al., (In Prep.). The mean of multiple
measurements from single individuals was used in species and population
comparisons.
Statistical Analysis

Abdomen size (as determined by weight), was divided by wing length to
get a relative measurement of the relationship between the two values for P
canadensis and P glaucus females. Using these measurements, I compared the
variances using an F test (to determine homogeneity) and then Students t-test to
determine it the means were statistically different for the two species.

I also analyzed the covariance between wing length and abdomen weight
using species as a cofactor. I compared three models: a simple model that did

not take species into account, an additive model with species as a cofactor and a

60

 

multiplicative model, again with species as a cofactor. Models were compared
using adjusted coefﬁcient of determination (Adj. R2), the Akaike information
criterion (AIC; Akaike, 1978) and the Schwarz Bayesian lnfonnation Criterion
(BIC; Schwarz, 1978).

Yellow coloration was compared for both hue and chroma between the two
species using either Student’s t test when variances were statistically equal or the
Welch two sample t test when variances were unequal. Variance was examined
using an F test. Hue and chroma were also compared between populations
using an analysis of variance (ANOVA), followed by Tukey’s honest signiﬁcant
difference (HSD) for painivise comparisons. For all analysis an alpha value <

0.05 was considered signiﬁcant

Results
Abdomen Size

Summary statistics for the relative comparison between abdomen weight
and wing length are displayed in table 4.1. Variances between P canadensis
and P glaucus were homogeneous. There was a statistically signiﬁcant
difference between these two species ['I’(46)=-4.433,p<0.001].

In contrast to the relative abdomen size comparison described above, my
analysis of covariance revealed that there was not a signiﬁcant effect of
population for my regression models (table 4.2, ﬁg. 4.1). The slopes and y-

intercepts are statistically equivalent, and the simple model which does not take

61

Table 4.1. Summary statistics for the relative relationship between forewing

length (mm) and abdomen weight (g) in female P glaucus and P canadensis.

 

 

 

 

 

 

 

 

Species Observations Mean Standard
DeVIation
P canadensis 24 0.450 0. 11 5
P glaucus 24 0.628 0.159

 

 

62

Table 4.2. Results from analysis of covariance (ANCOVA) for a comparison
between P canadensis and P glaucus females for wing length and abdomen
weight with species as a cofactor. The simple model which does not take into
account species appears to have the greatest predictive value and suggests that
the two butterﬂies do not differ signiﬁcantly in their basic allometric relationship
between wing length and abdomen weight. Bolded numbers represent the
values for the preferred model in the three different analyses (Adj. R2, AIC and

BIC). See text for more details.

 

 

 

 

 

 

 

Model Species Model Equation Ag! AIC BIC
. weight = 0.016 * _ _
Simple NA wi n g_l e n gth _ 0.536 0.597 263.01 119.17
P weight = 0.015 *
d ' ' I h - .4
Additive cane 9"” w'"9- engt 0 69 0.591 -261.39 -115.69
weight = 0.015 *
P' g’aucus wing_length - 0.451
P weight = 0.012 *
. . _ canadensis wing_length - 0.337
Multiplicative _ ht 0 017 0.588 -260.12 -112.55
weig = . *
P’ glaucus wing_length - 0.580

 

 

 

 

 

 

 

 

63

 

 

 

 

 

 

 

A
0.45— A P. glaucus
. A
o P. canadensrs ‘ A t
A 040- ,
39 A A
4—11 ‘ A
g) 0 35 ~ ‘ a
.5 o
A
g 0 30 1 0 A
C O
G) O o O 0
_. 0
g 025 o A
‘C _ O O
.o 020 o o
< o 8 8 o ‘ A
o 15 — o o o o
O
A
0-10 T l i i i
45 50 SS 60

Wing Length (mm)

Figure 4.1. Comparison of wing length (mm) and abdominal weight (g) in P
canadensis and yellow morph, P glaucus females. Regression lines are drawn
using the simple model (See table 4.2). The slopes are not signiﬁcantly different.

See text for more details.

species into account was determined to be the best based on adjusted R2, AIC
and BIC values.
Yellow Coloration

Summary statistics for species’ hue and chroma comparisons are
presented in tables 4.3 a&b, and for population comparisons in tables 4.4 a&b.
The variance between P glaucus and P canadensis for hue was homogeneous
and the difference was signiﬁcant [T(32)=4.463,p<0.001]. For chroma however,
the variances were not homogonous, and the means were not signiﬁcantly
different [T(19.654)=-1.217,p=0.238].

In line with the results found in my species analysis, analysis of variance
showed that there was a signiﬁcant difference between populations for hue
[F(3,30)=9.998,p<0.001], but not for chroma [F(3,30)=0.998,p=0.407]. For hue,
observed signiﬁcance was the result of differences between the Florida P
glaucus samples and samples from the two P canadensis populations (Alaska
and Michigan). The P canadensis samples were not signiﬁcantly different from
the Ohio P glaucus samples (P values from Tukey’s HSD are presented in table

4.5).

Discussion

I examined two possible morphological differences between P canadensis
and P glaucus that may explain asymmetrical male preferences for P glaucus
females relative to P canadensis females. These were abdomen weight (as a

proxy for ‘size’) and yellow coloration. I found that P glaucus females have

65

Table 4.3a. Summary statistics for hue value in female P glaucus and P

 

 

 

canadensis
. . Standard
Species Observations Mean Deviation
P canadensis 17 0.245 0.022
P glaucus 17 0.265 0.066

 

 

 

 

 

 

Table 4.3b. Summary statistics for chroma value in female P glaucus and P

 

 

 

canadensis
. . Standard
Species Observations Mean Deviation
P canadensis 17 1.276 0.082
Pglaucus 17 1.123 0.115

 

 

 

 

 

 

66

Table 4.4a. Summary statistics for hue value in populations of female P

canadensis (Alaska and Michigan) and P glaucus (Ohio and Florida)

 

 

 

 

 

 

 

 

 

. . Standard

Population Observations Mean Deviation
Alaska 9 1 .266 0.084
Michigan 8 1.287 0.085
Ohio 6 1.202 0.109
Florida 11 1 .08 0.096

 

Table 4.4b. Summary statistics for chroma value in populations of female P

canadensis (Alaska and Michigan) and P glaucus (Ohio and Florida).

 

 

 

 

 

 

. . Standard

Population Observations Mean Deviation
Alaska 9 0.246 0.026
Michigan 8 0.243 0.019
Ohio 6 0.245 0.028
Florida 11 0.276 0.078

 

 

 

 

67

 

 

Table 4.5. Signiﬁcance (p) values from Tukey’s HSD for all painivise comparisons

between populations in the yellow color analysis between P canadensis (Alaska

and Michigan), and P glaucus (Ohio and Florida). Values on the lower left side

of the ﬁgure represent chroma comparisons and values on the upper right side of

the ﬁgure represent hue comparisons. Bolded values in the dark gray boxes are

signiﬁcant (p50.05). See text for more details.

 

 

 

Florida

 

 

 

5‘ .9" It,~.'\'.v“.~.“'. ’3 .‘l
- ' _ ‘-t'i‘ “37‘s.," .')}I.

' 7 ' 1.1:

. .7 ..

 

 

 

 

 

 

enH

 

 

Alaska Michigan Ohio
Alaska 0.965 0.571
Michigan 0.999 0.346
Ohio 1.000 1.000
Florida 0.545 0.483 0.613
Chroma

68

relatively heavier abdomens (in relation to wing length), but an absolute
comparison of the two traits revealed a similar allometric relationship between
wing length and abdomen weight for both species. This suggests that similar
sized individuals (as indicated by wing length) should have similar abdomen
sizes regardless of species. Why then the relative differences were signiﬁcant is
unclear. One explanation may be that the observations for each species were
not spread out evenly across the complete distribution, but rather most of the P r.
glaucus observations occurred at longer wing lengths and heavier abdomens

weights, whereas most of the P canadensis observations were for shorter wing

 

lengths and correspondingly lighter abdomen weights.

For yellow, the mean hue value was signiﬁcantly different between P
canadensis and P glaucus females, but not chroma. This was mostly due to
Florida P glaucus observations deviating from the other three populations. This
suggests that the fundamental ‘color’ is different between the two species, or at
least between Florida and the other populations. Color has immense importance
for many other components of ﬁtness (e.g. thennoregulation; Ellers and Boggs,
2003), and how males respond to females of different colors is unknown at this
time.

To that end, both of these examinations are extremely preliminary and
should be considered mostly in the context of future areas for study.
Nonetheless, in both cases results appear promising for potential differences
between the two species that may help to explain male mating preferences. Of

the two, abdomen size is arguably the more likely candidate for a mate

69

preference cue, even if it is not strongly different between the two species. A
large abdomen likely would suggest more eggs and correspondingly more
potential offspring to a male, or else larger eggs that may result in offspring more
likely to survive the eariy stages of their life (Fischer et al. 2002). It may be
possible that a female with a smaller abdomen would be less pursued by males
or even rejected outright. Careful study of how these two traits vary between
Papilio species and populations as well as how they inﬂuence male mate
preferences will likely reveal a great deal of useful information regarding both the
tiger swallowtail butterﬂies speciﬁcally and the evolution of mating systems in

general.

70

CHAPTER 5:
SUMMARY AND CONCLUSIONS

The divergence of tiger swallowtail butterﬂies is a fascinating example of
how many evolutionary inﬂuences can interact to generate new species.
Vicariance, sexual selection, adaptation to divergent climates, host use
divergence, mimicry and hybridization have likely all played important roles in the
diversiﬁcation of this small group.

I examined one component of sexual selection in tige‘r swallowtails in the
hope of better understanding the role of male mate recognition and preferences
in their diversiﬁcation. Speciﬁcally, I examined variation in coloration between
two common species, the Canadian tiger swallowtail, Papilio canadensis and the
eastern tiger swallowtail, P glaucus.

I found that the amount of dorsal blue coloration in females of P glaucus
was signiﬁcantly more extensive then this same coloration on male P glaucus
and both male and female P canadensis. I also found that this coloration was
strongly UV reﬂective; making it likely that it is highly visible to other tiger
swallowtail individuals.

Previous research found that males of both species preferred to mate with
P glaucus females when given the choice between these and similarly sized P
canadensis females. This observation, combined with my ﬁnding that P glaucus
females have a large amount of blue coloration, suggested that this blue was a
good candidate for a trait males may use to distinguish and select females. I

investigated this possibility using one- and two-choice mate preference

71

experiments and found to my surprise that males do not appear to select mates
based on the extent of their dorsal blue coloration

Perhaps the most important lesson that can be drawn from this work is the
necessity of conducting rigorous experimentation before assigning utility to a
morphological characteristic. Certainly, in most cases the hypothesized function
of a trait will be found to be correct, but there may also be other, ancillary
functions or, as in the case of blue coloration on tiger swallowtail females, the
postulated purpose will turn out to be completely incorrect.

Another interesting ﬁnding of this work is in the illumination of areas where
there is clear morphological divergence between P canadensis and P glaucus.
Although it was qualitatively recognized that blue coloration appeared to be more
extensive in P glaucus females, this observation has now been quantiﬁed and
veriﬁed. Additionally, it was previously unknown how these butterﬂies appeared
in the ultraviolet spectrum, something which has now been assessed.

Although I ultimately disproved the importance of dorsal blue coloration in
male mate selection, I have opened up many paths for future research on the
inﬂuence of blue in other areas such as mimicry, thennoregulation and sexual
conﬂict with conspeciﬁcs or other species. It may be particularly illuminating to
investigate the apparent inconsistency between the visual blue coloration and the
UV coloration of dark morph P glaucus females. Furthermore, the problem of
what traits males were cuing in on that generated overwhelming preference for P

glaucus females remains. I have shown that variation in yellow coloration and

72

abdomen size may be good candidates to explain this, but both remain to be
tested.

Overall, tiger swallowtails compose a diverse, complicated and fascinating
clade that have the potential to maintain scientiﬁc inquiry for many years to
come. There is little doubt that this inquiry will lead to a better understanding of
how variation in organic beings inﬂuences and is inﬂuenced by evolutionary
processes. Likewise, it will generate much general information that will

contribute to our knowledge of the natural world.

73

Appendix 1

Record of Deposition of Voucher Specimens“

The specimens listed on the following sheet(s) have been deposited in the named museum(s) as
samples of those species or other taxa, which were used in this research. Voucher recognition
labels bearing the Voucher No. have been attached or included in ﬂuid-preserved specimens.

Voucher No.: 2010-06
Title of thesis or dissertation (or other research projects):
THE INFLUENCE OF FEMALE WING COLOR VARIATION ON MALE MATE CHOICE IN
TIGER SWALLOWTAIL BUTTERFLIES (PAPILIO SPP.)
Museum(s) where deposited and abbreviations for table on following sheets:
Entomology Museum, Michigan State University (MSU)

Other Museums:

lnvestigator’s Name(s) (typed)
Matthew L. Aardema

 

 

Date AUG. 3. 2010

*Reference: Yoshimoto, C. M. 1978. Voucher Specimens for Entomology in North America.
Bull. Entomol. Soc. Amer. 24: 141-42.

Deposit as follows:
Original: Include as Appendix 1 in ribbon copy of thesis or dissertation.

Copies: Include as Appendix 1 in copies of thesis or dissertation.
Museum(s) ﬁles.
Research project ﬁles.

This form is available from and the Voucher No. is assigned by the Curator, Michigan State
University Entomology Museum.

74

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pages

 

 

 

of1

 

 

 

Appendix 1.1

 

Voucher Specimen Data

Page 1

 

 

75

 

 

 

 

 

 

 

 

 

 

 

 

 

\v - mumm— .5
QSQXSSAV N Somme: 98
B. Em
E2025 99m $9222 65 c_ .883
Le mcmEeeam 02m: 06% 85 emzcomm mEmEm< 32:22
8-28 .62 6898 @6862 8969882.
Locoso>
$6380: : mucosa 6:03:33 32
Lentom .52.. 38-57..
9.3 5:22 =N§>o=m>>m .69... cmﬁmcmo
Ems. m .00 cmmxonwco Z<O_IO=2 wsmtmbmcmo 0&le
xomEom .1 88-5%-:
281 eummucmw __E>>o__m>>mcloq_._. cmﬁmcmo
392 m .00 coHdEEmm h2021m> $959250 0:..le
Lentom .52.. moomicm—zém
823cm 9me seem >3. ammo .530:mb LcmHEEmmm
392 m .00 >>m._ <D_mO.E 96sz OLLLQmQ
Lmntow .52.. ooowem/xé
mzommm 99m 920w >9. .muco €50:mb .03. 59me
:ms. a co 56.. 3&9“. 635% Siam
m w 10 0+ 8 e
u are ..u are s s m “hp a 5 “32.836 new new:
e e s h u n P w 9 . coxmtcﬁo Lo 36on
sh t u u u _ m 9 Lo moooomcmEGcamB m cm .
me o M M P Wu E a... . EBA:
so .8252

LITERATURE CITED

Abramoff MD, PJ Magelhaes, SJ Ram (2004) Image processing with ImageJ.
Biophotonics lntemational 11:36-42.

Acorn J (2002) Assessing ultraviolet reﬂections of Iepidopterans with video and
digital cameras. News of the Lepidopterists’ Society 44:60.

Akaike H (1978) A Bayesian analysis of the minimum AIC procedure. Annals of
the Institute of Statistical Mathematics 3029-14.

Andolfatto P, JM Scriber and B Charlesworth (2003) No association between
mitochondrial DNA haplotypes and a female-limited mimicry phenotype in
Papilio glaucus. Evolution 57:305-316.

Arms K, P Feeny, and RC Lederhouse (1974) Sodium: stimulus for puddling
behavior by tiger swallowtail butterﬂies, Papilio glaucus. Science 185:372-
374.

Ayres MP and JM Scriber (1994) Local adaptation to regional climates in Papilio
canadensis (Lepidoptera: Papilionidae). Ecological Monographs 64:465-
482.

Beck J and K F iedler (2009) Adult life spans of butterﬂies (Lepidoptera:
Papilionoidea + Hesperioidea): broadscale contingencies with adult and
larval traits in multi-species comparisons. Biological journal of the Linnean
Society 96:166-184.

Beldade P and PM Brakeﬁeld (2002) The genetics and evo—devo of butterﬂy wing
patterns. Nature Reviews Genetics 3:442-452

Bissoondath CJ and C Wiklund (1996) male butterﬂy investment in successive
ejaculates in relation to mating system. Behavioral
Ecology and Sociobiology 39:285-292.

Boggs CL, WB Watt, and PR Ehrlich (2003) Butterﬂies: Ecology and Evolution
Taking Flight. The University of Chicago Press, Chicago, USA.

Bonduriansky R (2001) The evolution of male mate choice in insects: a synthesis
of ideas and evidence. Biological Review 76:305-309.

Bosi SG, J Hayes, MCJ Large and L Poladian (2008) Color, iridescence and
thennoregulation in Lepidoptera. Applied Optics 47:5235-5241.

Bossart JL and JM Scriber (1995) Maintenance of ecologically signiﬁcant genetic

76

 

variation in the tiger swallowtail butterﬂy through differential selection and
gene ﬂow. Evolution 49: 1163-1 171 .

Briscoe AD (2000) Six opsins from the butterﬂy Papilio glaucus: molecular
phylogenetic evidence for paralogous origins of red-sensitive visual
pigments in insects. Journal of Molecular Evolution 51 :110-121.

Brower JVZ (1958) Experimental studies of mimicry in some North American
butterﬂies: part II. Battus philenor and Papilio Troilus, P polyxenes, and P
glaucus. Evolution 12:123-136.

Brower LP (1959) Speciation in butterﬂies of the Papilio glaucus group. II.
Ecological relationships and interspeciﬁc sexual behavior. Evolution
1 3:21 2-228.

Brower LP and JVZ Brower (1962) Relative abundance of model and mimic
butterﬂies in natural populations of Battus philenor mimicry complex.
Ecology 43: 154-158.

Brunton CFA and MEN Majerus (1995) Ultraviolet colours in butterﬂies: intra- or
inter-speciﬁc communication? Proceedings of the Royal Society B
260:199-204.

Chamberlin NL, RI Hill, DD Kapan, LE Gilbert, and MR Kronforst (2009)
Polymorphic butterﬂy reveals the missing link in ecological speciation.
Science 326:847-850.

Chen D, JS Collins, and TH Goldsmith (1984) The UV receptor of bird retinas.
Science 225:337-339.

Church SC, ATD Bennett, IC Cuthill and JC Partridge (1998) Ultraviolet cues
affect the foraging behaviour of blue tits. Proceedings of the Royal Society
B 265:1509-1514.

Cuthill IC and ATD Bennett (1993) Mimicry and the eye of the beholder.
Proceedings of the Royal Society B 243:203-204.

Deering MD (1998) Preferentail mate selection by males as a reproductive
isolating mechanism between swallowtail species; Papilio glaucus and P
canadensis (Lepidoptera, Papilionidae). MS Thesis, Michigan State
University.

Deering and Scriber (1998) Heterospeciﬁc mating behavior of Papilio palamedes
in Florida (Lepidoptera: Papilionidae). Holarctic Lepidoptera 5:49-51.

Deering MD and JM Scriber (2002) Field bioassays show heterospeciﬁc mating

77

preference asymmetry between hybridizing North American Papilio
butterﬂy species (Lepidoptera: Papilionidae) Journal of Ethology 20:25-33.

Douglas MM (1981) Therrnoregulatory signiﬁcance of thoracic lobes in the
evolution of insect wings. Science 211 :84-86.

Ehrlich PR and SE Davidson (1960) Techniques for capture-recapture studies of
Lepidoptera populations. Journal of the Lepidopterists Society 14:227-
229.

Ellers J and CL Boggs (2003) The evolution of wing color: male mate choice
opposes adaptive wing color divergence in Colias butterﬂies. Evolution
57:1100—1106.

Endler JA (1990) On the measurement and classiﬁcation of colour in studies of
animal colour patterns. Biological Journal of the Linnean Society 41 :315—
352.

Fischer K, BJ Zwaan and PM Brakeﬁeld (2002) How does egg size relate to body
size in butterﬂies? Oecologia 131:375-379.

Fitzpatrick S, Berglund A, and G Rosenqvist (1995) Ornaments or offspring: costs
to reproductive success restrict sexual selection processes. Biological
Journal of the Linnaean Society 55:251-260.

Ghiradella H (1991) Light and color on the wing- structural colors in butterﬂies
and moths. Applied Optics 30:3492-3500.

Grant PR (1986) Ecology and evolution of Danivin’s ﬁnches. Princeton University
Press, Princeton, NJ, USA.

Grbning J and A Hochkirch (2008) Reproductive Interference Between Animal
Species. The Quarterly Review of Biology 83:257-282.

Hagen RH (1990) Population structure and host use in hybridizing subspecies of
Papilio glaucus (Lepidoptera: Papilionidae). Evolution 44:1914-1930.

Hagen RH and JM Scriber (1989) Sex-linked diapause, color, and allozyme loci
in Papilio glaucus glaucus- linkage analysis and signiﬁcance in a hybrid
zone. Journal of Heredity 802179-185

Hagen RH and JM Scriber (1991) Systematics of the Papilio glaucus and P
Troilus species groups (Lepidoptera, Papilionidae)- inferences from
allozymes. Annals of the Entomological Society of America 84:380-395.

Hagen RH and JM Scriber (1995) Sex chromosomes and specition in the Papilio

78

glaucus group. In: Scriber JM, Y Tsubaki, and RC Leaderhouse (eds) The
Swallowtail Butterﬂies: Their Ecology and Evolutionary Biology, Scientiﬁc
Publishers, Gainesville, FL, USA, pp 211-228.

Hastings IM (1994) Manifestations of sexual selection may depend on the
genetic basis of sex determination. Proceedings of the Royal Society B
258:83—87.

Heegaard E (1997) Morphological variation within Andreaea blyttii in relation to
the environment on Hardangervidda, Western Nonivay: a quantitative
analysis. The Bryologist 100:308-323.

Hereau H and JM Scriber (2003) Interspeciﬁc copulation of a dark morph Papilio
glaucus female and a male P polyxenes (Papilionidae): observation and
signiﬁcance. Journal of the Lepidopterists Society 57:71-74.

Holm S (1979) A Simple Sequentially Rejective Bonferroni Test Procedure.
Scandinavian Journal of Statistics, 6:65 -70.

Hunt J, MW Blows, F Zajitschek, MD Jennions and R Brooks (2007) Reconciling
strong stabilizing selection with the maintenance of genetic variation in a
natural population of black ﬁeld crickets (Teleogryllus commodus).
Genetics 177:875-880.21

Jemvall J, JP Hunter and M Fortelius (1996) Molar tooth diversity, disparity, and
ecology in Cenozoic ungulate radiations. Science 274:1489-1492.

Kaufman PK, FD Enﬁeld and RE Comstock (1977) Stabilizing selection for pupa
weight in Tribolium castaneum. Genetics 87:327-341.

Kotanen PM and J Bergelson (2000) Effects of simulated grazing on different
genotypes of Bouteloua gracilis: how important is morphology?
Oecologica 123:66-74.

Krebs RA (1988) The mating behavior of Papilio glaucus (Papilionidae). Journal
of Research on the Lepidoptera 26:27-31.

Kunte K, C Shea, ML Aardema, JM Scriber, PAndolfatto, T Juenger, LE Gilbert,

and MR Kronforst (In Prep.) Sex chromosome mosaicism and hybrid
speciation among tiger swallowtail butterﬂies. Science

Landini G Threshold_Colour, Version 1.12 for ImageJ. Available from
http:llwww.dentistiy.bham.ac.ukllandiniglsoftware. Accessed on 10 May,
2010.

Lederhouse RC (1995) Comparative mating behavior and sexual selection

79

In North American swallowtail butterﬂies. In: Scriber JM, Y Tsubaki and RC
Lederhouse (eds) Swallowtail Butterﬂies: Their Ecology and Evolutionary
Biology. Scientiﬁc Publishers, Gainesville, FL, USA.

Lederhouse RC and JM Scriber (1996) Intrasexual selection constrains the
dorsal color pattern of male black swallowtail butterflies, Papilio polyxenes.
Evolution 50:717-722.

Lederhouse RC, MP Ayres, and JM Scriber (1989) Evaluation of spennatophore
counts in studying mating systems of Lepidoptera. Journal of the
Lepidopterists Society 43293-101.

Lederhouse RC and JM Scriber (1987) Ecological signiﬁcance of a postrnating
decline in egg viability in the tiger swallowtail. Journal of the Lepidopterists
Society 41:83-93.

Levin MP (1973) Preferential mating and the maintenance of the sex-limited
dimorphism in Papilio glaucus: evidence from laboratory matings.
Evolution 27:257-264.

Lewis 2 and N Wedell (2007) Effect of adult feeding on male mating behaviour in
the butterﬂy, Bicyclus anynana (Lepidoptera: Nymphalidae). Journal of
Insect Behavior 202201-213.

Linhart YB and MC Grant (1996) Evolutionary signiﬁcance of local genetic
differentiation in plants. Annual Review of Ecology and Systematics
27:237-277.

Luebke HJ, JM Scriber, and BS Yandell. (1988) Use of multivariate discriminant
analysis of male wing morphometrics to delineate a hybrid zone for Papilio
glaucus glaucus and P g. canadensis in Wisconsin. American Midland
Naturalist 119:366-379.

Lyytinen A, L Lindstrom, and J Mappes (2004) Ultraviolet reﬂection and predation
in diurnal and nocturnal Lepidoptera. Behavioral Ecology 15:982-987.

McLain DK and AE Pratt (1999) The cost of sexual coercion and heterospeciﬁc
sexual harassment on the fecundity of a host-speciﬁc, seed-eating insect
(Neacoryphus bicrucius) Behavioral Ecology and Sociobiology 46: 164-

1 70.

Majerus MEN, CFA Brunton, and J Stalker (2000) A bird’s eye view of the
peppered moth. Journal of Evolutionary Biology 13:155-159.

Mallet J and M Joron (1999) Evolution of diversity in warning color and mimicry:

80

polymrophisms, shifting balance and speciation. Annual Review of
Ecology and Systematics 30:201-233.

Martinsen GD, TG Whitham, RJ Turek, and P Keim (2001) Hybrid populations
selectively ﬁlter gene introgression between species. Evolution 5521325-
1335.

Mayr E (2005) Fonrirard. In: Hallgn’msson B and BK Hall (eds) Variation: a central
concept in biology. Elsevier Academic Press, Burlington, MA, USA, p xvii.

Mercader RJ, ML Aardema, and JM Scriber (2009) Hybridization leads to host
use divergence in a polyphagous butterﬂy sibling species pair. Oecologia
1 58:651-662.

Mercader RJ and JM Scriber (2007) Diversiﬁcation of host use in two
polyphagous butterﬂies: differences in oviposition speciﬁcity or host rank
hierarchy?

Mercader RJ and JM Scriber (2008a) Divergence in the ovipositional behavior of
the Papilio glaucus group. Insect Science 15:361-367.

Mercader RJ and JM Scriber (2008b) Asymmetrical thermal constraints on the
parapatric species boundaries of two widespread generalist butterﬂies.
Ecological Entomology 33:537-545.

Morris WA (1952) The chipmunk as a predator of the adult yellow swallowtail
butterﬂy. Journal of Mammalogy 34:510-511.

Mullen SP, EB Dopman, and RG Harrison (2008) Hybrid zone origins, species
boundaries, and the evolution of wing-pattem diversity in a polytypic
species complex of North American admiral butterﬂies (Nymphalidae:
Limenitis). Evolution 62: 1400-141 7.

Naisbit RE, CD Jiggins, and J Mallet (20010 Disruptive selection against hybrids
contributes to speciation between Heliconius cydno and Heliconius
melpomenei. Proceedings of the Royal Society B 268:1849-1854.

Nielsen MG and WB Watt (2000) Interference competition and sexual selection
promote polymorphism in Colias (Lepidoptera, Pieridae). Functional
Ecology 14:718-730.

Nowak C, CM Brown, and JJ Hellmann (2009) Microsatellite markers for the
hybridizing tiger swallowtails, Papilio glaucus and P canadensis, and their
applicability to historic and congeneric species. Molecular Ecology
Resources 9:800-803.

81

Otis GW, B Locke, NG McKenzie, D Cheung, E MacLeod, P Careless, and A
Kwoon (2006) Local enhancement in mud-puddling swallowtail butterﬂies
(Battus philenor and Papilio glaucus). Journal of Insect Behavior 19:685-
698.

Panhuis TM, R Butlin, M Zuk, and T Tregenza (2001) Sexual selection and
speciation. Trends in Ecology & Evolution 16:364-371.

Peterson MA, BM Honchak, SE Locke, TE Beeman, J Mendoza, J Green, KJ
Buckingham, MA White, and KJ Monsen (2005) Relative abundance and
species-speciﬁc reinforcement of male mating preference in the
Chrysochus (Coleoptera: Chrysomelidae) hybrid zone. Evolution 5922639-
2655.

Platt AP, SJ Harrison, and TF \Mlliams (1984) Absence of differential mate
selection in the North American Tiger Swallowtail Papilio glaucus. In:
Vane-Wright, RI and PW Ackery (eds) The Biology of Butterﬂies. Academic
Press, London. pp. 245-250.

Prum RO, T Quinn and RH Torres (2006) Anatomically diverse butterﬂy scales all
produce structural colours by coherent scattering. The Journal of
Experimental Biology 209:748-765.

Putnam AS, JM Scriber, and P Andolfatto (2007) Discordant divergence times
among z-chromosome regions between two ecologically distinct
swallowtail butterﬂy species. Evolution 61 :912-927.

Rawson GW and PF Bellinger (1953) Sparrows feeding on congregating Papilio.
Lepidopterists’ News 7:27.

Ritland DB and JM Scriber (1985) Larval development rates of 3 putative
subspecies of tiger swallowtail butterﬂies, Papilio glaucus and their hybrids
in relation to temperature. Oecologia 65:185-193.

Saint-Laurent R, M Legault and L Bematchez (2003) Divergent selection
maintains adaptive differentiation despite high gene ﬂow between
sympatric rainbow smelt ecotypes (Osmerus mordax Mitchill). Molecular
Ecology 12:315-330.

Schulter D (2000) The ecology of adaptive radiation. Oxford University Press
Inc., New York, USA.

Schwarz G (1978) Estimating the dimension of a model. Annals of Statistics
6;461-464.

Scriber JM (1988) Tale of the tiger: beringial biogeography, binomial

82

classiﬁcation, and breakfast choices in the Papilio glaucus complex of
butterﬂies. In: Spencer K (ed) Chemical Mediation of Coevolution.
Academic Press, San Diego, CA, USA, pp 241-301.

Scriber JM (1990a) Interaction of introgression from Papilio glaucus canadensis
and diapause in producing spring form eastern tiger swallowtail butterﬂies,
P glaucus (Lepidoptera, Papilionidae). Great Lakes Entomologist 23:127-
1 38.

Scriber JM (1990b) Two new aberrant forms of tiger swallowtail butterﬂy from the
great lakes hybrid/transition zone (Lepidoptera: Papilionidae). The Great
Lakes Entomologist 23:121-126.

Scriber (1995) Overview of swallowtail butterﬂies: taxonomic and distributional
latitude. In: Scriber JM, Y Tsubaki and RC Lederhouse (eds) Swallowtail
Butterﬂies: Their Ecology and Evolutionary Biology. Scientiﬁc Publishers,
Gainesville, FL, USA.

Scriber JM (1996a) Tiger tales: natural history of native North American
swallowtails. American Entomologist 42:19-32.

Scriber JM (1996b) A new ‘cold pocket’ hypothesis to explain local host
preference shifts in Papilio canadensis. Entomologia Experimentalis et
Applicata 80:315-319.

Scriber JM (2002) Latitudinal and local geographic mosaics in host preferences
as shaped by thermal units and voltinism in Papilio spp. (Lepidoptera:
Papilionidae). European Journal of Entomology 99:225-239.

Scriber JM and M Ayres (1988) Observations on the puddling behavior of the
Canadian tiger swallowtail butterﬂy Papilio glaucus canadensis R&J in
northern Michigan. Great Lakes Entomologist 21 :141-142.

Scriber JM, RH Hagen and RC Lederhouse (1996) Genetics of mimicry in the
tiger swallowtail butterﬂies, Papilio glaucus and P canadensis
(Lepidoptera: Papilionidae). Evolution 50:222-236.

Scriber JM, K Keefover, and S Nelson (2002) Hot summer temperatures may
stop movement of Papilio canadensis butterﬂies and genetic introgression
south of the hybrid zone in the North American Great Lakes Region.
Ecography 20225-33.

Scriber JM, RL Lindroth, and J Nitao (1989) Differential toxicity of a phenolic

glycoside from quaking aspen to Papilio glaucus butterﬂy subspecies,
hybrids, and backcrosses. Oecologia 81 :186-191.

83

Scriber JM, GL Lintereur and MH Evans (1982) Foodplant suitabilities and a new
oviposition record for Papilio glaucus canadensis (Lepidoptera:
Papilionidae) in northern Wisconsin and Michigan. The Great Lakes
Entomologist 1 5: 39-46.

Scriber JM and GJ Ording (2005) Ecological speciation without host plant
specialization; possible origins of a recently described cryptic Papilio
species. Entomologia Experimentalis et Applicata 115:247-263.

Scriber JM, GJ Ording and RJ Mercader (2008) lntrogression and parapatric
speciation in a hybrid zone. In: Tilmon KJ (ed) Specialization Speciation
and Radiation: The Evolutionary Biology Herbivorous Insects. University of
California

Stump AD, FAH Sperling, A Crim, and JM Scriber (2003) Gene ﬂow between
great lakes region populations of the Canadian tiger swallowtail butterﬂy,
Papilio canadensis, near the hybrid zone with P glaucus (Lepidoptera:
Papilionidae). Great Lakes Entomologist 36: 41-53.

Svensson El, J Abbott and R Hardling (2005) Female polymorphism, frequency
dependence and rapid evolutionary dynamics in natural populations. The
American Naturalist 165:567-576.

Takahashi Y and M Watanabe (2010) Female reproductive success is affected by
selective male harassment in the damselﬂy Ischnura senegalensis. Animal
Behaviour 79:211-216.

Thomas JA, JJ Welch, R Lanfear and L Bromham (2010) A generation time effect
on the rate of molecular evolution in invertebrates. Molecular Biology and
Evolution 27:1173-1180.

Tinbergen N (1948) Social releasers and the experimental method required for
their study. The Wilson Bulletin 60:6-51.

Tremblay RL, JD Ackerrnan, JK Zimmerman, and RN Calvo (2005) Variation in
sexual reproduction in orchids and its evolutionary consequences: a
spasmodic journey to diversiﬁcation. Biological Journal of the Linnean
Society 84:1-54.

Turelli M and NH Barton (2004) Polygenic variation maintained by balancing
selection: pleiotropy, sex-dependent allelic effects and G x E interactions.
Genetics 166:1053-1079.

Winter CB and AH Porter (2010) AF LP linkage map of hybridizing swallowtail
butterﬂies, Papilio glaucus and Papilio canadensis. Journal of Heredity
1 01 :83-90.

Xu J and Q Wang (2009) Male moths undertake both pre- and in-copulation mate
choice based on female age and weight. Behavioral Ecology and
Sociobiology 63:801-808.

Yeaman S and A Jarvis (2006) Regional heterogeneity and gene ﬂow maintain

variance in a quantitative trait within populations of lodgepole pine.
Proceedings of the Royal Society B 273:1587—1593.

85

 

"'ilillljlllij @11lein ljillljlijljijlllies