z .;

J....uhn§.:.~w.xm.rnm
in? x 31:.
Wfiufih

‘ 1 .. 2.. J.»-
’A.V :3: 2..
3.... .v t.
:1!

5.3.1

2'!
34.1% .
.e.

.n. a...
2

nluauu

35:11:19....

1!: .

,¥LMMJ.man: _.‘ ..,.
I

L.
a :2 I) I. . .
a! . 3. z 3...: li!!!...
(«Liar < . 1.19:}
13:27... 1.35
than. .

g.

”gnu-ah.
3.05
brash-

g

infirm“:
91

6hr.
.%

ibmshammnwm , ".2. h.
fimmuulwfinflwpwgw ...
.iflnufiubg
‘1." o

. zzufiumw
1...- .3...
mmufinda c Hugs
3»
“II-oninvfin

12.4. at.)
.nxi IMF”.
an" .

I .1
It}! I)! up! 11.5%.... [thrifltf
3145mm? gaiturm
. .94. t 2:...
.adafir... .

1" l .0
:I!S. : 05.5.55.
3!. .5 3.1».iili ityl§.~l.i.
.u

t 112...?! I

.3 251:7...
(ft... 2,155 . ,Jflao
I..fi.le..:ann ‘ .alib u. .

5.49

0.273.133 11.52!
.3, :1...

£34 3.1.. €8.11).

Ell“... fink. our

374......
1:6... :7

:

 

:2: 71...}.
A} $12? :.
‘13:

It}
. 1..) 23193;?
t 4!}..7 : ‘3 III.-
xllfixininiahutriuias H
I... nit... .9)’.t}?.1¢.{
.. fig giant.” {:3 is:
.vrllvt . h . . . ..

xrsllu
(551.1: 33:81:53.3”
. 0!? 54.; 1...”.2!
3|: :9.
litafisl
.I.\5tl.‘tal§n
if‘}? 11‘)
(1.9.2.5... .1:
:51}:

r...

1...";
17:..b9...
13.5.3.)

Ill!

THESIS

.1 3.000

 

LIBRARY
University

 

Michigan State

 

 

This is to certify that the

thesis entitled

Lack of cryptic reproductive isolation between
Papilio canadensis and Papilio glaucus; and

 

population genetics near their hybrid zone

 

presented by

Aram Daniel Stump

has been accepted towards fulfillment

of the requirements for

M' S ° degree in

 

Entomology

 

0/

I

/ I' ,1 ~'
.. x'i/
i.

(I/4

((4 {72:77

 

[,3

Date 9‘?) At"! 2/0

\5.

0-7 639

Major professor

MS U i: an Affirmative Action/Equal Opportunity Institution

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

 

MA? ll l§0334

 

 

 

 

 

 

 

 

 

 

 

 

 

 

woo mus-p.14

LACK OF CRYPTIC REPRODUCTIVE ISOLATION BETWEEN PAPILIO
CANADENSIS AND PAPILIO GLA UC US; AND POPULATION GENETICS NEAR
THEIR HYBRID ZONE

By

Aram Daniel Stump

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Entomology
and
Ecology, Evolutionary Biology and Behavior Program

2000

ABSTRACT
LACK OF CRYPTIC REPRODUCTIVE ISOLATION BETWEEN PAPILIO
CANADENSIS AND PAPILIO GLA UCUS; AND POPULATION GENETICS NEAR
THEIR HYBRID ZONE
By

Aram Daniel Stump

The objectives of this thesis relate to the maintenance of species differences
across a hybrid zone between the swallowtail butterflies Papilio canadensis and Papilio
glaucus. The first objective was to determine if there is physiological (postpairing,
prezygotic) isolation between these species. Heterospecific pairings between canadensis
and glaucus were no less likely than conspecific pairings to last at least 30 minutes, result
in sperrnatophore deposition, result in oviposition, or result in production of larvae.
When females were mated to more than one male, there was no preferential use of sperm
from conspecific males (conspecific sperm precedence). Together, these indicate that
there is no physiological isolation between canadensis and glaucus.

The second general objective was to study gene flow, both within canadensis and
between species. Allozymes indicate high gene flow between canadensis populations,
with F ST-values less than 0.01 for all four polymorphic enzyme loci used. Introgression
of glaucus alleles into canadensis populations was found at two different types of loci:
the X-linked nuclear Pgd gene, and in mitochondrial DNA. However, introgression was
found only in canadensis populations nearest hybrid zone areas, indicating some genetic
structure. When present, introgressed alleles were at low frequency, and individuals

carrying introgressed alleles at one locus rarely carried introgressed alleles at other loci.

To my parents

iii

ACKNOWLEDGMENTS

Thanks to my major advisor, Dr. Mark Scriber for his guidance, enthusiasm, and
patience. Thanks to my guidance committee, Dr. Guy Bush, Dr. Cathy Bristow, Dr.
Suzanne Thiem, and Dr. Jim Smith for helpful suggestions and revisions. Thanks also to
Drs. Bush and Smith for welcoming me into their laboratory to conduct the molecular
biology parts of these studies.

Thanks to Dr. Felix Sperling for providing information on PCR primers and
restriction sites. Dr. Wayne Wehling introduced me to allozyme electrophoresis and to
Papilio lab techniques and natural history. Amber Crim and Kathi Caulkins introduced
me to PCR methods.

Finally, thanks to all of the graduate students of the Scriber Lab, for practical
assistance and moral support in too many ways to list: Jennifer Donovan, Gabriel Ording,
Dylan Parry, Piera Giroux, Chip Francke, Heather Govenor, Mark Deering, and Cheryl

Frankfater.

iv

TABLE OF CONTENTS

LIST OF TABLES .................................................................................. vii
LIST OF FIGURES .................................................................................. x
CHAPTER 1:
INTRODUCTION .................................................................................... 1
Barriers to Gene Flow Between Species ........................................................... 2
Physiological isolation with singly-mated female insects .............................. 3
Conspecific sperm precedence ............................................................. 5
Population Genetics of Hybrid Zones .............................................................. 7
Gene flow within species ................................................................... 7
Interspecific introgression .................................................................. 8
Isolation in Tiger Swallowtails: Papilio canadensis and Papilio glaucus. . . . . .. . .. . . ..10
Objectives ............................................................................................ 1 5
CHAPTER 2:
ARE HETEROSPECIFIC » PAIRINGS BETWEEN PAPILIO CANADENSIS AND
PAPILIO GLA UC US LESS SUCCESSFUL THAN CONSPECIFIC PAIRINGS? ....... 16
Introduction .......................................................................................... 16
Methods ............................................................................................... 18
Results ................................................................................................ 21
Discussion ............................................................................................ 42
CHAPTER 3:
DOES CONSPECIFIC SPERM HAVE PRECEDENCE IN PAPILIO CANADENSIS OR
P. GLA UC US? ....................................................................................... 46
Introduction .......................................................................................... 46
Methods ............................................................................................... 47
Results ................................................................................................ 50
Discussion ............................................................................................ 58
CHAPTER 4:
HIGH LEVELS OF GENE FLOW BETWEEN POPULATIONS OF THE CANADIAN
SWALLOWTAIL, PAPILIO CANADENSIS ................................................... 61
Introduction .......................................................................................... 61
Methods ............................................................................................... 63
Results ................................................................................................ 66
Discussion ............................................................................................ 73

CHAPTER 5:
INTROGRESSION OF PAPILIO GLA UCUS GENES INTO P. CANADENSIS

POPULATIONS .................................................................................... 75
Introduction .......................................................................................... 75
Methods ............................................................................................... 77
Results ................................................................................................ 78
Discussion ............................................................................................ 83
CHAPTER 6:

SUMMARY AND CONCLUSIONS ............................................................. 86
APPENDIX 1:

RECORD OF DEPOSITION OF VOUCHER SPECIMENS ................................. 91
APPENDIX 1.1:

VOUCHER SPECIMEN DATA .................................................................. 93
LITERATURE CITED ............................................................................. 98

vi

LIST OF TABLES

Table 1.1. Barriers to gene flow between species (adapted from Campbell 1993) ......... 3
Table 1.2. Species differences between canadensis and glaucus (Hagen et al. 1991). . ..12

Table 2.1. Success of pairings grouped by mating duration. Proportions i s.d. are.
followed by sample sizes in parentheses. The pairings were of various types
(conspecifrc, heterospecific, backcross, and F2), involved various Papilio species
(canadensis, glaucus, eurymedon, rutulus, multicaudatus, and troilus), and were between
lab-reared females and either lab-reared or wild-caught males ............................... 22

Table 2.2. Success of pairings grouped by whether a spermatophore had been deposited
or not. Proportions i s.d. are followed by sample sizes in parentheses. Chi-square values
compare the success of pairings resulting in spermatophore deposition with those not
leaving sperrnatophores. The pairings were of various types (conspecific, heterospecific,
backcross, and F2), involved various Papilio species (canadensis, glaucus, eurymedon,
rutulus, multicaudatus, and troilus), and were between lab-reared females and either lab-
reared or wild-caught males ........................................................................ 22

Table 2.3. Chi-square values from ANOVA of the proportions of pairings involving
canadensis and glaucus males and females lasting at least 30 minutes, out of all pairings
that lasted a minimum of five minutes. The model was a 2x2x2 factorial design with
effects being the species of the male, the species of the female (canadensis or glaucus for
each), the origin of the male (wild or lab), and their interactions ............................. 24

Table 2.4. Chi-square values from ANOVA of proportions of pairings involving
canadensis and glaucus males and females resulting in spermatophore deposition, out of
all pairings lasting at least 30 minutes. The model was a 2x2x2 factorial design with
effects being the species of the male, the species of the female (canadensis or glaucus for
each), the origin of the male (wild or lab), and their interactions ............................. 25

Table 2.5. Chi-square values from ANOVA of proportions of pairings involving
canadensis and glaucus males and females leading to oviposition, out of all pairings with
spermatophore deposition. The model was a 2x2x2 factorial design with effects being
the species of the male, the species of the female (canadensis or glaucus for each), the
origin of the male (wild or lab), and their interactions ....................................... 26

Table 2.6. Chi-square values from ANOVA of proportions of pairings involving
canadensis and glaucus males and females leading to production of larvae, out of all
pairings leading to oviposition. The model was a 2x2x2 factorial design with effects
being the species of the male, the species of the female (canadensis or glaucus for each),
the origin of the male (wild or lab), and their interactions .................................... 28

vii

Table 2.7. F-values (Type III SS) from ANOVA of proportions of hatching eggs out of
all eggs laid, averaged over all pairings involving canadensis and glaucus males and
females producing at least one hatching larva. The model was a 2x2x2 factorial design
with effects being the species of the male, the species of the female (canadensis or
glaucus for each), the origin of the male (wild or lab), and their interactions .............. 29

Table 2.8. Types and numbers of pairs, each with a complete suite of data, and their
reproductive success ................................................................................ 36

Table 2.9. Egg hatchability of broods producing larvae ....................................... 37

Table 2.10. Calculation of combined index of mating success. For each pairing type, the
proportion of pairings producing hatching larvae (Table 2.8) is multiplied by the average
hatchability of broods producing larvae (Table 2.9) ............................................ 42

Table 3.1. Types and numbers of double-paired lab-reared females and reproductive
success following remating ........................................................................ 51

Table 3.2. Types and numbers of remated wild-caught females and reproductive success
following remating .................................................................................. 52

Table 3.3. Transmission and expression of PGD in larval offspring. PGD is X-linked in
Papilio, where females are XY and males are XX. Homozygotes of PGD cannot be
distinguished from hemizygotes in larvae because their sex is unknown ................... 54

Table 3.4. Sperm precedence (the proportion of offspring that were sired by the second
male; P2) for double-paired lab-reared females. Also indicated are the number of larvae
produced after remating that had paternity determined (N), origin of the male used for
remating, the days between pairings, the duration of the second mating, and the number
of sperrnatophores present in the female at death ............................................... 56

Table 3.5. Sperm precedence (the proportion of offspring that were sired by the second
male; P2) for remated wild-caught females. Also indicated are the number of larvae
produced after remating that had paternity determined (N), origin of the male used for
remating, the days between collection of the female and remating, the duration of the
remating, and the number of sperrnatophores present in the female at death ............... 57

Table 4.1. Enzymes resolved and running conditions used ................................. 65
Table 4.2. Chi-square values from tests for linkage of enzyme loci. The null hypothesis
was Ho: genotypes at one locus are distributed independently from genotypes at the other

locus ................................................................................................ 66

Table 4.3. P—values from tests of Hardy-Weinberg equilibrium. For each locus in each
population, Genepop tests the null hypothesis of Hardy-Weinberg equilibrium. . . . . . . . ....67

viii

Table 4.4. Wright’s F -statistics for six canadensis populations through the Great Lakes
region. Standard errors were obtained by jackknifing over populations, and are indicated
in parentheses ........................................................................................ 68

Table 5.1. Verification of diagnostic mtDNA haplotypes for canadensis and glaucus as
visualized by PCR-RFLP. Frozen specimens had been stored at -80°C, dried specimens
had been stored’pinned in drawers at room temperature. The canadensis haplotype (—) is
indicated by the absence of a T an restriction site in the 294bp PCR fragment, the glaucus
haplotype (+) is indicated by the presence of a T an restriction site in the same

fiagment .............................................................................................. 80

Table 5.2. Individuals from 1998 samples carrying introgressed alleles or haplotypes.
Introgressed alleles are underlined ................................................................ 83

ix

LIST OF FIGURES

Figure 1.1. Ranges of canadensis and glaucus (adapted from Hagen & Scriber 1991)...11

Figure 1.2. Schematic of female genitalia of ditrysian Lepidoptera; lateral cross section
of posterior half of abdomen. BC: bursa copulatrix; S: sperrnatheca; O: ostium
oviductus; B: ostium bursae; D: ductus seminalis; Ov: ovaries. Modified fi‘om
Drummond (1984) .................................................................................. 14

Figure 2.1. Out of all pairings that lasted a minimum of five minutes, the proportion
lasting at least 30 minutes. Error bars are +1 s.d. and numbers in bars are sample sizes.
Female canadensis (C) or glaucus (G) were paired to male canadensis or glaucus, with
the female listed first. Filled bars indicate that the males used were wild-caught, open
bars indicate that the males used were lab-reared ............................................... 24

Figure 2.2. Out of all pairings lasting at least 30 minutes, the proportion resulting in
spermatophore deposition. Error bars are +1 s.d. and numbers in bars are sample sizes.
Bars not sharing a letter are significantly different from each other at p=0.05. Female
canadensis (C) or glaucus (G) were paired to male canadensis or glaucus, with the
female listed first. Filled bars indicate that the males used were wild-caught, open bars
indicate that the males used were lab-reared .................................................... 25

Figure 2.3. Out of all pairings with spermatophore deposition, the proportion leading to
oviposition. Error bars are +1 s.d. and numbers in bars are sample sizes. Bars not
sharing a letter are significantly different from each other at p=0.05. Female canadensis
(C) or glaucus (G) were paired to male canadensis or glaucus, with the female listed first.
Filled bars indicate the males used were wild-caught, open bars indicate the males used
were lab-reared ....................................................................................... 26

Figure 2.4. Out of all pairings leading to oviposition, the proportion leading to
production of larvae. Error bars are +1 s.d. and numbers in bars are sample sizes. Bars
not sharing a letter are significantly different from each other at p=0.05. Female
canadensis (C) or glaucus (G) were paired to male canadensis or glaucus, with the
female listed first. Filled bars indicate the males used were wild-caught, open bars
indicate the males used were lab-reared ......................................................... 28

Figure 2.5. For all pairings producing at least one hatching larva, the mean proportion of
viable eggs (hatching/total eggs). Values expressed are means, error bars are +1 s.e. and
numbers in bars are numbers of pairings. Bars not sharing a letter are significantly
different from each other at p=0.05. Female canadensis (C) or glaucus (G) were paired
to male canadensis or glaucus, with the female listed first. Filled bars indicate the males
used were wild-caught, and open bars indicate the males used were lab-reared. . ........29

Figure 2.6. Out of all pairings lasting at least 30 minutes, the proportion of pairings
resulting in spermatophore deposition. Error bars are +1 s.d. and ntunbers in bars are
sample sizes. Female glaucus (G), canadensis (C), or troilus (T) were paired to wild-
caught male glaucus, with the female listed first ................................................ 30

Figure 2.7. Out of all pairings with spermatophore deposition, the proportion leading to
oviposition. Error bars are +1 s.d. and numbers in bars are sample sizes. Bars not
sharing a letter are significantly different from each other at p=0.05. Female glaucus (G),
canadensis (C), or troilus (T) were paired to wild-caught male glaucus, with the female
listed first ............................................................................................. 31

Figure 2.8. Out of all pairings leading to oviposition, the proportion of pairings leading
to production of larvae. Error bars are +1 s.d. and numbers in bars are sample sizes. Bars
not sharing a letter are significantly different from each other at p=0.05. Female glaucus
(G), canadensis (C), or troilus (T) were paired to wild-caught male glaucus, with the
female listed first .................................................................................... 31

Figure 2.9. For all pairings producing at least one hatching larva, the mean proportion of
viable eggs (hatching/total eggs). Values expressed are means, error bars are +1 s.e. and
numbers in bars are numbers of pairings. Bars not sharing a letter are significantly
different from each other at p=0.05. Female glaucus (G), canadensis (C), or trailus (T)
were paired to wild-caught male glaucus, with the female listed first ....................... 32

Figure 2.10. Out of all pairings that lasted a minimum of five minutes, the proportion of
pairings lasting at least 30 minutes. Error bars are +1 s.d. and numbers in bars are sample
sizes. Female canadensis (C), glaucus (G), or hybrid canadensis x glaucus (C x G) were
paired to wild-caught male canadensis, with the female listed first .......................... 33

Figure 2.11. Out of all pairings lasting at least 30 minutes, the proportion of pairings
resulting in spermatophore deposition. Error bars are +1 s.d. and numbers in bars are
sample sizes. Female canadensis (C), glaucus (G), or hybrid canadensis x glaucus (C x
G) were paired to wild-caught male canadensis, with the female listed first ............... 33

Figure 2.12. Out of all pairings with spermatophore deposition, the proportion of
pairings leading to oviposition. Error bars are +1 s:d. and numbers in bars are sample
sizes. Bars not sharing a letter are significantly different from each other at p=0.05.
Female canadensis (C), glaucus (G), or hybrid canadensis x glaucus (C x G) were paired
to wild-caught male canadensis, with the female listed first ................................. 34

Figure 2.13. Out of all pairings leading to oviposition, the proportion of pairings leading
to production of some larvae. Error bars are +1 s.d. and numbers in bars are sample
sizes. Bars not sharing a letter are significantly different from each other at p=0.05.
Female canadensis (C), glaucus (G), or hybrid canadensis x glaucus (C x G) were paired
to wild-caught male canadensis, with the female listed first .................................. 34

xi

Figure 2.14. For all pairings producing at least one hatching larva, the mean proportion
of viable eggs (hatching/total eggs). Values expressed are means, error bars are +1 s.e.
and numbers in bars are numbers of pairings. Bars not sharing a letter are significantly
different from each other at p=0.05. Female canadensis (C), glaucus (G), or hybrid
canadensis x glaucus (C x G) were paired to wild-caught male canadensis, with the
female listed first .................................................................................... 35

Figure 3.1. Reproductive success of multiply-mated females following remating as a
function of female origin and success before remating. A) Proportion of females laying
eggs afier remating. B) Of females laying eggs, the proportion producing larvae. Error
bars are +1 s.d., numbers within bars are number of females, and bars with the same letter
are not significantly different at the p=0.05 level. (Note: b is significantly different than
b’ at p=0.0918) ....................................................................................... 55

Figure 3.2. Niunber of larvae sired by each male and the days following remating of their
production by females producing mixed broods: A) Brood 12328; B) Brood 14252; C)
Brood 14288 .......................................................................................... 59

Figure 4.1. Sample sites. 1: Cook Co., Minnesota; 35 males. 2: Gogebic Co., Michigan;
36 males, 1 female. 3: Dickinson Co., Michigan; 48 males, 20 females. 4: Charlevoix
Co., Michigan; 50 males, 18 females. 5: Mason Co., Michigan; 50 males, 15 females. 6:
Isabella Co., Michigan; 50 males, 14 females ................................................... 64

Figure 4.2. GPI allozyme frequencies for six sampled canadensis populations.
Populations not sharing a letter are significantly different at p=0.05. The P-value for the
test of overall allele differentiation is 0.004 ................................................... 69

Figure 4.3. PGM allozyme frequencies for six sampled canadensis populations.
Populations not sharing a letter are significantly different at p=0.05. The P—value for the
test of overall allele differentiation is 0.561 ..................................................... 70

Figure 4.4. HBDH allozyme frequencies for six sampled canadensis populations.
Populations not sharing a letter are significantly different at p=0.05. The P-value for the
test of overall allele differentiation is 0.185 ..................................................... 71

Figure 4.5. PGD allozyme frequencies for six sampled canadensis populations.
Populations not sharing a letter are significantly different at p=0.05. The P-value for the
test of overall allele differentiation is 0.060 ..................................................... 72

Figure 5.1. Sites of six sampled canadensis populations and one sampled glaucus
population collected in May and June 1998. Sampled canadensis populations. 1: Cook
Co., Minnesota; 35 males. 2: Gogebic Co., Michigan; 36 males, 1 female. 3: Dickinson
Co., Michigan; 48 males, 20 females. 4: Charlevoix Co., Michigan; 50 males, 18
females. 5: Mason Co., Michigan; 50 males, 15 females. 6: Isabella Co., Michigan; 50
males, 14 females. Sampled glaucus populations. 7: Lawrence Co., Ohio; 22 males. . ..79

xii

Figure 5.2. Diagnostic molecular marker frequencies for six canadensis populations and
one glaucus population sampled in 1998. Filled portions indicate frequencies of
canadensis alleles (for PGD) or haplotypes (for mtDNA) and open portions indicate
frequencies of glaucus alleles or haplotypes. The left column indicates PGD frequencies,
the right column mtDNA. Numbers next to pies indicate the number of alleles or
haplotypes sampled ................................................................................. 82

xiii

CHAPTER 1:

INTRODUCTION

Closely related species with parapatric or sympatric distributions represent an
interesting problem for biologists. Differences between them are generally maintained,
even though they often still share significant similarity in their reproductive systems.
This problem is especially interesting when two species meet at a hybrid zone, an area
where they meet and interbreed (Barton & Hewitt 1985), and differences are maintained
even in the face of hybrid production.

Such species allow the study of key questions in evolution and ecology. They
represent an-important stage of speciation: differentiation between the two groups, while
limited enough to allow hybrid production, is complete enough to isolate the two (Hewitt
1988). They also allow the study of species boundaries, and why they lie where they do
(Hoffman & Blows 1994).

This thesis addresses questions relating to the maintenance of species differences
across a swallowtail butterfly hybrid zone. Because unhindered gene flow between two
differentiated populations will quickly homogenize the two, in cases where species
differences are maintained, there must be barriers to gene flow (Barton 1979). I
examined potential barriers to gene flow that occur afier mating begins but before hybrid
zygote formation. Also, because in some cases gene flow between two species does not
stop entirely (Barton & Hewitt 1985), I studied patterns of gene flow near and across the

hybrid zone.

Barriers to Gene Flow Between Species

Darwin’s (1859) consideration of the maintenance of species differences was
limited to a discussion of the inviability and sterility of many hybrids. The Biological
Species Concept (BSC), promoted by Dobzhansky (1951) and Mayr (1963), states that
speciation has occurred when two groups of organisms are no longer capable of
exchanging genes, and calls this state reproductive isolation. Both authors included
classifications and examples of factors that can cause reproductive isolation in their
treatments of the BSC.

The BSC has been criticized for a number of reasons (Mallet 1995, Harrison
1998), including a questioning of the need for a complete cessation of gene flow between
recently diverged species. Many other species concepts have been proposed, but
regardless of what concept individual researchers adhere to, the BSC has been important
by focusing attention on factors that restrict gene flow between two species (Harrison
1998)

Mayr (1963) divided barriers to gene flow into two general categories: premating
and postrnating. Alternatively, these can be grouped as either prezygotic or postzygotic
(Table 1.1), a more informative classification because broadly speaking, prezygotic
barriers can potentially be selected for (Dobzhansky 1951), whereas postzygotic barriers
cannot. This thesis will focus on physiological isolation between species: barriers to gene
flow that occur after copulation has started, but before eggs are fertilized (Table 1.1).
These barriers can be caused by divergence in genitalic physiology, by cryptic female
choice (Eberhard 1996), or by competition between sperm from different males

(Birkhead & Moller 1998).

Table 1.1. Barriers to gene flow between species (adapted from Campbell 1993).

 

 

 

Category Description
Prezygotic: Prevents the production of hybrids
1. Geographic Isolation Species live in different areas
2. Temporal Isolation Species mate at different times
3. Behavioral Isolation Species meet, but do not attempt to mate
4. Mechanical Isolation Species attempt to copulate, but cannot
5. Physiological Isolation“ Species copulate, but sperm does not reach egg
Postzygotic: Reduces the fitness of hybrids
l. Zygote Mortality Eggs are fertilized but do not hatch
2. Hybrid Inviability Hybrid individuals die before sexual maturity
3. Hybrid Sterility Hybrids do not produce functional gametes
4. Hybrid Breakdown Offspring of hybrids have reduced viability or fertility
5. Ecological Selection Hybrids are poorly adapted to certain habitats

 

* Refers to postpairing, prezygotic isolation, including: cryptic female choice,
incapacitation of sperm, and conspecific sperm precedence.

Physiological isolation with singly-mated female insects

Difi‘erences in genitalia are often found between closely related insect species,
and this observation has produced the ‘lock-and-key’ hypothesis: genitalic differences
should mechanically prevent males from being able to inseminate females of other insect
species (DufOur 1844). However, few examples have been found where the lock-and-key
hypothesis holds (Dobzhansky 1951, Porter & Shapiro 1990). Recently though, other
forms of postpairing‘, prezygotic isolation between species have been found.

Differences in genitalia may still play a part in species isolation, even when they
do not mechanically prevent successful mating. Genitalic differences between species of
scarab beetles appear to allow females to exercise choice about whether to allow full

insemination by a male (Eberhard 1992). Courtship during copulation could have a

 

' For the purposes of this thesis, any factor that is referred to as postpairing will be something that occurs
after a copulation has started (including events after copulation), and any factor called postcopulatory will
refer to something occurring only afier a copulation has ended.

similar effect (Eberhard 1994). In these cases, species differences are potential cues that
females can take advantage of, leading to physiological isolation via female choice.

Physiological isolation can also be due to poor sperm performance. In some
ladybird beetles, heterospecific crosses produce a lower percentage of hatching eggs than
conspecific crosses (Nakano 1985), caused by a partial incapacitation of heterospecific
sperm in the female’s reproductive tract (Katakura 1986). In some ground crickets
(Gregory & Howard 1993) and katydids (Shapiro 2000), females mated to conspecific
males produce more eggs than those mated to heterospecific males, which could indicate
conspecific sperm produces a stronger oviposition response.

Physiological isolation can also be asymmetric between species. In the cricket
genus Gryllus, hybrid pairings between G. firmus females and G. pennsylvanicus males
are unsuccessful, whereas pairings between G. peimsylvanicus females and G. firmus
males are fully successful (Harrison 1983).

There can also be geographic variation in physiological isolation, as in two
species of green lacewings with ranges that overlap in some areas and do not overlap in
others (Albuquerque et a1. 1996). When individuals taken from areas of sympatry are
mated, heterospecific sperm fails to transfer to the sperrnatheca of the female. However,
when individuals from areas where the other species is not found are mated,
heterospecific sperm is transferred to the sperrnatheca and is used to fertilize eggs. This
points toward reinforcement of premating isolation (Butlin 1989), although it should be
remembered that these species share broad stretches of their ranges, allowing selection on

many individuals. Where species meet at narrow hybrid zones, most individuals never

meet heterospecific individuals, so most individuals are never under any selection for
reproductive isolation.

It is important to note that there may be many examples of insects where there is
no postpairing, prezygotic isolation. In several species of longwing butterflies,
heterospecific matings are just as successful as conspecific matings (McMillan et al.

1997), and there may be many other examples where this is true.

Conspecific sperm precedence

Parker (1970) introduced the idea that when insect females mate with multiple
males and store sperm, sexual selection can continue even after copulation. This has
usually been presented as sperm competition, where the sperm from one male competes
with the sperm from other males to fertilize the eggs of the female (postcopulatory
intrasexual selection) (Birkhead & Muller 1998). However, female choice, where a
female chooses what sperm to use based on some characteristic of the males or their
ejaculates (postcopulatory intersexual selection), may also play an important role
(Eberhard 1996). Sperm precedence is the general term given to patterns of sperm usage
by a female that has mated to more than one male (Simmons & Siva-Jothy 1998).

Different types of sperm precedence have been found in insects, with the most
common being for the female to use a mixture of sperm from her various mates, biased to
some degree towards the most recent male (Gwynne 1984). Mechanisms for this type of
sperm precedence (‘last-male’) include the ejection of previous spermatophores (e. g.
DeVilliers & Hanrahan 1991) or the repositioning or displacement of previous sperm

(e.g. Siva-Jothy & Tsubaki 1994, Eady 1994). Another type is ‘first-male’ sperm

precedence, where all or almost all offspring continue to be sired by the first male, even
after later copulations with one or more different males (Gwynne 1984). Large
spermatophores acting as mating plugs can lead to this pattern (Lorch et al. 1998). There
are also cases where a female will preferentially use sperm from a certain male based on
some other factor (female choice), such as spermatophore size in arctiid moths
(LaMunyon & Eisner 1994). It has been noted that there is often much variability in
patterns of sperm usage from pairing to pairing (Simmons & Siva-Jothy 1998).

Most studies on sperm precedence have looked at postcopulatory sexual selection
between individuals belonging to the same species. However, several cases have been
found where sperm competition can result in a barrier to gene flow between species
(Howard 1999). In ground crickets (Gregory & Howard 1994), flour beetles (Wade et al.
1994), a grasshopper (Hewitt et al. 1989), and some Drosophila (Price 1997), a female
mated only to a heterospecific male will produce many hybrid offspring. However, when
mated to a conspecific male and a heterospecific male, such a female will produce
offspring sired almost exclusively by the conspecific male, regardless of the order of
copulations. This is called conspecific sperm precedence, and it has been shown that
when females mate multiply, it can be a potent barrier to gene flow between Species
(Gregory & Howard 1994, Howard et al. 1998).

Postcopulatory, prezygotic isolation has been called cryptic, because it is ignored
by traditional Darwinian measures of success, which focus on male success in achieving
mating access to females (Eberhard & Cordero 1995). However, physiological barriers

could be important and under-appreciated in restricting gene flow between insect species.

Population Genetics of Hybrid Zones

Hybrid zones are clines maintained by the opposing forces of dispersal and gene
flow on the one hand, and reproductive isolation and selection against hybrids on the
other hand (Barton & Hewitt 1985). This means that to understand how species
differences are maintained, one must understand not just the barriers to gene flow, but
also the potential of both species for dispersal and gene flow. This can be done either by
studying gene flow within both species, or by studying gene flow between species

(introgression).

Gene flow within species

Gene flow is defined as the movement of genes between populations (Slatkin
1985). It is important because it tends to homogenize populations, counteracting drift
and local adaptation, as well as spreading advantageous alleles (Slatkin 1987). If gene
flow is high within two closely related species, the barriers to gene flow between them
must be quite strong to maintain the differences between the two (Barton 1979).
However, if gene flow within both species is low, the barriers between the two need not
be as strong.

There are two basic approaches to studying gene flow within Species: direct and
indirect (Slatkin 1987). Direct measures of gene flow are based on observations of
dispersing individuals. One weakness with this approach is that gene flow may be
sporadic, and occasional events of high gene flow could be enough to homogenize
populations significantly. Unless observations were made during these times, a direct

approach would underestimate the effective gene flow that occurs over an evolutionary

time scale. The other basic approach is indirect: estimating gene flow by looking at the
geographic patterns of allele and genotype distribution. Wright (1931) provided the
earliest statistical tools for indirect measurement of gene flow, F -statistics. The
development of allozyme electrophoresis provided many potentially neutral, codominant
markers that are widely dispersed through the genome (Avise 1994), making them highly
compatible with Wright’s method.

Indirect methods of estimating gene flow based on allozymes have their
limitations. One is that in some cases allozymes may provide a reflection of historical
patterns of dispersal rather than current levels of gene flow (Bossart & Prowell 1998).
However, historical patterns, if they reflect long-term potential for gene flow, might be
more important to evolution than the pattern of relatively few recent years. Another
limitation is that allozymes do not reveal much of the DNA variation that is present, even
in the genes for the enzymes (Richardson et al. 1986). They may miss genetic
discontinuities between populations that other markers may reveal (Karl & Avise 1992).
However, while methods more powerful at detecting variation are continually being
developed, allozymes remain the most accessible and cost-effective way of surveying
genetic variation at many variable loci in a large number of individuals (Richardson et al.

1986).

Interspecific introgression
Introgression is a special case of gene flow: gene flow across species boundaries
(Harrison 1993). There is debate about the importance of introgression to the evolution

of parental species (Arnold et al. 1999), but in any case introgression can be informative

about the nature and completeness of barriers to gene flow between species. Hybrid
zones are typically characterized by short, steep clines flanked by long tails of
introgression on either side (Barton & Hewitt 1985). Measuring the length of tails, and
the frequency of introgressed alleles in those tails, allows the calculation of the strength
of selection against interspecific alleles as a function of distance from a hybrid zone
(Porter et al. 1997).

Tails of introgression are rarely the same for all loci. In some groups it has been
found that mitochondrial DNA introgresses more readily than nuclear genes (Barton &
Jones 1983, Powell 1983), and selection may maintain differences in diagnostic traits
while allowing significant gene exchange at other loci (Barton & Bengtsson 1986). Tails
of introgression of mtDNA or enzyme loci may be longer than clines in quantitative traits
such as morphological characters (Barton & Hewitt 1985), or narrower than other
quantitative traits such as host use abilities (Hagen 1990). The width of clines and the
length of tails of introgression should reflect the strength of selection against these
introgressed characters.

If introgression varies at different geographic points along a hybrid zone, or if it
changes through time, it could be informative about what is causing species boundaries
and barriers to gene flow. For example, abnormally warm years might allow increased
introgression of genes from southern species into closely related neighboring northern
species. Introgression might also be asymmetric between species (Sperling & Spence
1991).

The genotypic pattern that introgression takes is important as well. When it is

present, whether it is found at homozygous or heterozygous loci, and whether

introgression at one locus tends to be coincidental within individuals with introgression at

other loci can indicate how recent the gene flow across the hybrid zone was.

Isolation in Tiger Swallowtails: Papilio canadensis and Papilio glaucus

The Eastern Tiger Swallowtail, Papilio glaucus L., and the Canadian Tiger
Swallowtail P. canadensis Rothschild & Jordan (Lepidoptera: Papilionidae) are closely
related butterfly species with parapatric distributions (Figure 1.1). Until fairly recently,
canadensis was considered to be a subspecies of glaucus, but based on morphological,
molecular, and ecological differences (Table 1.2), it was given separate species status
(Hagen et al. 1991). Several of the differences follow a pattern in Lepidoptera: a
disproportionately high number of diagnostic traits (th, Pgd, diapause induction, dark
color suppression, Hagen & Scriber 1989) are X-linked (Sperling 1994).

Allopatric speciation has been hypothesized for canadensis and glaucus:
speciation during the Pleistocene ice age, with the proto-glaucus populations spending
the last 40,000 years south of the glaciation and the proto-canadensis populations isolated
in the unglaciated pocket of Beringia, now Alaska (Scriber 1988). However, there is
higher variability in allozymes in Michigan canadensis populations than in Alaskan
canadensis populations (Hagen & Scriber 1991). Because ancestral populations often are
more genetically variable than dispersed populations, one could hypothesize parapatric
speciation, although population size or introgression could also account for the
differences seen.

Both canadensis and glaucus are found over wide geographic ranges (Figure 1.1),

covering a number of different ecological habitats. Local adaptation to regional habitats

10

 

 

Figure 1.1. Ranges of canadensis and glaucus (adapted from Hagen & Scriber 1991).

Table 1.2. Species differences between canadensis and glaucus (Hagen et al. 1991).

 

 

 

 

Characteristic canadensis glaucus
(Morphological)
White transverse bands on lst instar larvae 3 l
Forewing underside submarginal yellow Band Spots
Hindwing upperside anal cell black band Wide Narrow
Adult size Small Large
(Ecological/Physiological)
Obligate diapause (X-linked recessive) Present Absent
Melanie gene (Y —linked) Absent Present
Melanie suppressor gene (X-linked) Present Absent
Tulip tree use ability Low High
Quaking aspen use ability High Low
(Molecular)
Hk (autosomal) allozymes I-IK 110 HK 100
th (X-linked) allozymes LDH 40, 80 LDH 100
Pgd (X-linked) allozymes PGD -80, -125 PGD -50, -100
mtDNA Tan site in C01 gene Absent Present

 

has been found in both species: canadensis to thermal climates (Ayres & Scriber 1994),
and glaucus to regional hostplants (Scriber 1986, Bossart & Scriber 1995). Local
adaptation in glaucus contrasts with a finding, based on allozyme distributions, of high
gene flow between widely separated populations (Bossart & Scriber 1995) and evidence
of high dispersal potential of glaucus individuals (Lederhouse 1982, Scriber et al. 1998).
The ranges of canadensis and glaucus meet at a narrow hybrid zone (Hagen et al.
1991, Hagen 1990, Luebke et a1. 1988). Study of the hybrid zone in New York using
allozymes found no evidence of assortative mating in the zone (Hagen 1990). F1 hybrids
are viable, fertile (Hagen & Scriber 1995), and able to survive on the hostplants of either
parental species (Scriber et al. 1995). Long-range dispersal of a hybrid out of the hybrid
zone has been documented (Scriber et al. 1998). One focus of research on these Papilio

species is to determine what keeps them distinct.

12

A couple of potential barriers to gene flow have been found. Deering (1998)
found behavioral isolation between the two species, with a caveat. In Florida, glaucus
males choose to court and copulate with glaucus females rather than equally sized
canadensis females. However, canadensis males in northern Michigan also prefer
glaucus females over canadensis. In a hybrid zone, the preference of glaucus males for
conspecific females would reduce hybridization, but the preference of canadensis males
for heterospecific females would increase hybridization.

A common pattern in species hybridization is the Haldane Effect: hybrids of the
heterogametic sex (females in the Lepidoptera) often have lower fitness than hybrids of
the homogametic sex (Coyne & Orr 1989). A Haldane effect is seen in one of the types
of crosses between canadensis and glaucus (Hagen & Scriber 1995). When a glaucus
female is paired to a canadensis male, female offspring have higher pupal mortality than
males, but when a canadensis female is paired to a glaucus male, there is no increase in
pupal mortaility in either sex. This Haldane effect is apparently due to canadensis X-
linked genes, when combined with glaucus genes, disrupting pupal development (Hagen
& Scriber 1995). Slight endogenous reduction of hybrid fitness such as this could be
expected to reduce gene flow between the species, but would not prevent it. The search
is still on for barriers to gene flow between canadensis and glaucus.

A possible barrier to gene flow might be postpairing, prezygotic isolation. These
butterflies might be good candidates for such isolation for two reasons: 1) females mate
more than once (Lederhouse et al. 1989), so even if a female copulates with the wrong
kind of male, her reproductive potential can be rescued by finding a better male; and 2)

sperm are passed in large, possibly nutritious spermatophores (Lederhouse 1995), which

13

could act as a cues to females about the quality and appropriateness of males.
Postcopulatory isolation might also be facilitated by the female reproductive system of
butterflies. Butterfly females, as in most of the Lepidoptera, are ditrysian, meaning they
have two genital openings: one for copulation, one for oviposition (Figure 1.2). When
mating, a male secretes a spermatophore, sperm, and other ejaculate into the bursa
copulatrix. Sperm then leaves the spermatophore and travels through the ductus
seminalis to the spermatheca Sperm from the spermatheca is then used to fertilize eggs
as they pass through the oviduct on the way to be oviposited. This physiology may make
it more likely for heterospecific sperm to be incompatible with the female reproductive
tract, or it may allow the female increased postcopulatory choice (T schudi-Rein & Benz
1990). Postpairing, prezygotic isolation could appear as a reduction of success of
heterospecific pairings relative to conspecific pairings, or as conspecific sperm
precedence.

Another focus of research on Papilio is their population genetics. Gene flow has

already been studied in glaucus populations (Bossart & Scriber 1995), however there is

sax. .

BC

9

V

Figure 1.2. Schematic of female genitalia of ditrysian Lepidoptera; lateral cross section of
posterior half of abdomen. BC: bursa copulatrix; S: spermatheca; O: ostium oviductus; B:
ostium bursae; D: ductus seminalis; Ov: ovaries. Modified from Drummond (1984).

14

no companion study for canadensis. Also, introgression of diagnostic allozymes has
been documented (Hagen et al. 1991), but introgression of maternally inherited
mitochondrial genes has not yet been investigated. Previous work with these species has
created a supply of genetic markers ready to be used to continue the study of their

population genetics (Hagen & Scriber 1991, Sperling 1994).

Objectives

My objectives for these studies were: 1) to determine if heterospecific pairings
between canadensis and glaucus are less successful than conspecific pairings; 2) to
determine if there is conspecific sperm precedence in either canadensis or glaucus; 3) to
indirectly determine levels of gene flow between isolated canadensis populations using
allozyme electrophoresis; and 4) to determine levels of interspecific introgression of
glaucus genes into canadensis populations using both allozymes (for nuclear

introgression) and PCR-RF LP (for mitochondrial introgression).

15

CHAPTER 2:
ARE HETEROSPECIFIC PAIRINGS BETWEEN PAPILIO CANADENSIS AND

P. GIA UC US LESS SUCCESSFUL THAN CONSPECIFIC PAIRINGS?

Introduction

Hybrid zones, geographic areas where individuals of different species meet and
interbreed, are not uncommon in nature (Barton & Hewitt 1985). Generally, the
distinctness between the species at large is still maintained, even in the face of hybrid
production. Even where hybrids are found, if heterospecific matings are less successful1
than conspecific matings, fewer hybrids will be produced and gene flow will be reduced
between species. This type of postpairing, prezygotic barrier to gene flow could arise
either from divergent reproductive physiologies or from cryptic mate choice (Eberhard
1996)

Hybrids with high viability and fertility are common in Papilio, and it has been
suggested that because of this, prezygotic barriers between species should be important
(Sperling 1990). These could include postpairing barriers in addition to premating
isolation. Lab matings between Papilio canadensis and P. glaucus produce viable and
fertile hybrids (Scriber et a1. 1995), but it is not known if heterospecific pairings are less
successful than conspecific pairings.

The complicated reproductive tract of females of the ditrysian Lepidoptera means
that there are a number of stages at which a mating can fail. Hand-paired Papilio

butterflies will often struggle against each other immediately after being paired, and at

 

' Mating success will refer to a number of factors: pairing duration, spermatophore deposition, oviposition,
and egg hatchability. Of course, the final measure of mating success is production of offspring.

16

this point they can break apart easily. This may be a response to their mate, or it could be
a response to significant human handling. As Clarke and Sheppard (1956) observed,
after several minutes the two seem to lock together and some pulling will not separate
them. Coincident with this locking, the head and thorax of the male relax, and both
individuals become still. At this point, pairs typically remain together for upwards of half
an hour. Lab pairings with glaucus have found that pairings must last at least 30 minutes
(at least in the lab) for the male to transfer a spermatophore (Lederhouse et al. 1990).
Thus a pairing can be unsuccessful due to breaking up prematurely either before locking
occurs (in the first few minutes) or after locking (the pairing lasts more than a few
minutes, but fewer than 30 minutes). The first is best studied in conditions as natural as
possible to minimize handling effects, but the second could legitimately be studied using
hand-pairings in lab conditions.

Even if a pairing lasts the minimum amount of time, it is not successful if it does
not result in spermatophore deposition or if it fails to spur the female to lay eggs. Even if
these successfi'rlly occur, if sperm is not moved to the spermatheca in significant
numbers, females may lay only unfertilized eggs or a low percentage of fertilized eggs.
Heterospecific pairings may fail more frequently than conspecific pairings at any of these
points.

I investigated postpairing, prezygotic reproductive isolation between Papilio
canadensis and P. glaucus. First, I asked some basic questions relating to mating
success: 1) does the duration of a copulation affect its chance of being successful; and 2)
can a mating be successful without spermatophore deposition? Second, I used

canadensis and glaucus to compare heterospecific mating success to conspecific mating

17

success with respect to copulation duration, spermatophore deposition, oviposition, and
production of larvae. Third, I investigated the effect of increased phylogenetic distance
on mating success by pairing females of the more distantly related species Papilio troilus
to glaucus males. Finally, I investigated the success of matings involving a type of

interspecific hybrid by pairing canadensis x glaucus hybrid females to canadensis males.

Methods

Lab-reared, virgin females were used for all pairings. Males had either been
reared in the lab or caught in the wild. Females were fed a 20% honey in water solution.
Males were fed a 20% honey solution containing electrolytes and amino acids to increase
virility (Lederhouse et a1. 1990). Lab-reared males were not paired for at least two days
following adult eclosion to allow reproductive maturation.

All matings were initiated by hand (Clarke & Sheppard 1956), and pairing
duration was recorded. All pairings breaking apart before the individuals locked together
(usually about five minutes in) were either re-paired or disregarded. After pairing,
females were placed in plastic oviposition arenas lined with hostplant foliage following
Scriber (1993) to stimulate oviposition, an established technique for facilitating egg
production in our lab. Eggs were counted daily and placed in a grth chamber.
Hatching larvae were also counted daily. Dead females were stored in a freezer, and later
dissected to determine if a spermatophore was present (Lederhouse et a1. 1989). For
various reasons, not all data were collected for some pairings.

To determine the minimum parameters for a successful mating, I combined

pairings of various types (conspecific, heterospecific, backcross, and F2), involving

18

various Papilio species (canadensis, glaucus, eurymedon, rutulus, multicaudatus, and
troilus), and using lab-reared females and lab-reared or wild-caught males, into several
general comparisons. I compared pairings of different durations with respect to
likelihood to lead to spermatophore deposition, oviposition, and production of hatching
larvae. I also compared pairings where spermatophore deposition had occurred to those
where it had not with respect to likelihood to lead to oviposition and larval production.
These proportions were compared using a chi-square analysis (PROC FREQ; SAS
Institute Inc. 1990). The proportion of pairings leading to larvae, out of all those leading
to oviposition, was also calculated.

To compare the success of conspecific and heterospecific pairings, I paired lab-
reared female canadensis and glaucus to lab-reared and wild-caught male canadensis and
glaucus. For each of the eight pairing types, I calculated: the proportion of copulations
lasting at least 30 minutes, out of all pairings reaching a locked state; the proportion of
copulations resulting in spermatophore deposition, out of all pairings that lasted at least
30 minutes; the proportion of pairings leading to the female laying at least one egg, out of
all pairings that resulted in a spermatophore being deposited; and the proportion of
pairings leading to at least one hatching larva, out of all pairings that had led to
oviposition. The effects of female and male species (canadensis or glaucus) and male
origin (lab-reared or wild-caught) on these measures were determined using a
contingency table analysis (PROC CATMOD; SAS Institute Inc. 1990). For each pairing
resulting in at least one hatching larva, the number of larvae was divided by the number
of eggs laid to give the proportion of hatching eggs. Then these proportions were

averaged for each of the eight pairing types, to give average egg hatchability. The effects

19

of female and male species and male origin on egg hatchability were determined using an
analysis of variance (PROC GLM; SAS Institute Inc. 1990).

To determine the effect of increased phylogenetic distance on heterospecific
mating success, pairings between wild glaucus males and glaucus and canadensis
females were compared to pairings between wild glaucus males and females of the more
phylogenetically distant Papilio troilus (Hagen & Scriber 1991, Caterino & Sperling
1999). I also paired canadensis x glaucus hybrid females to wild canadensis males and
compared them to pairings between wild canadensis males and canadensis and glaucus
females to determine if pairings with a type of hybrid female Show reduced success.
Pairing success for these comparisons was analyzed as above, except instead of looking
at species effects, the pairing types were simply compared.

Because the preceding comparisons break matings down into individual
components, they do not provide an overall picture of reproductive success. For this, all
pairings for which the first four measures were known (pairing duration, spermatophore
deposition, whether eggs had been laid, and whether larvae had been produced) were
compiled. The numbers successful in each measure were compared for each pairing type,
and the proportion of all of these pairings that produced larvae was calculated for each
pairing type. Then, to combine this with egg hatchability data, that proportion of pairings
that produced any larvae was multiplied by the average egg hatchability for each pairing
type. This gave an overall index of mating success that could be compared for each of

the types of pairings.

20

Results

No pairings lasting for fewer than 30 minutes resulted in spermatophore
deposition (Table 2.1). About half of the pairings 30-39 minutes long produced
spermatophores, and almost all pairings lasting 40 minutes or longer resulted in
spermatophores.

Some pairings from all duration divisions led to female oviposition (Table 2.1).
Pairings lasting for fewer than twenty minutes were less likely to stimulate oviposition,
but those only 20-29 minutes long produced females that were just as likely to oviposit as
those copulating for longer.

Of pairings that resulted in oviposition, none lasting for fewer than 30 minutes
resulted in hatching larvae, but most of those lasting 30 minutes or longer did result in
larvae (Table 2.1). Thus while pairings lasting for' fewer than 30 minutes can spur a
female to oviposit, they do not result in spermatophore deposition or larval production. It
appears that the minimum time for a completely successful mating is 30 minutes, at least
at room temperature (75-85 F°).

Pairings resulting in spermatophore deposition were only marginally more likely
to result in oviposition than those not leading to spermatophores (Table 2.2). However,
pairings resulting in spermatophore deposition were significantly more likely to lead to
hatching larvae after oviposition than those not. Out of 24 pairings without
spermatophores that led to oviposition, two led to hatching larvae. Thus females can
easily be induced to lay eggs without being provided a spermatophore, however larval
production with no spermatophore (while possible) is rare. This result of larval

production without a spermatophore is surprising, but it has been observed previously in

21

Table 2.1. Success of pairings grouped by mating duration. Proportions :1: s.d.l are
followed by sample sizes in parentheses. The pairings were of various types
(conspecific, heterospecific, backcross, and F2), involved various Papilio species
(canadensis, glaucus, eurymedon, rutulus, multicaudatus, and troilus), and were between
lab-reared females and either lab-reared or wild—caught males.

 

Duration Proportion of pairings Proportion of pairings Proportion of pairings with

 

(minutes) with a spermatophore with eggs larvae (of those with eggs)
5-19 0.00 i 0.00 (4) 0.33 i 0.19 (6) 0.00 :1: 0.00 (2)
20-29 0.00 :1: 0.00 (6) 0.75 :1: 0.12 (12) 0.00 :1: 0.00 (9)
30-39 0.59 :t 0.12 (17) 0.75 :1: 0.09 (24) 0.41 i 0.12 (17)

40.49 0.94 :t 0.03 (54) 0.78 i 0.05 (70) 0.66 :t 0.06 (53)
5059 0.98 i 0.02 (62) 0.77 i 0.05 (73) 0.64 :1: 0.07 (47)
60-69 1.00 :1: 0.00 (54) 0.73 :1: 0.05 (74) 0.51 :t 0.07 (53)

70—79 1.00 :1: 0.00 (19) 0.70 :1: 0.08 (30) 0.60 :t 0.11 (20)
80-89 0.94 i 0.06 (16) 0.76 :1: 0.08 (25) 0.61 :t 0.11 (18)
90-99 0.62 i- 0.17 (8) 0.69 i 0.12 (16) 0.38 :1: 0.17 (8)
100-109 0.78 :1: 0.14 (9) 0.73 :1: 0.13 (11) 0.38 :1: 0.17 (8)
110-119 1.00 i 0.00 (4) 0.83 i 0.15 (6) 0.60 i 0.22 (5)
120-129 1.00 :1: 0.00 (4) 1.00 i 0.00 (7) 0.57 i 0.19 (7)
2130 0.60 i 0.22 (5) 0.86 :1: 0.13 (7) 0.00 i 0.00 (5)

 

l s.d.=‘l(proportionx(1-proportion)/sample size)

Table 2.2. Success of pairings grouped by whether a spermatophore had been deposited
or not. Proportions i s.d.l are followed by sample sizes in parentheses. Chi—square
values compare the success of pairings resulting in spermatophore deposition with those
not leaving spermatophores. The pairings were of various types (conspecific,
heterospecific, backcross, and F2), involved various Papilio species (canadensis, glaucus,
eurymedon, rutulus, multicaudatus, and troilus), and were between lab-reared females
and either lab-reared or wild-caught males.

 

Proportion of pairings Proportion of pairings with

 

with eggs larvae (of those with eggs)
Spermatophore absent 0.62 :1: 0.07 (45) 0.08 :1: 0.06 (24)
Spermatophore present 0.75 d: 0.03 (274) 0.57 :t 0.04 (189)
df=1, xz=3.322# df=1, xz=l9.867"**

 

# P5010; **** PS0.001
‘ s.d.=‘/(proportionx( l -proportion)/sarnple size)

22

Papilio (Lederhouse et al. 1989) and other insects (George & Howard 1968), and could
be due to deposition of fi'ee sperm by the male (which seems more likely than
parthenogenesis by the female). There was another surprising result: one singly-mated
female was carrying two spermatophores, which has also been seen previously in Papilio

(Lederhouse et al. 1989). Of all laboratory hand-pairings leading to oviposition, 0.54 i

0.03 (proportion i s.d.) led to hatching larvae (N=339).

The proportions of “locked-together” pairings (lasting for longer than about five
minutes) involving canadensis and glaucus females and males that lasted at least the
minimum 30 minutes are shown in Figure 2.1. Almost all of these pairings lasted at least
30 minutes, and there were no significant differences between pairing types. There were
no significant effects on pairing duration (Table 2.3).

The proportions of pairings resulting in spermatophore deposition, out of all
pairings lasting at least 30 minutes are shown in Figure 2.2. There were no significant
differences between most pairing types. Lab-reared canadensis males were slightly less
effective in depositing a spermatophore when mated to glaucus females, but there was no
reduction in success when the canadensis male was from the wild. There were no
significant effects on spermatophore deposition (Table 2.4).

The proportions of pairings leading to oviposition, out of all pairings resulting in
spermatophore deposition are shown in Figure 2.3. There were again no significant
differences between most pairing types. There was a significant effect of the species of

the female, with canadensis females less likely to oviposit than glaucus females (Table

23

 

[lwild male Dlab maleJ

 

 

 

 

a)
or
.E
I-
"6
n.
"6
C
3
1:
o
8
E 72 16
CxC GxG CxG GxC

Pairing Type

Figure 2.1. Out of all pairings that lasted a minimum of five minutes, the proportion lasting at
least 30 minutes. Error bars are +1 s.d. and numbers in bars are sample sizes. Female
canadensis (C) or glaucus (G) were paired to male canadensis or glaucus, with the female
listed first. Filled bars indicate that the males used were wild-caught, open bars indicate that the
males used were lab-reared.

Table 2.3. Chi-square values from ANOVA of the proportions of pairings involving
canadensis and glaucus males and females lasting at least 30 minutes, out of all pairings
that lasted a minimum of five minutes. The model was a 2x2x2 factorial design with
effects being the species of the male, the species of the female (canadensis or glaucus for
each), the origin of the male (wild or lab), and their interactions.

 

 

Source of variation (If X2
female species 1 0.00 _
male species 1 —'
female species*male species 1 0.01
male origin 1 0.00
female species*male origin 1 0.00
male species*male origin 1 0.00

female species*male species*male origin 1 0.00

l 12 not calculated (df=0) due to near fixation of success in pairings involving glaucus
males.

24

 

[Iwild male Dlab male]

1‘ ab ab a ab a ab ab
0.8 7 b
0.6 _
0.4 -
0.2 7
0 _ 34 . 13 . 15 . 8 .
CXC GxG CXG GXC

Pairing Type

 

 

Proportion of Pairings

Figure 2.2. Out of all pairings lasting at least 30 minutes, the proportion resulting in spermato-
phore deposition. Error bars are +1 s.d. and numbers in bars are sample sizes. Bars not
sharing a letter are significantly different hour each other at p=0.05. Female canadensis (C) or
glaucus (G) were paired to male canadensis or glaucus, with the female listed first. Filled bars
indicate that the males used were wild-caught, open bars indicate the the males used were lab-
reared.

Table 2.4. Chi-square values from AN OVA of proportions of pairings involving
canadensis and glaucus males and females resulting in spermatophore deposition, out of
all pairings lasting at least 30 minutes. The model was a 2x2x2 factorial design with
effects being the species of the male, the species of the female (canadensis or glaucus for
each), the origin of the male (wild or lab), and their interactions.

 

 

Source of variation ’ (If X2

female species 1 0.00
male species 1 0.00
female species*male species 1 0.00
male origin 1 0.00
female species*male origin 1 0.00
male species*male origin 1 0.00

p—s

female species*male species*male origin 0.00

25

 

 

a) 1-

2’ ab

30.8“ b

a

n.

“50.6-

50.44

‘3

5.0.2-

0- 04 38
CXC

 

 

Iwild male 1] lab male

 

 

 

 

 

 

  

 

 

ab ab 3 ab
b ab
25 21 13
‘ —'|
6x6 CxG GXC
Pairing Type

F igtne 2.3 . Out of all pairings with spermatophore deposition, the proportion leading to ovipo-
sition. Error bars are +1 s.d. and numbers in bars are sample sizes. Bars not sharing a letter
are significantly different fi'om each other at p=0.05. Female canadensis (C) or glaucus (G)
were paired to male canadensis or glaucus, with the female listed first. Filled bars indicate the
males used were wild-caught, open bars indicate the males used were lab-reared.

Table 2.5. Chi-square values from ANOVA of proportions of pairings involving
canadensis and glaucus males and females leading to oviposition, out of all pairings with
spermatophore deposition. The model was a 2x2x2 factorial design with effects being
the species of the male, the species of the female (canadensis or glaucus for each), the
origin of the male (wild or lab), and their interactions.

 

 

Source of variation df x2
female species 1 3.85*
male species 1 0.31
female species*male species 1 0.00
male origin 1 0.31
female species*male origin 1 0.04
male species*male origin 1 0.93
femalgrecies‘male species*male origin 1 2.93#

 

# P5010; * PS0.05

26

2.5). There was also a marginally significant three-way interaction between male
species, female Species, and male origin.

The proportions of pairings leading to production of larvae, of those leading to
oviposition are shown in Figure 2.4. There was a significant effect of male origin (wild-
caught males were more successful than lab-reared), and there was a significant
interaction between male species and male origin (lab-reared canadensis males were less
successful than wild-caught canadensis males, whereas lab-reared glaucus males were
equivalent to wild-caught glaucus males) (Table 2.6). There was also a significant
interaction between female species and male species, with heterospecific pairings slightly
more likely to produce larvae than conspecific pairings.

The average egg hatchabilities of clutches containing at least one hatching egg are
shown in Figure 2.5. There was a significant effect of male species (glaucus males led to
greater average egg hatchability than canadensis males) and a significant effect of male
origin (wild-caught males led to greater average egg hatchability than lab-reared males)
(Table 2.7).

Thus in pairings involving canadensis and glaucus males and females, only one
component of mating success (proportion of pairings producing larvae, out of all leading
to oviposition; Figure 2.4) showed a significant difference between heterospecific and
conspecific success (as shown by a significant female species*male species interaction;
Table 2.6). However, heterospecific pairings were more likely to produce larvae than
conspecific pairings. None of these components of mating success had heterospecific

pairings significantly less successful than conspecific pairings.

27

 

Iwild male 1:11ab male

 

 

 

           

 

 

 

 

 

   

 

 

 

 

m 1-

g 08 a a

E ' ab ab

wo- 0.6 '-

5. 0.4 - bc

5

n. 02 - a

3 O 42 ' . 18 . 57
0x0 6x6 0x6 GxC

Pairing Type

Figure 2.4. Out of all pairings leading to oviposition, the proportion leading to production of
larvae. Error bars are +1 S11 and numbers in bars are sample sizes. Bars not sharing a letter
are significantly difi'erent from each other at F005- Female canadensis (C) or glaucus (G)
were paired to male canadensis or glaucus, with the female listed first Filled bars indicate the
males used were wild-caught, open bars indicate the males used were lab-reared.

Table 2.6. Chi-square values fiom ANOVA of proportions of pairings involving
canadensis and glaucus males and females leading to production of larvae, out of all
pairings leading to oviposition. The model was a 2x2x2 factorial design with effects
being the species of the male, the species of the female (canadensis or glaucus for each),
the origin of the male (wild or lab), and their interactions.

 

 

Source of variation . df 11.2
female species 1 0.63
male species 1 2.28
female species*male species 1 606*
male origin 1 906*"
female species*male origin 1 2.71#
male species*male origin 1 914*"
female species*malegecies‘male origin 1 1.15

 

# P50.10; * P5005; *"‘* P50.005

28

 

 

Iwild male Ellab male

 

 

     

 

    

 

 

 

 

 

 

 

1 _

(D

a aC

g 0.8 ac ac

o 0.6 - bcd Cd

.5 a

t 0.4 -

a T

g 0.2 -

°' 0 7 18 10 4

1 . —-r

CXC GXG CXG GXC

Pairing Type

Figure 2.5. For all pairings producing at least one hatching larva, the mean proportion of viable
eggs (hatching/total eggs). Values expressed are means, enor bars are +1 s.e. and numbers in
bars are numbers of pairings. Bars not sharing a letter are significantly different fiom each other
at p=0.05. Female canadensis (C) or glaucus (G) were paired to male canadensis or
glaucus, with the female listed first Filled bars indicate the males used were wild-caught, and
open bars indicate the males used were lab-reared.

Table 2.7. F -values (Type 111 SS) from ANOVA of proportions of hatching eggs out of
all eggs laid, averaged over all pairings involving canadensis and glaucus males and
females producing at least one hatching larva. The model was a 2x2x2 factorial design
with effects being the species of the male, the species of the female (canadensis or
glaucus for each), the origin of the male (wild or lab), and their interactions.

 

 

Source of variation - df F
female species 1 1.48
male species 1 538*
female species*male species 1 0.73
male origin 1 6.86"
female species*male origin 1 0.89
male species*male origin 1 0.76
female species*male species*male ogin l 1.52

 

* P5005; “ P5001

29

When comparing troilus x glaucus pairings to glaucus x glaucus and canadensis
x glaucus pairings, all pairings that locked together lasted at least 30 minutes. There
were also no significant differences in proportions of pairings resulting in spermatophore
deposition (Figure 2.6) or leading to oviposition (Figure 2.7). Out of all pairings leading
to oviposition, the two types of heterospecific pairings were significantly more likely to
lead to production of larvae than were the conspecific pairings (Figure 2.8). The troilus x
glaucus pairings led to a lower average egg hatchability than the other two types of

pairings (canadensis x glaucus and glaucus x glaucus) (Figure 2.9).

 

Proportion of Pairings

 

GxG CxG TxG
Pairing Type

Figure 2.6. Out of all pairings lasting at least 30 minutes, the proportion of pairings resulting in
spermatophore deposition. Error bars are +1 s.d. and numbers in bars are sample sizes.
Female glaucus (G), canadensis (C), or troilus (T) were paired to wild-caught male glaucus,
with the female listed first.

30

 

 

m 1 2

2’

E 0.8 -

It“

'6 0.6 -

I:

.2 0.4 -

E

g- 0.2 ~

5

n. 0 j 10 36 8
6x6 0x6 TxG

Pairing Type

Figure 2.7. Out of all pairings with spermatophore deposition, the proportion leading to ovipo-
sition. Error bars are +1 s.d. and numbers in bars are sample sizes. Bars not sharing a letter
are significantly different from each other at p=0.05. Female glaucus (G), canadensis (C), or
troilus (T) were paired to wild-caught male glaucus, with the female listed first.

 

   

1 ' b
E: - b
E 0.8 -
a“
946 0.6 -
c a
.2 0.4 -
t
8
2 0.2 -
IL

0 -

GxG CxG TxG
Pairing Type

Figure 2.8. Out of all pairings leading to oviposition, the proportion of pairings leading to
production of larvae. Error bars are +1 s.d. and numbers in bars are sample sizes. Bars not
sharing a letter are significantly different from each other at p=0.05. Female glaucus (G),
canadensis (C), or troilus (T) were paired to wild-caught male glaucus, with the female listed

1-

 

 

a

g 0.8 -
m a
"6 0.6 -
C
.2
1g 0.4 - b
E 0.2 - ——V

0 ' I _

6x6 0x6 TxG

Pairing Type

Figure 2.9. For all pairings producing at least one hatching larva, the mean proportion of viable
eggs (hatching/total eggs). Values expressed are means, error bars are +1 s.e. and numbers in
bars are nmnbers of pairings. Bars not sharing a letter are significantly different from each other
at p=0.05. Female glaucus (G), canadensis (C), or troilus (T) were paired to wild-caught
male glaucus, with the female listed first.

There were no differences between pairings involving hybrid canadensis x
glaucus females and pairings involving females of either parental species (when paired to
canadensis males) in the proportion lasting 30 minutes (Figure 2.10) or the proportion
resulting in spermatophore deposition (Figure 2.11). There also were not differences 1
between pairings involving hybrid canadensis x glaucus females and pairings involving
females of either parental species in the proportion leading to oviposition (Figure 2.12) or
the proportion leading to larvae (Figure 2.13). Pairings between hybrid females and wild
canadensis males did not lead to a significantly lower egg hatchability than pairings with
conspecific canadensis females, however both led to a lower hatchability than pairings
with the heterospecific glaucus females (Figure 2.14).

The proportion of pairings leading to larvae, out of all pairings for which all of the

first four measures were known (pairing duration, spermatophore deposition, whether

32

0.8 7
0.6 -
0.4 -
0.2 7

0 1

 

 

Proportion of Pairings

CxC (CxG)xC ch
Pairing Type
Figure 2.10. Out ofall pairingsthat lasted aminimurn offive minutes, theproportion ofpairings
lasting at least 30 minutes. Error bars are +1 s.d. and numbers in bars are sample sizes. Fe-

male canadensis (C), glaucus (G), or hybrid canadensis x glaucus (C x G) were paired to
wild-caught male canadensis, with the female listed first.

 

 

Proportion of Pairings

CXC (CxG)xC GXC
Pairing Type

Figure 2.1 l . Out of all pairings lasting at least 30 minutes, the proportion of pairings resulting in
spermatophore deposition. Error bars are +1 s.d. and numbers in bars are sample sizes.
Female canadensis (C), glaucus (G), or hybrid canadensis x glaucus (C x G) were paired to
wild-caught male canadensis, with the female listed first.

33

 

   

"5.1 ab
C
E 0.8 - a
:1"
Mo- 0.6 _
C
.2 0.4 —
E
g 0.2 —
n. 0 _
CxC (CxG)xC GxC
Pairing Type

Figure 2.12. Out of all pairings with spermatophore deposition, the proportion of pairings
leading to oviposition. Error bars are +1 s.d. and numbers in bars are sample sizes. Bars not
sharing a letter are significantly different from each other at p=0.05. Female canadensis (C),
glaucus (G), or hybrid canadensis x glaucus (C x G) were paired to wild-caught male
canadensis, with the female listed first.

 

 

on 1 '
2’
"c 0.8 - a
t a a
Mo- 0.6 "
C
.2 0.4 -
1:
8.
2 0.2 -
n.

0 ..

CXC (CxG)xC GXC
Pairing Type

Figure 2.13. Out of all pairings leading to oviposition, the proportion of pairings leading to
production of some larvae. Error bars are +1 s.d. and numbers in bars are sample sizes. Bars
not sharing a letter are significantly different from each other at p=0.05. Female canadensis
(C), glaucus (G), or hybrid canadensis x glaucus (C x G) were paired to wild-caught male
canadensis, with the female listed first.

—L
I

 

 

m
g 0.8 ' b
ur
“6 0.6 .
B
.2
t 0.4 -
o
8
a 0.2 -

o -

CxC (CxG)xC GXC
Pairing Type

Figm'e 2.14. For all pairings producing at least one hatching larva, the mean proportion of
viable eggs (hatching/total eggs). Values expressed are means, error bars are +1 s.e. and
numbers in bars are numbers of pairings. Bars not sharing a letter are significantly difi‘erent fi'om
each other at p=0.05. Female canadensis (C), glaucus (G), or hybrid canadensis x glaucus
(C x G) were paired to wild-caught male canadensis, with the female listed first.

eggs had been laid, and whether larvae had been produced; Table 2.8) were multiplied by
the average egg hatchabilities of each type of pairing (calculated in Table 2.9), to give a
combined index of mating success (calculated in Table 2.10). Heterospecific pairings
have the two highest index values, and conspecific pairings have two of the three lowest
index values, so heterospecific pairings are not at a disadvantage with respect to this
combined index of mating suCcess (Table 2.10). One of the surprisingly low values of
this index is for glaucus x glaucus wild pairings, with an index value of 0.08, the lowest
value for any pairing type involving wild-caught males (Table 2.10). This low value is
mainly due to the low frequency of pairings leading to larvae, out of those with
oviposition (Table 2.8), which could have been affected by the low sample number for

this pairing type.

35

.2 5:85

_.N 853. E 28 3 .N swap—5 mm 852m 5 vows—2: 0.3 .3838 26% 05 we 29: 8 28 89a 83 wee—oa— .mw=Eeg 3.8235

9:3 U x A0 x 08
23, wwueoheeeo x

 

N3 : 2 a a : Agnew x 5.3822

3:3 O x .C

no.0 o h w a a 2:5 3.933% x .352»

32 U x 9

Ed m a N— 2 3 an. Maneuvers x $8sz

3:3 u x 9

$6 m— 3 om mm mm 23» SERVES x 3.033%

32 O x B

mmd A. m“ 2 .m a as 9.833% x mmwnoheeeo

3:3 O x B

mmd m fl 2 mm em on 23> .383er x ammemnezeo

32 O x 2

93 2 cm mm 4N mm as 488% x 483m

8:3 O x Q

:6 ~ c m e a 2:5 .303er x §o=e~m

33 O x 8

:6 m R on :4 3 n2 uwmzmnezeo x Sweeneeeo

com? 0 x UV

Ned m _ 5 mm mm 3” 2m? amazonezeo x 2.5.5338

oatm— oafifl :oummoqgo eoummoaoe 8358 on $328 w

mafia“..— 55 weEBa: 5m? econaouaficonm «$2 8 means “meo— E wash:

wwgn 359.8 55 838:: .8982" 05 5m? BASE .8285: =a .mmeEan Afiwcnmu x m v

we 8:535 05 .805 m0 .805 HO 2: .805 mo 05 .805 .5 mo “BEA—Z 09$. mega

 

.3083 26268ro :2: e5 .33 .«o 83» 829:8 a 5; nose .33 .«o Eon—Ea use 89¢. .w.m 2an

36

Table 2.9. Egg hatchability of broods producing larvae.

 

 

 

 

Pairing Type Number Number Proportion Average
(S2 x cinigin) of eggs of larvae of eggs hatchability

laid hatching hatching for all broods
canadensis x canadensis wild 23 8 0.35 0.49 i 0.06,
(C x C wild) 5 2 0.40 N=17

51 48 0.94

20 7 0.35

31 5 0.16

74 64 0.86

63 35 0.56

119 37 0.31

10 6 0.60

91 29 0.32

104 8 0.08

55 29 0.53

59 33 0.56

48 46 0.96

21 8 0.38

49 36 0.74

14 3 0.21
canadensis x canadensis lab 37 3 0.08 0.40 i 0.12,
(C x C lab) 24 6 0.25 N=7

73 ’ 19 0.26

49 22 0.45

39 31 0.80

68 59 0.87

56 6 0.11
glaucus x glaucus wild 34 17 0.50 0.72 i 0.10,
(G x G wild) 219 104 0.48 N=5

450 445 0.99

89 80 0.90

25 19 0.76

 

37

Table 2.9 (cont’d).

 

 

Pairing Type Number Number Proportion Average
(9 x &rigin) of eggs of larvae of eggs hatchability
laid hatching hatching for all broods
glaucus x glaucus lab 66 31 0.47 0.63 i 0.07,
(G x G lab) 75 65 0.87 N=18
72 l 0.01
107 13 0.12
17 5 0.29
389 261 0.67
188 161 0.86
85 66 0.78
15 4 0.27
27 22 0.82
84 63 0.75
126 121 0.96
9 7 0.78
63 61 0.97
70 63 0.90
54 22 0.41
68 36 0.53
45 40 0.89

 

38

Table 2.9 (cont’d).

 

 

 

Pairing Type Number Number Proportion Average
(9 x &figin) of eggs of larvae of eggs hatchability
laid hatching hatching for all broods
canadensis x glaucus wild 15 11 0.73 0.61 i 0.06,
(C x G wild) 209 62 0.30 N=22
51 37 0.72
39 19 0.49
184 166 0.90
5 5 1.00
5 2 0.40
82 52 0.63
1 l 7 0.64
24 24 1.00
73 48 0.66
30 3 0.10
94 61 0.65
94 61 0.65
66 55 0.83
36 14 0.39
37 24 0.65
21 3 0.14
1 13 1 1 1 0.98
40 9 0.22
40 19 0.48
91 74 0.81
canadensis x glaucus lab 91 71 0.78 0.47 i 0.10,
(C x G lab) 50 35 0.70 N=10
50 48 0.96
96 5 0.05
28 1 0.04
36 9 0.25
14 5 0.36
62 22 0.36
26 22 0.85
32 13 0.41

 

39

Table 2.9 (cont’d).

 

 

Pairing Type Number Number Proportion Average
(9 x &rigin) of eggs of larvae of eggs hatchability
laid hatching hatching for all broods
glaucus x canadensis wild 42 7 0.17 0.66 i 0.04,
(G x C wild) 205 196 0.96 N=37
398 293 0.74
146 102 0.70
115 86 0.75
83 37 0.45
148 128 0.86
102 42 0.41
132 16 0.12
39 36 0.92
95 41 0.43
125 79 0.63
129 126 0.98
70 34 0.49
23 5 0.22
47 37 0.79
113 94 0.83
86 79 0.92
97 94 0.97
99 52 0.52
89 26 0.29
97 95 0.98
350 225 0.64
203 184 0.91
82 22 0.27
29 19 0.66
102 102 1.00
56 6 0.11
170 139 0.82
50 44 0.88
163 126 0.77
67 54 0.81
27 5 0.18
124 74 0.60
209 157 0.75
271 230 0.85
254 224 0.88

 

40

Table 2.9 (cont’d).

 

 

 

 

Pairing Type Number Number Proportion Average
(9 x &rigin) of eggs of larvae of eggs hatchability

laid hatching hatching for all broods
glaucus x canadensis lab 185 66 0.36 0.28 i 0.09,
(G x C lab) 17 8 0.47 N=4

40 1 0.02

29 8 0.28
troilus x glaucus wild 54 13 0.24 0.25 i- 0.03,
(T x G wild) 98 38 0.39 N=6

30 6 0.20

96 27 0.28

40 10 0.25

135 18 0.13
(canadensis x glaucus) 100 29 0.29 0.37 i 0.06,

x canadensis wild 3O 9 0.30 N=l7

((C x G) x C wild) 9 9 1.00

83 9 0.1 1

111 86 0.78

32 15 0.47

29 9 0.31

109 54 0.50

81 10 0.12

120 58 0.48

99 43 0.43

114 31 0.27

57 12 0.21

185 58 0.31

38 l 0.03

132 63 0.48

86 21 0.24

 

41

Table 2.10. Calculation of combined index of mating success. For each pairing type, the
proportion of pairings producing hatching larvae (Table 2.8) is multiplied by the average
hatchability of broods producing larvae (Table 2.9).

 

 

 

Pairing Type Out of pairings lasting Average egg Combined
(9 x &figm) at least 5 minutes, the hatchability of index of
proportion producing pairings mating
larvae producing larvae success
C x C wild 0.42 0.49 0.20
C x C lab 0.11 0.40 0.04
G x G wild 0.11 0.72 0.08
G x G lab 0.40 0.63 0.25
C x G wild 0.52 0.61 0.32
C x G lab 0.33 0.47 0.16
G x C wild 0.59 0.66 0.39
G x C lab 0.19 0.28 0.05
T x G wild ' 0.67 0.25 0.17
(C x G) x C wild 0.52 0.37 0.19
Discussion

I observed no reduction in pairing success (as indicated by copulation duration,
spermatophore deposition, oviposition, and egg hatchability) for heterospecific pairings
between canadensis and glaucus, indicating there is no postpairing, prezygotic
reproductive isolation between these species when females have mated once. Pairings
between troilus females and glaucus males were only less successful in average egg
hatchability. The ability of species as phylogenetically separate as troilus and glaucus to
pair successfully, and to do so with fairly high frequency, is quite impressive. These two
species are probably behaviorally isolated in the wild (but see Deering & Scriber 1998;
documents observation of a courtship and copulation between a tethered canadensis
female and wild male of Papilio palamedes, a member of the P. troilus species group),

but once behavior is superseded, even considerable physiological differentiation does not

42

prevent successful mating. The reduced egg hatchability that was observed (Figure 2.9)
could be due either to an inability to fertilize many eggs (prezygotic isolation) or low egg
viability (postzygotic isolation). Pairings between hybrid females and canadensis males
were also very successful, with only slight reduction in egg hatchability, indicating very
little reduction in fertility in hybrids (at least of this type).

These results do not address courtship or mate recognition early in a pairing,
before the two have locked together. Reproductive isolation at this point would be best
addressed under more natural conditions to minimize the effects of human handling.
However, once pairs have locked together, they seem to progress well despite the
artificial environment of the laboratory.

This study found that for larval production, the minimum copulation duration is
30 minutes, which matched the result of a previous study (Lederhouse et al. 1990).
However, it is possible that the minimum duration could be shorter in nature. The
pairings for this study were carried out at room temperature (75-85 F"), but in the wild,
Papilio butterflies are usually found mating in the early afternoon, during the hottest part
of the day. Warmer conditions might speed up the physiological processes of copulation,
shortening the time required to mate successfully.

Lab-reared males have previously been found to be less reproductively successful
than wild-caught males (Lederhouse et al. 1990). It was concluded that adult nutrition
was to blame, and it was recommended that lab-reared males be fed honey water
supplemented with amino acids and salts to provide the nutrients that males in the wild
presmnably obtain by puddling. However, in this study lab males were fed this solution,

but they were still less successful than wild males in egg hatchability (Table 2.6, Table

43

2.7). Either the honey solution is still missing some important nutrient, or males in the
wild benefit from some other factor that lab males do not get. This effect is also not due
to inbreeding because most lab-reared individuals used in our lab are the offspring of
wild-caught females.

Many females laid only unfertilized eggs, and even females that laid some
fertilized eggs also laid many that were not fertile. This wastefulness of eggs is
surprising, but in line with previous findings with Papilio, both for hand-paired
butterflies (Clarke & Sheppard 1956) and for wild-caught females (Lederhouse & Scriber
1987). This suggests that females are dependent on males to provide adequate
spermatophores and sperm in pairings, and that great variation in male (or male ejaculate)
quality exists (Drummond 1984). However, they seem to have little ability to measure
male quality (at least afier a copulation has progressed) because females will lay eggs
even with no spermatophore present or after short pairings. This seems surprising, but it
is wise to avoid what Eberhard (1996) calls “fertilization myopia”, the thinking that in the
wild all copulations will lead to offspring and all eggs that females lay will be fertile.
This will rarely be the case, so lab findings of ‘wasted eggs’ should not necessarily be
shrugged off as the result of lab conditions.

Since females of both canadensis and glaucus often mate more than once in the
wild (Lederhouse and Scriber 1989; Lederhouse 1995), another aspect of mating that
could be very important is sperm competition (Birkhead & Muller 1998). Some insect
females that have mated to two males, one conspecific and one heterospecifrc, will
produce only conspecifrc offspring (Howard 1999). This potent reproductive barrier will

be investigated in the following chapter.

44

In conclusion, heterospecifrc copulations between canadensis and glaucus were
not less successful than conspecifrc, so there does not appear to be a postcopulatory,

prezygotic barrier to gene flow between these species in singly-mated females.

45

CHAPTER 3:
DOES CONSPECIFIC SPERM HAVE PRECEDENCE IN PAPILIO

CANADENSIS OR P. GLA U C US?

Introduction

There are postcopulatory, prezygotic barriers to gene flow that do not appear in
singly-mated females. Groups of species of insects have been found where heterospecific
pairings are no less successful than conspecifrc pairings when females mate only once,
but when a female is paired to both a heterospecifrc male and a conspecific male, she
produces only conspecifrc offspring, regardless of the order of the pairings (Howard
1999). This is called conspecific sperm precedence, and it can be a potent barrier to gene
flow provided females can be expected to mate with multiple males (Howard et al. 1998).

The multiple-mating swallowtail butterfly species Papilio glaucus and P.
canadensis interbreed to form viable, fertile hybrids (Lederhouse et al. 1989, Scriber et
al. 1995). In the lab, heterospecifrc pairings are no less successful than conspecific I
pairings (Chapter 2). These two species can also form viable hybrids with the more
distantly related western Papilio species P. rutulus, P. ewymedon, and P. multicaudatus
(Scriber et a1. 1995). Weak postzygotic barriers to gene flow between these species may
indicate that prezygotic barriers isolate them (Sperling 1990).

The female reproductive system of the ditrysian Lepidoptera (Figure 1.2) might
facilitate conspecifrc sperm precedence. Because males do not place sperm directly into

the spermatheca of the female, they cannot directly displace the sperm of previous males

46

(Drummond 1984). Females may also be able to choose what sperm is sent to the
spermatheca (Eberhard 1996).

Sperm precedence for an individual doubly-mated female can be expressed as P2,
the proportion of offspring produced afier a second mating that was sired by the second
male (Gwynne 1984). When there is first-male sperm precedence, P2 will be close or
equal to zero for most double pairings, and if last-male sperm precedence is the rule, P2
will usually be close or equal to one. However, with conspecific sperm precedence, P2
will be high when the last male was conspecific and low when the last male was
heterospecific.

To look for conspecific sperm precedence in glaucus and canadensis, I paired
virgin females twice, once to a conspecific male and once to a heterospecific male, and
determined the paternity of offspring using allozyme electrophoresis. I also paired wild-
caught females (that had presumably already mated in the wild to conspecifrc males) to
heterospecific males. In addition to females and males of canadensis and glaucus, we
also used males of the more distantly related species rutulus, eurymedon, and

multicaudatus.

Methods

Both wild—caught and lab-reared male and female butterflies were used for
pairings. Females and males of glaucus and canadensis were used, and males of rutulus,
eurymedon, and multicaudatus were used. Through adulthood, female butterflies were
fed a 20% honey solution and males were fed a 20% honey solution supplemented with

amino acids and salts to increase fertility following Lederhouse et al. (1990). Lab-reared

47

males were not paired for at least two days following adult eclosion to allow reproductive
maturation.

Lab-reared females were hand-paired to males, allowed to oviposit in plastic
oviposition arenas lined with hostplant foliage (Scriber 1993), and remated after two to
six days, again by hand-pairing, to a male of a different species. Females were then
allowed to oviposit again. Table 3.1 shows the number and types of double-pairs made.
Additionally, wild-caught females were allowed to oviposit, then remated after one to
five days by hand-pairing to a male of a different species, and allowed to oviposit again.
Table 3.2 shows the number and types of wild female rematings. Only females that were
actively laying eggs were remated, and the duration of lab pairings were recorded.

Larvae from eggs laid both before and after rematings were collected and reared
on black cherry (Prunus serotina) foliage, a common favorite of tiger swallowtail
species. After reaching approximately the third instar, larvae were frozen at -80°C.
Mothers and male mates were also stored frozen after death. Females were later
dissected to determine how many spermatophores were present at death.

Lab and wild females producing larvae before remating were compared to those
not producing larvae before remating with respect to success in laying eggs and
producing larvae following remating. Data were analyzed using a contingency table
analysis (PROC CATMOD; SAS Institute Inc. 1990).

Allozyme electrophoresis, following Hagen and Scriber (1991), was carried out
on thin-layer cellulose acetate plates (Titan 111, Helena Laboratories, Beaumont TX).
Small larvae were homogenized whole in buffer, and the head and thorax of larger larvae

were homogenized in buffer. With adult males, the distal half of the abdomen was used,

48

and with adult females, the proximal half of the abdomen was used (to avoid including
male allozymes from spermatophores). The enzyme 6-phosphogluconate dehydrogenase
(PGD) was stained for to determine paternity because there are diagnostic differences
between the species of the P. glaucus species group in PGD allozymes (Hagen and
Scriber 1991). There are other enzyme loci with diagnostic differences between species
as well. Lactate dehydrogenase (LDH) and hexokinase (HK) can also be used to
differentiate glaucus and canadensis, but LDH staining was faint for larvae and HK
staining was uninterpretable for larvae. Staining of PGD was fainter for larvae than for
adults, but it was clear and interpretable.

I verified the inheritance of Pgd as well as its expression in larvae. For the
sixteen broods shown in Table 3.3, PGD allozymes were determined for the female, the
first male to mate, and five to ten larvae produced'before the female was remated.
Expected offspring allozymes and proportions were compared to the actual offspring
allozymes and numbers.

The paternity of offspring produced after remating was established by
determining PGD allozymes of larvae produced after remating, several larvae produced
before remating, and both of the males mated (in several cases the males were lost and
not able to be checked). Sperm precedence for each brood was expressed as P2, the
proportion of larvae produced after the remating that were sired by the male used for
remating. For several very large broods, I only determined the paternity of about twenty
larvae produced after the remating rather than the entire brood. Broods where a female
produced no hatching larvae before remating but did produce larvae after are not included

in the tables of results, although the paternity of those larvae was determined.

49

When a doubly-mated female produces a brood of mixed paternity (0<P2<1), the
pattern of sperm use is of interest because it may be a clue for the mechanics of sperm
replacement. For broods of mixed paternity where the production of larvae was spread
out over more than one day, the numbers of larvae sired by each male for each day

following remating was compared.

Results

The number of lab-reared females for each type of double-pairing is shown in
Table 3.1. Out of 82 females, only 32 produced larvae both before and after being
remated. The other 50 either laid no eggs after being remated, laid no hatching eggs after
remating, or had laid no hatching eggs before being remated. Of the 32 females
producing larvae both before and after being remated, five had been mated to males that
shared allozymes (some interspecific introgression is found at the Pgd locus; Hagen et al.
1991 , Chapter 5), making determining paternity of offspring impossible, leaving 27
broods where P2 was determined.

The number of remated wild females is shown in Table 3.2. Out of 27 females,
20 produced larvae both before and after being remated. Six females laid no eggs after
being remated. Almost all females that laid eggs afier remating had larvae hatching from
those eggs. All of the remated wild females produced larvae before being remated. Only
one female that had produced larvae after being remated had undeterrninable P2, leaving
19 broods where P2 was determined.

The comparison of allozymes of parents (a once mated female and her male mate)

to larval offspring allozymes found that larvae had allozymes corresponding to their

50

 

 

5N Nm 9 Nu Nw 30H

onmxev

fl — _ _ ~ 836% x 9:53 x 336%

onszV

o N _ _ _ .38sz x “SENSSEBE x .3983

Azxoxev

~ ~ _ ~ ~ gaenzeufizs x 3.633% x 333%

onmeV

N N N v v 43:63:66 x 2985,98 x fimzmneeeu

axuxov

N N N m n 2636:8236 x £36358 x amazoheeeu

6 x U x 9

m w m— f ON axonsm x £38333 x 3633M

G x O x g

c c a 2 S fianohgeo x 333% x .3283

onoxov

m m o N_ 2 Maneuvers x .36qu x «arouses.»

onoxov

m c w E 3 9.835% x Maneuvers x awmeoweneo
Bamgofiu 09 2:3 mafia?— wfifiafi mesa—E: 838$
um 80:? means—2 Sam 28 Eamon not“ oatm— baa wwwo 3:3

not“ 28 283 outa— oafifi unwed—Ea memos—Goa mar?— -oS=ov mo 3% x m6 x av

mafia—tea 338m any—«Bum moEEom magnum 8:852 25 merged—9:09

 

.wquag wage—Bu 8038 252689: 23 838$ 6688-32 3:365:29 me 8098:: 23 89C. ._.m 033.

51

 

a oN 0N _N N 38.

 

E x 83 Q

m m m m o .4553 x 25 833M

:2 x 23, B

m m m n v 88338338 x 23» 833%

G x 25 Q

m c o o n 88558 x ES, 833%

. 6 x 83 B

N N N N N 83283828 x 23» 833%

6 x 25 B

o o o b w 833% x E?» 338388
888.538 on 2:00 98388 $8388 8388 838$

«m 08:3 8388 82.3 83 803p 8cm 8P3— 83 .3me Emu—8.2m?
833 83 283 88— oatfi 98260.8 8325.8 m8?— 8388 6 x E? mv
$82698 8383 838m 8380a 838cm Mo 85852 25 8383—

 

.wauaEB w830=$ 883 0808882 83 838$ 8930.23, 8388 me 898:: 33 893. .N.m $3.“.

52

parents (Table 3.3). The sample numbers are not high enough to be able to expect to see
the actual frequencies in a large population, but based on these broods there is no reason
to suspect non-Mendelian inheritance.

There was no significant difference in likeliness to lay eggs after being remated
between lab females laying hatching eggs, lab females laying non-hatching eggs, and
wild females laying hatching eggs before being remated (Figure 3.1). However, of
females laying eggs afier remating, females that had laid fertile eggs before the remating
were significantly more likely to lay fertile eggs afier the remating than females that had
laid no hatching eggs before being remated. Wild-caught females were slightly more
likely to lay hatching eggs after being remated than lab females that had been laying
hatching eggs (p=0.09l8).

Most of the broods that P2 was determined for had P2=0 (34 of 46 broods) (Table
3.4, Table 3.5). These broods were spread out through the different‘pairing types and
female origins. However, there were cases of P2>0 through most of the double pairing
types. There were seven broods of mixed paternity.

Durations of second pairings were recorded. No second pairing lasting for fewer
than 30 minutes resulted in sperm replacement (Table 3.4, Table 3.5). No female found
to be carrying only one spermatophore showed any sperm replacement either. However,
most second pairings lasted for longer than 30 minutes, and most females were found
carrying two (or more for wild caught females) spermatophores, and even in many of

these cases P2 was equal to zero.

53

 

82.: x 82- 2

82 32.8 H 82- 88.8

 

” 8:822“ ” 82-82-: ” 8:82-23 ” 82-52238 2 2-82- 2. 22;
82.2 H 82.2 82-58 H 82-23
”82-82-: ” 8282-8 “82-82-828 u 82.52.88 82-82- m2- 222 m x o
82.3 ” 82.32.:- 8228 H 82-82-28 282- 82- ~22
82.2. H 82.32.: 82-2.8 ” 82-82-28 2.82- 82. £2;
82$ ” 8282-2.” 82-2.8 H 82-32-28 «2-82- 82- 32; u x 0
82-2. ” 8282-3 82-28 x 82-82-28 82-82- m2- 322
82-2 H82-82% 82-82 M 82-82-28 82-52- 2. 22;
82.: M 82-32% 82-82 ” 82-32-28 82-82- 2- ~82 o x o
82-: 82-2 82-82- 82. 222
82-: 82-2 82-82- 82- 22;
82-: 82-2 82-82- 82- 22; e x G
83.2 H 82.3 H 82822.. 882.8 N 82-2.8 ” 83-32-23 2-32- 2- 22;
82-: 82-2 «2-32- 2- :2;
82-2. 82-2 2.32. 2- 822 o x o
888% 8 x 8%
80288 23 88823 82225.8 23 88823 DOA 898.8 25
8: 288° .32 858% no.2 mate-.20 8.898 808 222 0.582 238 238

 

.5388: 2 xom 235 0238: 88— 8 8823822 8o$ 8.3sz38 on 2888 90m .20 88980823 .Xx
03 838 83 >X 0.3 838$ 823 6.88% 8 8287K 2 90m 2828-2-20 382 8 man no 803888 83 82883882. .m.m 23H

54

A) Laying eggs after remating

1
0.8 a a a
0.6
0.4
0.2
0 -

 

 

lab female, lab female, wild female,
eggs but no eggs and eggs and
larvae larvae larvae

B) Producing larvae afler remating

1 b‘
0.8 b
0.6
0.4 a
0.2
o - 42 -

 

Proportion of Females

 

lab female, lab female, wild female,
eggs but no eggs and eggs and
larvae larvae larvae

Female Origin and Success Before Remating

Figure 3. 1. Reproductive success of multiply-mated females following remating asa a fimction
of female origin and success before remating. A) Proportion of females laying eggs afier
remating. B) Of females laying eggs, the proportion producing larvae. Error bars are +1 s.d.,
numbers within bars are number of females, and bars with the same letter are not significantly
different at the p=0.05 level. (Note: b is significantly different than b’ at p=0.091 8)

55

Table 3.4. Sperm precedence (the proportion of offspring that were sired by the second
male; P2) for double-paired lab-reared females. Also indicated are the number of larvae
produced after remating that had paternity determined (N), origin of the male used for
remating, the days between pairings, the duration of the second mating, and the number
of spermatophores present in the female at death.

 

 

Double- Female Days Male Duration Spermatophores P2 N

pairing type number between origin of second present

(S? x (31 x pairings mating

62) (minutes)

C x C x G 13088 3 lab 65 2 1 15
13093 3 lab 15 1 0 2
13100 3 lab 100 2 O 27
14100 2 lab 26 1 O 61
14197 3 lab 60 2 0.2 5

C x G x C 13077 6 lab 108 2 0 1
14278 3 lab 59 2 0 7
14279 3 lab 35 1 0 22
14284 2 lab >43 2 0 12
14093 4 wild 57 2 1 21

G x G x C 14280 5 lab 93 1 O 16
14281 2 lab >36 2 0 21
14287 2 lab >38 2 O 23
14288 3 lab 62 2 0.36 11
14289 3 lab 63 2 0 26
14321 3 lab 106 1 0 14

G x C x G 12328 3 lab (?) 2 0.82 11
14103 2 lab 65 . 2 0 1
14192 4 lab 85 2 0 72
14085 2 wild 87 2 0 26
14086 4 wild >30 2 0 23

CxCx E 14251 4 wild >41 2 0 21
14252 2 wild 73 2 0.93 14

C x E x C 14256 2 wild >85 1 O 21
14259 2 wild >91 2 1 5

G x G x M 14277 4 wild >48 2 0 3

GxRx G 14381 1 wild 99 2 0 19

 

56

Table 3.5. Sperm precedence (the proportion of offspring that were sired by the second

male; P2) for remated wild-caught females. Also indicated are the number of larvae

produced after remating that had paternity determined (N), origin of the male used for

remating, the days between collection of the female and remating, the duration of the
remating, and the number of spermatophores present in the female at death.

 

 

Remating Female Days Male Duration Spermatophores P2 N

type number until origin of present

(S2 wild x (3) remating remating

(minutes)

C wild x G 14000 1 wild >45 2 l 2
14004 3 wild 1 15 3 0 1
14005 3 wild 42 2 O 7
14010 4 wild 33 (?) 0 14
14017 3 wild 6O 3 0.14 7
14024 3 wild 72 3 O 5

G wild x C 14330 2 lab 42 1 O 20
14331 2 lab 49 2 O 20

G wild x E 12483 1 wild >40 2 O 3
12484 1 wild >40 4 0.07 14
12485 1 wild 54 2 0.12 8
14301 4 wild 64 1 0 22
14294 4 wild >75 2 O 21

G wild x M 12487 5 wild >40 2 1 1
12488 5 wild >40 (?) 0 2
12496 5 wild >40 3 O 4

G wild x R 12490 5 wild 66 3 O 2
12494 1 wild 27 1 O 11
12590 2 wild (?) 2 0 3

 

57

There were eleven females that produced larvae after remating, but had produced
no larvae before remating. For three of those females, P2 could not be determined. For
seven of the remaining eight females, P2 was equal to one. The eighth, female number
14133, was a glaucus female who had been paired first to a wild canadensis male, laid 37
infertile eggs, and was then paired to a lab glaucus male. After the remating, she
produced a brood with P2=O.27. This indicates that in most of the cases where a female
laying no fertile eggs mates again and starts to produce fertile eggs, she will exclusively
be using the sperm of the most recent male. However, in some cases she might be using
sperm from the earlier male as well, even though before remating that sperm was not
being successfully utilized.

Of the seven mixed broods, three could be divided up by the day that offspring
were produced (Figure 3.2). All three of these had one larva produced the first day
following remating that was sired by the first male, but two of the three had larvae
produced on later days that had been sired by the first male as well. One brood, 14252,
appears to follOw the model of the first egg produced following remating being fertilized

by the first male, followed by eggs fertilized by the second male.

Discussion

I found sperm replacement to be possible in remated Papilio females, but more
commonly females continued to exclusively use sperm from the original mating.
Heterospecific males were no less likely to replace sperm from a previous mating than
conspecific males (and conversely, remated females were just as likely to continue to use

heterospecific sperm from a first mating as conspecific sperm), meaning there was no

58

 

 

 

 

 

 

El sired by first male

A) BFOOd 12323 Isired by second male
12
1o —

8 ..

6 _

4 ..

2 .-

0 ..

firstday second third day fourthday
day
B)Brood14252
12

10y
r—1,l__-,

 

l L l

 

 

ON§QQ
A

 

Number of Larvae

firstday second third day fourthday
day

C) Brood 14288

 

 

 

 

 

T I

first day second third day fourthday
day

Days following remating

Figure 3.2. Number of larvae sired by each male and the days following remating of their
production by females producing mixed broods: A) Brood 12328; B) Brood 14252; C) Brood
l 42 8 8.

59

evidence for conspecific sperm precedence in canadensis or glaucus. The general pattern
seems to be first-male sperm precedence, although with the variability in precedence
found in many insects (Simmons & Siva-Jothy 1998). F irst-male priority is further
supported by the fact that female reproductive success after remating was influenced by
success before remating (Figure 3.1).

The failure of second matings to sire offspring is likely due to many of the same
limitations on pairing success found in first pairings (Chapter 2). Other factors may
enhance this as well. Large spermatophores may act as temporary mating plugs, as they
seem to in some insects (Lorch et al. 1993), similar to the permanent mating plugs some
Papilio males produce (Orr 1995). Large spermatophores might in the wild prolong the
time until a female solicits another mating, as in bushcrickets (Wedell 1993). Thus
increased time between pairings might increase the success of second matings. The use
of lab males also may have reduced replacement success, although wild males were
generally quite unsuccessfiil at replacing paternity as well.

In conclusion, I did not find that conspecific sperm has precedence in either
canadensis or glaucus. Along with the results of Chapter 2, this means that there is no

evidence for postpairing, prezygotic barriers to gene flow between these species.

60

CHAPTER 4:
HIGH LEVELS OF GENE FLOW BETWEEN POPULATIONS OF THE

CANADIAN SWALLOWTAIL, PAPILIO CANADENSIS

Introduction

Hybrid zones are clines maintained by a balance between gene flow and barriers
to gene flow (Barton & Hewitt 1985). This means that in addition to studying
reproductive isolation between the species involved, it is important to study dispersal and
gene flow within both species as well as across the hybrid zone. Potential gene flow
between species (if reproductive isolation, habitat differences, and any other barriers to
interbreeding were to suddenly vanish) is equivalent to the actual gene flow within each
of the species. If the potential for gene flow between species is high, then the barriers
that isolate them must be quite strong.

The swallowtail butterflies Papilio canadensis and P. glaucus have ranges that
meet at a narrOw hybrid zone, and have overlapping flight times. There is no postpairing,
prezygotic reproductive isolation (Chapters 2 & 3), and male behavior (canadensis males
are more attracted to glaucus females than canadensis females, Deering 1998) might
even increase gene flow between species. Hybrids are viable and fertile (Hagen &
Scriber 1995), so as yet no strong barriers to gene flow have been found. However, if
gene flow within both of the species is low, the differences between them could be
maintained by weaker barriers (Barton & Hewitt 1985).

Evidence for high gene flow between widely-distributed glaucus populations has

been found (Bossart & Scriber 1995), but it has not yet been investigated exclusively in

61

canadensis populations. Because of the biology of canadensis, there may be lower gene
flow between its populations than between glaucus populations. Individuals of
canadensis are typically smaller (possibly indicating lower resources for dispersal
flights), undergo obligate pupal diapause (resulting in only one generation per year), and
face a more time-limited growing period (Scriber 1994), all of which could reduce
dispersal and gene flow in canadensis relative to glaucus.

One way to test the strength of gene flow is to sample populations separated by
natural barriers. For example, in checkerspot butterflies gene flow is limited between
populations in mountain areas, but not between plateau populations (Britten et a1. 1995).
The most significant natural barriers in the Great Lakes region are the lakes themselves.
Lakes Michigan and Huron have been found to reduce gene flow between populations of
the butterfly Limem’tis arthemis (Waldbauer & Stemburg 1988), and may do so in other
insect species as well.

A popular approach to studying gene flow has been to estimate it from geographic
patterns of allele distribution (Slatkin 1987). A classical technique for this has been to
use F -statistics (Wright 1931). Using some codominant, genetic characteristic, allele and
genotype frequencies are determined fiom samples of individuals from several
populations. If all individuals are treated as members of a single breeding population, the
reduction in heterozygosity relative to Hardy-Weinberg equilibrium is calculated and
expressed as F n. If there are fewer heterozygotes than expected, F 1r will be positive, and
if there are more than expected, it will be negative. F n can be broken down into two
components: F15, which is the reduction in heterozygotes within the subpopulations, and

F ST, which is the reduction in heterozygotes due to the population being divided into

62

subpopulations. If F ST is significantly larger than zero, it is an indication of reduced gene
flow between subpopulations. Another indirect method to estimate gene flow is to
statistically compare allele frequencies across populations (Raymond & Rousset 1995).
To estimate levels of gene flow between canadensis populations, I sampled
populations throughout the Great Lakes region. Allozyme electrophoresis was used to
determine allele and genotype frequencies at four enzyme loci, and F -statistics and other
statistical methods were used to look for genetic structure (reduced gene flow) between

populations.

Methods

Individuals were collected from six locations in the range of canadensis
throughout the Great Lakes region: one from northeast Minnesota (Cook Co.), two from
the Upper Peninsula of Michigan (Gogebic Co., Dickinson Co.), and three from the
Lower Peninsula of Michigan (Charlevoix Co., Mason Co., Isabella Co.) (Figure 4.1).
All specimens Were collected between 14 May and 23 June of 1998 (peak flight time for
canadensis in Michigan), and stored at -80°C.

Allozyme electrophoresis protocols followed Hagen and Scriber (1991). Samples
were prepared by grinding the distal half of the abdomen for males or the proximal half
of the abdomen for females (to avoid including spermatophore proteins from male mates)

in lOOuL buffer (0.1M tris, 1.07mM EDTA, 0.15mM NAD, 0.13mM NADP, 35.75mM

2-mercaptoethanol, pH 7.0) and centrifuging for 10 minutes at 16,000 x g. Allozymes

were separated by electrophoresis on thin layer cellulose acetate plates (Titan 111, Helena

63

 

(
”a. .-
° 4

o 5 6

 

Figure 4.1. Sample sites. 1: Cook Co., Minnesota; 35 males. 2: Gogebic Co., Michigan; 36
males, 1 female. 3: Dickinson Co., Michigan; 48 males, 20 females. 4: Charlevoix Co.,
Michigan; 50 males, 18 females. 5: Mason Co., Michigan; 50 males, 15 females. 6: Isabella
Co., Michigan; 50 males, 14 females.

64

Laboratories, Beaumont, TX). The four enzymes used (GPI, PGM, HBDH, and PGD)
and the running conditions for each are shown in Table 4.1.

Enzyme stains followed Richardson et a1. (1986). Gels were scored as in Hagen
& Scriber (1991). The most common allozyme for each enzyme was given score ‘100’,
the origin (where samples had originally been applied) was given score ‘0’, and all other
allozymes were given a score corresponding to their location relative to these two points.
Every sample plate was run with at least two previously scored samples to act as internal
standards. These relative migration distance scores were then used as names for different
alleles at the enzyme gene locus.

The program Genepop v3.1 (Raymond and Roussett 1995) was used to test for
linkage disequilibrium, Hardy-Weinberg equilibrium, and allele frequency differences )
between locations. The program F stat v2.8 (Goudet 1997) was used to calculate Wright’s

F-statistics and standard errors.

Table 4.1. Enzymes resolved and running conditions used.

 

 

Enzyme Name (EC. Number) Buffer" Origin Voltage Time

GPI Glucose phosphate I cathode 275V 45 min.
isomerase (5.3.1.9)

PGM Phosphoglucomutase I cathode 27 SV 45 min.
(2.7.5.1)

HBDH Hydroxybutyrate D anode or 300V 90 min.
dehydrogenase (1.1.1.30) cathode"

PGD 6-Phosphogluconate D anode or 300V 90 min.
dehydrogenase (1.1.1.44) cathode"

 

*Buffers (as in Richardson et al. 1986): I=25mM tris, 192mM glycine, pH 8.5; D=15mM
tris, SmM EDTA, 10mM MgClz, 5.5mM boric acid, pH 7.8.
“Under these conditions, HBDH and PGD migrated towards the center of the plate

regardless of origin.

65

Results

When two genes are located close to each other on the same chromosome, they
will tend to be inherited together, acting like a single gene. Within individuals, if certain
alleles at one gene tend to be associated with certain alleles at other genes, they are
probably being inherited as a single unit. This state is called linkage disequilibrium
between those genes. If two genes indicate similar geographic patterns of allele
distribution, they are only independent sources of information if there is not linkage
disequilibrium between those genes. There were no significant p-values from chi-square
tests of linkage disequilibrium between the four enzyme loci used here (Table 4.2),

indicating that these four loci can be taken as independent sources of information.

Table 4.2. Chi-square values from tests for linkage of enzyme loci. The null hypothesis
was Ho: genotypes at one locus are distributed independently from genotypes at the other
locus.

 

Locus pair X2 df P-value

 

Gpi & Pgm 9.845 12 0.630
Gpi & Hbdh 10.307 12 0.589
Pgm & Hbdh 4.386 12 . 0.975
Gpi & Pgd 14.229 12 0.286
Pgm & Pgd 9.555 12 0.655
Hbdh & Pgd 17.025 12 0.149

66

Hardy-Weinberg equilibrium is a neutral state for polymorphic genes: one allele

is not more likely to be selected against (either by natural or sexual selection) or enter or

leave an area than other alleles. P-values for deviations fiom Hardy-Weinberg

equilibrium are nonsignificant at p=0.05 in most locations for all loci (Table 4.3).

However, there are significant deviations from equilibrium in populations for both PGM

(in Cook Co., W and Dickinson Co., MI) and HBDH (in Gogebic Co., MI and Isabella

Co., MI), meaning that equilibrium cannot be assumed for all loci in all populations.

However, in these cases, it is the frequency of genotypes involving rare alleles that

deviates from Hardy-Weinberg equilibrium, meaning that these deviations might be

affected by sample sizes. None of the loci deviate from equilibrium in all populations,

and no population deviates from equilibrium at all loci. This means that forces asting on

these loci (selection, assortative mating, etc.) that would confound gene flow measures

are weak or nonexistent.

Table 4.3. P-values from tests of Hardy-Weinberg equilibrium. For each locus in each
population, Genepop tests the null hypothesis of Hardy-Weinberg equilibrium.

 

 

Location GPI PGM HBDH PGD
Cook Co., MN 0.889 0.014 0.054 1
Gogebic Co., MI 1 0.429 0.035 1
Dickinson Co., MI 0.123 0.004 1 0.397
Charlevoix Co., MI 0.312 0.270 0.576 1
Mason Co., MI 0.742 0.845 0.598 0.076
Isabella Co., MI 0.492 0.777 0.006 1

 

67

Allozyme frequencies are similar in all six canadensis populations for all four loci
(Figure 4.2, Figure 4.3, Figure 4.4, Figure 4.5). However, some fi'equency differences
are seen (e.g. the frequency of GPI'00 is near 90% in Isabella Co., MI but less than 84%
in the other locations, Figure 4.2), and there are significant differences between some
populations at all four loci. Significant overall allele fi'equency differences are found for
GPI, and marginally significant overall differences are found for PGD. Nevertheless,
there is no general pattern of populations separated by the lakes being significantly
different, and neighboring populations are as likely to be different as separated
populations. Two populations with allele fiequencies significantly different at one locus
are generally not different at other loci.

Wright’s F—statistics for these six populations are shown in Table 4.4. All F sr
values are less than 0.01. F ST for PGM was calculated to be less than zero, and for the
other three enzymes, F51 was within its standard error’s range of zero. This indicates that
there is little significant reduction in heterozygosity due to population subdivision. There
may still be genetic structure in these populations (because F ST for three of the four loci is

greater than zero), but if so, it is probably slight.

Table 4.4. Wright’s F -statistics for six canadensis populations through the Great Lakes
region. Standard errors were obtained by jackknifing over populations, and are indicated
in parentheses.

 

Locus F15 (s.e.) Fs'r (s.e.) Fn‘ (s.e.)

GPI 0.002 (0.015) 0.009 (0.009) 0.01 1 (0.016)
PGM 0.051 (0.044) -0.004 (0.002) 0.046 (0.043)
HBDH 0.108 (0.076) 0.002 (0.008) 0.1 10 (0.078)
PGD -0.016 (0.039) 0.005 (0.007) -0.01 1 (0.039)

68

 

 

 

 

 

 

Figure 4.2. GPI allozyme frequencies for six sampled canadensis populations. Populations not
sharing a letter are significantly different at p=0.05. The P—value for the test of overall allele
differentiation is 0.004.

69

 

 

 

[21100

I115 5

El other I

 

 

 

 

J . ‘

Figure 4.3. PGM allozyme frequencies for six sampled canadensis populations. Populations
not sharing a letter are significantly different at p=0.05. The P-value for the test of overall allele
differentiation is 0.561. .

70

 

 

I100
I160
[345

I170 L

Bother ‘

Figure 4.4. HBDH allozyme frequencies for six sampled canadensis populations. Populations
not sharing a letter are significantly different at p=0.05. The P-value for the test of overall allele
differentiation is 0. 185.

 

 

 

 

 

 

71

 

 

0 til 6

 

I-80
121-150

I other ‘

a

 

 

 

Figure 4.5. PGD allozyme frequencies for six sampled canadensis populations. Populations not
sharing a letter are significantly different at p=0.05. The P-value for the test of overall allele
differentiation is 0.060.

72

Discussion

I found little evidence of genetic structuring in Great Lakes area canadensis
populations, which suggests high gene flow between populations. This is in line with
many other results with Papilio species: high gene flow has been inferred in P. hospiton
(Aubert et a1. 1997), P. machaon (Aubert et al. 1997, Hoole et al. 1999), P. glaucus
(Bossart & Scriber 1995), and P. zelicaon (Tong & Shapiro 1989), although isolation has
been found between subspecies of P. troilus (Margraf & Scriber in prep).

The result of high gene flow within Papilio species appears at odds with the local
adaptation that is often found (Bossart & Scriber 1995, Tong & Shapiro 1989, Ayres &
Scriber 1994). For these and other reasons, inferring gene flow from F -statistics based on
allozyme data has been criticized (Bossart & Prowell 1998). However, local adaptation
need not be inconsistent with high gene flow. If the selection on some character is weak,
that selected character can be unlinked from other loci (such as enzyme loci), producing
no allele differentiation at most loci with high differentiation at a few (selected) loci.
This seems possible for such traits as hostplant use efficiency (immigrant individuals will
still be able to survive on the new local host, just with lower efficiency). Still, it is wise
to follow Da1y(1989), who recommends treating a result of high gene flow inferred from
allozymes as a hypothesis of high gene flow, not a concrete conclusion.

These results produce a working hypothesis of high gene flow between
canadensis populations, even those separated by the Great Lakes. This matches a
previous result of high gene flow between glaucus populations (Bossart & Scriber 1995).
Together, these imply that potential gene flow between the two species could be quite

high, which would mean that to produce a hybrid zone as narrow as is found, and to

73

maintain the differences that are found between species (Hagen et al. 1991), barriers to
gene flow between canadensis and glaucus must be quite strong. However, as yet few

strong barriers have been found (see Chapters 2 & 3).

74

CHAPTER 5:
INTROGRESSION OF PAPILIO GLA UC US GENES INTO P. CANADENSIS

POPULATION S

Introduction

Introgression is a special case of gene flow: the passage of alleles from one
species to another, which comes about as a result of successful hybridization. There is
debate as to the importance of introgression to evolution (Arnold et al. 1999), but
regardless of its importance it is informative as to the strength and completeness of
barriers to gene flow between species. Hybridization and introgression can be difficult to
detect based on morphology, but molecular markers can be very powerful in this respect
(Scriber et al. 1995).

Between Papilio canadensis and P. glaucus, introgression has been detected at all
three diagnostic allozyme loci (Pgd, th, and Hk) (Hagen et al. 1991‘). It is thought to be
partially responsible for the appearance of the “spring form” of glaucus: canadensis-like
individuals that appear in early spring glaucus populations (Scriber 1990). However, the
extent of introgression at other loci is unknown. Studying introgression at mitochondrial
genes is of particular interest because it would track maternal inheritance. Recent
phylogenetic studies on Papilio based on mtDNA gene sequences have found sequence
differences between individuals of different species, and these could yield diagnostic
mtDNA markers (Sperling 1993).

In some insect species, it has been found that mtDNA introgresses more readily

than nuclear genes (Aubert & Solignac 1990, Powell 1983). Mitochondrial DNA in

75

Papilio might follow this pattern, introgressing more readily than nuclear enzyme alleles,
or it may be found that there is very limited mtDNA introgression, possibly due to a
Haldane effect weeding out female hybrids more strongly than male (Hagen & Scriber
1995)

Hybrid zones are typically characterized by short, steep clines maintained by
strong selection, flanked on either side by long tails of introgression (Barton & Hewitt
1985). In Chapter 4 I examined allozyme frequencies for PGD, which has fixed
differences for glaucus and canadensis (Hagen & Scriber 1991). This means that for
these populations, introgressed allele fiequencies are already known, providing
information on the length of tails of introgression for PGD. This can provide the basis
for comparisons of introgression of nuclear and cytoplasmic genes. It also allows us to
determine if mtDNA introgression tends to be found in individuals that also carry
introgressed nuclear genes.

I first used canadensis and glaucus individuals from a number of different
geographic locations to verify the fixation of alternate mtDNA haplotypes as revealed by
PCR-RFLP (Polymerase Chain Reaction, followed by Restriction Fragment Length
Polymorphism). The resulting diagnostic molecular marker was then used to compare
mitochondrial introgression to nuclear introgression at the Pgd gene locus in the
canadensis population samples from Chapter 4, plus a glaucus population sampled the
same year. Finally I determined if introgression at one gene tended to be coincidental

within individuals with introgression at other genes.

76

Methods

Fifteen canadensis individuals and seventeen glaucus individuals from a number
of geographic locations, all collected prior to 1997, some stored at -80°C and others
stored as pinned specimens at room temperature (Table 5.1) were used to verify the
consistency of the PCR primer sites and the restriction site that was used. DNA
extraction methods followed Sperling & Hickey (1995). From each specimen, two legs
were plucked and macerated in 800 ii of Lifion buffer (0.2M sucrose, SOmM EDTA,
100mM Tris, and 0.5% SDS). Samples were vortexed and left at room temperature for
30 minutes. Then 100 [.1 8M KoAc was added and each sample was inverted and put on
ice for 60 minutes. Samples were centrifuged for 20 minutes and the supernatant was
transferred to a new tube. Samples were extracted once with phenol and once with
chloroform/isoamyl alcohol (24:1). Samples were then precipitated in isopropanol,
washed with 70% ethanol, then dried and resuspended in 200 pl 1X TE buffer.

The PCR primers that were used had sequences 5’ ATA ATT GGA GGA TTT
GGA AAT TG 3’ and 5’ ATT GTA GTA ATA AAA TTA ATT GCT CC 3’, provided
by F.A.H. Sperling (University of California, Berkeley). These primers were produced as
a result of sequencing work on canadensis and glaucus mitochondrial C01 and C011
genes (Caterino & Sperling 1999), and were expected to produce a DNA fragment 294
base pairs long. Within this fragment were five potentially diagnostic restriction sites,
also provided by Dr. Sperling. I chose a T an restriction site anticipated to be present in
glaucus individuals and absent in canadensis individuals.

PCR was carried out using the above primers in a total reaction volume of 100 ii

using AmpliTaq Gold DNA polymerase in a Perkin Elmer GeneAmp 9600 Cycler. PCR

77

products were verified by running them out on a 2% agarose gel along with a 100bp
DNA ladder, visualized by ethidium bromide (EtBr) under ultraviolet light. PCR
products were then digested by T an restriction enzyme incubated at 65°C for 120
minutes, and digested DNA was also run out on a 2% EtBr agarose gel with a 100bp
DNA ladder for comparison.

To compare cytoplasmic and PGD introgression, the six canadensis populations
from Chapter 4, plus a glaucus population from southern Ohio also sampled in May
1998, were used (Figure 5.1). PGD allozyme determination is described in Chapter 4.
The PGD allozymes from the six canadensis populations described in that chapter were
compared to those determined for the glaucus population. Twelve individuals from each
of the seven populations were randomly chosen, and for these twelve individuals the
mtDNA haplotype (as revealed by T an PCR-RFLP) was determined. Additionally, all

individuals carrying PGD interspecific introgression were haplotyped as well.

Results

All but one of the 32 individuals picked to verify PCR-RFLP had successful PCR
products (Table 5.1). This included both frozen and dried specimens. The one specimen
for which PCR was unsuccessful was a dried canadensis specimen. It is unknown if the
PCR for this individual was unsuccessful due to degraded DNA, lack of primer
correspondence, or an unsuccessful DNA extraction. All other 31 specimens had a PCR
product slightly shorter than 300bp long, exactly as long as would be expected based on

sequencing. No individual produced two PCR fragments.

78

 

 

 

 

Figure 5.1. Sites of six sampled canadensis populations and one sampled glaucus population
collected in May and June 1998. Sampled canadensis populations. 1 : Cook Co., Minnesota;
35 males. 2: Gogebic Co., Michigan; 36 males, 1 female. 3: Dickinson Co., Michigan; 48
males, 20 females. 4: Charlevoix Co., Michigan; 50 males, 18 females. 5: Mason Co., Michi-
gan; 50 males, 15 females. 6: Isabella Co., Michigan; 50 males, 14 females. Sampled glaucus
populations. 7: Lawrence Co., Ohio; 22 males.

79

Table 5.1. Verification of diagnostic mtDNA haplotypes for canadensis and glaucus as
visualized by PCR-RFLP. Frozen specimens had been stored at -80°C, dried specimens
had been stored pinned in drawers at room temperature. The canadensis haplotype (—) is
indicated by the absence of a T an restriction site in the 294bp PCR fragment, the glaucus
haplotype (+) is indicated by the presence of a T an restriction site in the same fi'agment.

 

 

Species Origin Storage mtDNA
haplotype
canadensis S? Fairbanks, Alaska 6/95 frozen (—)
canadensis SB Fairbanks, Alaska 6/95 frozen (—)
canadensis 6 Fairbanks, Alaska 6/95 frozen (—)
canadensis 6 Fairbanks, Alaska 6/87 dried (—)
canadensis 9 Thunder Bay, Ontario 6/95 frozen (-)
canadensis 6 Thunder Bay, Ontario 6/95 frozen (—)
canadensis Q Pancake Bay, Ontario 6/ 95 frozen (—)
canadensis 6 Bayfield Co., Wisconsin 6/95 frozen (—)
canadensis 6 Forest Co., Wisconsin 6/95 frozen (—)
canadensis 6 Lincoln Co., Wisconsin 6/85 dried (—)
canadensis 6 Ontonagon Co., Michigan 6/87 dried *
canadensis 6 Mackinac Co., Michigan 6/96 frozen (—)
canadensis Q Charlevoix Co., Michigan 6/95 frozen (—)
canadensis 6 Manistee Co., Michigan 6/95 frozen (—)
canadensis £2 Isabella Co., Michigan 6/96 frozen (—)
glaucus S? dark Dane Co., Wisconsin 8/83 dried (+)
glaucus S? yellow Dane Co., Wisconsin 8/83 dried (+)
glaucus 6 St. Joseph Co., Michigan 7/95 frozen (+)
glaucus 6 St. Joseph Co., Michigan 7/95 frozen (—)
glaucus 6 Adams Co., Ohio 7/85 dried (+)
glaucus S? dark Lawrence Co., Ohio 9/95 frozen (+)
glaucus 92 yellow Lawrence Co., Ohio 9/95 frozen (—)
glaucus 6 Wise Co., Virginia 8/94 frozen (+)
glaucus 6 Wise Co., Virginia 8/94 frozen (+)
glaucus 6 Clarke Co., Georgia 5/87 dried (+)
glaucus 6 Clarke Co., Georgia 8/95 frozen (+)
glaucus S2 dark Clarke Co., Georgia 8/95 frozen (+)
glaucus 9 yellow Clarke Co., Georgia 8/95 frozen (+)
glaucus 6 Highlands Co., Florida 4/82 dried (+)
glaucus 6 Highlands Co., Florida 9/95 frozen (+)
glaucus SB dark Highlands Co., Florida 9/95 frozen (+)
glaucus S2 yellow Highlands Co., Florida 9/95 frozen (+)

 

"‘ No DNA amplified.

80

None of these fourteen canadensis specimens with successful PCR had a
fragment that was cut by the T an restriction enzyme (Table 5.1). Fifteen of the
seventeen glaucus specimens had PCR fragments that were cut by the T an restriction
enzyme, producing a fragment slightly longer than 200bp long and another fragment that
was not visualized by EtBr (probably because of its size, there is not enough DNA to
fluoresce brightly enough under the UV). Two glaucus individuals had PCR fiagments
uncut by T an. This means that the presence of a T an restriction site in this DNA region
can be taken as a mitochondrial marker for glaucus, and the absence of this site can be
taken as a marker for canadensis.

For the 1998 population samples, relative frequencies of canadensis and glaucus
PGD alleles and mtDNA haplotypes are shown in Figure 5.2. PGD introgression was
found in the three lower peninsula canadensis populations and in the one glaucus
population, all at frequency lower than 0.1 (0.018 in Charlevoix Co., 0.009 in Mason Co.,
0.035 in Isabella Co., 0.068 in Lawrence Co.). Introgression at mtDNA was found only
at Mason and Isabella counties (one out of twelve individuals, 0.083 for both). No .
introgression was found in either Michigan Upper Peninsula population or the northern
Minnesota population.

There were two canadensis individuals carrying mtDNA introgression. One
carried no introgression at any of the diagnostic allozyme loci (Pgd, th, Hk), and the
other carried introgression only at Hk, and was heterozygous at that locus (Table 5.2).
There were eight individuals with introgressed Pgd alleles, and seven of them had no
other introgressed alleles at either of the other enzyme loci or in their mtDNA. The

eighth carried introgression also at Hk (again heterozygous there) but not at the other loci.

81

 
 
  

5’
PQD mtDNA

and MP

 

Figure 5. 2. Diagnostic ' ‘ ruarkcr‘ , ' for six
glaucus population sampled in 1998. Filled portions indicate frequenciesr of canadensis alleles
(for PGD) or haplotypes (for mtDNA) and open portions indicate frequencies of glaucus alleles
or haplotypes. The left column indicates PGD frequencies, the right column mtDNA. Numbers
next to pies indicate the numbers of alleles or haplotypes sampled.

82

Table 5.2. Individuals from 1998 samples carrying introgressed alleles or haplotypes.
Introgressed alleles are underlined.

 

 

 

 

 

Individual mtDNA PGD LDH HK
canadensis Charlevoix MI 644 (—) ~125/;5_0 80/80 1%“ 10
canadensis Charlevoix MI S? 10 (—) M 80 1 10/1 10
canadensis Mason MI 62 (—) -125/fl0 80/80 110/110
canadensis Mason MI 621 (i) -125/-125 80/80 110/110
canadensis Isabella MI 6 5 (fl -125/-125 80/80 100/110
canadensis Isabella MI 6‘ l3 (—) -125/3flQ 80/ 80 110/1 10
canadensis Isabella MI 6 43 (—) -125/;1_m 80/80 110/110
canadensis Isabella MI SB 10 (—) M 80 110/110
canadensis Isabella MI 9 13 (—) L199 40 110/110
glaucus Lawrence OH 613 (+) iii/40° 100/100 100/100
Key ,
Species mtDNA PGD LDH I-IK
canadensis (—) -125 80, 40 1 10
ngaucus (+) ~50, -100 100 100

 

Within individuals, introgression at one locus was usually not coincidental with
introgression at other loci. This means that this introgression was old, rather than the

result of primary hybridization.

Discussion

I found the DNA extractions and PCR reactions to be quite reliable, even when
using small amounts of tissue (from plucked legs) and specimens that had been dried and
stored at room temperature for over twelve years. The Tan restriction site was found to
be almost absent in canadensis populations and almost fixed in glaucus populations
(Table 5.1, Figure 5.2). Individuals of one species carrying the haplotype of the other

were found in both species. This can still represent a diagnostic character, because 1)

83

such individuals were rare, and 2) they had been collected from areas near hybrid zone
areas, so these cases could be explained as introgression.

PGD introgression was found in three canadensis populations and in one glaucus
population sampled in 1998. Because of the high sample numbers for those populations
(44 alleles sampled for the glaucus population, over 100 alleles sampled for these
canadensis populations), the frequencies determined here are probably indicative of what
they were in the wild, and PGD introgression in these populations for that year is
concluded to have been present at low frequencies.

In the 1998 sample, mtDNA introgression was only found in the two
southernmost canadensis populations. However, because of the sample sizes (only
twelve haplotypes sampled for each population) the actual fi'equencies in the populations
cannot be estimated with confidence. Introgression at mtDNA might be at higher
frequency than at PGD in these populations, but a larger sample would be needed to
determine this. Although no introgressed mtDNA was found in the glaucus population in
the 1998 sample, it was found in two glaucus individuals in the initial survey: one
individual was from St. Joseph County in Michigan (which is near the hybrid zone), the
other was from Lawrence County in Ohio (where no mtDNA introgression was found in
1998). Southern Ohio is quite far from the Michigan hybrid zone, but it is near a tail of
canadensis hybridization that extends southward along the Appalachian mountain range.

Introgressed alleles (both in Pgd and in mtDNA) in the 1998 canadensis
populations were only found in the lower peninsula populations (Figure 5.2). This is
evidence for genetic structure between canadensis populations: some reduction in gene

flow between populations. However, the genetic structure is slight (because introgressed

84

alleles are found at low frequency). This means that the small F gT-values found in
Chapter 4 cannot be interpreted as being equal to zero. There is some small reduction in
gene flow between populations. However, gene flow between populations is still quite
strong.

Within individuals, introgression at either mtDNA or Pgd was generally not
coincidental with introgression at other loci (Table 5.2). This indicates that most
introgression was not recent, giving time to separate loci. The introgressed alleles now
may be under negative selection, or they may be merely acting like any other rare alleles.

If introgressed molecular markers are typically noncoincidental within
individuals, then probably introgressed ecological characters (diapause, oviposition
preference, host use ability) and morphological characters (size, wing morphometrics,
laraval characters) will also have become unlinked to other introgressed characters. This
means that if an individual in a canadensis population such as Isabella or Charlevoix
Counties is found with glaucus-like oviposition preference or host use ability, there is no
reason to expect to find other glaucus-like characters.

The presence of introgressed PGD alleles and mtDNA haplotypes indicates that
barriers to gene flow are not complete. However, because it is limited in both frequency
and distance from interspecific populations, the barriers to gene flow that are present

must be quite strong.

85

CHAPTER 6:

SUMMARY AND CONCLUSIONS

I found evidence from allozymes of high gene flow between species of
canadensis through the Great Lakes region (Chapter 4), matching a previous result
showing high gene flow in glaucus (Bossart & Scriber 1995). Introgression of glaucus
genes into canadensis populations and vice-versa was found at both nuclear and
mitochondrial loci (Chapter 5). Mitochondrial introgression indicates that some
introgression is female-mediated, despite potential Haldane effects against female hybrids
(Hagen & Scriber 1995). However, introgression was very limited, both in frequency
within populations and in the length of the tails of introgression.

High gene flow and limited introgression indicates that there must be either strong
barriers to gene flow between species or strong selection against hybrids. So what
maintains the species differences between canadensis and glaucus? ’

There does not appear to be postpairing, prezygotic isolation, at least once mates
have locked together (Chapters 2 & 3). However, it would be worthwhile to investigate
mate recognition very early in the mating, before locking occurs. Some pairings separate
within a couple minutes of the start of the pairing, and if heterospecific pairings do so
more often than conspecific, hybrid production would be reduced. This isolation would
fit with the content of Chapter 2, but it would be difficult to study in the lab using hand-
pairings. Rather, because such isolation would probably be strongly affected by
behavior, it would be better to study naturally initiated matings occurring in more natural

conditions.

86

The other aspect of prezygotic isolation that has not yet been addressed with these
species is female choice during courtship. Male choice has been studied (it might be an
important barrier in glaucus, but it would appear to increase hybridization into
canadensis populations, Deering 1998), and it makes sense to do so in Papilio because of
the costliness of male ejaculates (Gwynne 1984, Lederhouse 1995). However, females
should still be the more discriminating of the two (Darwin 1871), and it has been
observed that glaucus females are able to spum potential (conspecific) male mates (Krebs
1988). Studying female choice of conspecific males versus heterospecific males should
be very important for understanding maintenance of species differences. Intraspecific
mate choice of Papilio glaucus females has been studied in large flight cages by Krebs
(1988), and this approach could be used to study interspecific mate choice as well.

Endogenous selection against hybrids of these species appears to be weak (Hagen
& Scriber 1995). Hybrids are viable and fertile, and the only Haldane effect so far
identified is a slightly higher mortality of glaucus X canadensis female pupae (Hagen &
Scriber 1995). Chapter 2 found that canadensis >< glaucus hybrid females pair with
success equal to pure species canadensis females paired to conspecific males. This is a
measure of hybrid fitness not previously studied in Papilio. However, hybrid breakdown
(endogenous weakness of backcross individuals or F2 hybrids) remains incompletely
studied.

Ecological selection against hybrids is another potential barrier to gene flow
(Sperling 1990). Diapause might be critical: because the canadensis obligate diapause
gene is recessive, most individuals in a canadensis population introgressed for that gene

will not diapause, which would likely be a fatal error in a time-limited growing season.

87

In glaucus populations (which normally undergo two or more flights per year), entering
diapause after the first flight could leave the resulting pupa open to increased predation or
parasitism for the remainder of the summer (West & Hazel 1982). There are a number of
other ecological factors of potential interest: 1) are hybrids sexually attractive; do host
use abilities break down upon backcrossing and other crossing; and are the diagnostic
allozyme loci actually adapted to their respective ranges? Exogenous selection against
hybrids could come in the form of weak selection on a combination of these traits.

There are many important ecological and evolutionary aspects to the study of
hybrid zones (Harrison 1993, Howard & Berlocher 1998). One central area of interest,
which was a focus of this thesis, is the identification of the barriers to gene flow that
maintain differences across clines. Another is the distribution of traits diagnostic for the
two species, and how genes for these traits move within and between populations,
another focus of this thesis. These two swallowtail butterflies provide an excellent
example for the study of the maintenance of species differences acress hybrid zones.

This story is especially interesting because of the intriguing behavior of the males
(Deering 1998), and the high fitness of the hybrids (Hagen & Scriber 1995). Another
advantage this system offers is the number of ecologically important differences between
these species that have been identified (Scriber et al. in press). Continued study of
potential barriers to gene flow between canadensis and glaucus, as well as of clines of
multiple traits, as was done with PGD and mitochondrial DNA in this thesis, will

continue to improve our knowledge of this unique system.

88

APPENDICES

89

APPENDIX 1:

RECORD OF DEPOSITION OF VOUCHER SPECIMENS

90

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.: onnn n1:
VJ

hvvv

Title of thesis or dissertation (or other research projects):

Lack of cryptic reproductive isolation between Papilio
canadensis and Papilio glaucus; and population genetics
near their hybrid zone

 

Museum(s) where deposited and abbreviations for table on following sheets:

Entomology Museum, Michigan State University (MSU)

Other Museums:

Investigator's Name (5) (typed)

.__Aram_Daniel_S£ump

 

 

 

Date 21 August 2000

 

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

Deposit as follows:

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

Copies: Included as Appendix 1 in c0pies 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.

91

APPENDIX 1.1:

VOUCHER SPECIMEN DATA

92

APPENDIX 1.1
voucher Specimen Data

Pages

__"__

of

I

Page

     

\ OuwD
nmwwmvxnmeW\.\dw rlxurru: \\
.8
huamuo>acb eumuw an «sea: any 8% “mouse
new messaeoam veuuaa o>oae onu vo>wooom
morooow .oz uosuso>

 

doom um=u=q _~ mama

 

 

 

nabum Hoaamn 5324
Aboahuv Amvmamz m.u0umwuumo>eH

Amummmoooa ma mucosa Hmeowuuvvm away

 

 

 

 

 

 

 

 

 

 

 

 

wewuoon .z

8882 22:8 on

.00 ooaouamq ono

"afiwfiuo Hmauoumx

am: a ofiqufi wooum voumouinma

3:22:03 .3

umefi um=w=< 0H

.00 mucouama ono

"afiwwuo Hmauouwz

am: a mfifimfi vooum woumoulnmg

wafiuooo .2

2882 um=m=< cm

.00 mucousmq ono

"awwfiuo Hmauoumz

am: 2 wmamfi vooum woumoulnma

zoamusmz .h

Nam2 um=w=< 32

pm: 2 .oo «sumac <Hemomo .2 msoamam osfiaamm
Jun? vmuqmonev mam mom: no vmuooHHoo coxmu umcuo no moauoqm

m e r r m m e .m w mcoEHuoam new wumv Henna
exams-amps”...
w.n e t t .d .o u m. a .8
deiOAAPNLE

 

 

"mo wonssz

 

 

93

APPENDIX 1.1
Vbucher Specimen Data

Pages

L

of

_2_

Page

can:

noumuao

 

hufinuubwa: ouuum ammunouz any a2 admoaov

 

oooN um=w=< 28 «2.8

 

 

.aaom:=_mwoaoaoucm

 

you ndofiaouau vuuoHH m>onw any vo>2uuum

 

maaum Hmaamn amu<

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

melooom .oz nosoao> Avunhuv Auvuauz m.uouam2uno>du
. Ahuunmuuoa m2 muuwnm HQGOfiufivvu umav
waaummn .2
8882 82:8 28
.oo meHmno um 2230mm2z
“cwwfiuo 2maumumz
am: 2 momc2 vooum vmummulnmq
8882 82:8 28
.09 mmaumzo um wusommfiz
"afiwwuo Hmcumumz
pm: 2 ~2mq2 vooum vmummulnmA
8882 82:8 88
muu< mane muw>2m «02:8
:8: N .08 nammow .28 :882202x .2 8822022 02228m8
0232000: .a
8882 88: RN
um>2m mumwnm
am: 2 .oo :om22828 «2:20222mo
ma22smz .m.m
nam2 hash mm hnuHM
am: 2 ~ .00 comuowmoh ovmnoaou maumvsm02u258 0222mmm
Jun? wouwuonov can vane no vuuucHHou noxuu vogue no «vacuum
m e r r m m e .m u unusauuam you muwv Henna
e r 0.0 e .m .1 a w s
“mammdmwm...”
M v.04; n._A .A D. .L nu
"mo umpsbz

94

APPENDIX 1.1
Vbucher Specimen Data

Pages

of

3

Page

uuua

acuwuao

 

8888 888888 28 8888

 

 

madmuu>apa uuuum cowficv2z any ad uwmonuv

 

.Ebum:z_mw02050ucm

 

you mfiofiwuuno vuuuqa w>oau may vu>aouum

 

monooom .02 .2052; 2.0983

meaum 2026mm amu<

Anvmamz m.uoumw8umu>=H

Amummmmuma ma muumsm Hacoauwvvu many

 

sz

sz

88:

am:

N
p.

«228000: .9

8882 888888

suaoaaoz

woo £20m commuo
228309 .m

8882 8288 8

mmmm muoaom

.00 0:02 mwauom2amo
”damage Hmaumumz

omm22 mooum moummulnwA
228309 .m
8882 82:8 88
m2auom22wo
Ham3on .m
8882 8288 88
88 8838828
mfiauom22mu

. OD OUMHOQ HQ

.08 onwuoa 2m

 

mmUSA m=2=u=u 0822nmm

 

mausu aovmahusm afiafinmm

 

Museum
where

depos-
ited

 

Other

Adults 6'

Pupae
Nymphs

 

 

 

 

 

Adults 9

Larvae

 

Eggs

vauuuonov van van: no kuuoaaou
060520080 you aunt Hanna

 

coxuu nunuo no uoquoam

 

 

"we nunaaz

 

 

95

APPENDIX 1.1
Vbucher Specimen Data

 

Pages

.4._

of

Page

 

cyan

mugmum>2ab ouuum unwasofiz any :2 uwuonuv

 

Houmuao

 

8888 888888828 8888

 

 

.eaoonz 8&02050ucm

 

now maoauummu vouu82 u>opu uzu v0>2uoum

 

qaaum 202ama amu<

monocou .oz uuzono> Amunmuv Amvusmz m.uouww«umu>uH

xsou2uA£mmuuoa «22 3002.3 2920222228286 33

 

.8

8821

amt

sz

H

0882 «can N2

vmom Hoosom cam

.88 88aam z<82882z
“:2m220 Hmauuumz

822m vooum vuummulnmg
8822883 .3

8882 82:8 8

.oo cowNGOuao z<02moH2
":2w2uo Hmaumumz

momw2 vooum woumwulnmg
xsoufiu .m

oma2 wash «2

humumaou zwwmmo

.88 8882888: z<82882z
“afiwfiuo Hacksaw:

mm2m vooum vmummulnmq
umnfiuom .2.8

@882 scam Q2

.88 82288882 2882882:
ucfiwauo Hmauoumz

«M2N2 wooum wmummulnmq

ammuow w vafisumnuom
wfimamvmcmo owawmmm

 

Museum
where

 

depos-
ited

Other

 

 

 

 

Adults 6'
Adults 9
Pupae
Nymphs

 

Larvae
Eggs

 

wouquonov can can: no wouumaaou
sausaouam you 8288 Huang

 

Gonna umcuo no uo2uonm

 

"mo noneaz

 

 

 

96

LITERATURE CITED

97

LITERATURE CITED

Albuquerque, G.S., C.A. Tauber & M.J. Tauber. 1996. Postmating reproductive
isolation between Chrysopa quadripunctata and Chrysopa slossonae:
Mechanisms and geographic variation. Evolution 50: 1598-1606.

Arnold, M. L., M. R. Bulger, J. M. Burke, A. L. Hempel & J. H. Williams. 1999.
Natural hybridization: How low can you go and still be important? Ecology 80:
371-381.

Aubert, J ., B. Barascud, H. Descimon & F. Michel. 1997. Ecology and genetics of
interspecific hybridization in the swallowtails, Papilio hospiton Géné and P.
machaon L., in Corsica (Lepidoptera: Papilionidae). Biol. J. Linn. Soc. 60: 467-
492.

Aubert, J. & M. Solignac. 1990. Experimental evidence for mitochondrial DNA
introgression between Drosophila species. Evolution 44: 1272-1282.

Avise, J .C. 1994. Molecular markers, natural history and evolution. Chapman and Hall,
New York.

Ayres, M.P. & J .M. Scriber. 1994. Local adaptation to regional climates in Papilio
canadensis (Lepidoptera: Papilionidae). Ecological Monographs 64: 465-482.

Barton, NH. 1979. The dynamics of hybrid zones. Heredity 43: 341-359.

Barton, N.H. & B.O. Bengtsson. 1986. The barrier to genetic exchange between
hybridizing populations. Heredity 56: 357-3 76.

Barton, N.H. & G.M. Hewitt. 1985. Analysis of hybrid zones. Ann. Rev. Ecol. Syst. 16:

113-148.

Barton, N.H. & J .S. Jones. 1983. Mitochondrial DNA: new clues about evolution.
Nature 306: 317-318.

Birkhead, T.R. & A.P. Mallet (eds.). 1998. Sperm competition and sexual selection.
Academic Press, San Diego.

Bossart, J .L. & D.P. Prowell. 1998. Genetic estimates of population structure and gene
flow: Limitations, lessons and new directions. Trends Ecol. Evol. 13: 202-206.

Bossart, J .L. & J .M. Scriber. 1995. Maintenance of ecologically significant genetic

variation in the tiger swallowtail butterfly through differential selection and gene
flow. Evolution 49: 1163-1171.

98

 

"3L

r“h_‘“ 1.28.1
I . __ .

 

Britten, H. B., P. F. Brussard, D. D. Murphy & P. R. Ehrlich. 1995. A test for isolation-
by-distance in central Rocky Mountain and Great Basin populations of Edith’s
checkerspot (Euphydryas editha). J. Hered. 86: 204-210.

Butlin, R. 1989. Reinforcement of premating isolation. in D. Otte & J .A. Endler (eds.),
Speciation and its consequences. Sinauer, Sunderland, Massachusetts.

Campbell, NA. 1993. Biology. Benjamin/Cummings, Redwood City, California.

 

Caterino, M.S. & F.A.H. Sperling. 1999. Papilio phylogeny based on mitochondrial
cytochrome oxidase I and II genes. Mol. Phylogenet. Evol. 11: 122-137.

Clarke, C.A. & P.M. Sheppard. 1956. Hand-pairing of butterflies. Lepid. News 10: 47-
53.

 

Coyne, J .A. & H.A. Orr. 1989. Two rules of speciation. in D. Otte & J .A. Endler (eds.)
Speciation and its consequences. Sinauer, Sunderland, Massachusetts.

Daly, J. C. 1989. The use of electrophoretic data in a study of gene flow in the pest
species Heliothis armigera (Hiibner) and H. punctigera Wallengren (Lepidoptera:
Noctuidae). in H. D. Loxdale & J. den Hollander, (eds.) Electrophoretic studies
on agricultural pests. Clarendon Press, Oxford.

Darwin, C. 1859. The origin of species by means of natural selection. Murray, London.
Darwin, C. 1871. The descent of man, and selection in relation to sex. Murray, London.

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

Deering, M.D. & J .M. Scriber. 1998. Heterospecific mating behavior of Papilio
palamedes in Florida (Lepidoptera: Papilionidae). Holarctic Lepidoptera 5: 49-
51.

De Villiers, P.S. & S.A. Hanrahan. 1991. Sperm competition in the Namib desert beetle,
Onymacris unguicularis. J. Insect Physiol. 37: 1-8.

Dobzhansky, T. 1951. Genetics and the origin of species. Columbia University Press,
New York.

Drummond 111, BA. 1984. Multiple mating and sperm competition in the Lepidoptera.
in R.L. Smith (ed.), Sperm competition and the evolution of animal mating
systems. Academis Press, Orlando, Florida.

99

Dufour, L. 1844. Anatomic generale des dipteres. Ann. Sci. Nat. 1: 244-264.

Eady, P. 1994. Sperm transfer and storage in relation to sperm competition in
Callosobruchus maculatus. Behav. Ecol. Sociobiol. 35: 123-129.

Eberhard, W.G. 1992. Species isolation, genital mechanics, and the evolution of
species-specific genitalia in three species of Macrodactylus beetles (Coleoptera,
Scarabeidae, Melolonthinae). Evolution 46: 1774-1783.

Eberhard, W.G. 1994. Evidence for widespread courtship during copulation in 131
species of insects and spiders, and implications for cryptic female choice.
Evolution 48: 711-733.

Eberhard, W.G. 1996. Female control: Sexual selection by cryptic female choice.
Princeton University Press, Princeton, New Jersey.

Eberhard, W.G. & C. Cordero. 1995. Sexual selection by cryptic female choice on male
seminal products — a new bridge between sexual selection and reproductive
physiology. Trends Ecol. Evol. 10: 493-495.

George, J .A. & M.G. Howard. 1968. Insemination without spermatophores in the
oriental fruit moth, Grapholitha molesta (Lepidoptera: Tortricidae). Can. Ent.
100: 190-192.

Goudet, J. 1995. FSTAT (version 1.2): A computer program to calculate F -statistics.
86: 485-486.

Gregory, P.G. & D.J. Howard. 1993. Laboratory hybridization studies of Allonemobius
fasciatus and A. socius (Orthoptera: Gryllidae). Ann. Entomol. Soc. Am. 86: 694-
701.

Gregory, P.G. & D.J. Howard. 1994. A postinsemination barrier to fertilization isolates
two closely related ground crickets. Evolution 48: 705-710.

Gwynne, D.T. 1984. Male mating effort, confidence of paternity, and insect sperm
competition. in R.L. Smith (ed.), Sperm competition and the evolution of animal
mating systems. Academis Press, Orlando, Florida.

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

Hagen, R.H., R.C. Lederhouse, J .L. Bossart & J .M. Scriber. 1991. Papilio canadensis
and P. glaucus (Papilionidae) are distinct species. J. Lepid. Soc. 45: 245-258.

100

 

Hagen, R.H. & J .M. Scriber. 1989. Sex-linked diapause, color, and allozyme loci in
Papilio glaucus: Linkage analysis and significance in a hybrid zone. J. Heredity
80: 179-185.

Hagen, R.H. & J .M. Scriber. 1991. Systematics of the Papilio glaucus and P. troilus
species groups (Lepidoptera: Papilionidae): Inferences from allozymes. Annals
Entomol. Soc. Am. 84: 380-395.

Hagen, R.H. & J .M. Scriber. 1995. Sex chromosomes and speciation in tiger
swallowtails. in Scriber, J .M., Y. Tsubaki, & R.C. Lederhouse (eds.).
Swallowtail butterflies: Their ecology and evolutionary biology. Gainesville, FL:
Scientific Publishers.

Harrison, R.G. 1983. Barriers to gene exchange between closely related cricket species.
1. Laboratory hybridization studies. Evolution 37: 245-251.

Han'ison, R.G. 1993. Hybrids and hybrid zones: historical perspective. In R.G. Harrison
(ed.) Hybrid zones and the evolutionary process. Oxford University Press, New
York.

Harrison, R.G. 1998. Linking evolutionary pattern and process: the relavence of species
concepts for the study of speciation. In DJ. Howard & S.H. Berlocher, (eds.)
Endless forms: Species and speciation. Oxford University Press, New York.

Hewitt, G. M. 1988. Hybrid zones - natural laboratories for evolutionary studies.
Trends Ecol. Evol. 3: 158-167.

Hewitt, G.M., P. Mason & R.A. Nichols. 1989. Sperm precedence and homogamy
across a hybrid zone in the alpine grasshopper Podisma pedestris. Heredity 62:
343-353.

Hoffman, A.A. & M.W. Blows. 1994. Species borders: ecological and evolutionary
perspectives. Trends Ecol. Evol. 9: 223-227.

Hoole, J .C., D.A. Joyce & A.S. Pullin. 1999. Estimates of gene flow between
populations of the swallowtail butterfly, Papilio machaon in Broadland, UK and

implications for conservation. Biological Conservation 89: 293-299.

Howard, DJ. 1999. Conspecific sperm and pollen precedence and speciation. Annu.
Rev. Ecol. Syst. 30: 109-132.

Howard, D.J. & S.H. Berlocher, (eds.). 1998. Endless forms: Species and speciation.
Oxford University Press, New York.

101

 

 

Howard, D.J., P.G. Gregory, J. Chu & M.L. Cain. 1998. Conspecific sperm precedence
is an effective banier to hybridization between closely related species. Evolution
52: 5 1 1-5 16.

Karl, S.A. & J .C. Avise. 1992. Balancing selection at allozyme loci in oysters:
implications from nuclear RF LPs. Science 256: 100-102.

Katakura, H. 1986. Evidence for the incapacitation of heterospecific sperm in the female
genital tract in a pair of closely related ladybirds (Insecta, Coleoptera,
Coccinellidae). Zool. Sci. 3: 115-121.

Krebs, RA. 1988. The mating behavior of Papilio glaucus (Papilionidae). J. Res.
Lepid. 26: 27-31.

LaMunyon, C.W. & T. Eisner. 1994. Spermatophore size as determinant of paternity in
an arctiid moth (Utethesia ornatrix). Proc. Natl. Acad. Sci. USA 91: 7081-7084.

Lederhouse, RC. 1982. Factors affecting equal catchability in swallowtail butterflies,
Papilio polyxenes and P. glaucus. Ecol. Entomol. 7: 379-383.

Lederhouse, RC. 1995. Comparative mating behavior and sexual selection in North
American swallowtail butterflies. in Scriber, J .M., Y. Tsubaki, & R.C.
Lederhouse (eds.). Swallowtail butterflies: Their ecology and evolutionary
biology. Gainesville, FL: Scientific Publishers.

Lederhouse, R.C., M.P. Ayres & J .M. Scriber. 1989. Evolution of spermatophore counts
in studying mating systems of Lepidoptera. J. Lepid. Soc. 43: 93-101.

Lederhouse, R.C., M.P. Ayres & J .M. Scriber. 1990. Adult nutrition affects male virility
in Papilio glaucus L. Funct. Ecol. 4: 743-751.

Lederhouse, R.C. & J .M. Scriber. 1987. Ecological significance of a postmating decline
in egg viability in the tiger swallowtail. J. Lepid. Soc. 41: 83-93.

Lorch, P.D., G.S. Wilkinson & P.R. Reillo. 1993. Copulation duration and sperm
precedence in the stalk-eyed fly Cyrtodiopsis whitei (Diptera: Diopsidae). Behav.
Ecol. Sociobiol. 32: 303-311.

Luebke, H.J., J .M. Scriber & B.S. Yandell. 1988. Use of multivariant discriminant
analysis of male wing morphometrics to delineate a hybrid zone for Papilio
glaucus glaucus and P. g. canadensis in Wisconsin. American Midl. Nat. 119:
366-3 79.

Mallet, J. 1995. A species definition for the modem synthesis. Trends Ecol. Evol. 10:
294-299.

102

 

Mayr, E. 1963. Animal species and evolution. Harvard University Press, Cambridge.

McMillan, W.O., C.D. Jiggins & J. Mallet. 1997. What initiates speciation in passion-
vine butterflies? Proc. Natl. Acad. Sci. USA 94: 8628-8633.

Nakano, S. 1985. Effect of interspecific mating on female fitness in two closely related
ladybirds (Henosepilachna). Kontyu 53: 112-119.

Orr, AG. 1995. The evolution of the sphragis in the Papilionidae and other butterflies.
in Scriber, J .M., Y. Tsubaki, & R.C. Lederhouse (eds.). Swallowtail butterflies:
Their ecology and evolutionary biology. Gainesville, F L: Scientific Publishers.

Parker, GA. 1970. Sperm competition and its consequences in the insects. Biol. Rev.
45: 525-567.

Porter, A. H., R. Wenger, H. Geiger, A. Scholl & A. M. Shapiro. 1997. The Pontia
daplidice-edusa hybrid zone in northwestern Italy. Evolution 51: 1561-1573.

Porter, A. H. & A. M. Shapiro. 1990. Lock-and-key hypothesis: Lack of mechanical
isolation in a butterfly (Lepidoptera: Pieridae) hybrid zone. Ann. Entomol. Soc.
Am. 83: 107-114.

Powell, J .R. 1983. Interspecific cytoplasmic gene flow in the absence of nuclear gene
flow: Evidence from Drosophila. Proc. Natl. Acad. Sci. USA 80: 492-495.

Price, C.S.C. 1997. Conspecific sperm precedence in Drosophila. Nature 388: 663-666.

Raymond, M & F. Rousset. 1995. GENEPOP (version 1.2): Population genetics
software for exact tests and ecumenicism. J. Hered. 86: 248-249.

Richardson, B.J., P.R. Braverstock & M. Adams. 1986. Allozyme electrophoresis: A
handbook for animal systematics and population studies. Academic Press, San
Diego.

SAS Institute Inc. 1990. SAS/STAT® User’s Guide, Version 6, Fourth Edition. SAS
Institute Inc., Cary, North Carolina, USA.

Scriber, J .M. 1986. Origins of the regional feeding abilities in the tiger swallowtail
butterfly: Ecological monophagy and the Papilio glaucus australis subspecies in

Florida. Oecologia 71: 94-103.

Scriber, J .M. 1988. Tale of the tiger: Beringial biogeography, binomial classification,
and breakfast choices in the Papilio glaucus complex of butterflies. Pp. 241-301
in K. C. Spencer, ed. Chemical mediation of coevolution. Academic Press, New
York.

103

Scriber, J .M. 1990. Interaction of introgression from Papilio glaucus canadensis and
diapause in producing “spring form” eastern tiger swallowtail butterflies, P.
glaucus (Lepidoptera: Papilionidae). Great Lakes Entomol. 23: 127-138.

Scriber, J .M. 1993. Absence of behavioral induction in oviposition preference of Papilio
glaucus (Lepidoptera: Papilionidae). Great Lakes Entomol. 26: 81-95.

Scriber, J .M. 1994. Climatic legacies and sex chromosomes: Latitudinal patterns of
voltanism, diapause, size, and host-plant selection in two species of swallowtail
butterflies at their hybrid zone. Pp. 133-171 in H. V. Danks, ed. Insect life-cycle
polymorphism: Theory, evolution, and ecological consequences for seasonality
and diapause control. Kluwer Academic, Dordrecht, The Netherlands.

Scriber, J .M., MD. Deering & A.D. Stump. 1998. Evidence of long range transport of a
dark morph swallowtail butterfly (Papilio glaucus) on a storm fi'ont into northern
Michigan (Lepidoptera: Papilionidae). Great Lakes Entomol. 31: 151-160.

Scriber, J .M., MD. Deering & A.D. Stump. In press. Hybrid zone ecology and
swallowtail speciation: Geographic and genetic distance influence behavioral,
biochemical, and ecological trait clines in North American Papilionid butterflies.
In C. Boggs, W. Watt & P. Ehrlich, eds. Ecology and evolutionary flight:
Butterflies as model study systems. University of Chicago Press, Chicago, IL.

Scriber, J .M., R.C. Lederhouse, & R.V. Dowell. 1995. Hybridization studies with North
American swallowtails. in Scriber, J .M., Y. Tsubaki, & R.C. Lederhouse (eds.).
Swallowtail butterflies: Their ecology and evolutionary biology. Gainesville, FL:
Scientific Publishers.

Shapiro, L.H. 2000. Reproductive costs to heterospecific mating between two
hybridizing katydids (Orthoptera: Tettigoniidae). Ann. Entomol. Soc. Am. 93:
440-446.

Simmons, L.W. & M.T. Siva-Jothy. 1998. Sperm competition in insects: mechanisms
and the potential for selection. In T.R. Birkhead & A.P. Moller (eds.) Sperm
competition and sexual selection. Academic Press, San Diego.

Siva-Jothy, M.T. & Y. Tsubaki. 1994. Sperm competition and sperm precedence in the
dragonfly Nanophya pygmaea. Physiological Entomology 19: 363-366.

Slatkin, M. 1985. Gene flow in natural populations. Ann. Rev. Ecol. Syst. 16: 393-430.

Slatkin, M. 1987. Gene flow and the geographic structure of natural populations.
Science 236: 787-792.

Sperling, F.A.H. 1990. Natural hybrids of Papilio (Insecta: Lepidoptera): Poor
taxonomy or interesting evolutionary problem? Can. J. Zool. 68: 1790-1799.

104

Sperling, F.A.H. 1993. Mitochondrial DNA variation and Haldane’s rule in the Papilio
glaucus and P. troilus species groups. Heredity 71: 227-233.

Sperling, F .A.H. 1994. Sex-linked genes and species differences in Lepidoptera. Can.
Entomologist 126: 807-818.

Sperling, F .A.H. & D.A. Hickey. 1995. Amplified mitochondrial DNA as a diagnostic
marker for species of conifer-feeding Choristoneura (Lepidoptera: Tortricidae).
Can. Entomol. 127: 277-288.

Sperling, F.A.H. & J .R. Spence. 1991. Structure of an asymmetric hybrid zone between
two water strider species (Hemiptera: Gerridae: Limnoporus). Evolution 45:
1370-1 3 83.

Tong, M.L. & A.M. Shapiro. 1989. Genetic differentiation among California
populations of the anise swallowtail butterfly, Papilio zelicaon Lucas. J. Lepid.
Soc. 43: 217-228.

Tschudi-Rein, K. & G. Benz. 1990. Mechanisms of sperm transfer in female Pieris
brassicae (Lepidoptera: Pieridae). Ann Ent. Soc. Am. 83: 1158-1164.

Wade, M.J., H. Pattemson, N.W. Chang & N.A. Johnson. 1994. Postcopulatory,
prezygotic isolation in flour beetles. Heredity 72: 163-167.

Waldbauer, G.P. & J .G. Stemburg. 1988. Lakes Michigan and Huron limit gene flow
between the subspecies of the butterfly Limenitis arthemis. Can. J. Zool. 66:
1790-1795.

Wedell, N. 1993. Spermatophore size in bushcrickets: Comparative evidence for nuptial
gifts as a sperm protection device. Evolution 47: 1203-1212.

West, D.A. & W.N. Hazel. 1982. An experimental test of natural selection for pupation
site in swallowtail butterflies. Evolution 36: 152—159.

Wright, S. 1951. The genetical structure of populations. Ann. Eugen. 1: 323-334.

105

ll11111111111111111111|