MSU RETURNING MATERIALfiS:
Place in book drop to
LJBRARJES remove this checkout from

n. your record. FINES will
be charged if book is
returned after the date
stamped be10w.

 

 

 

 

 

 

_

(3"5953- C3Co

 

a. .‘ '. 4 ‘~ ‘i'
r- ‘0 ‘1‘; .‘

,1

ill " -' 3;;
JUN 2 3 2012

d" .

 

 

 

 

 

 

THE BIOCHEMICAL GENETICS OF HOST RACE FORMATION AND
SYMPATRIC SPECIATION IN RHAGOLET IS POM ONELLA
(DIPTERA: TEPHRITIDAE)

Jeffrey Lee Feder

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Zoology

1989

545 4(033

ABSTRACT

THE BIOCHEMICAL GENETICS OF HOST RACE FORMATION AND
SPECIATION FOR RHAGOLET IS POMONELLA

By

Jeffrey Lee Feder

The allopatric separation of populations is widely thought to be a
prerequisite for speciation in animals. However, true fruit flies in the
Rhagoletis pomonella (Walsh) sibling species group have been proposed to
speciate sympatrically (that is, in the absence of geographic isolation) as a
consequence of shifting and adapting to new host plants. The recent shift of
the species R. pomonella from its native host hawthom (Crataegus mollis
Scheele) to introduced domestic apple (Malus pumila Mill) provides a test
of whether host specialization is sufficient to differentiate fly populations
without geographic barriers to gene flow.

A test of the sympatric speciation hypothesis was conducted
by examining hawthorn and apple infesting populations of R.
pomonella for evidence of genetic differentiation using the
technique of starch gel electrophoresis. Six allozyme loci (Mali c
enzyme, Aconitase-Z, Mannose phosphate isomerase, NADH-diaphorase-Z,
Aspartate aminotransferase-Z and Hydroxyacid dehyrodenase) consistently

showed significant allele frequency differences at paired apple and

hawthorn sites across the eastern United States. These six loci mapped to
three different regions of the genome and linkage disequilibrium was
observed between non-allelic genes within each region. Allele frequencies
for five of the six loci displaying host associated differences also co-varied
significantly with latitude. Inter-host genetic differentiation is therefore
superimposed on clinal patterns of variation within the host races such that
the magnitude of host associated divergence is a frmction of latitude. In
addition, the allozyme studies suggested that ambient temperature is an
important environmental parameter influencing host-parasite interactions
for R. pomonella as allele frequencies for Malic enzyme and Aconitase-Z
for hawthorn flies correlated closely with the configuration of irregular
iso-thermal clines in Michigan and Wisconsin. Further support for the
temperature/development hypothesis came from the fact that frequency
clines for both host races were steepest across three regions in the Midwest
that corresponded with major ecological transition zones. The results
from this Dissertation therefore confirm that partially reproductively
isolated "host races" can rapidly evolve in sympatry and implicate
differences in host recognition and host associated development (related to
ambient temperature) as key traits restricting gene flow between hawthorn
and apple populations.

Dedication

This Dissertation is dedicated to Benjamin Walsh, the father of sympatric
speciation for phytophagous insects. Also to my family and to Emma Wang
for their long-standing support, encouragement and love, without which I

would not have been able to complete this study.

iv

Acknowledgments

I would like to thank my major professor Dr. Guy L. Bush for
stimulating my interest in evolutionary biology and for involving me in
research on Rhagoletis fruit flies. Special mention must also be made for
my Doctoral committee consisting of Dr. Donald 0. Straney, Dr. Thomas
Friedman, Dr. Wes Brown and Dr. J. Mark Scriber for having the patience
to allow me to mature and develop as a research scientist and person.
Charley Chilcote and Dan Howard played key roles in perfecting the
electrophoretic techniques and discussing the experimental strategies which
were used in this Dissertation. I would also like to thank Scot M. Williams,
Stewart H. Berlocher and Bruce A. McPheron for reading and making
useful comments on an earlier draft of this Dissertation. Jill Millett, Bruce
RJ. Angus MacMillan, Tim A. "Slim" Hunt, Miss Anne Wertheirner, '
Kristy Placido, Jo Ann Palma, Sue Jamrog, Angel Lake, James Edward
Zablotny ("Z"), Todd Bierbaum, Scot "Phineas" Williams, Matthew D.B.
Liebold, Syd Hess, Sue Bourbeau, and Don Swiderski assisted in numerous
aspects of the study including fly collecting, fly rearing and gel running.
Ron Prokopy and Sylvia S. Cooley kindly supplied me with flies from
Massachusetts, Stewart H. Berlocher, Bruce A. McPheron and D. Courtney
Smith with flies from Illinois, Ann L. Averill and Harvey Reissig with flies

from New York and Willis Neilson with flies from Canada. Bruce S. Weir
provided me with useful information on F-statistics and linkage analysis.
Fred Numberger, Doug Clark, Wayne M. Wendland and Ken Scheeringa
were gracious in giving me access to weather station data from the states of
Michigan, Wisconsin, Illinois and Indiana, respectively. I would also like to
thank Ed and Jim Hansen and Mr. & Mrs. Melvin Turner for the use of
their land. Suzanne Tolin, Angela Sorrells and Lynda Belisle lent much
appreciated secretarial help. I am also grateful to Karen Cebra for drawing
Figures 1, 2, 3, 5, and 9. Dorie Bush played a critical part in assuring that I
never lacked for nutritional fortitude or friendly banter. Have I missed
anybody? If so, thanks. Finally, a small but special honorable mention
must also go to incipient Dr. Ronald W. DeBry for finishing his dissertation

on schedule with mine.

vi

Table of Contents

List of Tables ................................................................................ xi
List of Figures .............................................................................. xix
Introduction .................................................................................. 1
Modes of Speciation ............................................................... 1
Allopatric Models of Speciation ................................................ 2
Sympatric Models of Speciation ................................................ 5
Summary of Similarities and Differences between Sympatric
and Allopatric Modes of Speciation .......................................... 8
Rhagoletis pomonella as a Model System of Host Race Formation
and Sympatric Speciation ......................................................... 9
Starch Gel Electrophoresis of Soluble Proteins ........................... 11
Life History of R. pomonella ................................................... 12
F-statistics for the Analysis of Population Structure ................... 12

vii

Organization of the Body of the Thesis ...................................... 13

Chapter 1 Inheritance and Linkage Relationships for Allozymes

in Rhagoletis pomonella ............................................. 15
Introduction ......................................................................... 15
Materials and Methods ........................................................... 16

Crossing of R. pomonella ............................................. 16

Electrophoretic Methods ............................................. 17

Genetic Nomenclature ................................................. 19

Calculation of Linkage Disequilibrium ............................ 19
Results and Discussion ............................................................ 22

Segregation Studies ...................................................... 22

Linkage Results ........................................................... 25

Linkage Disequilibrium and Genetic Variation in the
Laboratory Stock ........................................................ 31

Analysis of Disequilibrium between Laboratory and
Natural Populations ................................................... 36

Chapter 2 Genetic Differentiation between Sympatric Host Races

of the Apple Maggot Fly, Rhagoletis pomonella ............ 40
Introduction ........................................................................ 40
Materials and Methods ........................................................... 41

Results ................................................................................ 41

Discussion ............................................................................ 50

Chapter 3 The Geographic Pattern of Genetic Differentiation

between Host Races of Rhagoletis pomonella ................... 54
Introduction .......................................................................... 54
Materials and Methods ............................................................ 60
Results .................................................................................. 60
Discussion ............................................................................. 82

Chapter 4 Gene Frequency Clines for Rhagoletis pomonella

Host Races in the mid-Westem United States .................. 107
Introduction ........... V .............................................................. 107
Materials and Methods ........................................................... 111
Results ................................................................................. 1 17
Discussion ........................................................................... 138

Chapter 5 Local and Microgeographic Genetic Variation between
Host Races of Rhagoletis pomonella ............................. 161

Introduction ......................................................................... 161

ix

Results .............................................................................. 170
Acon-Z ................................................................... 170

Me and Mpi ............................................................ 177

Act-2 and Did-2 ........................................................ 182

Had ...................................................................... 1 84

Discussion .......................................................................... 184
Chapter 6 Concluding Summary and Remarks ............................ 200
List of References ....................................................................... 207

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

List of Tables

Enzyme systems resolved for the laboratory strain of R.
pomonella. Isozymes that migrated nearest to the cathode
were designated 1, the second nearest 2, etc. ...................... 18

Null hypotheses (HI-H4) concerning tests of linkage
disequilibrium within gametes (Dw) and between united
gametes (Db). ............................................................... 21

Segregation results from test crosses for six newly

resolved allozymes for the apple maggot. Data for Aat-I

are given separately according to sex (F=female, M=ma1e).

The number of test crosses pooled for each cross type

is given in parentheses following the numerical designation

for the cross type. .......................................................... 23

Map distances (in centimorgans) calculated between loci
from single pair crosses. ................................................. 29

Linkage disequilibrium between pairs of non-allelic genes

in the laboratory stock of R. pomonella. Results are given

for disequilibrium within gametes (Dw), between united
gametes (D1,) and for a composite measure of Dw and D1,

(A, Burrows value). See Table 6 for explanations of test
hypotheses H1-H4. ...................................................... 32

xi

Table 6.

Table 7.

Table 8.

Table 9.

Table 10.

Table 11.

Table 12.

Table 13.

Observed and expected heterozygosity and the presumed
number of alleles lost in the laboratory stock compared to
natural apple maggot populations. ................................. 34

Correlation coefficients calculated between pairs of non-
allelic genes for the laboratory stock and a natural apple fly
population collected from Grant, Michigan. ...................... 37

Allozymes resolved for the host races of R. pomonella.

Enzyme abbreviations are given in parentheses. Isozymes

that migrated the nearest to the cathode are designated

system 1, the second nearest 2, etc. ................................ 44

Allele frequencies for Me, Acon-Z, Mpi, Aat-Z, Did-2 and
Had for flies collected from hawthorn and apple trees at the
Grant,Michigan study site. n = sample size. ...................... 46

G-contingency tests for allele frequency differences for

Me, Acon-Z, Mpi, Aat-Z, Did-2 and Had at the Grant,

Michigan site. Degrees of freedom are given in

parentheses. ................................................................ 47

Correlation coefficients (rAB) calculated between pairs of
non-allelic genes based upon disequilibrium values. .......... 49

Collecting sites for R. pomonella across the eastern United
States and Canada. A=apple, H=hawthorn, sym = paired

site where hawthorn and apple trees sampled were within

100 meters or less of one another. ................................ 61

G-contingency tests for allele frequency heterogeneity at

paired sites between hawthom and apple infesting

populations of R. pomonella. Degrees of freedom for tests

are given in parentheses. .................................... 65

xii

Table 14.

Table 15.

Table 16.

Table 17.

Table 18.

Table 19.

Table 20.

Table 21.

Table 22.

G-contingency tests for intra-host allele frequency
heterogeneity across years and sites. Degrees of freedom

for tests are given in parentheses. .................................

FST values across years for apple and hawthorn

populations at site 1 (Ephraim, Wisconsin). ......................

FST values across years for apple and hawthorn

populations at site 5 (Grant, Michigan). ...........................

Fs'r values across apple and hawthorn populations
calculated by the method of Weir and Cockerham (1984).

66

67

68

Negative Fs'r values are reported as 0.0000. .................... 70

Results of linear regression analysis of arcsin
transformed allele frequencies and latitude across the 10
paired apple and hawthorn sites. r2=explained genetic
variation. Asterisk(s) denote significant deviations of
regression coefficients (i.e., slopes of regression lines)

from 0. ......................................................................

Correlation coefficients between non-allelic genes
(Cockerham and Weir, 1977) for apple populations across

the eastern United States and Canada. ...........................

Correlation coefficients between non-allelic genes
(Cockerham and Weir,l977) for hawthorn populations

across the eastern United States and Canada. ......................

Allele frequencies for Me and Had for hawthorn (H) and
apple (A) populations of R. pomonella across the eastern

United States and Canada. ...........................................

Allele frequencies for Acon-Z for hawthorn (H) and apple
(A) populations of R. pomonella across the eastern United

73

80

States and Canada. ...................................................... 96

Table 23.

Table 24.

Table 25.

Table 26.

Table 27.

Table 28.

Table 29.

Table 30.

Table 31.

Allele frequencies for Mpi for hawthorn (H) and apple (A)
populations of R. pomonella across the eastern United
States and Canada. ..................................................... 97

Allele frequencies for Aat-Z and Did-2 for hawthorn (H)
populations of R. pomonella across the eastern United
States and Canada. ..................................................... 98

Allele frequencies for Act-2 and Did-2 for apple (A)
populations of R. pomonella across the eastern United
States and Canada. ..................................................... 99

Allele frequencies for Pep-2 for hawthom (H) and apple
(A) populations of R. pomonella across the eastern United
States and Canada. .................................................... 100

Allele frequencies for Aat-I for hawthorn (H) and apple
(A) populations of R. pomonella across the eastern United
States and Canada. .................................................... 101

Allele frequencies for Pgi for hawthorn (H) and apple (A)
populations of R. pomonella across the eastern United
States and Canada. .................................................... 102

Allele frequencies for Pgm for hawthorn (H) and apple (A)
populations of R. pomonella across the eastern United
States and Canada. .................................................... 103

Allele frequencies for Acy for hawthom (H) and apple (A)
populations of R. pomonella across the eastern United
States and Canada. .................................................... 104

Allele frequencies for Idh for hawthom (H) and apple (A)
populations of R. pomonella across the eastem United
States and Canada. .................................................... 105

xiv

Table 32.

Table 33.

Table 34.

Table 35.

Table 36.

Table 37.

Allele frequencies for Ak for hawthorn (H) and apple (A)
populations of R. pomonella across the eastem United
States and Canada. .................................................... 106

Numerical designations and descriptions of field sites.

Sites were generally organized into five latitudinal

transects (see Figure 5 and Table 34). Latitude and

growing degree days base 50°F are also given for

sites. ...................................................................... 1 14

G-contingency tests for allele frequency heterogeneity

between paired apple and hawthom populations. Degrees

of freedom for tests involving each locus are given in
parentheses following locus abbreviations. At site 15 both
adults (Ad) and larvae (L) were tested for inter-host

variation. .................................................................. 1 l8

G-contingency tests for allele frequency heterogeneity
among apple and hawthorn populations along five
latitudinal transects in the midwestern United States. ........ 121

F57 values (Weir and Cockerham, 1984) across the 28

paired apple and hawthorn populations analyzed in the

study. R refers to the PST value calculated by pooling

all alleles remaining for a locus besides those listed in

the table. .................................................................... 122

First order linear regression analyses between arcsin
transformed allele frequencies and either latitude or

growing degree days base 50°F (GDD) for apple

populations along either Michigan/Indiana or

Wisconsirr/lllinois transects. r2 = amount of genetic

variation explained by the linear regression. ................... 124

XV

Table 38.

Table 39.

Table 40.

Table 41.

Table 42.

First order linear regression analyses between arcsin
transformed allele frequencies and either latitude or

growing degree days base 50°F (GDD) for hawthorn
populations along either Michigan/Indiana or

Wisconsin/Illinois transects. r2 = amount of genetic

variation explained by the linear regression. ................... 125

Allele frequencies for Acon-Z for hawthorn (H) and apple

(A) populations of R. pomonella across the mid-westem

United States. At sites 4, 15 and 17 where both larvae and
adults were collected, (L) refers to larvae and (Ad) to

adults. Sites are arranged within the five transects

according to latitude from north to south. .................... 151

Allele frequencies for Me, Had and Did-2 for hawthom (H)

and apple (A) populations of R. pomonella across the mid-
westem United States. At sites 4, 15 and 17 where both

adults and larvae were collected, (L) refers to larvae and

(Ad) to adults. Sites are arranged within transects

according to latitude from north to south. .................... 153

Allele frequencies for Mpi for hawthorn (H) and apple (A)
populations of R. pomonella across the mid-westem United
States. At sites 4, 15 and 17 where both adults and larvae

were collected, (L) refers to larvae and (Ad) to adults.

Sites are arranged within transects according to latitude

from north to south. .................................................... 155

Allele frequencies for Aat-Z for hawthorn (H) pOpulations

of R. pomonella across the mid-westem United States. At

sites 4, 15 and 17 where both larvae and adults were

collected, (L) refers to larvae and (Ad) to adults. Sites are
ananged within the five transects according to latitude

from north to south. .................................................... 157

xvi

Table 43.

Table 44.

Table 45.

Table 46.

Table 47.

Table 48.

Table 49.

Table 50.

Allele frequencies for Act-2 for apple (A) populations of

R. pomonella across the mid-westem United States. At

sites 4, 15 and 17 where both larvae and adults were

collected, (L) refers to larvae and (Ad) to adults. Sites are
arranged within the five transects according to latitude

from north to south. .................................................... 159

G-contingency tests for intra- and inter-host allele
frequency heterogeneity at site B. ............................... 171

G-contingency tests for intra-host allele frequency
heterogeneity across years at site B and among local

apple (A) and hawthorn (H) populations. Degrees of

freedom for tests are given in parentheses. ..................... 173

Fs'r values calculated by the method of Weir and
Cockerham (1984) across years for apple and hawthorn
trees at site B. ............................................................... 174

Distribution of f values (2 F13 of Wright, 1951) for

individual populations calculated by the method of

Barker, East and Weir (1986). The number of f values
significantly greater or less than 0 are given in

parentheses. ............................................................... 176

PM values for local apple and hawthorn populations
calculated by the method of Weir and Cockerham (1984).... 178

Fs'r values for hawthorn and apple trees at site E in 1985
and 1987 calculated by the method of Weir and Cockerham
(1984). .................................................................... 179

Allele frequencies for Me, Acon-Z and Mpi for individual

apple (A) trees collected at site B from 1984 to 1987. See

Figure 9c for a diagram of site B which includes apple

tree designations. N = sample size. ............................... 194
xvii

Table 51 .

Table 52.

Table 53.

Table 54.

Table 55.

Allele frequencies for Me, Acon-Z and Mpi for individual
hawthorn (H) trees collected at site E from 1984 to 1987.

See Figure 90 for a diagram of site B which includes

hawthorn tree designations. N = sample size. ............... 195

Allele frequencies for Did-2, Ant-2 and Had for individual

apple (A) trees collected at site B from 1984 to 1987.

See Figure 9c for a diagram of site E which includes

apple tree designations. N = sample size. ..................... 196

Allele frequencies for Did-2, Aat-Z and Had for individual
hawthorn (H) trees collected at site B from 1984 to 1987.

See Figure 9c for a diagram of site B which includes

hawthorn tree designations. N = sample size. ................ 197

Allele frequencies for Me, Acon-Z and Mpi for local apple
(A) and hawthorn (H) populations. See Figure 9a,b for
study site designations. N = sample size. ........................ 198

Allele frequencies for Did-2, Aat-Z and Had for local apple

(A) and hawthonr (H) populations. See Figure 9a,b for
study site designations. N = sample size. ........................ 199

xviii

Figure 1.

Figure 2.

Figure 3.

List of Figures

Linkage group assignments and map distances (in

centimorgans) between loci for R. pomonella.

Assignments and distances given in parentheses are based

on data from Berlocher and Smith (1983) ...................... 27

Scale diagram of study site located near the town of Grant,

in western Michigan. The site is an old field which has
remained fallow since at least 1922 (E. Hansen, pers.

comm.) Host tree designations (A=apple, H=hawthorn) are
given along with collecting dates and inter-tree

distances (in meters). ...................................................... 43

Collecting sites for R. pomonella across the eastern United
States and Canada (see Table 12 for descriptions of study
sites). Dotted portion of map designates the region where
flies were that observed attacking apples. Dotted and
shaded areas encompass the distribution of apple flies when
the first published report of the race was made by Ward
(1866). Years are given for the first observation of the
apple race in a region. The current distributions of the
apple and hawthorn races are indicated by the solid and
dashed lines, respectively. Figure 1 is a modification of
Figure 3 from Bush (1969a). ........................................... 57

xix

Figure 4.

Figure 5.

Figure 6.

Average arcsin transformed allele frequency for M e1 00 ,
Acon-295 ,Mpi37, Die-2100 and Had100 for paired

apple and hawthorn sites plotted against latitude. Site 8
(Kenton, Ohio) is not included because data for Dia-2100

are missing for this site. The linear regression equation is

given along with the amount of explained variation (r2) and
probability levels (13) for the AN OVA F-test of the null
hypothesis that the slope of the regression line = 0. ........... 75

Collecting sites for R. pomonella in the midwestem United
States (see Table 33 for detailed descriptions of study sites).
Solid lines indicate isothermal Clines for growing degree
days base 50°F (GDD). GDD values for Illinois and
Indiana sites are given in parentheses. Dashed lines connect
sites comprising each of the five latitudinal transects
collected in the study. 0: site where the frequency of
Acon-275 was > 0.100 for either apple or hawthorn flies.
0 = site where the frequency of Acon-275 was > 0.100
for both apple and hawthorn flies. A = site where the
frequency of Acon-289 was > 0.200 for hawthorn flies
and/or > 0.100 for apple flies. .................................... 113

Average arcsin transformed allele frequency for Me] 00,
Acon-295, Mpi3 7, [lat-2100, Did-2100 and Had] 00 for

apple and hawthom populations along Michigan/Indiana

and Wisconsin/Illinois transects plotted against growing

degree days base 50°F. The first order linear regression
equations are given for apple and hawthorn flies along with

r2 values (the amount of genetic variation explained by the
regression equation) and the probability level (P) for the

AN OVA F-test of the null hypothesis that the slope of the
regression line = 0. .................................................... 128

XX

Figure 7.

Figure 8.

Figure 9.

Allele frequencies for M e1 00, Acon-295 and Mpi3 7 for

apple and hawthorn populations along transect 1
(Michigan/Indiana) and transect 4 (W isconsin/lllinois).

Sites are arranged by decreasing latitude from left to right
along the x-axis. Site 10 is included as the northern most

site for transect 1. .................................................... 131

Allele frequencies for [lat-2100, Did-2100 and Had] 00 for
apple and hawthorn populations along transect 1
(Michigan/Indiana) and transect 4 (W isconsin/Illinois).

Sites are ananged by decreasing latitude from left to right
along the x-axis. SitelO is included as the northern most

site for transect l. .................................................... 133

Collecting sites for R. pomonella in western Michigan.

Panel b is a diagram of local populations sampled within a

60 km2 area near the town of Grant, Michigan.

0 = apple tree, A = hawthorn tree. Panel c gives the
distribution of apple and hawthorn trees at site B.

A=apple, H=hawthom. ............................................... 167

xxi

Introduction

"In dim outline evolution is evident enough. But that particular and
essential bit of the theory of evolution which is concerned with the
origin and nature of species remains utterly mysterious."

William Bateson 1922
Evolutionary faith and modern doubts
Science 55 :55-61 .

Modes of Speciation

Although written almost 70 years ago Bateson's statement is still
largely valid today. Empirically demonstrating the mechanistic basis for
speciation (the cladistic splitting of evolutionary lineages in sexually
reproducing organisms) remains one of the most difficult tasks facing
evolutionary biologists. As enumerated by MJ.D. White (1978), a fully

documented history of speciation must include:
1.) A precise map of the present geographic distributions of taxa.

2.) Geological and climatological evidence suggesting past
distributions.

3.) A detailed morphological description of the taxa.

4.) Ecological data on preferred habitats and the life history of the
organisms including information on any daily, seasonal or

1

yearly cycles.

5.) Behavioral information regarding any etlrological isolating
mechanisms or habitat selection for the taxa.

6.) A detailed genetic description of the taxa based on cytological,
biochemical and molecular data.

7.) An extensive study of morphological and genetic geographic
variation.

8.) Results of experimental hybridization studies in the laboratory.

9.) Information on the frequency of hybridization in nature,
provided of course, that the taxa co-occur in at least some
portion of their respective ranges.

Allopatric Models of Speciation

Needless to say there is no example of speciation for which all of this
information is available. As a result, much attention has been focused on
the theoretical basis for speciation. The traditional view of speciation first
advanced by Wagner (1868, 1889) and later more clearly defined by Jordan
(1896) requires the complete geographic separation (allopatry) of con-
specific populations as a pre-requisite for divergence. This classic
"dumbbell" model of speciation (Mayr, 1942) involves three different
stages. 1.) The range expansion of a population into a new area. 2.) The
development of a geographic barrier between the two areas. 3.) The
evolution of genetic modifications within populations in the two areas such
that when the populations again come into contact they will be kept apart by

genetic isolating mechanisms. The emphasis on the need for the complete

3

geographic isolation of populations stems from the view that the genome is
a highly integrated genetic system maintained at the status quo by a balance
between epistatic interactions among genes and inter-demic gene flow. As
stated by Mayr (1963, pp. 518, 521):

"The real problem of speciation is not how to produce
difference but rather how to escape from the cohesion of the gene
complex. No one will comprehend how formidable this problem is
who does not understand the power of the cohesive forces that are
responsible for the coadapted harmony of the gene pool ...... The
importance of complete isolation becomes evident as soon as the
extensive epistatic effects of genes are properly taken into
consideration. As a consequence, a population cannot change
drastically as long as it is exposed to the normalizing effects of gene
flow."

As a consequence, Mayr (1963, pg. 525) believes that " genetic
differences between populations are built up not by the independent changes
of gene frequencies at different loci, but by the harmonious restructuring of
the total genotype." Speciation is therefore the process whereby
reproductive isolation develops between populations as a by product of the
breakdown and restructuring of the genetic system within geographically

isolated demes experiencing different selection pressures.

"As soon as they are separated, two populations will drift apart
in their genetic contents for a number of reasons. The probability
is nil that the same mutations will occur in the two populations in
the same sequence. Each incorporated mutation changes the genetic
background of the population and thus affects the selective value of
all subsequent mutations. Furthermore, recombination will
produce different genotypes in the two gene pools, and, since the
same gene may have different selective values in different
genotypes, this will likewise lead to a gradual shifting of gene
frequencies ...... The selective agents to which the two separated
populations are exposed are also not the same, since there are no
two places on the face of the earth where even the physical

4

environment is quite identical. Where a population is completely
isolated , the composition of the biota is invariably different and
this shift of the biotic environment results in an additional powerful
difference in selection pressure. This is true for competition,
predation, and whatever other ecological interactions may exist.

These local conditions exert selection pressures leading to a
steady change of gene contents and to the development of numerous
new adjustments. Having to become adapted to a new environment
may lead to a sufficiently drastic genetic reconstruction to affect the
nature of isolating mechanisms, including potential sterility
barriers with sister populations."

E. Mayr (1963, pp. 527,528)

An important aspect of the classic "dumbbell" model is that
population sizes are assumed to be large during all stages of speciation.
Consequently, the evolution of genetic isolating mechanisms between
allopatric populations is a slow and gradual process. However, range
expansions of populations during stage 1 of allopatric speciation may
involve only a few colonists. This consideration and earlier theoretical
work by Wright (1940a,b), led Mayr (1942) to propose an auxillary form
of allopatric speciation based on the "founder principle" in which a new
isolate formed by only a few individuals could rapidly undergo the genetic
turnover necessary for speciation. As discussed by Mayr (1942, 1954,
1963):

"The founder population differs from the parental population
not only in the drastic reduction of the diversity of its gene pool,
but also by being exposed to a totally new constellation of
environmental factors, biotic as well as genetic ones. The most
important of these is the sudden conversion from an open to a
closed-and at that a small closed-population ...... After conversion of
the population into a closed one, homozygosity increases
inevitably ...... and as a consequence of their increased frequency in

the founder population, homozygotes will be much more exposed
to selection and those genes will be favored which are specially

5

viable in homozygous condition. Thus the soloist is now the
favorite, rather than the good mixer. We come thus to the
important conclusion that the mere change of the genetic
environment may change the selective value of a gene very
considerably (and particularly the change from an open to a closed
population). This change, in fact, is the most drastic genetic change
(except for polyploidy and hybridization) which may occur in a
natural population, since it may affect all loci at once. Indeed, it
may have the character of a veritable genetic revolution.
Furthermore, this genetic revolution, released by the isolation of
the founder population, may well have the character of a chain
reaction. Changes in any locus will in turn affect the selective value
at many other loci, until finally the system has reached a new state
of equilibrium ...... Therefore, during a genetic revolution, the
population will pass from one well integrated and stable condition
through a highly unstable period to another period of balanced
integration ...... I use the term genetic revolution advisedly, because
the genetic reconstitution was in all these cases far too drastic to be
described in terms of a change of gene frequencies. Surely, a
change of gene frequencies also occurred, but it was associated with
a sufficiently drastic replacement of supergenes that it resulted in
the acquisition of a new epigenotype."

Carson (1959, 1965, 1968, 1971) has further developed Mayr's idea
by suggesting that a series of sudden population increases ("flushes") and

decreases ("crashes") greatly enhances the probability of speciation by the

founder principle (although in retrospect, Carson's ideas on the founder-
flush hypothesis can also be directly traced back to Wright, 1940a,b).

Sympatric Models of Speciation

Although there are few doubts about the reath of allopatric

speciation, questions have been raised over the universality of geographic

isolation as a pre-requisite for speciation in sexually reproducing animals

and about the genetic basis underlying the speciation process (Fisher, 1930;

Lewis, 1962, 1966; Bazykin, 1965, 1969; Maynard—Smith, 1966; Bush,

6

1966, 1969a,b, 1975a; White, 1968, 1978; Murray, 1972; Endler, 1977;
Templeton, 1981). One non-allopatric mode of speciation which has gained
considerable attention is termed "sympatric speciation" and is based on the
premise that populations can diverge as a consequence of adapting to
different ecological niches or habitats which co-occur within the same
geographic area. Sympatric models of speciation owe their origin to
Benjamin Walsh (1864, 1867) who, five years after Darwin (1859)
published "0n the Origins of Species", postulated that reproductive
isolation among certain species of phytophagous insects was the result of
these insects sympatrically shifting and adapting to new host plants.
Mathematical models of sympatric speciation can be traced back to
Maynard Smith's (1966) application of the work of Levene (1953) on the
effects of selection in heterogeneous environments to the question of
population divergence.

The possibility of sympatric speciation has been criticized by Mayr
(1963, pp. 451,452, 527) who has stated that:

"There is no evidence to support the notion of ecological races
as a category distinct from that of geographical races. A theory of
ecological speciation would have to prove that a population of a
species established itself in a new kind of habitat within the cruising
range of individuals of the parental populations and merely by
conditioning and habitat selection segregated itself in the new
habitat until it had reached species level ...... Since there are
normally very numerous ecological niches within the dispersal area
of a deme, niche specialization is impossible without continued new
pollution in every generation by migrants ..... Contiguous
populations of a species are therefore held together by such a close
tie of genetic cohesion that it is impossible to divide this essentially
single gene pool into two. There is no mechanism (except for
instantaneous speciation through polyploidy) by which the cohesion
might be loosened to the extent that it would permit the sympatric
development of two independent gene complexes which would give

7
reproductive isolation and ecological compatibility."

Nevertheless, during the century following publication of Walsh's
theory, several putative cases of host shifts supporting a sympatric mode of
divergence have appeared in the literature (Brues, 1924; Thorpe, 1930;
Smith, 1941; Bush, 1966; see review by Diehl and Bush, 1984). Perhaps the
most well known example is that for certain species groups of true fruit
flies in the genus Rhagolen's (Diptera: Tephritidae). Because Rhagoletis
larvae are highly specific, internal frugivorous parasites of their host plants
and mate selection in these flies is directly coupled to host selection, Bush
(1966, 1969a,b, 1975a) has argued that variation for traits related to host
recognition and host associated survivorship could act as premating
barriers to gene flow and result in the sympatric divergence of fly
populations specialized on different host plants.

Bush (1969a,b; 19753) contends that Mayr has overlooked some
critical points with regard to sympatric speciation and the genetics of
speciation. First, Mayr neglects to distinguish between migration and
effective gene flow. Individuals occupying different habitats are likely to
be under different selection pressures. Therefore, although migration may
occur between con-specific populations occupying different habitats,
effective levels of gene flow may still be quite low due to selection acting
against immigrants entering a niche to which they are not well adapted.
Second, mating in many phytophagous insects and parasites occurs only
within preferred habitats. Such organisms therefore have mating systems
which are inherently biased toward the development of positive assortative
mating following a habitat or host shift. Consequently, any trait conferring
differential habitat usage simultaneously produces a pre-mating isolating

8

barrier which further reduces inter-demic gene flow. Third, genomes are
not necessarily highly "co-adapted gene complexes" and speciation may not
require or be initiated by a " genetic revolution". For instance, although
traits related to host recognition are undoubtedly polygenetic (J aenike,
1987) changes in just a few genes or even a single locus could have major
effects on proper host identification (Huettle and Bush, 1972; Carson and
Ohta, 1981; Thompson 1988). Populations in which mating is confined to
specific habitats could therefore diverge rapidly due to fixation of
alternative allelic variants at only a limited number of key polymorphic loci
controlling habitat choice.

Summary of Similarities and Differences between Sympatric and

Allopatric Modes of Speciation

Perhaps it would be good to digress for a moment to summarize and
further clarify the debate concerning the geographic context of speciation.
As written by Mayr (1963, pg. 451,453):

"Theories of sympatric and geographic speciation agree in their
emphasis on the importance of ecological factors in speciation.
They differ in the sequence in which the steps of the speciation
process follow each other. The theory of geographic speciation lets
an extrinsic event separate the single into several gene pools, with
the ecological factors playing their major role after the populations
have become geographically separated. According to the theory of
sympatric speciation the splitting of the gene pool itself is caused by
ecological factors, and whatever spatial isolation of the thus formed
populations ensues is a secondary, later phenomenon ...... The
question to be solved, then, is whether there are any natural
phenomena known that can be interpreted as cases of incipient
species in a process of sympatric speciation. The complex of
phenomena usually labeled biological races has indeed been cited as
evidence for such sympatric speciation and must therefore be

studied carefully".
Or as succinctly stated by Maynard Smith (1966):
"The crucial step in sympatric speciation is the establishment of
a stable polymorphism in a heterogeneous environment. Whether
this paper is regarded as an argument for or against sympatric
speciation will depend on how likely such a polymorphism is
thought to be, and this in turn depends on whether a single gene

difference can produce selective coefficients large enough to satisfy
the necessary conditions."

Rhagoletis pomonella as a Model System of Host Race Formation

and Sympatric Speciation

As mentioned before, the biology and natural history of true fruit
flies in the genus Rhagoletis are ideally suited for studies of population
divergence and speciation. In particular, members of the R. pomonella
sibling species group offer a unique opportunity to directly test various
tenants of sympatric speciation theory. Members of the R. pomonella
complex are sympatric or broadly parapatric in their geographic
distributions across North America (Bush, 1966). The four described and
one undescribed species comprising the group are all indigenous to North
America and each species attacks a different, non-overlapping set of host
plants (Bush, 1966). Although several pairs of R. pomonella group species
can be crossed in the laboratory, these species are genetically distinct and
rarely hybridize in nature (Bush, 1966; Berlocher and Bush, 1982; Feder et
al., in press; Feder and Bush, submitted; Williams et al., manuscript in
prep.) Most importantly, the species R. pomonella is actively generating
new, sympatric host associated populations (Ward, 1866; Walsh, 1867;
Herrick, 1920; Shervis et al., 1970; Prokopy and Bush, 1972; Prokopy and

10

Berlocher, 1980). The most famous and best documented host shift for R.
pomonella is that from its native host hawthorn (C rataegus spp.) to
introduced, domestic apple (Malus pumila). The known historical
timeframe (~150 years ago; see Walsh, 1867; Illingworth, 1912; O'Kane,
1914; Porter, 1928; Bush, 1966) and probable location of the shift (Hudson
River Valley, New York) allow a direct analysis of the host specialization
process from a context that excludes the possibility of prior geographic
isolation between apple and hawthorn infesting populations. Therefore, if
behavioral and/or genetic differences exist between hawthorn and apple
populations at sympatric field sites, then this would strongly support a
sympatric mode of divergence for these flies.

The objective of this dissertation is to examine hawthorn and apple
infesting forms of R. pomonella for evidence of genetic differentiation.
Three general questions will be addressed. First, do genetic differences
exist between hawthorn and apple flies and, if so, do these differences
indicate that gene flow is restricted between the two host associated
populations? In other words, are hawthorn and apple flies partially
reproductively isolated 'host races" as predicted by sympatric speciation
theory? Second, what is the nature of any observed genetic variation
between sympatric apple and hawthorn populations. Is genetic
differentiation widespread throughout the genome, as indicative of the
"genetic revolution" of Mayr, or are host related differences restricted to
allele frequency changes at only a limited number of loci located in specific
regions of the genome? Third, what is the geographic pattern of genetic
variation for R. pomonella, and does this pattern reveal anything about
possible factors differentiating apple and hawthorn populations? Is

selection a possibility or is the pattem of intra- and inter-host divergence

ll

suggestive of stochastic processes such as genetic drift?
Starch Gel Electrophoresis of Soluble Proteins

Starch gel electrophoresis of soluble proteins forms the basis for the
survey of genetic polymorphism in R. pomonella populations. The
technique was first used for analyzing genetic variation in natural
populations of Drosophila melanogaster by Lewontin and Hubby (1966)
and in humans by Harris (1966), and has since become a standard
biochemical tool in the study of evolution. In protein gel electrophoresis,
an individual is ground in a stabilizing buffer solution and the homo genate
is subjected to an electric field in a starch, cellulose acetate or acrylanride
gel medium. Protein variants can be distinguished because amino acid
substitutions (due to changes in the DNA sequence encoding the protein) can
cause net charge or conformational changes in the enzyme which alter its
rate of migration in the electric field. Specific histochemical stains are used
to visualize the mobilities of a protein or class of proteins in the gel.
Mobility differences among or within individuals for the same enzyme
system are usually indicative of allelic variation at the locus coding for the
protein. Unfortunately, not all genetic variation is detected by starch gel
electrophoresis of soluble proteins (Lewontin, 1974). However, studies
suggest that the technique does give a fair approximation of the extent of
genetic polymorphism in populations (Rarnshaw et al., 1979). In this
dissertation, the term allozyme refers to enzyme loci resolved by different
histochenrical stains and isozyme refers to differentially migrating enzyme

systems which were detected using the same histochemical stain.

12
Life History of R. pomonella

Before continuing further, I would like to present some additional
remarks on the life history of R. pomonella and on F -statistics which I hope
will clarify the subsequent presentation of the data. First, R. pomonella
generally has a univoltine life history cycle across its range (Bush, 1966;
Dean and Chapman, 1973). Adult females lay their eggs directly into the
flesh of the host fruit while the fruit is still on the tree. Eggs hatch within
two days and immatures develop through three larval instars over a period
of 3-5 weeks. A larvae feeds and completes development only within the
single host fruit in which it hatched. After the fruit abscises from the host
plant and falls to the ground, late third instar larvae leave the fruit and
burrow a few inches into the soil where they pupate after 3-4 days. Flies
therefore overwinter as pupae in the soil beneath the tree which they
infested as larvae during the preceding summer. Most flies terminate
diapause and eclose as adults the next summer, although some pupae require
two winters of chilling in the field before completing development (Phipps
and Dirks, 1933b).

F-statistics for the Analysis of Population Structure

F-statistics were invented by Sewall Wright as a method for
quantifying population structure with respect to the amount of inbreeding
within populations and subpopulations (Wright, 1943, 1951). Wright
developed separate inbreeding coefficients for an individual relative to its
subpopulation (F13), for an individual with respect to the total population
(F11), and for the correlation between randomly chosen gametes within

subpopulations relative to the total population (PST). The three inbreeding

13
coefficients are related algebraically by the formula:

Fsr = (Frr - Frs) / (1 - Frs)

An alternative concept of F-statistics has been developed by Nei
(1977) based on characterizing the degree of genetic differentiation among
existing populations rather than on inbreeding or coancestry within
populations. Nei's method has certain advantages over Wright's in that it
does not make explicit assumptions about the nature of factors influencing
gene and genotype frequencies. In contrast, mutation, selection and
migration are assumed to have minimal effects on the genetic structure of
populations under Wright's calculations. However, because of problems
related to sampling theory, direct comparisons of Nei's FST values between
populations is not always advisable. Since in many instances in this
dissertation I use Fs'r values for comparative purposes, I calculated F-
statistics using the formulae of Weir and Cockerham (1984); a
methodology which retains Wright's original conceptual definition of F -

statistics as inbreeding coefficients.
Organization of the Body of the Thesis

The body of the thesis is organized into six chapters. The first five
chapters are presented in scientific format, with each Chapter being
subdivided into an introduction, materials and methods, results, and
discussion sections. Chapter 6 is a concluding summary. For ease of
reference, tables containing large amounts of allele frequency data appear
at the end of Chapters 3, 4 and 5 following the discussion sections.

Chapter 1 establishes a Mendelian mode of inheritance as well as

linkage group associations and relative map positions for 12 different

l4

polymorphic allozymes in R. pomonella. Chapter 2 reports the observation
of significant allele frequency heterogeneity for allozymes between
hawthorn and apple infesting populations of R. pomonella at a sympatric
field site in westem Michigan. Chapter 3 addresses the question of
geographic variation for apple and hawthorn populations through a survey
of allozyme variation across a major portion of R. pomonella's range in
eastem North America. Chapters 4 and 5 present a more fine grained
analysis of spatial and temporal genetic differentiation between hawthorn
and apple host races. Chapter 4 explores the regional pattern of divergence
on a county by county basis in the nridwestem United States by analyzing
populations along five latitudinal transects which run from Michigan
through Indiana and from Wisconsin through Illinois. Chapter 5 dissects
the pattenr of local and rrricrogeographic genetic variation within a 60 km2
area surrounding the original study site near the town of Grant, in western
Michigan. Chapter 5 therefore assesses whether patterns observed on a
broad geographic scale across eastern North America are spatially and
temporally consistent on a very fine level down to individual host trees,
over years, within an old field. Chapter 6 is a concluding summary which
includes some brief comments on my future research plans for the R.

pomonella group.

Chapter 1

Inheritance and Linkage Relationships for
Allozymes in Rhagoletis pomonella

Introduction

The apple maggot fly, Rhagoletis pomonella (Walsh) (Diptera:
Tephritidae) is of economic importance because it is a major pest of
domesticated apples throughout North America. Apples (Malus pumila)
are not, however, the native host of R. pomonella. Apple populations of the
fly are thought to have originated within the past 150 years via one or more
host shift(s) from hawthoms (Crataegus spp.; Bush, 1966, 1969a,b, 1975a).
The historical context of the shift(s) makes R. pomonella interesting to
evolutionary biologists as well as to applied entomologists as a model
system for the study of sympatric host race formation and speciation.

Our understanding of the population genetics and phylogenetic
relationships of R. pomonella and related species has primarily come from
starch gel electrophoresis of soluble proteins (Berlocher, 1976; Berlocher
and Bush, 1982). Allozymes have also provided genetic markers toward
the construction of a linkage map for the apple maggot (Berlocher and
Smith, 1983) which is crucial for future research on the genetics of host

15

l6

selection in these flies. The scope of electrophoretic studies has been
limited, however, by the number of enzyme loci which can be stained and
reliably scored in R. pomonella.

The objective of this Chapter is to resolve and characterize new
allozyme systems in the apple maggot. Previous studies on a laboratory
stock of R. pomonella established a Mendelian mode of inheritance for
allozymes at 12 electrophoretic loci (Berlocher, 1980; Berlocher and
Smith, 1983). Ten of these loci were subsequently mapped to four different
linkage groups. Here, I expand on the original results by verifying a
genetic basis for six additional allozymes and determining map positions
for five of these loci in the same laboratory fly stock. Linkage
disequilibrium is also examined between pairs of non-allelic genes for the
laboratory strain and compared to values observed in natural R. pomonella
populations.

Materials and Methods

W

In nature, R. pomonella has a univoltine life cycle, spending the
winter months in a facultative pupal diapause. For this reason, crosses were
performed on a non-diapausing laboratory strain of the apple maggot that
has a generation time of approximately three months. The strain originated
from an apple population collected near Geneva, New York in the mid
1970's and has gone through approximately 40 generations in the
laboratory.

Fifty single-pair crosses were started by placing pairs of virgin flies

into cages containing Red Delicious apples. The apples were removed from

17
the cages every other day and replaced with fresh apples. After six weeks

the crosses were terminated and surviving parents frozen at -80°C until
they could be analyzed electrophoretically. F1 progeny were reared to
adulthood from the infested apples and also stored at -80°C. In 15 of the
original 50 matings parents were frozen alive and crosses produced
sufficient progeny (>20) to warrant genetic analysis. Eighteen progeny
were initially run from each of these 15 crosses with additional progeny

being examined from crosses involving parents with especially informative

genotypes.

WM

Standard starch gel electrophoretic techniques were used (Berlocher
and Bush, 1982). Individual flies were sexed and homogenized with a glass
rod in 20 pl ofa 0.1 M Tris, 1 mM EDTA, 0.13 mM NADP, 0.15 mM
NAD, 0.3 M beta-MeEtOH (pH 7.0) grinding buffer. When possible,
parents and their progeny were examined on the same gel to facilitate the
correct scoring of allozymes. Gels were made from 11% electrostarch (O.
Hiller and Co.) using one of three different buffer systems: system I , 3.7
mM Citric acid, 0.1% v/v N-(30 amino-propylmorpholine)(pH 6.1);
system 2, 8 mM Tris, 3 mM Citric acid (pH 6.7); system 3, 76 mM Tris, 5
mM Citric acid (pH 8.7). Electrode nmning buffer for system I was a 10x
concentrated form of gel buffer 1, for system 2 was 0.22 M Tris, 86 mM
Citric acid (pH 6.3), and for system 3 was 0.3 M Borate (pH 8.2). Gels
were sliced and stained for the 11 different enzymes listed in Table 1 along
with their abbreviations, enzyme commission numbers and gel buffer

systems.

18

£82 62958808 a 3.8%.: *

 

M 3.2 om 8E
N 3.0m em 3
m 33% Om 3.1%..
N 3.? em a:
m 9.4.2 um 22
_ «3.: em :2
m 8.2; em 3:
N 3.3 om alumina
3: 2.3 um 31.2
m 42.3 um 3<
N 3.3 um Ne: 83.
cute; 30 cinem— eem acouanBn’x

ogeeagswefimonm
cannon—Sm emcee—monumeam
888$

omEoEemm Sufi—mean 08:52
9335 one—z

0353638 oabmoom— ,
Become—.38 Seguefizméon
sepaaafimn<z
goumegeeg Became.
033885

3.3834

g

.80 .N 528: 388 o... A 3.23.9.8 Be? 8058 on. 9 .830:
BEES 35 guofl 33.38% .x we :39. b88092 05 e8 3282 8.89? 2:35 ._ 03$.

19

MW

Sixteen different allozymes were resolved in the study, 12 of which
were polymorphic (Table 1). Isozymes that migrated the nearest to the
cathode were designated 1, the second nearest 2, etc. Six of the
polymorphic loci resolved (Pgm, Had, Idh, Acon-Z, Aat-Z and Pgi) have
been previously characterized in R. pomonella (Berlocher, 1980;
Berlocher and Smith, 1983) and their original allele designations were
retained whenever possible. Alleles for six previously unresolved
polymorphic loci (Me, Mpi, Acy, Pep-2, Did-2, Aat—I ) were named in the
conventional manner with the most common electromorph for a locus
designated 100 and all other alleles assigned according to their relative
anodal mobility to the common type. The most common allele at each of
the new loci was determined from a parallel electrophoretic study of 11
natural R. pomonella populations distributed across eastern North America
from Door Co., Wisconsin to Nova Scotia, Canada (see Chapter 3).

Cl 1' [1.1 12' il'!’

Linkage disequilibrium refers to deviations from non-randomness in
the associations among non-allelic genes within gametes and in the union of
these gametes in forming zygotes. When coupling and repulsion double
heterozygotes can be distinguished it is possible to partition linkage
disequilibrium into within gamete (Dw) and between united gametes (Db)
components (Cockerham and Weir, 1977; Weir, 1979). If f 11 denotes the
frequency which two non-allelic electromophs A and B are found together
on the same gamete within the same individual and g11 is the frequency
which A and'B are found on different gametes within the same individual,
then:

20
I3w==fll ' 811 and

Db=gll “m,

where p and q are the respective frequencies of alleles A and B in the
population (Note: Dw + Db is equal to the conventional measure, D, of
linkage disequilibrium). A series of four different null hypotheses
concerning Dw and Db can be tested by Chi-square goodness of fit statistics
and are explained in Table 2.

When coupling and repulsion heterozygotes cannot be distinguished
or if loci are unlinked, a composite measure of linkage disequilibrium
(AAB) proposed by Peter Burrows (Cockerham and Weir, 1977; Weir,
1979) can still be calculated where,

AAB=Dw+2Db=f11+g11-2pq
A test statistic of the hypothesis AAB = 0 (H4; Table 2) is provided by,
x20 do = N<AAB>2 / [p(1-p>+DAAi iq(1—q)+DBB]

where N is the sample size and DAA and DBB measure deviations of
observed from expected frequencies of individuals homozygous for the
non-allelic genes A and B, respectively. Direct comparisons of AAB
between populations are complicated, however, by the fact that AAB is gene
frequency dependent. Therefore, for inter-populational comparisons we
calculated correlation coefficients as suggested by Laurie-Ahlberg and
Weir (1979) in which,

N = AAB/ [p(1-p)+ DAA] [q(1-q) + DEB].

 

21

Table 2. Null hypotheses (I-Il-H4) concerning tests of linkage
disequilibrium within gametes (Dw) and between united

gametes (Db).

MWW

H1 Db==0 Test of random association of non-allelic

genes between uniting gametes.

HZ Db-i-Dw=0 Test of random association of non-allelic
genes within gametes. Db+Dw is equal
to the conventional measure of linkage
disequilibrium (D). H2 is, therefore, the
standard test for gametic disequilibrium.

H3 Dw=0 Test of non-allelic genes being associated
more strongly within gametes than
between uniting gametes.

H4 Dw+2Db=O Test of whether composite measure of

linkage disequilibrium involving non-
allelic gene associations within and
between uniting gametes is significantly
different than 0.

22

Results and Discussion

W

Alleles for all 12 polymorphic loci were found to have a genetic basis
and all F1 progeny had genotypes that could be transmitted from their test
cross parents (see Table 3 for segregation results; data for Acon-Z, Act-2,
Pgm, Idh, Pgi and Had are omitted because Mendelian segregation for
alleles at these loci has already been demonstrated by Berlocher and Smith,
1983). Only one locus, Aat—I , consistently failed to fit Mendelian
expectations for an autosomal locus. Heterozygote males were not found
for Aat-I . This suggests that Aat-I is sex-linked and that males are the
heterogametic (XY) sex in R. pomonella. Confinnation of sex-linkage for
Aat-I came from crosses between heterozygous females and single banded
males (males which were presumed to be hemizygous). Female progeny
from such crosses were either homozygous for the allele possessed by their
fathers or heterozygous. Male offspring, however, were always single
banded for one of the Aat-I alleles possessed by their mothers and were,
therefore, hemizygous (Table 3).

Three "cross types" showed significant deviations from Mendelian
expectations (Table 3; Note: G-contingency tests indicated that genotypic
frequencies were homogeneous among test crosses pooled to form all of the
cross types analyzed in the study.). An excess of Did-21ml Dia-290
progeny was observed in cross type 6 for Did-2 while cross types 6 and 7
for Pep-2 produced an excess of Pep-2100] Pep-2106 heterozygotes.
Although the three significant deviations out of a total of 62 tests could be
due to any number of factors including selection, they are most easily

explained by random type I error.

23

6 *5;
Amv***cw.w_
AN6 62
CV 86
CV 6w;
Amv *mmd
ANV Rum
CV 36
5 RN
:6 NN6
AS aN.N
A: hmd
ANV 6c.N
A: 56
3.3 NM

v 6— ~52 2 625: a 6252 m 6252
2 6952 N H“.652 MN 6252 2 6252
2 6252 wN 6252 a 6252

5N 6252 hN 6252

Nm H2 ~52 NN 6252

g 6252

vw 6252

v 6656 62 6&9: 2 66266 6 6252
a 6656 Nm 625m 2 6252

6N 625m 62 633
N6 625x 6». 6252

m 6266 2 ”mafia
6w 6252
2 M352 2 6252
62 6252
6Q ”252 N6 6252
52 6252
m 6252 Z 6w52 N 6w5w
6» 6652 g 6w5w
w2 6w56

625: x 6252

352 x 62\8~
6252 x 66292
6252 x 6252
2 262 x 6252
6252 x 6252
6252 x 6252

252 x 6952
352 x 6652
35w x 352
6252 x 252
665w x mafia
6252 “56:82
352 x 6252

6252 x 8252

252 x 6252
c252 x6252

352 x 6652
352 x 85m
85m x 665”

8 h
S e
6 m
5 v
6 m
8 N
a:

AN6 6
A8 6
ANV v
A3 m
A: N
A8 _

ANV N
33 2

3 N
2: _

A: m
36 N
A66 2

$5

Néa

>o<

a:

22

.25 «8.5 05 .56 365.386 3058:: 05 wage—3.. women—E23 5 53% mm
25 898 :80 8.6 6289 mommob an. «e .3682. BE. Ao—afiz .ofifiomumv x8 9 asp—08a banana :03»
2a 23.. .86 8.5 Swans 03% 05 8.. ago—3 3382 :30: “a 8.. 830.8 83 :86 93%.: couampHMom

.m 033.

24

Seaman 0255.2 65 025: 8.2 222.5228 256 203 2.8< .52 8382325
.83. 03:23-25 .3 BEE—206 2 2338.25 0292.28» 896 8023256 .5o62ew2m 28.6 w m .2... 266 w M2 2. 36 -v. m ...

 

A26 Nm.N
22v 8.2
AC EN
226 2 26
A26 mud
22v 86
22v 3.2
2 26 EN
6.2.3 NM

2. 622- v2 62- 2
6622-52- m 622-522- "2
m2 622- MN 62- 2
2 622-52- : 62.52- «2
m #6- 2 62- 22
662-56- 662-52- "2
6 66- 22 62- S2
2.62.56- N H26.56- "2
M6 62- 52

m6 62.52- "2
gun—23%

$2 22-

2 62-

22 c2-

2V6

S2 62-

x 2 2.52- .2

x 2 2.52- n2

x 26.562- .2

x V6.52- ”2

x 62.52- "2

226 m

2N6 v

23 m

226 N

22.6 2

2 -23.

3% add-Inga

2.6.8on

.m use.

25

Confirmation of a genetic basis of inheritance for Me, Dia-Z, Acy,
Mpi, Pep-2 and Aat-I brings to 18 the number of polymorphic allozymes
characterized in R. pomonella. Extensive biochemical surveys of natural
apple maggot populations are now possible on a scale approaching those of
other study organisms. The combination of genetic, behavioral and
ecological data now available for the apple maggot makes the animal a
unique organismal system for the study of evolutionary processes within a
historical timeframe.

Wes-ultr-

Restricted crossing over in Rhagoletis males (Berlocher and Smith,
1983) makes it possible to determine whether two loci are linked based on
the pattern of assortment observed in crosses between double heterozygous
males and homozygous females. Our results continued that exchange is
limited in males in that only one recombinant genotype was observed out of
a total of 660 cases in which crossing over in males could be detected. We
therefore used the simple method of following the assortment pattern in
double heterozygote male crosses to determine linkage groups. Linkage
assignments for Acon-Z, Idh, Aat-Z, Had, Pgi and Pgm were in agreement
with the relationships previously reported for the apple maggot (Figure 1;
see Berlocher and Smith, 1983). In addition Did-2, Me, Mpi and Pep-2
were mapped to linkage groups I, II, II, and 111, respectively. Aat-I was
placed in linkage group V which was designated by Berlocher and Smith
(1983) as containing the sex-determining factor(s). Acy was not mapped to
any specific linkage group. However, I can rule out its association with loci
on groups I and II due to its random assortment with Idh, Aat-Z, Did-2 and

Acon-Z in test crosses involving double heterozygote males. Also, Acy did

26

Figure 1. Linkage group assignments and map distances (in
centimorgans) between loci for R. pomonella. Assignments
and distances given in parentheses are based on data from
Berlocher and Smith (1983).

27

.2 8:5

H .3 22.2 2.20 5 8c "34

H 3 2 965 5 6: "23.:

H .6 222 .22 .2 95.5 E .o: 6263
3250 tonnes 3.22:5

 

 

 

62.230 .352 0: 2H
53055.28 .8» ._ -8< H
Ed H
ill-mfimlllt
" u x L. E
m-eed ea:
. .17. t
322 7.651 .r . H2
1% €53-32-
illnmémllll
.AN null]. Flint-2m 9i. “ H

 

-.e< ~65 e... :3 sad

 

82.230 32 one 60-. 955 oooxcj

 

28
not show sex linkage so it is not in linkage group V.

Map distances and the relative gene order of linked loci can be
determined from crosses involving female parents that are heterozygous
for multiple loci. The results for linkage group I differed slightly from
those of Berlocher and Smith (1983) in that in our study the crossover
distance between Idh and Aat-Z was greatly reduced (10.3 centimorgans
(CM) compared to 30.3; Table 4). Recombination rates between Idh and
Aat-Z were heterogeneous, however, among the test crosses of Berlocher
and Smith (1983). We also found cross to cross variation in map distances
between Act-2 and Idh ranging from O to 22.2 cM (Table 4). The lower
map distance observed in this study could therefore be the result of
calculating exchange rates from crosses involving a higher percentage of
inversion heterozygote females than in the previous work. Support for this
interpretation comes from the finding of significant gametic disequilibrium
between pairs of non-allelic genes encoded at Act-2 and Idh (see discussion
in next section on Linkage Disequilibrium and Genetic Variation in the
Laboratory Stock). Unfortunately, extensive ectopic pairing of polytene
chromosomes in Rhagoletis (Bush,1966) prevents us from directly
verifying the existence of inversion polymorphism.

Map distances were also determined between Had and Pep-2 in
linkage group III and Acon-Z, Mpi and Me in linkage group II (Table 4,
Figure l). Surprisingly, not a single recombinant genotype was observed
among Acon-Z, Mpi and Me in 162 progeny scored.

Five allozyme linkage groups are now known for R. pomonella
which presumably correspond to 5 of the 6 chromosomes comprising the
haploid karyotype of these flies (Bush, 1966). Any biochemical,
morphological or behavioral trait can therefore be mapped to a specific

Tmme‘fi“

chi

Linkage_gzgun_l
Dia-Z/Aat-z
Dia-Z/Aat-z
Aat-Z/Idh

Idh/Pgm
Linkage_g:222_111
Pep-Z/Had

29

thHfimmmxsGnomuhmngum)mflmflmwdbmwwznkxifiouummfle
pair crosses.

 

 

Gregg #
16 .29 21 ii 11 A221
0.0 3.7 11.1 0.0 0.0 3.2
11.1 7.4 11.1 5.6 0.0 7.1
11.1 11.1 22.2 5.6 0.0 10.3
-- -- -- -- 38.9 38.9
2Q 11 AXE;
22.4 27.8 23.9

30

linkage group. Unfortunately, we are currently unable to associate specific
linkage groups with their corresponding chromosomes due to the poor
quality of polytene chromosome preparations in R. pomonella. However,
we assume, like previous investigators (Berlocher and Smith, 1983), that
the unmarked sixth linkage group represents the small dot chromosome
observed in metaphase preparations (Bush, 1966).

Linkage groups appear to be surprisingly well conserved between
Rhagoletz's and its distant confamilial relative the medfly, Ceratitis capitata.
For example, the loci Pgm, Idh and Aat are all linked in the medfly as in R.
pomonella (see Figure 1 and Gasperi et al., 1986). Also, Mpi and Adh map
to the same chromosome in the medfly while Had, Pgi, Mpi-Adh, and Aar-
Idh-Pgm constitute four different linkage groups (Gasperi et al., 1986).
Genetic studies of additional allozyme loci and Tephritid species will
therefore provide valuable insight into the evolution of specific gene
clusters.

Aat-I represents the first sex-linked marker mapped in R.
pomonella. The segregation pattern for Aat-I is the first evidence that
males are the heterogametic (XY) sex in the species. Hemizygosity for Ant-
1 in males implies that the X and Y chromosomes are not completely
homologous. However, morphologically distinct sex chromosomes have
not been found for R. pomonella (Bush, 1966). It is possible that an
undetected translocation between an autosome and the sex determining
chromosome (linkage group V) was responsible for generating sex-linkage
for Aat-I. XY: autosome translocations have been found for R. tabellaria
(McPheron and Berlocher, 1985) and in the R. suavis group (Berlocher,
1984a), however, translocated loci were not hemizygous in these cases.

Determination of linkage relationships for Aat-I in additional Rhagoletis

31
species will help resolve this point.

L°k-é'Dlr-.-_or1_.11- 9 1.11.“..11. _1‘ him .0 01k

Progeny testing of double heterozygote parents allows complete
identification of coupling and repulsion two locus genotypes. Therefore,
Dw (linkage disequilibrium within gametes) and Db (disequilibrium
between united gametes) could be calculated between pairs of linked, non-
allelic genes for the R. pomonella laboratory strain (Table 5). Hypothesis 1
(Db = 0; see Table 2 for explanation of hypotheses 1-4) was true for 35 of
37 pairwise tests. Only MeIOO/Aeon-289 (xi = 3.9, 2 5 0.05, 1 d1) and
Mpi100/ Aron-2100 (x2 = 6.71, 2 _<_ 0.01, 1 df) showed significant
deviations from non-random union of gametes. These two significant tests
may be explained by random type I error.

When gametes unite at random and survivorship is equal for
coupling and repulsion genotypes, Dw is a direct measure of linkage
disequilibrium within gametes and H2 (Dw + Db = 0) becomes equivalent to
H3 (Dw = 0; Cockerham and Weir, 1977). Of the 35 tests in which H1 (Db
= 0) was true, 23 were significant for either H2 or H3 (Table 5). Non-
random associations were found between non-allelic electromorphs for
Me/Acon-Z, Idh/Aat-Z, Idh/Dia-Z, Aat—Z/Dia-Z and Had/Pep-Z. Linked
genes, therefore, showed substantial amounts of genetic covariation in the
non-diapause line.

Burrows disequilibrium values (AAB) were calculated between
pairwise combinations of unlinked genes in the laboratory stock. Only 6 of
140 tests of the hypothesis AA3 = 0 were significant and no pattem was
discemable among loci displaying significant tests (data not presented).

The general lack of disequilibrium between alleles at unlinked loci indicates

32

 

 

Table 5- Linkage disequilibrium between pairs of non-allelic genes in the
laboratory stock of R. pomonella. Results are given for
drseq' uilibrium within gametes (Dw), between united gametes (Db)
and for a composite measure of Dw and Db (A, Bunows value). See
Table 6 for explanations of test hypotheses Hl-H4.
Loci/Alleles Bu 1}, NB Significant Hypotheses
1B1? 00 0.1000 -0.0305 0.0389 112** 113*
bloc/W5 -0.0330 —0.0111 -0.0555 114*
1210611111189 -0.0667 0.0416 0.0167 111* 113*
11911 /Acon1°° 0.0667 -0.0411 -0.0150 111** 113*
1611100 ' -285 -0.1670 0.0225 -0.1220 112*** 113*** 114**
range/0164100 0.1670 —0.0333 0.1000 112*** 113** 114*
1611 8fluid-2‘35 0.1830 -0.0186 0.1460 112*** 113*** 114***
161111 Ania-210° -0.1830 0.0277 -0.1280 112*** 113*** 114**
1611104/13121-2100 0.0167 0.0055 0.0277 112* 114*
r 1°°/Aat-221 0.1167 -0.0216 0.0733 112*** 113**
Idh °° Ant-2'33 -0.0833 0.0366 -0.0100 113*
Idh1 Bat-2 -0.0833 0.0366 -0.0100 112** 113* 114*
161111 8/Aat-250 -0.0667 0.0144 -0.0377 112*
161111 8/Aat-275 -0.1167 0.0128 -0.0911 112*** 113** 114*
Idhll /Aat-275 0.1000 —0.0400 0.0200 113*
Idh mat-2123 0.0677 0.0058 0.0783 112** 113* 114*
Din-235 hat-25° -0.0667 0.0011 -0.0644 112** 114**
Din-235/Aat-221 -0.1833 0.0394 -0.1044 112*** 113*** 114***
via-23S/Aat-275 0.1500 -0.0300 0.0900 112*** 113** 114*
Din-2306112152123 0.0667 -0.0092 0.0483 112** 113*
via-21 Aat-250 0.0833 -0.0111 0.0611 112*** 114*
via-21 oo/Aat-zzl 0.1500 -0.0277 0.0944 112*** 113** 114**
Din-21 Aat-275 -0.1500 0.0330 -0.0833 112*** 113** 114*
Die-21 /Aat-2123 -0.0500 0.0000 -0.0500 112* 114*
Pap-210° 10° -0.0667 -0.0064 -0.o790 112** 114*
pep-21 /Had122 0.0500 -0.0006 0.489 112*

 

=1 12 _<,0.05 1* 12 s 0.01 m 12 5 0.001 Chi-square test (1 d1).

33

that genetic interactions across chromosomes are not strong in the non-
diapause strain.

The non—diapausing line has a surprisingly high level of
heterozygosity (Table 6). Ten of the 12 loci analyzed showed a trend for an
excess of heterozygosity over expectation. Several loci in the non-
diapausing strain also displayed significantly higher levels of
heterozygosity than were observed in 11 natural populations of R.
pomonella (Table 6). Small population sizes associated with laboratory
rearing should theoretically lead to a loss of neutral allozyme variability

according to the equation:

Ht/Ho = (1-(1/2Ne))t .

where Ht is the heterozygosity of a population after t generations in the
laboratory, H0 is the heterozygosity of the founding population and Ne is
the effective size of the laboratory population (Wright, 1931). If Ne=150
and t=40 (reasonable approximations for the laboratory stock), the ratio of
Ht to Ho should theoretically be 0.875. Although we do not know Hg for
the original New York population, an estimate of 0.406 can be made based
on the average level of heterozygosity observed in 11 natural populations
(Table 6). The average Ht for the laboratory population is 0.439 so that the
actual value of Ht/Ho is 1.08. The laboratory line therefore has a slightly
higher, rather than a lower, overall level of heterozygosity compared to
natural populations.

Heterozygosity is not, however, the only aspect of genetic variation
which can be affected by a sudden reduction in population size. Allelic

diversity is also a sensitive indicator of whether a population experienced a

34

 

 

Table 6. Observed and expected heterozygosity and the presumed number of
alleles lost in the laboratory stock compared to nann'al apple maggot
populations.

Ave 11 natural
9229131112115

1.193.113 11.5212: 21.9119... 81.191.92.129; HM

Me 0.333 0.278 0 0.472 t 0.066

Mp1 0.133 0.125 4 0.247 t 0.076

Acon-Z 0.767 0.654 4 0.561 t 0.164

Aat-Z 0.800 0.738 7 0.737 t 0.035

Dia-Z 0.633 0.506 0 0.369 1 0.050

Idh 0.467 0.529 3 0.422 t 0.033

Had 0.433 0.406 2 0.373 t 0.089

Pgi 0.400 0.444 5 0.182 t 0.052

Aat-l 0.533 0.424 2 0.299 t 0.075

Acy 0.066 0.064 2 0.348 1 0.092

Pep-2 0.533 0.475 0 0.503 t 0.160

Pgm 0.167 0.156 6 0.344 t 0.091

All loci 0.439 0.400 2. 0.406

 

35

recent bottleneck (Nei et al., 1975; Maruyama and Fuerst, 1985). This
point certainly holds true for R. pomonella as the laboratory strain appears
to have lost almost half the number of electromorphs normally found in
wild apple populations (Table 6). However, we can only assume that alleles
missing from the laboratory stock were lost because we have no
information concerning the genetic composition of the original apple
population from which the non-diapausing line was derived. Nevertheless,
our rough estimate of a 40% decline in allelic diversity for the laboratory
line agrees reasonably well with the colony having gone through a
bottleneck of around 150 individuals (see Table 3 of Nei et al., 1975). The
slightly higher level of heterozygosity in the non-diapausing strain is
therefore surprising because of the obvious effect that the population
bottleneck had on allelic diversity.

What factors, then, are responsible for maintaining gametic
disequilibrium and the higher than expected levels of heterozygosity in the
laboratory population? Founder effects and drift related to the smaller size
of the laboratory population are undoubtedly involved in generating
gametic disequilibrium. However, several of the loci displaying
disequilibrium are separated by a number of map units. For example, Idh
and Aat-Z are 10.3 cM apart and Had and Pep-2 are separated by 23.9 cM.
Recombination should be common among these loci and result in the rapid
decay of linkage disequilibrium. Consequently, founder effects and drift
cannot be the only factors responsible for gametic disequilibrium in the
laboratory line. The heterogeneity of recombination rates observed among
allozyme markers in linkage group I suggests that inversion
polymorphisms exist in the laboratory population. It is therefore possible

that heterosis associated with inversion heterozygote genotypes is

36
maintaining gametic disequilibrium and genetic heterozygosity in the
laboratory stock.

‘ 1... i fDi f! 111111-11 11‘ 5111 11m -.to -. 1 N 1.1: P 0. tion

Comparisons of correlation coefficients (rAB) between the
laboratory and a natural apple maggot population from western Michigan
indicate a drastically different pattern of linkage disequilibrium for
Me/Acon-Z , Pep-Z/Had and Dia-Z/Aat-Z (Table 7). For instance,
MeIOO/Acon-295, Pep-2100/Had100 and Dia-285/Aat-221 are highly
positively correlated in nature while being highly negatively correlated in
the laboratory strain. Besides Me/Acon-Z, Had/Pep-Z, Dia-Z/Aat—Z and
Me/Mpi, however, no other pair of loci show a consistent pattern of linkage
disequilibrium in the natural population.

Why is the pattern of linkage disequilibrium between non-allelic
genes different in nature than in the laboratory? Geographic
differentiation between populations from Geneva, New York (the original
source of the laboratory stock) and Michigan could account for the
differences. However, analysis of other natural R. pomonella populations
from across the northeastern United States (Chapter 3) has revealed the
same pattern of disequilibrium as seen in the Michigan population ruling
out geographic variation as a possible explanation.

Differences in the breeding structure of the laboratory and natural
populations could also account for the observed differences in linkage
disequilibrium. Unfortunately, the composite nature of rAB does not allow
us to determine the extent to which the non-random association of genes
within gametes (Dw) and the non-random union of gametes (Db) affect

linkage disequilibrium in the natural population. If Dw is the same in

37

Table 7. Correlation coefficients calculated between pairs of non-allelic genes
for the laboratory stock and a natural apple fly population collected

from Grant, Michigan.
AW W1 W
MelOO/Acon95 -0.374* 0.379***
MelOO/Mpiloo 0.069 -0.214*
up11°°/Acon1°° -0.010 -0.242*
IdthO/Dia-235 -0.530** 0.044
IdthO/Dia-2100 0.499* -0.055
Idh113/01a-235 0.631*** -0.018
Idh113/oia-2100 -0.570** 0.018
Idh104/oia-2100 0.373* 0.098
IdthO/Aat-2123 -0.399* -0.036
IdhllB/Aat-zzl -0.431* 0.131
Idh113/Aat-2123 0.454* 0.058
Dia-285/Aat-250 -0.473** 0.203
01a-235/Aat-221 —0.604*** 0.250**
via-285/Aat-2100 -0.218 -0.333***
via-21°0/Aat-250 0.463* -0.307***
via-21°0/Aat-221 0.564** -0.456***
oia-zloo/Aat-zloo -0.163 0.380***
Dia-zloo/Aat-275 -0.428* 0.254**
oia-zloo/Aat—2123 -0.366* 0.031
pep-21°0/uad100 -0.457* 0.486***

 

* B S 0.05 *" B _<_ 0.01 "* 2 30.001 Chi-square test of Burrow's values.

38

nature as it is in the laboratory then selection, differential fecundity or
fertility, deviations from non-random mating and/or gene frequency
differences between sexes, could all be significant in the natural population.
Almost all single locus Hardy-Weinberg tests, however, were in
equilibrium for the Michigan population (see site 5 in Tables 21-32). This
suggests that mating is random and single locus selection is not strong
within the wild apple maggot fly population. Also, gene frequencies were
not significantly different between males and females in the western
Michigan population (data not shown). These findings all suggest that it is
Dw, rather than Db, which differs between laboratory and natural R.
pomonella populations.

Although we cannot completely discount non-random union of
gametes in nature, two observations make it tempting to speculate that
selection for high fecundity and/or rapid development has favored
alternative combinations of non-allelic genes for Me/Acon-Z, Pep-2/Had
and Dia-Z/Aat-Z in the laboratory population. First, test crosses between
laboratory strain flies generally produce about twice as many F1 offspring
as those between wild type parents (J .L. Feder, personal observation).
Second, allele frequencies for Me, Acon-Z and Had correlate with the
timing of diapause termination and adult eclosion in natural R. pomonella
populations (Feder et al., manuscript in preparation).

Regardless of its causal basis, the results from this study indicate that
the non-diapausing line has an altered genetic constitution. Genetic,
physiological and behavioral changes associated with laboratory rearing
are now known for a number of insect species including Drosophila
melanogaster (Laurie-Ahlberg and Weir, 1979); the screw worm,
Cochliomyia hominivbrax (Bush and Neck, 1976; Bush et al., 1976; Singh,

39

1984); the codling moth, Laspeyresia pomonella (Bush, 1975b); the olive
fruit fly, Dacus oleae (Prokopy et al., 1975; Bush and Kitto, 1979;
Economopoulous, 1980; Economopolous and Zervas, 1982; Loukas et al.,
1985); and the med fly, Ceratitis capitata (Saul, 1986; for a more extensive
review of the insect literature see Menken and Ulenberg, 1987). Although
we are not aware of any studies directly comparing the behavior of
laboratory and wild apple maggot flies, the genetic changes observed for
the non-diapausing line give a warning that the colony may also be
behaviorally modified.

Chapter 2

Genetic Differentiation between Sympatric Host
Races of the Apple Maggot Fly, Rhagoletis
pomonella

Introduction

True fruit flies in the Rhagoletis pomonella species group have been
proposed to speciate sympatrically (that is, in the absence of geographic
isolation) as a consequence of shifts to previously unexploited host plants
(Bush, 1966, 1969a,b, Bush, 1975a). Because Rhagoletis larvae are host
specific fruit parasites and mate selection in these flies is directly coupled to
host plant recognition (Prokopy et al., 1971,1972), variation for larval
survivorship and host preference traits can act as genetically based barriers
to gene flow and subsequently result in the sympatric divergence of fly
populations adapted to alternative hosts. The shift of R. pomonella fr0m its
native host hawthorn (Crataegus spp.) to domestic apples (Malus pumila) in
eastern North America within the last 150 years (Walsh, 1867) provides an
opportunity to determine whether host specialization is sufficient to
differentiate populations without prior periods of geographic isolation. In
this Chapter, I report finding significant genetic differentiation between co-

occurring hawthorn and apple populations of R. pomonella at a field site

40

41

near Grant, Michigan. The result confirms that hawthorn and apple flies
represent partially reproductively isolated "host races" and is consistent
with a sympatric mode of divergence for these flies.

Materials and Methods

R. pomonella flies were collected from an old field located near the
town of Grant (Newaygo Co.) in western Michigan (Figure 2). In 1984 and
1985 flies were sampled as larvae from infested fruit and reared to
adulthood in the laboratory. In contrast, adults were captured directly off
of host fruits by sweep net in 1986. Standard horizontal starch gel
electrophoretic techniques were used (see Chapter 1 for details and also
Berlocher and Bush, 1982) and a total of 29 different allozyme systems
were resolved (Table 8). Alleles were numbered according to their relative
anodal mobilities, with the most common allele for a locus assigned a value
of 100 and used as a standard. Isozyme systems that migrated the nearest to
the cathode were designated 1, the second closest 2, etc. Thirteen loci were
polymorphic (frequency of the common allele less than 0.95; Table 8) and,
excluding Aat-I which is sex-linked, only 15 significant deviations from
Hardy-Weinberg equilibrium were observed for these loci in 290 tests (data
not shown). The significant deviations displayed no regular pattern and

their number does not differ from random expectation due to type I error.

Results

Six allozyme loci, Aconitase-Z (Acon-2), Malic enzyme (Me),
M annose phosphate isomerase (Mp1), Aspartate amino-transferase (Am-2),
NADH-Diaphorase-Z (Did-2) and beta-Hydroxy acid dehydrogenase (Had),

42

Figure 2. Scale diagram of study site located near the town of Grant, in
western Michigan. The site is an old field which has remained
fallow since at least 1922 (E. Hansen, pers. comm). Host tree
designations (A=apple, H=hawthom) are given along with
collecting dates and inter-tree distances (in meters).

43

.N 2:me

    

 

 

 

 

  

 

nmxmxo
83.3 $2?»
838
33.3
2
$33 68.
883 e
.6253
65.
$33.
83.3 . 62.. .4 838
858 :1 $53

 

 

¢m\n_\m

 

 

 

.08 S8222 4:05 0:. .0 80:08.08: 88.60.”. 30:38:10.... 800.38% 835;... 800. a 8.8.0:. .

nod A 30:0 85:50 0... .8 88:00.... 0... .23 802 605.. £8.88 :0 8.865 .4

 

.805
86
area
.82.
.68
2.5
.885
.338
cm .0 73$
33
888$
.85

0808002888.... at 08.088. 0.0.38.3 08E.
080$ 08038885 8.: 08—20....
$00,880.. Gem. 08.28:. 028.85

8.8.88. 928.9... 36982 a e . an: m 0 78.88...
08.0.8 202 .a a. 208 N 8 7.88883... 9%:

gouaczsoc 0.8.88— 35 8830x0m
0888.628 egaxegmé A806 8.08:".
~-ae£%$8§ a 8 .85. m 8 788938-83...

a 8 . 8.88.50 65:8 288.8 62$ 08.82
85.0. 92.53. .a 8 23: N 6.. 88.200802... .282

«08.803. 2-803 708.203.

8.88:5 283 088.38.... 300-3582

.0.0 .N .0800: 0:88 0... 4 80.93 00.08.80 0... 00058 0:. o. .08.. 0... 8.0%.... .05 8.508.
8858...... 5 :03» 0.... 80830.30 088.5 450.853 .x no 800.. .8: 0... 8.. 00208.. 8880:... .w 030,—.

45

showed significant allele frequency differences between flies collected from
sympatric hawthorn and apple trees at the Grant, Michigan site (Tables 9,
10). Allele frequencies for Acon-Z and Me were significantly different
between hawthorn and apple flies within each of the three years sampled
and did not vary significantly among either apple or hawthorn trees within
any year (Tables 9, 10). Aat-Z, Did-2 and Mpi also showed host related
differences. However, gene frequencies varied among flies from different
apple trees for 1401-2 in 1985 and for Did-2 in 1984 and 1985 as well as
among hawthorn trees for Mpi in 1985 (Tables 9, 10). Allele frequencies
for Had differed significantly between hawthorn and apple flies in 1985 but
not in 1984 or 1986. Had also varied significantly among apple trees in
1984. None of the other 7 polymorphic loci resolved in this study showed
significant inter- or intra-host frequency differences.

Allele frequencies were reasonably constant between 1984 and 1985. Of
the 13 polymorphic loci, only apple tree #1 for Me (G=9.67, B s 0.01, 1 (If)
and hawthorn tree #1 for Dia-Z (G=4.66, E, _<_ 0.05, 1 df) had significant
frequency differences between 1984 and 1985 (data for Acon-Z, Me, Mpi,
Aat-Z, Did-2 and Had are presented in Table 10). However, eight out of a
total of 20 tests for Me, Acon-Z, Mpi, Aat-Z and Dia-Z showed significant
differences between 1985 and 1986. Allele frequencies for both apple and
hawthorn flies displayed a general trend to become more "apple-like" in
1986 compared to previous years (Table 9). The shift in 1986 allele
frequencies could be explained by selection either in the field or during
laboratory rearing as flies collected as larvae in 1985 and adults in 1986
represent different life history stages of the same generation. However,

frequency shifts between 1985 and 1986 were in the same direction and of

46

.8000088 .8.....0.... .0 0.9.0. .800... 0... 00.30... 8.0.... .0 .083. 0... 8.0.80 0.... 00.0.8500 m. 0 0.00...

 

000.0 00 ~00... 00 000.0 000.0 00 000.0 000.0 00 000.0 000.0 00 3.0.0 00 0130.. .3002.
000.0 00 000.0 00 000.0 000.0 00 0.00 000.0 00 000.0 03.0 00 000.0 00 0130... 0000
000.0 00 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 00 0130..
000.0 00 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 00 0130.. .0030...
000.0 00 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 000.0 .0 000.0 00 0130: 0000
.0080...
000.0 000 000.0 000 30.0 000.0 000 000.0 000.0 000 000.0 000.0 000 000.0 000 0130... 0000
000.0 00 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 00 0100.09. .3085
000.0 00 000.0 00 000.0 000.0 0.. 000.0 000.0 00 000.0 000.0 00 000.0 00 0100.50 0000
000.0 00 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 00 01009....
000.0 00 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 00 0100054 .0080.
000.0 00 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 00 0100.09. 0000
000.0 00 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 «0 0100.090 .080.
000.0 00 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 000.0 00 000.0 00 0100...... 0000
gagnflgqmgnagnan .0030.
8i jar... lg]: lgll Jall Idal a... a

.00.... 0.08 n .. .0... 0.00.0 000.00.24.80
0... .0 82. 0.0.... 0.8 08.03.... .8... 00.00000 8.... 8.. 002 000 0-0.5 .080. .32 0.800. .0: 8.. 8.000000... 0.0.2. .0 0.00...

47

.2 00.. 33.-00.0.... ..0.0. .00..

.200 0... .0. .0.0. 00.00.. 0... 30.000 0.00.. 00.. 0.. 00.. 0 ..0 00.000000 0.03 0.00. .00....0.... .0000. 0 .0. 00.3.0000 003 3.0000822.
.00..-0..... 800......00 00:3 ..30..-0.....< 0003.00. :00. 0.00. .00....0.... .0. 000.. 0.03 0.0.0. 00.00.. 000... 0:0 000.. .00.. 000.00
00.00.. 0.03 0.00. $00-00... 8000.00.0-000 .0. 00.000000... 00000 .00. 00000000090 .00... w m 3...... .0... w m 3... 00... w m 3

 

.0. 00.0 3.0. 00.3 .0. 0... .0. 00.0 .0. 00.3 .0. 00.0 0 8.03. .095.
.0. 00.0 .0. 00.0 .0. 00.0 .0. 00.. .0. 00.0 .0. 0.... 0 8.00.3. 0000\300.
.0. 00.0 .0. 00.0 .0. 0...... .0. 8.0 .0. 00.0 3.0. 00.0 . me... 0&0 .0300
a. 00.0 .0. 9.0 .3. 00.0 .0. 8.0 3.0. 00.0 3... 00.0 0 0.5.3. 3.60
.0. 00.0 .0. 00.0 .0. 8.0 .0. fl...- 3... 00.0. .0. 2.0 0 800:0. .85..
.0. 00.0 3.0. 00..- 33. 8.... 3.0. 00.0. .0. 00.0 .0. 8.0 0 on... 0&0 0000000.
.0. 00.. .0. 00.0 .0. 00.0 3.0. 03.0 .0. 00.0 3... 00.0 0 on... 0&0 50.80

.0. 00 0 3.0. 00.0 .- .- u u 03....- 8... 0&0

3.0. 00 0 33.0. 00.00 -. n .- u .30... up... 0&0
u -. 33.... 00.00 33.0. 03 00 33.0. 00.00 33.0. 00 00 3.0.0.3. 5980 .83....
3.0. 00.0 3... 30.0 .3. 00.0 .0. 00 0 .0. 00.0 .0. 00 0 In... is .00.

.- 33... 00.0. 3.... 8.0. u n u 3...... 88. 3&0

u a. .1. 33.... 00.00 . n u ml... 8.... 0&0

-. .0. S 0 .0. 03.0 -. - -. an... «a... 0&0

u u .- 310. 00.0. .- u 0 3... 30803.

.- -. n 33.0. 8.0. u u 0 .2... 33.-8&0

u u .- 3... 0...- .- u 0 on... 3.0.8.3.

33... 00 00 u u n 33.0. 00.3 33.0. 00.00 $59.00
.0. 8.0 .0. 00 0 .0. 8.00 3?. 3.0. .0. 00.00 .0. 00.0 800...! was. .095.
.0. 00.0 3.0. 00 0 3.0. 0100 .0. 00.0 .0. 00.0 .0. 8.0 8.5 «0&0; 000.

.0. 3.0 33.0. 00.0. 3.0. 0...... 33.0. 00.00 33.0. 00.00 33.0. 00.0. 30036380
.0. 00.0 .0. 00 0 .0. 8.0 .0. 00.0 .0. 00.0 .0. 8.0 80.03.. was. 33.0.
.0. 0.... .0. 00 0 .3. 00.0 .0. 00.0 .0. 3.0 a. 00.0 .85 1&0 an! 080
3 0H... a 3. E a. 3 EH!»

00005020.. 0. ..0>.0 0.0 80000... .0 000.000. .0...
00020.2 000.0 0... .0 0.5 000 0-0.5 .043. 0.0:. .0083. .03. .0. 00000.0...0 0000000.. 0.0.... .0. 0.00. 0000000000-.. .0. 0.00...

48
similar magnitude for both hawthorn and apple flies (Table 9). Therefore,

if selection did occur, its effects were unifomr across host plants.

Allele frequencies for Acon-Z, Me, Mpi, Aat-Z, Did-2 and Had may not
be evolving independently. Hitchhiking effects for linked loci and epistatic
selection can coordinately change allele frequencies at a number of loci. To
examine these possibilities, correlation coefficients were calculated between
pairs of non-allelic genes based upon Burrows disequilibrium values
(Cockerham and Weir, 1977; Laurie-Ahlberg and Weir, 1979). Significant
disequilibrium was found between Me/Acon-Z, Me/Mpi, Acon-Z/Mpi and
Aar—Z/Dia-Z (Table 11). These results agree with the known genetic map of
R. pomonella as Aar-Z and Did-2 are 3.2 centirnorgans apart on linkage
group I while Acon-Z, Me and Mpi are very tightly linked on linkage group
II (see Chapter 1). Loci showing genetic differentiation between hawthorn
and apple flies therefore map to thee different regions of the genome, and
hitchhiking effects are likely among non-allelic genes within at least two of
these regions.

Interactions between unlinked loci were not readily apparent from the
correlation coefficients. Twenty-one significant tests of gametic
disequilibrium were observed in 314 pairwise comparisons between Mpi,
Acon-Z or Me and either Aat-Z or Did-2 (data not shown). Had, which has
been mapped to linkage group III (see Chapter 1 and Berlocher and Smith,
1983), was in gametic equilibrium with either Acon-Z, Me, Mpi, Aat-Z or
Dia-Z in 153 of 160 tests. Although inter-chromosomal levels of gametic
disequilibrium cannot be completely explained by random type I errors, no
consistent pattern was observed in either the loci or alleles involved in the
significant tests. Allele frequencies for unlinked loci, therefore, appear to
be evolving independently in both hawthorn and apple populations.

49

«eehmmm.o
cutomm¢.o
«ceHmv¢.o
«cthmm.o

«tebmnm.o
«eeanm.o
«temmom.o
«ehnno.o
«mnm~.o
«eHmmw.o

«eemmmw.c
«cemmb¢.o
«cemvom.o

«echH~¢.OI
«hHon.OI
ammo.OI
«HHnn.ct

mm¢N.oI
HmmH.OI
mm~N.or
omm~.0t
thH.OI
«teemne.0I

«tmn¢n.0|
mth.OI
«eahan.0t

nme.OI
«imvvn.OI
vaN.OI
mhoH.OI

«onvn.ot
thN.OI
«ochN.OI
nan.OI
hmmH.OI
«Chum.OI

«ommH.OI
whNH.OI
«MNQN.OI

cnth.o
«emmmn.o
HnNH.o
momH.o

«eHmvc.o
cecmmn¢.o
«bnm~.o
«oomm.o
unHH.o
nnoo.OI

«nona.o
«mamm.o
«athm.o

«eemmwm.o
«eHmm¢.o
«emmmn.o
«aHHnn.o

«Hmmn.o
cenmbn.o
«evnwv.o

stepmom.o

«ann.o

bmnm.c

«echb~.o
«eamhmv.o
«aeHnon.o

mmmrflafimuuua mmmrbgbooifi Sigooam: 8H 7804\83 mmmufioaxooam:

N
H

HNM

N
H

.mSHm>.m3onH:m

a 3m:
H ya:
wanna
magma

n.3mm
n.3m:
H 3a:
manna
onna
manna

H am:
«done
adage

mama

Managua uogggfiuméfifluugfldbflm 3o. war; 8. war. no. mm...

.muaauav
mama

Aun>uuuv
mama

Aun>umqv
«mad

\qmma

dos?» 8355336 com: v83 mono» 0223-5: .3 33 :853 33328 $25 358530 5:39.80 4 $35.

50

Discussion

Several factors, alone or in combination, could be responsible for the
observed allele frequency differences between hawthorn and apple flies.
Among the possibilities are: (1.) Post mating reproductive isolation
between hawthorn and apple flies; (2.) A genetic bottleneck associated with
the founding of the apple race; (3.) Differential larval survivorship
associated with the host fruit environment; (4.) Differential host
recognition by adult flies; and (5.) Temporal differences in the timing of
adult emergence.

It is very unlikely that post mating reproductive isolation is involved in
differentiating host populations. Hawthorn and apple flies readily mate in
the laboratory and produce viable F1 progeny (R. Prokopy, pers. comm.)
Percent egg batch is similar for apple x apple, hawthom x hawthorn and
hawthorn x apple test crosses (Reissig and Smith, 1978). Although data
from F2 and backcrosses are needed to completely rule out postrnating
isolation, the likelihood of reproductive incompatibility between these flies
is remote.

A genetic bottleneck during the colonization(s) of apple by R. pomonella
may have initially caused gene frequency differences between the host races
due to drift. However, for neutral genetic differences to persist requires
restricted gene flow between sympatric apple and hawthorn populations.
Variation for intrinsic biological factors (i.e., differential larval
survivorship, host recognition and/or adult emergence) are therefore
necessary to account for the continued maintenance of host race

differentiation regardless of its original cause.

51

Although differential selection on larval populations infesting hawthorn
and apple fruits could account for genetic divergence, larval selection alone
cannot explain the significant frequency differences observed between
adults captured directly from hawthorn and apple fruits in 1986 (Table 9,
10). Random dispersal of flies between host plants would homogenize adult
gene frequencies even if larval selection was intense (Felsenstein, 1981).
Furthermore, adult hawthorn and apple flies show the same pattern and
magnitude of host related differentiation as larvae. Gene flow, therefore,
must be restricted between hawthorn and apple populations indicating that
the host races are not randomly mating in nature.

Traits involved in host recognition and the timing of adult eclosion are
the most likely candidates responsible for reducing gene flow between
hawthorn and apple races. Host acceptance behaviors do differ between
hawthorn and apple flies (Prokopy et al., 1988) and prior adult experience
can affect host preference in R. pomonella (Prokopy et al., 1982). Field
studies have shown that hawthom and apple flies differ in their mean adult
emergence times by approximately a week and a half in the field (Diehl,
1984; Feder et al., manuscript in preparation) and by even more under
laboratory conditions (Smith, 1986, 1988; Feder et al., manuscript in
preparation). Also, the distribution of adults at the Grant site closely
follows the fmiting phenology of their host plants as early in the summer
flies are abundant on ripe apples but scarce on immature hawthoms while
the reverse is true 3-4 weeks later in the season when hawthoms ripen (see
Figure 2 for adult collecting dates). Differences in diapause termination
and host preference, together with adult conditioning, could all contribute

to the establishment of an assortative mating system with early emerging

52

adults tending to reproduce on apples and later emerging flies on
hawthoms.

The existence of genetic differentiation between sympatric hawthom and
apple populations of R. pomonella at the Grant, Michigan site confirms the
status of these flies as non-randomly mating host races. The recent
formation of the apple race indicates that these flies are diverging
sympatrically in the absence of geographic isolation. Additional research is
needed, however, to clarify which biological factors are most responsible
for creating and maintaining host related frequency differences at the
Grant, Michigan site.

The geographic pattern of genetic differentiation between apple and
hawthorn populations of R. pomonella appears to be complex, however.
McPheron et al. (1988a) have also documented significant inter-host
frequency differences at a sympatric site in Urbana, Illinois. Their results
differ slightly from mine, however, in the pattern of loci and frequencies of
alleles displaying inter-host differentiation. In addition, genetic analysis of
geographic variation across the eastern United States (see Chapters 3, 4)
indicates that north-south allele frequency clines exist for both apple and
hawthorn flies across their ranges for five of the six loci showing host
associated heterogeneity. Inter-host differences are therefore
superimposed on latitudinal patterns of variation within the hawthom and
apple races and these clines help to explain the genetic differences between
sites in Michigan and Illinois. The results from the geographic survey raise
the question of whether sympatric mechanisms alone account for the
development of complete reproductive isolation between R. pomonella
populations. We consequently feel that it is inappropriate to definitively

state that hawthorn and apple races represent "incipient" species. However,

53

the R. pomonella complex does contain a number of sympatrically
distributed sibling species specialized on different host plants which are
inter-fertile in laboratory crosses, yet remain distinct in nature (Bush,
1966; Berlocher and Bush, 1982; Feder and Bush, submitted; Feder et al.,
submitted).

Chapter 3

The Geographic Pattern of Genetic Differentiation
between Host Races of Rhagoletis pomonella

Introduction

True fruit flies belonging to the Rhagoletis pomonella sibling species
group are at the center of a long standing controversy concerning modes of
speciation. Speciation in animals has traditionally been assumed to require
the complete geographic separation of populations in order to eliminate the
homogenizing effects of gene flow (Mayr, 1963). However, Bush (1966,
1969a,b, 1975a) has proposed that certain species groups in the genus
Rhagoletis diverge sympatrically via shifts to new host plants. Because
Rhagoletis larvae are host specific fruit parasites and mate selection in these
flies is directly coupled to host plant recognition (Prokopy et al., 1971,
1972), variation for traits related to host preference and larval survivorship
can produce genetically based barriers to gene flow. This reduction in gene
flow can result in the sympatric formation of partially reproductively
isolated "host races" and eventually species adapted to different host plants.

The recent sympatric shift of the species R. pomonella from its native
host hawthorn (Crataegus spp.) to domestic apple (Malus pumila) provides
an opportunity to determine whether host specialization can differentiate
fly populations in the absence of prior periods of geographic isolation.

54

55

Rhagoletis pomonella was first reported attacking domestic apples in the
Hudson Valley, New York in the 1860's (Ward, 1866; Walsh, 1867).
Agricultural records document the subsequent spread of the "apple fly"
from New York throughout eastern North America over the last 120 years
(Illingworth, 1912; O'Kane, 1914; Porter, 1928; Bush, 1969a; see Figure
3). The implication that apple flies originated from just a single or
relatively few colonization events may be misleading. The spread of the
apple fly from New York could also have resulted from several
independent host shifts corresponding to expansions in the range of
domestic apple cultivation. It is clear, however, that hawthoms are the
original, endemic host plant of R. pomonella. Although several native
species of crabapple exist in North America, none support populations of
the fly (Illingworth, 1912; O'Kane, 1914; Porter, 1928; Bush, 1966). The
reason for this has not been adequately studied but it has been suggested that
fruits of native crabapple species ripen too late for R. pomonella to
complete development before the onset of winter (O'Kane, 1914). The
historical timeframe of the host shift is also reliable as domestic apples were
introduced into North America from Europe only within the last 350 years
(Chapman and Lienk, 1971). Therefore, at no time during the formation of
the "apple race" was it geographically separated from hawthorn infesting
populations.

In Chapter 2, I reported the finding of genetic differentiation
between sympatric apple and hawthorn populations of R. pomonella at a
field site in western Michigan. Six allozyme loci displayed significant host
related frequency differences for both adult and larval life history stages of
the fly (Tables 9, 10). These six loci mapped to three different regions of

the genome (Chapter 1) and significant disequilibrium was observed

Figure 3.

56

Collecting sites for R. pomonella across the eastern United
States and Canada (see Table 12 for descriptions of study
sites). Dotted portion of map designates the region where flies
were first observed attacking apples. Dotted and shaded areas
encompass the distribution of apple flies when the first
published report of the race was made by Ward (1866). Years
are given for the first observation of the apple race in a region.
The current distributions of the apple and hawthorn races are
indicated by the solid and dashed lines, respectively. Figure 3
is a modification of Figure 3 from Bush (1969a).

57

 

“0“ 3‘: '- ..I '
mudc uammd

 

’.

    

/
w

a

 

a

0.
’u.o‘fi

   

      

.Jfiexmt.

O
O.I‘I-

.7.I.I.l.l

J
4,
.-.H

'I...- \-

32.x!

I

-o!'o-o‘

/
I
o.
.
a

an...» -....-.-.-.. , a.

.\.C.\~ .I

  

ex

  

   

a-..“
a
«3.-.-..

v

-.-i\
at. ,

I.-.‘ f'.'--.I--.'.-.'.A

{Nd-

”ho-0d

I
v

-..-.---..'

O ‘ A
[riff H.
I v. "a \-
, p a .
um ow
. frk
. a
o x

’O-
’I

m"'

o
l

a
I
\
7"
r‘
.'
’

--.‘O'
r -.'-'-o‘.

.-.-.'.-.b .'.-.'.-

~ .U.-.I.'I.._I.

 

 

I

 

e

.0 g
.43
.

.-o--- .-o-.-.'.-.-

  

 

..--‘.--.-.

  

 

n'o--n-.- -. u.
' '0

 

 

 

 

58

between non-allelic genes within, but not between, two of these regions
(Table 11). Hawthom and apple populations at the Michigan site therefore
represent partially reproductively isolated "host races", a result supporting
a sympatric mode of divergence for these flies.

Several critical questions concerning the geographic pattern of
genetic variation for apple and hawthorn flies still need to be answered,
however. The first issue is whether host associated differentiation is
widespread across the range of R. pomonella. In this regard, McPheron et
al. (1988a) have also found significant inter-host allele frequency
differences between apple and hawthorn flies at a sympatric site in Urbana,
Illinois. Are these two sites representative of host related divergence for R.
pomonella across eastern North America or do apple and hawthorn flies
differ on just a local or regional scale?

The second question is whether any trends exist in the partitioning of
genetic variation between and among apple and hawthorn populations and,
if so, whether these patterns point to factors responsible for differentiating
the host races. For instance, do the same loci consistently display inter-host
heterogeneity between apple and hawthorn flies across sites or are different
genes involved at every site? Differences do exist in the loci and alleles
showing host associated differentiation between R. pomonella populations
from Michigan and Illinois (see Chapter 2 and McPheron et al., 1988a).

A For example, Diaphorase-Z displayed higher levels of inter-host
heterogeneity in Michigan than in Illinois populations, while the reverse
was true for Hydroxyacid dehydrogenase. In addition, Aconitase-Z 95
allele frequencies were reversed between apple and hawthorn populations
in Michigan and Illinois. Are these the patterns of variation expected if

differential selection were occurring between the host races or is genetic

59
drift a more plausible explanation? Significant linkage disequilibrium was
also observed among loci displaying inter-host differentiation at the
Michigan site. How common is linkage disequilibrium in R. pomonella
populations? If certain non-allelic genes are regularly in disequilibrium
does this mean that host related selection is causing specific regions of the
genome to co-evolve?

Finally, the apple race reportedly arose in the Hudson Valley, New
York and subsequently spread across the rest of eastern North America. Is
there any genetic evidence for this? For example, is the apple race less
variable across its range than the hawthorn race as would be expected if the
apple race owed its origin to just a single, or relatively few, founder events?
Do peripheral apple populations possess alleles characteristic of hawthorn
populations near the supposed home of the apple race in New York or are
they genetically more similar to local hawthom populations, as would be
predicted by the multiple shift hypothesis? Answers to all of the above
questions bear directly on the evolutionary significance of sympatric host
race formation in the R. pomonella group.

In this Chapter, I report the results of an electrophoretic study of
allozyme variation for hawthom and apple populations of R. pomonella
from across the eastern United States and Canada. The findings indicate
that significant genetic differentiation exists between hawthorn and apple
flies across most of the range of R. pomonella. The same set of six loci are
consistently responsible for host related differences, and five of these six
loci show a pattern of geographic variation indicative of latitudinal allele

frequency clines in the two host races.

60
Materials and Methods

Flies were collected as larvae from infested fruit at ten paired feral
apple (Malus pumila) and hawthorn (Crataegus mollis) sites across the
eastern United States as well as from an apple site in Kentville, Nova Scotia
(see Figure 3 and Table 12 for locations and descriptions of study sites).
Sympatric site 5 (Grant, Michigan) represents the same apple and hawthorn
populations I analyzed in Chapter 2. Likewise, flies surveyed from site 4
(Urbana, Illinois) were from the same populations, but were not the same
individuals, that McPheron et al. (1988a) studied. Multiple year samples
were collected for six of the ten paired sites (Table 12). At site 5 in 1987
and site 7 (Gas City, Indiana), larvae were dissected from infested fruit and
were immediately frozen in liquid nitrogen for later electrophoretic
analysis. Otherwise, we brought infested fruit back into the laboratory and
allowed larvae to pupate in moist vermiculite. Pupae were sifted from the
vermiculite and subsequently stored at 4° C to simulate winter diapause
conditions. After five months, we removed the pupae from the cold and
held them at 22° C, under a 15:9 hr. lightzdark cycle to terminate diapause.
Eclosing adults were collected and stored at -80° C until they were analyzed
electrophoretically.

Standard horizontal starch gel electrophoretic techniques were used
and are described fully in Chapter 1 and in Berlocher and Bush (1982). I
stained for a total of 21 different enzyme systems and was able to resolve 29
allozymes (see Table 8 for a list of resolved loci and their abbreviations).
Isozymes which migrated the nearest to the cathode were designated 1, the
secOnd closest 2, etc. Alleles at each locus were numbered according to

their relative anodal mobilities, with the most common allele at a locus

61

.hm.vv
.vN.N¢
.nm.~v
.mn.ov
.Hnoov
.H¢.N¢
.HN.nv
.mo.ov
.ON.Nv
.mm.¢v
.ho.m¢

In 4
o.” :2.
5% =.<
o.m :3
a.» mi
m4 :3
56 we
16 *3
n6. 2.4
3% :3
a3 m.<

Aaxv mocuuuao

mm
mm.mm
mm
mm
hm
mo.vm
bolvw
mo
mo.vm
mm.vm

mmtvm

3333333

«8:8 638m «62
£2328. 603.3 .98 «E8 .3

.85; .uflofia .9658 6mm: :65

.>.z .580 638.8 éa .3 .m.».z

9.30 .55 .38 .3 .3»:

.3 .38 98 .38 .8 if

.6? 6523 .m .998 .35 .3 £2:

£22 £56 .38 rpm 5":

.SH .235 .3958 .SH .35

.fiH 689.32 name .2. if

.032 .ofldfifld .53 5.3x

.033 .533 .33 :89.
fig

.5505 25 .3 m8— 8

€308 9: £55 803 3388 82. 03% Ba c.5532 223 27. REE u 8.3 66532":

damn}. £350 93 8.8m 33:: E880 05 «8.8a ~336qu .x 3.. new“ wage—BU

O
H
H

o o o o o o o o o o
O" N n v In ‘0 I‘ Q Os 3

i

.2 035.

62

assigned a value of 100 and used as a standard (Note: The relative mobilities
of electrophoretic alleles resolved for the 13 polymorphic loci are given in
Tables 21-32 which appear after the Discussion section on pages 91-102).
Significant deviations from Hardy-Weinberg genotypic expectation
were determined by G-tests. Alleles were pooled, if required, so that the
expected frequency of all genotypic classes were >1. The Levene
correction (Spiess, 1977) was applied if sample sizes were <100 or if
genotypic classes had expected numbers <1. G-contingency tests were used
to test for allele frequency heterogeneity between host populations and
across sites and years. Alleles for a locus were pooled, when necessary, to
ensure that each cell in the G—contingency test had an observed number 2 5.
F-statistics were calculated by the method of Weir and Cockerham (1984)
with variances estimated by jackknifing over populations, loci or years.
Corrections for unequal sample sizes in Weir and Cockerham's method can
result in negative FST values. In these instances we report Fs'r as 0.
Correlation coefficients (rAB) were calculated between pairs of non-allelic
genes based upon Burrow's disequilibrium values (Cockerham and Weir,
1977). To test the null hypothesis rAB=O, the correlation coefficients were

2 transformed to approximate a normal distribution with variance:
’22 = l /n-3.

Arcsine transformed allele frequencies for apple and hawthorn
populations were linearly regressed against latitude for the ten paired sites
(Note: At sites with multiple year samples I transformed the mean allele
frequency for the population averaged across years). I tested the null

hypothesis that the regression coefficient (i.e., slope of the linear regression

63

equation) = 0 by ANOVA using F—tests. Curvilinear regressions were also
performed based upon second order polynomials but F-tests revealed that
they did not significantly improve the fit of regression lines compared to

first order equations.

Results

Thirteen of the 29 allozyme loci resolved were polymorphic
(frequency of the common allele 5, 0.95, Table 8). A Mendelian mode of
inheritance has been established in R. pomonella for 12 of these
polymorphic loci with Aat-I being sex-linked (see Chapter 1; Berlocher
and Smith, 1983). Excluding Aat-I , 30 significant deviations from Hardy-
Weinberg equilibrium were observed in 670 tests (Tables 21-32). The
significant deviations displayed no regular pattern and were fewer than the
number expected due to random type I error (30 compared to 33.5).
Genotypic frequencies for autosomal loci were, therefore, generally in
Hardy-Weinberg equilibrium within both apple and hawthom populations.
This was not true for the sex-linked locus Aat-I in that 27 out of 33
populations showed significant departures from equilibrium proportions.
However, when females (which are the homogametic sex in R. pomonella;
see Chapter 1) were considered separately from males (which are
hemizygous and therefore should not be included in calculations of Hardy-
Weinberg equilibrium), females were generally in Hardy-Weinberg
equilibrium for Aat-I (2 significant deviations out of 33 total tests).

R. pomonella flies infesting apples were genetically different from
those infesting hawthoms. Although no fixed allele difference was found

for any of the 29 loci resolved, significant allele frequency differences were

64

observed between apple and hawthorn populations at nine of the ten paired
collecting sites (Tables 13, 21-32; Note: Paired sites 1-10 are organized by
latitude from north to south in these three Tables). Only site 8 near Kenton,
Ohio did not differ significantly for at least 1 of the 13 polymorphic loci
analyzed.

Host associated genetic variation displayed a distinct dichotomy. Six
loci (Me, Acon-Z, Mpi, Die-2, Aat-Z and Had) consistently showed
significant host related differences (60 out of a total of 113 G-contingency
tests were significant for these six loci; Table 13). In contrast, the
remaining seven polymorphic loci analyzed in the study had much lower
levels of inter-host variation. Only 10 out of 113 tests were significant for
these seven loci and there was no trend evident in either the loci or alleles
involved in the significant differences.

Inter-host differentiation was stable over time for the six sites where
multiple year collections were made. At site 5 (Grant, Michigan) for
instance, Me, Acon-Z, Mpi, Dia-Z and Aat-Z all showed significant allele
frequency heterogeneity between apple and hawthorn flies over a 4 year
period from 1984 to 1987 (Table 13). The only other locus which varied
significantly between host populations at this site was Had in 1985.

Allele frequencies within apple or hawthorn populations fluctuated
across years, however (Table 14). This was especially true for hawthom
flies in which 16 out of 83 G-contingency tests indicated significant
heterogeneity over time (Table 14). Intra-host temporal variation did not
overlap in frequency or compare in magnitude to inter-host differences,
however. F-statistics calculated across years for Me, Acon-Z, Mpi, Did-2,
Aat-Z and Had at site 1 and site 5 document this point (Tables 15, 16).

Overall FST values for these six loci across years between host populations

65

 

.28... Eng 10 Sod w m :5 .86 .v. m 3 mod w m ..
«AHV34 «a a on v
«Advau4 .ecaavwmm «a pm h
III mm m
ccAHVnUH « «ea mm
«Advanm « a *«e «« vm n
4 mm
£18789 .1 a: mu OH
is «a; 4 mm
«Aavdrunw «a «as ea w
a«AHV>04 .cadvatuna «a «a has «a «a; mm m
IIIII can use tea «a; «at no
IIIII ace «*4 «a: «a; «a; ma
«a sea is; «a: «a: cue mm
4% «ea ace «44 at: we m
«a aaa «a; «a; as« «a; mo
¥¥¥ i Gm N
It!!! a; «as see eye on
«Advaluma a «4 he; mm
a «a vm H

 

”83333342

Egg

 

«82.323 5 :03» 2a 98. c8 Song .6 mop—won 63.853 .x we Sagan—on 9.585
03% 93 50532 5223 mama gm 3 bmoeowohso: 5:395 22? he 33 xocowauaonué

.mm 038.

 

 

 

 

.53 593580 nu H86 w m :5 .86 w m .3 mod .v. m a
«a Amvnlu2.1m3~|§ 2213 25:3 :13 3:13 .25 3.3 :13 58.335.—
a: avalancfimv >233: 3.3 «lama .513 33 2.3 «a 83 :13 e: 83 :13 a?
gas
«135% .33. 33 Hanna 3
£3332 4 n
I
3.83.98 4 3
333.com is 1: a
4 D
«332 13 «13 «13 :13 :13 x
.313 a m
.13 1.3 13 m
«3 a ~
«1.3; as m
«SEE a a
lag rial Ilmuumal Ilnumudl Jail lug twnlu g g
g fix: g

@8055?“— 5 coin 8a sun. com 8895 .3 monumen—
aémm 93 E3» 828 bmocoweao: 5:033 0.23 .8:.nu£ .83 $8. accom=u590

.3 03am.

67

.5005.» 0080M 38080. Ea 005088 .8: 803.3
50080.5 05 .3 058.80.80.— 88805 am > x 88: 8283030 05 E 000205
0.3 «coma—anon 0:50 05 c.6532 £3 :05: 00:0; 5mm 08 8 «80> x 0.8: 0

.8. as 2.88 £38. - 838 .5680 a

.050» ~05 08830 000538.. - 83300 Bataan a

 

n0d0.0 H «0N0.0 0N00.0 H ¢m00.0 n5N00.0 H 0000.0

~v00.0 H mn00.0 nnH0.0 H 0000.0 HmH0.0 H nm00.0
0n00.0 H 5m00.0 50H0.0 H N500.0 5N00.0 H 0000.0
H000.0 H «500.0 5000.0 H 0000.0 5m00.0 H 0000.0
m000.0 H 0000.0 mH00.0 H 0000.0 5500.0 H 0000.0
n¢H0.0 H va0.0 0500.0 H «500.0 NNH0.0 H N¢00.0

38.0 H 88.0 88.0 a 38.0 038.0 a 88.0
flag at; a
.3

 

 

.Asmcoama .832an
_ 0:» .0 83.238 E0532 0.3 038 3.. «.80» $800 0032, 5mm

«:02
«:08
E:
«.82

.n g 030.5

68

80.35. 00800 .8880. as. 8.0.098 .09. 503.2.
80800.... 0... ..o 03.083080. 0.5.0.8... a. > x 86.. 83.0.8.8 0... a. 008.0...
0.... 80.3.2.9. 0.0.... .05. E052... 59. :0...» 00...; .00". 0.... 8 80> x 0.83 0

.8. .25 28...... 0.3.8. - 8.330.. .528 ..

83>. .96 0.8.8 00.5.8. - 8.830.. 888.0 a

 

00.3.0 H 5500.0 0v00.0 H 0000.0 ~000.0 H 5.30.0 .0395

0000.0 H 030.0 #0006 H 0000.0 0000.0 H NN00.0 3
0000.0 H 0000.0 0000.0 H 0000.0 0000.0 H 0000.0 Nina
0500.0 H 5900.0 0000.0 H 0500.0 0000.0 H 0N00.0 "IQ—d
00.3.0 H «000.0 0000.0 H 0000.0 0000.0 H 0000.0 Hg
00....0.0 H 300.0 0000.0 H 0000.0 0000.0 H 0000.0 05

 

 

88.0 0 886 286 0 $8... 88... a 286 a:
a Hung g 0.8.
£33qu
..§m...0§..§.0.

n 0.... .0 83.0.2.9. c.8538 .08 0.0% 8.. 80.. 08.60 00...? km". .0. 0.8....

69

were over 4 x greater at both sites 1 and 5 than values calculated for
hawthorn populations alone and 20 x greater than those for apple flies
(Tables 15, 16).

Geographic variation was observed among both hawthorn and apple
populations across the eastern United States and Nova Scotia. G-
contingency tests indicated significant geographic variation for Me, Acon-
2, Mpi, Did-2, Aat I & 2, Had and Pep-2 for both host races (Table 14).
Acy also varied significantly among apple populations. F-statistics were
calculated to quantify the degree of intra-host geographic variation (Table
17). Only data from a single year for hawthorn and apple populations from
each of the ten paired collecting sites were used in the calculations (1986
being chosen whenever possible, with 1985 results used for sites 2 and 9 and
1987 data for site 7). Overall FST values for all 13 polymorphic loci were
significantly greater than zero for both apple (0.0117 i 0.0054) and
hawthorn flies (0.0450 :3; 0.0208) as indicated by standard deviations
derived by jackknifing over loci. Geographic differentiation was much
greater among hawthorn than apple populations, however, as evidenced by
the nearly 4 fold larger overall Fs'r value for hawthorn flies (see above).

Substantial heterogeneity existed among loci in their Fs'r values for
populations across the eastern United States (Table 17). For example, Me,
Acon-Z and Had all had a Fs'r of over 0.100 for hawthorn flies while Pgm,
Acy, Ak, Idh and Aat-I all had a Fs'r of under 0.005 among the same
hawthorn populations. In fact, the same six loci (Me, Acon-2, Mpi, Did-2,
Aat-Z and Had) which consistently showed host related differences also
displayed the highest levels of geographic differentiation. If just these six
loci are considered, apple populations had a 1351‘ of 0.0166 i 0.0107 while
hawthorn populations had a Fs'r of 0.0821 i 0.0383 (Table 17). In

70

Table 17. PST values across apple and hawthorn populations
calculated by the method of Weir and Cockerham (1984).
hkgmfiveIfirrvahusaucreponedas(10000.

1.9.99: W W
via-2 0.0042 1 0.0049a 0.0676 1 0.0404
Ant-2 0.0020 1 0.0021 0.0122 1 0.0067
Me 0.0173 1 0.0094 0.1916 1 0.0557
Acon-Z 0.0378 1 0.0172 0.1225 1 0.0388
Hpi 0.0140 1 0.0063 0.0426 1 0.0185
Had 0.0208 1 0.0066 0.1098 1 0.0421
Pep-2 0.0137 1 0.0068 0.0063 1 0.0043
Idh 0.0000 1 0.0019 0.0033 1 0.0069
Ak 0.0000 1 0.0034 0.0016 1 0.0008
Pgm 0.0000 1 0.0027 0.0024 1 0.0031
991 0.0062 1 0.0070 0.0076 1 0.0078
Aat-l 0.0197 1 0.0150 0.0006 1 0.0082
Acy 0.0024 1 0.0029 0.0013 1 0.0025
A11 13 loci 0.0117 1 0.0054b 0.0450 1 0.0208
First 6 loci 0.0166 1 0.0107 0.0821 1 0.0383
Last 7 loci 0.0061 1 0.0042 0.0036 1 0.0019

 

a Standard deviation - jackknite estimate over populations

b Standard deviation - jackknite estimate over loci

71

contrast, F ST values for the remaining seven polymorphic loci were 0.0061
;I-_ 0.0042 for apple flies and just 0.0036 :1; 0.0019 for hawthorn flies.

Despite the observed spatial diversity among R. pomonella
populations, no obvious relationship existed between the geographic
location of a population and the number of electrophoretic alleles scored
for the population. Of the 11 sites we analyzed, site 10 (Amherst,
Massachusetts) was located the closest to the Hudson Valley, New York, the
supposed source of the apple race. Apple infesting flies were, in fact,
already present in western Massachusetts at the time of the first published
account of the new race by Ward in 1866 (Figure 3). Fifty-five
electromorphs were observed in the apple population on the campus of the
University of Massachusetts at Amherst for the 13 polymorphic loci
resolved in this study. In comparison, apple populations near the northern
and southern ends of the flies distribution in Door County, Wisconsin (site
1), and Urbana, Illinois (site 4), had 59 and 50 alleles, respectively, not
appreciably different from the number scored at site 10. In fact, the
number of alleles detected for apple populations 1, 10 and 4 was
proportional to the sample size at each site (Tables 21 -32). Furthermore,
apple populations at all three of these sites contained either more or the
same number of alleles as sympatric hawthorn populations (55, 53 and 50
for hawthorn populations 1, 10 and 4, respectively). The apple population
at Kentville, Nova Scotia (site 11), did show a reduced amount of genetic
variation (39 electromorphs). However, Acy was not scored for site 11.
Acy had an average of four alleles in apple populations so the total number
of alleles for Kentville should be adjusted to approximately 43. In addition,
the sample size at site 11 (39 individuals) was smaller than that of any other

population analyzed, which reduced the likelihood of detecting rare alleles

72

at the Kentville site. I also do not have a paired hawthorn population for
site 11 to verify whether the fewer number of alleles present in the
Kentville apple population is due to a general reduction in genetic diversity
for R. pomonella in this portion of its range. It therefore appears that if a
population bottleneck did occur during the formation of the apple race, then
either it was not severe enough to greatly reduce genetic variation in apple
populations or subsequent gene flow from hawthorn populations has
restored the original level of allelic diversity to the apple race.

Genetic variation was not randomly distributed among either apple
or hawthorn populations across the eastern United States, however. Linear
regression analysis revealed significant relationships between latitude and
allele frequencies for M e1 00 , Acon-295, Mpil 00, Had] 00and Pep-2100
across both host populations and for Did-2100 across hawthom populations
(Table 18). Except for Pep-2100, the absolute value of the slopes of the
significant regression equations were always greater for hawthorn than
apple flies (Table 18). For Pep-2 100 the slopes were essentially the same
between the host races (Table 18). Hawthorn flies, therefore, displayed
steeper north-south frequency clines for M72100, Acon-295, Mpi100, Did-2
100 and Had 100 than apple flies; a result indicative of the higher levels of
geographic differentiation observed in the hawthorn race.

The geographic pattern of inter-host differentiation is due in large
part to the differences in the slopes of latitudinal allele frequency clines
between hawthorn and apple populations. Figure 4 graphically depicts the
relationship between latitude and the average arcsine transformed allele
frequency for Mel 00, Acon-295 , Mpi3 7, via-21 00 and Had100 for 9 of the
10 paired sites in the eastern United States (Note: Site 8 is not included in
Figure 4 because data for Did-2100 are missing at this site). Inter-host

73

 

ummuuu c>oz4 Hoo.o v m «4. Ho.o w m 44 mo.o w m e
een.o « oe~.o ”no.6 ene.o a H>H.ou moo.o ooH >04
coo.” « oo~.H em~.o «66.6 « emo.ou ooo.o coat H-084
nem.o « noo.ou aoo.o mme.o « mmm.o oea.o cod and
oo~.H « oon.e «enoo.o pom.o « oom.a emmm.o cod on:
vee.o u oov.au emem.o mo¢.o « ooe.au eemae.o cod «teem
can.o « oo.~u eemoe.o «Hm.o « ooe.au emom.o ooH an:
ooH.H “_ooe.o eeeono.o cod.” « con.n eanm.o mm «.2004
oov.a « oom.e eemne.o m~¢.o a oom.~ eeeaam.o ooH a:
ann.o u e~v.o ~ea.o eon.o u ~nn.ou cmo.o ooa sum
mmm.o a oun.ou ovo.o nmn.o « omm.o eH~.o cod xc
aan.o « «He.ou can.o «om.o « noo.ou ooo.o ooH ccH
mno.o n nae.on onH.o mm¢.o a e~e.on ~mo.o cod muuoc
oo~.H a ooe.n eamm.o omv.o “ eme.o me~.o ooH «raga

Aqualung Jr 44.413 Iiiu _ 30.3%..”
WQOHMMAQQGQ GHOQMNMN WQGHHMHQQGQ Manda
6 Bob €25 consume

88:2 93 86:35 20% 35.8833 583 «o $ng comedy—we Soc: .«o 938%

do mono—m .66 336803 commmaoe Mo 80332. 588:»? Bacon Exam—33‘
donate, 0:05» 3:33ch 52% 50532 93 03% Ed 8 05 888

.m— 03.;

Figure 4.

74

Average arcsin transformed allele frequency for M e1 00 ,
Acon-295 , Mpi3 7, Did-2100 and Had] 00 for paired apple
and hawthom sites plotted against latitude. Site 8 (Kenton,
Ohio) is not included because data for Dia-2100 are missing
for this site. The linear regression equation is given along
with the amount of explained variation (r2) and probability
levels (_13) for the ANOVA F-tests of the null hypothesis that
the slope of the regression line = 0.

75

MDDHFS
.o v .m 6 .¢ 6 .m v .N v

i!
u
0

2920

HQDHWTEA

as. .35 .32 .353. .85 52335. mzmw nmzmommzés 253:.

.34 be.

0

EoSBmI K

.v oasmam

mm
60
.mm
10v

.mv

AONEHOEEH ENSE)

.om

 

.mm

76

differentiation is greatest at northern sites where allele frequencies for
Memo, Acon-295, Mpi3 7, Did-2100 and Hadloo are generally higher in
hawthorn than apple populations (Fables 21-25). At centrally located sites,
frequency clines for the two host races converge and apple and hawthorn
flies become less genetically distinct until at site 8 (latitude 40° 39') no
locus shows significant inter-host heterogeneity (Table 13). Host
associated allele frequency differences appear again at southern sites.
However, in the south, frequencies for M e1 00 , Acon-295, Did-21 00and
Hadlo0 tend to be lower in hawthom than apple populations, the reverse of
what was found for northern populations. The latitude of a site is,
therefore, an accurate predictor of the magnitude and pattem of inter-host
differentiation.

The fact that almost all of the loci showing host associated differences
also co-vary significantly with latitude suggests that allele frequencies for
these loci may not be evolving independently in fly populations. Me, Acon-
2, Mpi, Act-2, Did-2, Had and Pep-2 do, in fact, map to only three different
regions of the genome (Chapter 1; Berlocher & Smith 1983). Dia-2 and
Aat-Z are located 3.2 centimorgans apart on linkage group I while Me,
Acon-Z and Mpi are very tightly linked on linkage group II (no
recombination was observed among Me, Acon-Z and Mpi for 162 test cross
progeny scored in Chapter 1). Finally, Had is on linkage group [[1 at a
distance of 23.9 centimorgans from Pep-2.

The linkage relationships for Me, Acon-Z, Mpi, Did-2, Aat-2, Had
and Pep-2 suggest that hitchhiking effects and/or epistatic genetic
interactions are occurring among loci. To further examine this possibility,
I calculated correlation coefficients between all pairs of non-allelic genes
for these seven loci based on the method of Cockerham and Weir (1977).

77

Significant disequilibrium was found between M e/Acon-Z, Me/Mpz', Acon-
2/Mpi, Aat-Z/Dia-Z and Had/Pep-Z for most apple and hawthorn
populations (Tables 19 and 20). The patterns of disequilibrium between
non-allelic genes were the same both within and between the host races
throughout the eastern United States and Nova Scotia. Allele frequencies
for linked loci within each of the three genomic regions showing inter-host
differentiation are, therefore, co-evolving.

Linkage disequilibrium appears to exist only within chromosomal
regions displaying host associated variation, however. Idh maps 7.1
centimorgans away from Dia-Z and 10.3 centimorgans from Aat-Z on
linkage group I (Chapter 1). Nevertheless, Idh was in equilibrium with
Did-2 in all 19 tests conducted. In addition, allele frequencies for Idh did
not co-vary significantly with those for either Ak or Pgm which are located
8.9 and 38.7 centimorgans, respectively, from Idh on the opposite side of
linkage group I as Dia-Z and Aat-Z (Chapter 1; Berlocher and Smith, 1983).

Interactions between unlinked loci were also not readily apparent
from the correlation coefficients. Only 24 out of 627 correlation
coefficients were significantly different from zero between pairs of
unlinked loci for apple flies while 28 of 590 tests were significant for
hawthorn flies (data not shown). No pattern was observed for either the
loci or alleles involved in the significant tests. Allele frequencies for
unlinked loci, therefore, appear to be evolving independently in both
hawthorn and apple populations.

.uuco.o_»eooo co.u0.0aeou veeao~ucccu u .6 anon.“ >3 uec.0ceuoe 00:00.».co_m 000.0 N a sec 'o.o w u .6 mo.o w u o

 

78

 

...on.o ....m.o ...e¢.o ...e..o ..ne.o on.o oo,\oo. ~.eee\u.= .___
00¢~.0 o~.0 00~.0 00n~.0 0cn.0 no.0 moxhn
«v.0 «0.0. 00.0 50.0 0nn.0. 00.0. 00p\00p
000—n.0. pp.0. 0-.0. op.0. 000¢.0. e~.0. m0\00p ~.cou<\_ax
00n~.0. mp.0 0—~.0. hp.0. 00v¢.0u 0pe.0. 00—\00p .a:\ox
m—.0 ”0.0 s—.0 90.0. n0.0 0~.0 00p\00—
00.0. N0.0. “p.0- 000p¢.0u 00000.0. e~.0. 00«\00p
0h—.0 0N.0 000on.0 00009.0 000vn.0 00¢¢.0 mO\0op ~.cou<\ox ._—
00001.0 000~w.0 00000.0 0000n.0 00050.0 0N¢.0 00—\00—
000p¢.0 000om.0 000¢n.0 000mn.0 ¢—.0 0¢n.0 nh\00p
000—v.0. 000n¢.0. 000nm.0. 000~n.0- ¢~.0. 000pm.0. 0m\00p
000no.0. 000¢¢.0. 0000n.0. 000oe.0. 000om.0. 000—o.0. —~\00p ~uul<\~.o_0 .—
13.; 1a.; .33..” .33“ an. 33...... 3:43 : mluuee
oooxc_d
C35 «new 03?.
.3350 05 «850 BE: 6393 05 888 ”cove—anon came .50 2.52
.Hfifkrfixhw_H=2finH¥00flVvmRZRHWnaflQ=HTcfixH_hyo>RfiXumacfimomcuKXv~23UMapcfifiu .Q;.01£dh.

79

.0uc0_o.eeooo co_u0.0aeou noeaoeucntu u 06 u.0u.u >0 00c.000uou 0oc0o.e_co.m .00.0 w M 000 p0.0 w u 00 no.0 w u .

 

“v.0

0—.0
0p.0.

no.0.

0~.0
00.0.
00.0.

0000m.0
00epn.0
00000.0.
000nm.0.

«aqua

0~.0

00.0
~—.0.

00.0.

00000.0
0000.0.
p0.0
0Nn.0
op.0
0on.0.
0~.0.

00050.0

00.0

«p.0.

00.0.

0n.0
00.0.

000N0.0 000wo.0
... 00.0
~0.0. n~.0.
00.0. 0N.0
No.0. 00n.0
n—.0 ~p.0
0—.0. 00.0.
0p.0 “0.0.
00nn.0 0N.0
00nn.0 0000m.0
0000.0. 0hn.0.
000m0.0- 0000.0.
dung" «wqqqfl

 

08.6 80m 203..

00—x00p

noxsn
00p\00—
moxoop

00—\00—

00—\00—
00p\00—
mux00—

00—x00p
ms\00p
0mx00—
—~\00p

uuqqqqa

~.aea\u.= .___

~.cee<\_e:

_axxex

~.cou<\oz .__

~.u0<\~.0.0 .—

aou «aqua
0003s.;

00.2500 .3 030.0

80

auc0_u_0eoou 06.00.00000 toecoeucccu u 00 anew.» >n eoc.0eeuou 00:00.0.cu.u p00.0 w u 000

00.0 W M 00

 

000on.0

00n~.o
0~N.0
005N.o.

05.0-

50.0
000.0.
000on.0

0000n.0
000nn.0
00050.0.
0005n.0.

auaqq

0mn.0

no.0
0-.0
-.0.

o—.0.

50.0.
0000n.0.
0o~.0

—N.0
000n.0
00050.0.
0000m.0.

000nn.0

”v.0
00-.0
00~N.0.

0000~.o.

50.0.
0000n.o.
00000.0

0w~.o

nn.o
000mn.o.
000on.0.

000on.0

000~.o
09.0
n¢.o.

00N.0.

~0.o
0000n.o.
000nn.0

0000~.o
00n~.o

000~o.0.

000nn.0.

4000“ «aqua annual

o~.0

55.0
Np.0
00.0.

00.0

0000n.0
0N.0.
00~0.0

000—m.0
0nn.0

00005.0.
00~.0.

«aqua

 

cat .28 55%:

00p500—

mox5n
00—500—
noxoow

00—x00.

00—5009
00—500p
“0500'

00—x00p
n5x00p
ouxoop
pmxoow

«ududdd

~.eee\u.= ..__

~.ceu«\.ex

.axxex

~.coo<\o: .~_

~.u¢<\~.0.o .-

quaa .uumuuu
000xc_4

.0350 05 0880 0805 6830 20 888 80003060 505303 .50 2.53
503 05 83.80.0000 Eco» emu—=25: .8933 956530 coma—2.50

.oa 035.

35mm.

81

oucooo.»eooo co_u0.occou noecoeucaau u we u.0u.u >0 uoc.0aouov 00:00.._cu_m 000.0 N

m 000 90.0 W M 00 00.0 W M 0

 

000~0.0

000.0
00000.0.

——.0.

5—.0.
0p.o.
00.0

00005.0
00000.0
00050.0.
00000.0.

0~.0

0~.0
00.0

0~.0.

0p.0.
00.0.
——.0

00000.0
000.0
00000.0.

000.0.

0p.0

00.0

50.0

o~.o.

0~.0

0—0.0.
o~.o

0000.0

00.0.
«v.0

00.0

00.0
0o~.0.
pp.o

000—0.0
0000.0
000N0.0.
00000.0.

03 an as. a:
get 8.8 page:

000N0.0

00.0.
00.0.
~0.o.

0—.o.

No.0
000.0.
0Nn.0

n~.0

0N.0
00005.0.
00.0.

a...

oo—xoop

00550
oowxoop
00x00—

oo—xoop

00'500—
00—500—
005005

00p\00p
05500p
00\00p
pwxoop

uuqqqqa

~.aoa\u-= .___
~.ceu<\.ax

.axxex
~.coo<\o: .__

~.un<\~.0_0 .—

quua .nuauqu
000uc.4

3.83 .8 use...

82
Discussion

We can now answer many of the questions about geographic
variation posed in the introduction to this Chapter. First, is genetic
divergence widespread between apple and hawthorn infesting populations
of R. pomonella? The answer to this question is yes. Significant allele
frequency differences were found at 9 out of 10 paired apple and hawthorn
sites examined across the eastern United States. Host related differentiation
is, therefore, not just a local or regional phenomenon in R. pomonella, but
is common across eastern North America.

Second, do genetic differences between apple and hawthorn
populations display any patterns or trends? Yes, the same set of six loci
(Me, Acon-Z, Mpi, Did-2, Ant-2 and Had) are primarily responsible for
inter-host differentiation. In addition, five of these six loci show latitudinal
allele frequency clines for the host races. Furthermore, the six loci
displaying host associated variation are not randomly distributed
throughout the genome but, instead, map to three specific chromosomal
regions. Significant linkage disequilibrium exists in natural fly populations
between linked, non-allelic genes within each of these three regions (Had,
which maps to a separate chromosome from the other five loci, was tested
against Pep-2 which is located 23.9 centimorgans away from Had on linkage
group III), but genetic interactions between chromosomes are not readily
apparent from the data. It therefore seems that only a limited number of
genetic changes at specific locations in the genome are involved in the initial
host divergence process. Although I have probably not identified all of the

regions responsible for differentiating host populations, it appears that a

83

" genetic revolution" involving a restructuring of interactions across the
genome has not accompanied the shift(s) of R. pomonella to apples.

Does the pattern of geographic variation for R. pomonella point to
possible factors differentiating the host races? Theoretical models of
sympatric speciation (Maynard-Smith, 1966; Bush, 1969a, 1975a; Diehl
and Bush, 1989) stress that habitat specific mating is a key trait for the
formation of new host races. In Chapter 2, I showed that adult flies do not
randomly migrate between different species of host plants but, instead, tend
to attack the same species of host plant that they previously infested as
larvae (larval conditioning has not been found for R. pomonella, although
adult conditioning has been documented; Prokopy et al., 1982). Because
courtship and mating in R. pomonella occur almost exclusively on or near
the fruits of the host plant (Prokopy et al., 1971, 1972), these differences in
host preference produce genetically based, pre-mating barriers to gene
flow. There is, consequently, little doubt that differential host utilization by
hawthom and apple flies results in a system of positive assortative mating
which helps to maintain genetic variation between the host races.

Two observations make it unlikely that any of the six loci displaying
inter-host frequency differences are directly involved in the host
recognition process, however. First, five of these six loci also show
latitudinal allele frequency clines within the host races. This is not what
would be expected for genes encoding host preference behaviors, traits
which should be relatively invariant across the range of a host race.

Second, the allozymes we resolved in this study catalyze steps in general
metabolic pathways. Me and Acon, for instance, are part of the TCA cycle
while Aat converts glutamate into oxaloacetate which is used to replenish
TCA intermediates. Mpi facilitates the entry of mannose into glycolysis and

34

Had and Dia are both involved in fatty acid biochemistry. Had transforms
acetoacetate to D-beta-hydroxybutyrate during ketosis while Dia is a
member of a non-phosphorylating electron-transport chain which is active
during the desaturation of fatty acids. It is improbable that general
metabolic enzymes such as Me, Acon, Mpi, Dia, Aat and Had directly affect
behaviors associated with proper host recognition.

There is no a priori reason to presume, however, that any of the
allozymes showing inter-host variation are directly under selection,
however. Perhaps a more reasonable initial assumption is that these loci are
linked to genes being differentially selected for in the host races and that
selection is strong enough to produce hitchhiking effects for linked genes.
Inversions could play a contributing role in the co-evolution of such linked
arrays of genes by reducing recombination among loci sufficiently for
selection to cause genetic hitchhiking. Unforttmately, the poor quality of
polytene chromosome preparations in R. pomonella prevent us from
directly testing the inversion hypothesis.

Regardless of the form or nature of selection, its effects do not
appear to be uniform across the range of either apple or hawthorn flies.
Except for Act-2, allele frequencies for all loci showing host related
differences also co-vary significantly with latitude. This result suggests that
the environmental factor(s) responsible for selection is in some way
correlated with latitude. Furthermore, because allele frequency clines are
steeper among hawthom than apple populations, apple flies must either be
buffered from or constrained by environmental conditions to a greater
extent than hawthorn flies. As a consequence, geographic variation is

reduced among apple populations compared to hawthorn populations.

85

Natural selection is not, of course, the only possible factor
influencing the pattern of geographic variation in the host races. A genetic
bottleneck during the founding of the apple race could also help explain
why apple flies are less genetically variable than hawthorn flies. However,
two points argue that natural selection is more important than stochastic
processes in shaping the overall genetic structure of R. pomonella
populations.

First, not all polymorphic loci show reduced amounts of geographic
variation among apple compared to hawthorn populations. The overall
FST value for Idh, Ak, Act-1, Pgm, Pgi, Pep-2 and Acy across the eastem
United States is actually higher for apple (0.0061 i 0.0042; Table 17) than
hawthorn flies (0.0036 i 0.0019). It is, therefore, only those loci which
show host associated differentiation which are less variable in the apple
race. Allele frequencies for the six loci displaying inter-host variation are
not homogeneous among apple populations, however (Table 14). In fact,
Me, Acon-Z, Mpi, Aat-Z, Did-2 and Had are among the most geographically
heterogeneous loci for the apple race (Table 17). Stochastic processes such
as genetic drift and/or founder events are expected to have roughly
equivalent effects on neutral genes. The diversity of geographic patterns
observed among loci for apple and hawthom flies is, therefore, not
predicted by neutral theory and is strongly suggestive of selection.

Second, allele frequencies for Acon-Z indicate that at least some gene
flow may occur between the host races. Acon75 is a relatively rare allele
usually found at frequencies of 0.05 or less in fly populations (Table 22).
However, at site 6 (E. Lansing, Michigan) and to a certain extent at site 5
(Grant, Michigan), Acon75 frequencies are elevated in both hawthom and
apple populations. Even low levels of gene flow between local apple and

86

hawthorn populations would homogenize gene frequencies for neutral
alleles. This may account for the limited amount of geographic variation
observed for polymorphic loci not showing host related differentiation and
for why apple and hawthorn populations share even the rarest alleles in
common (Tables 21 -32). However, inter-host gene flow does not explain
the existence of or differences in latitudinal frequency clines observed for
apple and hawthorn flies. If the frequency clines in the apple race were
only the result of a genetic bottleneck followed by local gene flow from
hawthorn populations, then allele frequencies for apple populations should
converge with those of the hawthorn race over time. However, inter-host
differentiation was consistent across years at sites where multi-year
collections were made (Table 13). This result argues that differential
selection is strong enough to offset local levels of inter-host migration and
that selection is largely responsible for both the existence of, and
differences in, allele frequency clines for apple and hawthorn flies. But
what environmental factor is responsible for selection in R. pomonella?
Several lines of evidence suggest that ambient temperature is an
important environmental factor differentially affecting apple and hawthorn
flies. Because R. pomonella is univoltine and life expectancy of adults is
only 3.6 weeks in the field (J. L. Feder, pers. obs.; see also review by Boller
and Prokopy, 197 6), flies must eclose as adults at times closely matching the
fruiting phenologies of their host plants. Fruits of domestic apple varieties
favored by R. pomonella generally ripen and abscise 3-4 weeks earlier than
those of hawthoms. "Temporal windows" for infesting apples and
hawthoms are, therefore, different. Ambient temperature is, of course,
correlated with latitude and has a direct effect on the timing of fruit

maturation. Temperature has also been shown to be a key determinant in

87
initiating, maintaining and terminating diapause for R. pomonella
(Prokopy, 1968; Reissig et al., 1979). Apple and hawthorn flies could,
therefore, be differentially adapted to temperature as an indicator of the
phenologies of their respective host plants. The distribution of adults and
larvae does, in fact, follow the fruiting phenologies of apples and
hawthoms. Early in the season flies are abundant on ripe apples but scarce
on immature hawthoms while the reverse is true later in the season when
hawthoms ripen. Apple and hawthorn flies also show genetically based
differences in the timing of adult eclosion with apple flies emerging
significantly earlier than hawthorn flies under controlled laboratory
conditions and in the field (Diehl, 1984; Smith, 1986, 1988; Feder et al.,
manuscript in preparation).

The temperature/development hypothesis also provides a possible
explanation for latitudinal variation in R. pomonella. If larvae develop too
early in the summer (a potential problem for populations in southern
latitudes), then at least part of the population will not diapause. Instead,
these flies will emerge as a second generation of "non-diapausing" adults in
the fall when host fruit is no longer available. Almost all larvae do, in fact,
fail to diapause when held at a constant temperature above 28° C and a
lightzdark cycle of 17:7 hr. (Prokopy, 1968) and second generations of flies
have been reported in field studies (Caesar and Ross, 1919; Porter, 1928;
Phipps and Dirks, 1933b). Alternatively, if larvae deve10p too slowly then
they are exposed to predators and parasites for a longer period of time and
run the risk of freezing to death during the first frost. Freezing is a major
problem for flies in the north as we have commonly observed frozen larvae
in fruit toward the end of the field season. Clinal variation can, therefore,

be due to the premium placed on rapid larval development in northern

88

populations compared to slower development and/or constitutive diapause
in the south. Latitudinal variation for developmental traits have, in fact,
been reported for a number of other insect species (Istock and Weisberg,
1987; Scriber and Hainze, 1987). Egg transplant experiments performed in
the laboratory under controlled nutritional and environmental conditions
would provide a direct test of whether southern R. pomonella have slower
developmental rates or are more prone to diapause than northern flies.

A puzzling observation is still unaccounted for by the
temperature/development hypothesis. Why do hawthorn infesting flies
show greater north-south genetic variation than apple flies? One possibility
is that some important physiological, chemical or phenological
characteristic varies more with latitude for hawthorn than apple trees. For
example, hawthorn trees may show greater north-south diversity in the
period between or timing of fruit maturation and abscission than apples.
Apples are, after all, a commercially cultivated fruit crop introduced from
Europe. Consequently, apples should be less genetically and phenotypically
variable across their range compared to native hawthorn species. The
relative nutritional value of hawthoms and apples may also be an important
consideration. Evidence indicates that hawthoms provide a better host-
fruit environment for R. pomonella larvae than apples, as larvae of both
apple and hawthorn origin exhibited greater survivorship in hawthorn
fruits in egg transplant experiments (Prokopy et al., 1988). These results
are not surprising since hawthoms are the native host of R. pomonella.
Larval development may, therefore, be inherently faster in hawthorn
fruits. As a result, the failure to diapause could be as much, or more, of a
problem for hawthorn flies than apple flies at southem latitudes despite the
later maturation time for hawthorn fruits. Reciprocal larval transplant

89

experiments performed in the field between different hosts using apple and
hawthorn origin flies from northern and southern latitudes would help
resolve why allele frequency clines are steeper in the hawthom than in the
apple race.

The validity of the temperature/development hypothesis hinges, of
course, on a relationship between developmental rates and the allozyme loci
showing latitudinal variation. Preliminary data indicate that such a
relationship does exist as allele frequencies for Me, Acorn-2 and Had
correlate with the timing of adult emergence for hawthorn flies (Feder et
a1 ., manuscript in preparation). Further classical and quantitative genetic
experiments are needed, however, to clarify the exact relationship Me,
Acorn-2 and Had (or linked genes) have with development in R. pomonella.

Secondary contact could also account for allele frequency Clines in R.
pomonella. Northern hawthorn populations may have become isolated and
genetically differentiated from southern populations in refugia during
pleistocene glaciations. However, the secondary contact hypothesis has no
supporting data and it is much more likely that R. pomonella and their
Crataegus hosts recolonized northern latitudes from the south following the
retreat of ice sheets 10,000-20,000 years ago. In addition, secondary
contact does not account for the frequency clines in the apple race and why
these clines are more gradual than those existing for hawthom flies. As we
discussed earlier, our results strongly suggest that differential selection is
occurring both within and between the host races and is largely responsible
for the observed pattern of allele frequency Clines. Finally, a genetic
survey of hawthorn populations across latitudinal transects in the Midwest
(see Chapter 4) revealed that allele frequencies for Me and Acon-Z co-vary

with irregularities in local thermal conditions. This result agrees with the

90

temperature/development hypothesis and is not readily explained by
secondary contact.

Another question I posed in the inu'oduction is whether the
geographic pattern of genetic variation is indicative of the origin(s) of the
apple race? Unfortunately, the results of this study are ambiguous with
regard to whether the apple race formed from a single or multiple host
shifts. The reduced amount of geographic variation for apple flies is
superficially consistent with a genetic bottleneck occurring during a single
or relatively few colonizations of apple by R. pomonella. However, the
clinal pattern of geographic variation indicates that differential selection
occurs across the range of the apple fly. Selection is a complicating factor
which makes it difficult to pinpoint a "genetic origin" for the apple race. In
addition, gene frequencies for Acon-Z indicate that at least some gene flow
may be occurring on a local scale between the host races. Gene flow
between apple and hawthorn populations would further obscure any traces
of the genetic origin(s) of the apple race and could account for why even
peripheral apple infesting populations contain roughly the same number of
electromorphs as hawthorn or centrally located apple populations.

A possible analytical approach to investigating the origin(s) of the
apple race would be to construct "phylogenetic networks" among apple and
hawthorn populations based upon pairwise genetic distance measures
between populations (for a review of this subject see Felsenstein, 1988). If
the apple race formed from a single host shift then apple populations should
cluster together on the network. Alternatively, if the multiple shift
hypothesis is true then local apple and hawthom populations should
generally be closer together on the network than geographically distant
populations of either race. Unfortunately, evolutionary models forming

91

the underlying theoretical basis for network building algorithms are based
on the selective neutrality of genetic characters (Felsenstein, 1988), an
assumption which is clearly false for the host races. Furthermore, several
loci analyzed in this study were shown to be tightly linked (Chapter 1) and
co-evolving (Chapter 2; Tables 19, 20), which violates the requirement for
the independence of characters used to derive distance estimates
(Felsenstein, 1988). One possible solution to the above problems would be
to use only loci such as Pgm, Pgi, Ak, Idh, Act-2 and Acy which do not
appear to be under differential directional selection in R. pomonella
(although, of course, this does not rule out the possibility of stabilizing
selection) and are in linkage equilibrium with one another. However, allele
frequencies for these six loci are phylogenetically uninfonnative in the
sense that they display very little variation both between and among apple
and hawthom populations (Tables 13, 17), presumably due to the
homogenizing effects of gene flow among populations. Genetic data
therefore do not exclusively support or reject either the single or multiple
shift hypotheses. In retrospect, it may be impossible to resolve this issue for
the apple race probably may very well have originated and spread from the
Hudson River Valley, New York, but, in the process of expanding its range
into new areas, became subjected to repeated episodes of gene flow from
local hawthorn populations.

A critical issue not fully explored in the introduction is whether apple
and hawthorn populations represent incipient species. The results of this
study demonstrate that genetically differentiated host races can arise in
sympatry. In addition, the R. pomonella group contains a number of
sympatrically distributed sibling species specialized on different hosts,

many of which are inter-fertile in laboratory crosses, yet remain distinct in

92

nature (Bush, 1966; Berlocher and Bush, 1982; Feder et al., in press; Feder
and Bush, submitted). All of these observations are consistent with host
shifts representing the initial step in the speciation process. However,
significant geographic variation exists within both hawthorn and apple
races. The pattern of geographic variation suggests that host races are
adapted to local environmental conditions and do not evolve uniformly as
single panmictic units across their ranges. Whether or not divergent
evolutionary processes occur within host races is not central to the question
of sympatric speciation, however. The important consideration is whether
host races are distinct evolutionary entities. That is, does the shift of R.
pomonella to apples ensure that apple and hawthorn races are on separate
evolutionary pathways. If the processes occurring within one race have
little or no effect on evolution in other races, then host shifts will generate
potential species.

To determine whether host races are distinct evolutionary units
requires characterizing levels of effective gene flow between hawthorn and
apple populations. Although not an easy task, mark-recapture and larval
transplant experiments are possible. A series of mark-recapture studies at
sympatric sites along a latitudinal transect would determine whether host
fidelity is constant for apple and hawthorn flies throughout their respective
ranges. Larval transplant experiments at the same sympatric sites as the
mark-recapture studies would establish survivorship levels for the
offspring of inter-host migrants. Taken together, the results would provide
a quantitative estimate of effective gene flow between apple and hawthorn
races.

Site 8 (Kenton, Ohio) documents the need for additional studies

quantifying inter-host gene flow. No genetic differentiation was found

93

between hawthorn and apple flies at this site. Migration of adults between
apple and hawthorn trees may, therefore, be sufficient to counter balance
the effects of host associated selection. If true, then apple and hawthorn
flies could represent a mosaic of both evolutionary distinct host races and
freely inter-breeding populations across eastern North America.
Alternatively, site 8 might be located in an area where selection pressures
on the allozymes are similar for apple and hawthorn infesting flies. Host
recognition differences may, therefore, still exist between hawthorn and
apple flies at the site and inter-host migration be minimal. However,
because none of the allozymes resolved in this study are likely to be
involved in the host recognition process, host related polymorphism would
not be detected. Further studies measuring inter-host gene flow are,
therefore, critical for evaluating the fidelity of host utilization and the
evolutionary significance of genetic differentiation between hawthorn and
apple flies. .

Even in the absence of direct experiments on gene flow, the allozyme
results, plus recently obtained behavioral and emergence data, indicate that
apple and hawthorn flies are partially reproductively isolated across a
majority of their distributions. Prokopy et a1. (1988) have demonstrated
that differences exist in the host acceptance behaviors of apple and
hawthorn flies. Apple flies also terminate diapause and emerge as adults
significantly earlier than hawthorn flies under controlled conditions and in
the field (Diehl, 1984; Smith, 1986, 1988; Feder et al., manuscript in
preparation). I have also been able to qualitatively document that adult flies
do not randomly migrate between apple and hawthorn trees but, instead,
tend to attack the same species of host plant they infested as larvae (Chapter
2). Finally, the results of the current electrophoretic study imply that

94

selection is acting differentially on flies infesting hawthoms and apples. All
available evidence therefore indicates that hawthorn and apple flies are
diverging sympatrically on separate evolutionary pathways.

95

 

 

 

 

 

0208805 053503-503: 59¢ 0003550 0000:2535 .50 53.0 ~86 W N 000 8.0 W N 00 00.0 W N 0
555.0 20.0 00 < 505.0 550.0 500.0 00 3 555.0 505.0 00 < 52.0 555.0 00 m 05 0
055.0 505.0 20.0 55 < 550.0 555.0 00 = 25.0 055.0 55 < 0005.0 005.0 00 I 55 5
055.0 050.0 05 0. 555.0 505.0 2. m 55.0 550.0 00 < 05.0 250.0 50 m 05 5
505.0 550.0 55 < 555.0 500.0 05 m .500 055.0 55 < 055.0 055.0 05 m 05 5
25.0 550.0 55 < 552.0 25.0 55 = 005.0 050.0 55 < 52.0 555.0 55 m 05 5
505.0 555.0 _5 < 03.0 055.0 05 = 050.0 055.0 05 < 55.0 03.0 05 m 05 0—
055.0 055.0 05 < 005.0 0005.0 05 : 3.5.0 050.0 50 < 555.0 20.0 05 m 55 0.
25.0 550.0 500.0 552 < 055.0 500.0 000.0 53 = 055.0 3.0.0 502 < 23.0 055.0 3.— : 55 0
255.0 000.0 05— < 555.0 550.0 000.0 02 : 005.0 005.0 05. < 055.0 000.0 552 m 05 0
555.0 505.0 55 < 52.0 005.0 500.0 50. 2 0505.0 550.0 :— < 53.0 555.0 0: m 55 0
505. 505.0 550 < 02.0 055.0 000 a 055.0 050.0 005 < 550.0 505.0 500 I 55 5
02.0 005.0 85 < 02.0 055.0 505 x 055.0 250.0 505 < 055.0 23.0 005 m 05 5
025.0 055.0 000.0 3.— < 52.0 055.0 02 = 550.0 555.0 55— < 55.0 055.0 05. I 55 5
02.0 255.0 500.0 55— < 02.0 0055.0 500.0 .02 = 055.0 250.0 552 < 555.0 555.0 5— : 05 5
5550 055.0 000.0 552 < 52.0 505.0 500.0 552 = 53.0 555.0 552 < 0000.0 005.0 052 I 55 5
005 0 005.0 0 < 002.0 85.0 50 2.. 550.0 505.0 0 < 050.0 055.0 50 m 05 5
505.0 555.0 500.0 00— < 055.0 005.0 30.0 002 x 550.0 505.0 02 < 30.0 005.0 0: m 05 _
02.0 055.0 05 < 02.0 055.0 _5 = 200.0 055.0 55 < 005.0 02.0 _5 m 55 _
055.0 055.0 55 < 03.0 055.0 55 2 000.0 05.0 55 < 500.0 505.0 55 = 05 2
505.0 505.0 05 < t .- t .. 2 500.0 505.0 05 < ... t t I 55 2
55. 8. 5 2 an «a. as a z Jaunl anloullzlaar 9: 3.3.8255» “=5
5: a:
.3050 05 5385 0205
5253 05 $0.55 0:39:05 .55 .5 5.5030000 20 0.005 05 0.0 0.5532 5.. 55: 0:5 055 .50 50500002..— o_o=< .25 035.5

 

 

 

96

 

 

 

 

 

 

 

.8383 93503.53: 22. 8:038 .5825.» a. 3-0 .8... w 5 ...: 5.0 w m .. 8.0 w u .

20.0 55.0 550.0 2.0 00.0 000.0 500.0 00 < 550.0 500.0 500.0 500.0 3.0 20.0 20.0 00 m 05 0
0.0.0 3.0 505.0 500.0 0.0.0 50 .4. 050.0 050.0 005.0 050.0 050.0 .5 m 55 5

00.0 0.0.0 05.0 .500 50 < 03.0 .00.0 05.0 .500 50 I 05 0
050.0 .2.0 050.0 .0..0 550.0 050.0 5.0.0 05 < 000.0 03.0 000.0 05.0 .000 50 ... 05 5
5.0.0 03.0 005.0 00.0 550.0 000.0 50 < 20.0 05.0 500.0 50.0 050.0 20.0 55 m 00 5
0.0.0 030 505.0 550.0 550.0 0.0.0 55 < 500.0 00.0 5.0.0 555.0 050.0 00 .0 00 0.
050.0 050.0 000.0 555.0 02.0 00 .0. 500.0 555.0 .00.0 050.0 000.0 05 = 50 0.
050.0 03.0 000.0 5.5.0 050.0 50 ..0 500.0 .0. < 000.0 500.0 0.5.0 500.0 500.0 05.0 05. .0 50 0
5.0.0 05.0 050.0 00.0 000.0 50.0 05. < 000.0 .5..0 500.0 505.0 .000 00 ..0 00. ... 05 0
0.0.0 005.0 550.0 05.0 .000 50 < 000.0 05.0 .550 005.0 03.0 000.0 55 5. 05 0
000.0 53.0 000.0 00.0 050.0 000.0 500 < 000.0 050.0 005.0 050.0 500.0 050.0 050 m 55 0
500.0 00.0 .550 50.0 000.0 55.0 500.0 005 < 000.0 550.0 055.0 500.0 .500 500.0 505 I 00 0
000.0 05.0 39000.0 505.0 000.0 53.0 000.0 55. < 000.0 050.0 555.0 000.0 500.0 500.0 500.0 00. ... 05 5
5.0.0 05.0 055.0 00.0 000.0 000.0 55. < 000.0 3.0 505.0 000.0 050.0 500.0 .0. m 05 5
000.0 50.0 050.0 500.0 500.0 000.0 50. < 500.0 005.0 .000 050.0 000.0 05. 5. 50 5

50.0 000.0 055.0 000.0 050.0 0 < 30.0 500.0 23005.0 3.5.0 ....0 000.0 000.0 50 .0 05 5
000.0 52.0 35.0 03.0 050.0 30.0 000.0 00. < 000.0 000.0 005.0 005.0 55.0 000.0 ... .0 00 .

550.0 .550 500.0 0.0.0 0.0.0 00 < 000.0 005.0 000.0 050.0 0.0.0 .5 I 55 .
0.0.0 53.0 050.0 005.0 050.0 .5 < 0.0.0 550.0 000.0 .500 500.0 55 5. 00 .

550.0 3555.0 000.0 000.0 00 < I a .. I z I ... ... ... 05 ..
j 8. 50 J05l 55 55 z g.....5...._ .005 85 50 an 05 55 2 31.3% «55.

.8550 05 5830 0305
6380 2.. 388 3.26qu .55 ..o 80022.00 30 0.0.... 05 Q0 c.8535. 8.. 5.080. .80 8.8302". o.o=< .55 0.05...

97

.8058? 93503-53: as. 835 .5355.” a. 02-0 .80 mm .... 8.0 w m .. 3.0 w u .

 

 

550.0 0000 550.0 50 < 050.0 500.0 5.0.0 5.0.0 05 m. 00 0
000.0 000.0 050.0 000.0 05 < 550.0 050.0 550.0 00 ... 50 5
0500 .500 050.0 050.0 00 < .500 550.0 500.0 5500 50 .0 05 0
0000 0500 000.0 550.0 05 0. 000.0 000.0 000.0 500.0 50 m 00 5
0.0.0 5000 0500 0.0.0 .0 < 5000 0.00 0.0.0 0.0.0 55 ... 00 5
0.0.0 5000 0.0.0 000.0 000.0 .5 < 30.0 050.0 050.0 5000 500.0 00 .. 05 0.
0000 550.0 550.0 000.0 0.0.0 50 < 000.0 5.0.0 550.0 05.0 0.0.0 00 .0 50 0.
0500 5.00 550.0 550.0 0.0.0 05. < 0500 505.0 5 ...0 500.0 000.0 00. .. 50 0
0500 500.0 000.0 5.0.0 05. < 5.0.0 005.0 .5.0 500.0 050.0 5000 05. .0 00 0
0.0.0 555.0 000.0 0.00 000.0 0.0.0 55 < 0500 0000 500.0 050.0 5.0.0 5.0.0 50 .. 50 0
5000 000.0 0000 0.0.0 000 < 580 0:200 5000 0000 550.0 5.0 0. 50 5
..00 .000 0000 0.0.0 505 < 500.0 .0000 550.0 00.0 0500 505 .. 00 5
500.0 050.0 000.0 .500 000.0 03 < 050.0 505.0 500.0 05.0 000.0 500.0 05. .. 50 5
5000 000.0 000.0 .500 500.0 50. < 000.0 005.0 0500 05.0 500.0 .0. ... 00 5
000.0 555.0 000.0 .000 0000 55. < .000 000.0 500.0 00.0 05. .. 50 5
000.0 050.0 0 < 050.0 .550 050.0 05.0 50 m. 00 5
0.0.0 0500 050.0 000.0 050.0 00 < 000.0 505.0 050.0 00.0 050.0 00. .. 00 .
050.0 0000 000.0 000.0 0.0.0 05 < 050.0 005.0 000.0 00.0 .5 .. 50 .
0000 000.0 0000 0000 55 .0. 0.0.0 005.0 000.0 55.0 0.0.0 55 .. 00 .
000.0 .550 0.0.0 0500 050.0 05 < .. .. .. .. .. .. .. .. 50 ..
an. a. a5 55 .0 0. z .50... 55. 00. 05 55 .5 0W5. 3

 

0:

.8050 000 8.80
02.0: 5303 o... 8080 0000900.. .05 ..0 30.3.0000 .3 0.005 .05 A... 80530.. 00.. .000 00.. 8.000003. o.o..< .05 0.00...

 

 

 

 

805805 5.30.03-50.00 .50 80.0.8 .5000... a. 3.0 .000 mm ... .00 w 5 : 50.0 m u . _
005.0 005.0 0... .. 5.0.0 550.0 555.0 005.0 5.5.0 5....0 5.0.0 00 m 00 0
5.0.0 0.5.0 00 .. 550.0 555.0 55.0 550.0 85.0 85.0 0.. = 50 5
.- .. .- .. 550.0 005.0 50.0 00.0 05 m 0.. 0
005.0 005.0 55 = 30.0 550.0 5.5.0 000.0 02.0 00.0 0.. m 0.. 5
500.0 000.0 .5 .. 0.0.0 0.0.0 0.5.0 005.0 500.0 0.0.0 .80 550.0 55 .. x. 5
5.5.0 0.5.0 05 0 5.5.0 000.0 550.0 50.0 00.0 0.. .. 05 0.
000.0 00.0 0.. = 000.0 005.0 505.0 05.0 50.0 .5 ... 50 0.
055.0 055.0 05. m 58.0 000.0 500.0 055.0 500.0 55.0 500.0 550.0 00.0 000.0 05. a 55 0
000.0 :50 55. = 550.0 500.0 .5.5.0 5.0.0 3.0 08.0 05.0 05. = 00 0
550.0 55.0 5.. .. .500 500.0 005.0 53.0 500.0 05 m 55 0
8 005.0 005.0 500 .. 000.0 0.0.0 .000 .000 .550 55.0 500.0 .000 .000 0.0 .. 5.. ..-
9 5050 005.0 005 0. 500.0 550.0 550.0 005.0 500.0 05.0 050.0 500.0 555 .. 00 5
500.0 52.0 00. m 58.0 050.0 005.0 500.0 005.0 500.0 3.0 0.0.0 05.0.0 3.. 0. 5.. 5
550.0 2.0 00. a 550.0 500.0 505.0 58.0 050.0 500.0 00.0 500.0 .0. m 3. 0
000.0 5.5.0 5.... .. 500.0 500.0 500.0 050.0 555.0 58.0 000.0 80.0 500.0 05. = 55 5
005.0 0.5.0 .0 .. 000.0 .550 555.0 500.0 .....0 500.0 .....0 50 m 00 5
505.0 0.5.0 8. a 500.0 550.0 0.5.0 3.5.0 .30 500.0 550.0 58.0 50 .. 00 .
505.0 555.0 .5 a 0.0.0 050.0 .30 0.5.0 00.... 0.0.0 05.0 .5 m 50 .
055.0 0.5.0 55 .. 0.0.0 .500 005.0 55.0 550.0 08.0 55 .. 3 .
- .- - 0 .- .. .. .- .. .- .. .- -- .- .. .- - m 5.. ..
Iglluuluzlafll j. 5: 00. 0. 55 05 05 0? 55 .5 .0 z 3
I5...- .0 5.00
0.00000 .0.... 8.0.0

00...... E0003 0... 000.80 0030.005. .5 ..0 2.000.002. A... 60530.. 00.. 5.0.0. .0.... 5.000. .0.. 8.000002... 0.0.3.

 

.05 0.00...

99

 

 

 

 

 

 

800.805 5.30.03.50.00 0.00 0000.50 .§0.&.. .05 02.0 .000 mm .... .00 mm ... 50.0 mm .
000.0 0000 00 < 5.0.0 500.0 5.5.0 555.0 555.0 50.0 00 < 05 ..
050.0 .050 ... < 03.0 055.0 0.0.0 005.0 0.0.0 055.0 000.0 05 < 55 5
.. .. .. 5. 050.0 005.0 05.0 00.0 55 < 05 5
.0000 .550 ..5 < 550.0 0.00 055.0 005.0 05.0 58.0 00.0 .5 < 05 5
.550 0550 55 < 0.0.0 050.0 0.0.0 505.0 050.0 555.0 050.0 5:0 0.0.0 050.0 50.0 .5 < ..5 5
5500 5.5.0 0.. < 050.0 ..050 ...50 0.0.0 505.0 550.0 0.0.0 .5 < 05 0.
0.50 055.0 3 < 005.0 005.0 0.0.0 55.0 050.0 0.0.0 ..5.0 0.. < 55 0.
.500 005.0 0... < 500.0 0.0.0 0.0.0 505.0 500.0 .550 550.0 55.0 500.0 550.0 55.0 500.0 05. < 55 0
550 0 5550 05. < 580 550.0 3505.0 0000 5.5.0 550.0 55.0 500.0 500.0 005.0 05. < 05 0
0.0.0 005.0 55 < 5.0.0 055.0 555.0 055.0 05.0 55 < 55 0
000.0 0.5.0 550 < 580 050.0 .....50 :5550 .000 .05..0 .000 500.0 05.0 500.0 500 < 55 5
050.0 ...50 005 < 500.0 550.0 .050. .5050 500.0 00.0 500.0 500.0 05.0 5000 505 < 05 5
055.0 005.0 55. < 0.0.0 550.0 .000 055.0 5.0.0 505.0 0.0.0 5....0 0000 5.0.0 00.0 .000 55. < 55 5
555.0 555.0 55. < 0.0.0 550.0 005.0 500.0 555.0 500.0 55.0 500.0 500.0 55.0 500.0 55. < ..5 5
050.0 055.0 55. < 500.0 050.0 055.0 055.0 .000 55.0 500.0 05.0 000.0 55. < 55 5
555.0 555.0 0 < 050.0 850 50.0 555.0 050.0 0 < 05 5
.500 055.0 50. < 500.0 0.0.0 555.0 555.0 50.0 0.0.0 0.0.0 .0. < 05 .
005.0 005.0 05 < 0.0.0 050.0 005.0 055.0 0:0 050.0 0.0.0 00.0 050.0 05 < 55 .
.550 005.0 55 < 0.00 5550 055.0 50.0 0.00 55.0 55 < ..5 .
.550 005.0 05 0. 5.0.0 350 550.0 505.0 5.0.0 8.0 05 < 55 ..
jllzlfiul 05. 55. 5.. 00. 105‘ 55 05 05 A? 55 .5 .0 23
«.0.: dua<
000050 005

50.8w 000...... 50.000 0... 500.00 3.0.6.000 .5. ..0 80.00.0000 .5 0.000 .0.. ~65 000 N03. .50.. 50.00.0002. 0.2.0. .3 0.00...

100

.8080005 0300.03.30... 0.0.0 000030.. .0850? .0.. 500.6 .8... WM :5

 

 

 

 

80.55.: 8.03.. _

05.0 55.0 500.0 500.0 00 < 55.0 505.0 5050 5.0.0 00 = 05 ..

02.0 555.0 .500 5.0.0 5.. < 50.0 05.0 .0500 05 .. 55 5

005.0 500.0 505.0 0.0.0 .5 5. 505.0 00.0 505.0 55 5. 05 5

.. .. .. .. .. .. < -- .. .. .. .. -- .. = 05 5

55.0 005.0 0.5.0 500.0 0.0.0 55 < .550 005.0 0.5.0 0.0.0 55 .. ..5 5

000.0 005.0 0000 050.0 05 < 00.0 50.0 550.0 5000 55 .. 05 0.

055.0 :0550 055.0 500.0 00 < 05.0 555.0 5550 0.0.0 ..5 5. 55 0.

05.0 55.50 055.0 050.0 500.0 .5. < 500.0 ....0 005.0 .0550 050.0 ..00 .5. 0. 55 0

005.0 555.0 :0.50 050.0 08.0 05. < 055.0 50.0 505.0 050.0 0.. a 05 0

505.0 ..50 .80 0.0.0 55 < 555.0 05.0 550.0 5.. 0. 55 0

.. .. .. .. .. .. .. < .- .. .. .. .. .. .. = 55 5
.. .. .. .. .. .. .. < .. .. .. .. .- .. .. .. 05 5
0050 555.0 050.0 550.0 000.0 05. < 5:0 ..550 555.0 .000 08.0 00. = 55 5

055.0 05.0 050.0 .500 55. < .. ...5050 505.0 5000 050.0 00 .. ..5 5

30.0 555.0 505.0 555.0 050.0 55. 5. 555.0 055.0 5500 550.0 000.0 55. 5. 55 5
555.0 550.0 5 < 555.0 555.0 050.0 000.0 500.0 50 z ..5 5

.. .. .. .. .. .. .. < .. .. .. .. .. .. .. m 05 .
0.00 05.0 055.0 050.0 0000 05 < 050.0 .0.50 505.0 .500 03.0 .5 .. 05 .
05.0 555.0 :0550 050.0 0.0.0 .5 < 55.0 505.0 0.5.0 000.0 .5 z 05 .

05. .0 550.0 005.0 05 < .. .. .. .. .. .. .. m 55 . .

.05. 0.. 00. 00. J0 3.. 05. 0.. 00. 00. 00 «0'33.

5400.0
00050 .0.... 50.06 000...:

E9000 0... 550.80 3.0.8.000 .m ..0 500003000 30 0.000 000 a... 800.30.. 50.. «.00.. .0.. 00.000000... 0.0..<

.3 0.00...

101

0000805 0000.03.50.50 0.2. 800.30 .5005... .0. 02-0 .80 w u ... .00 w u .. 50.0 m u _.

 

 

 

366 000636 @666 66.6 66 < ...66 005.06 «666 026 66 .. on v
0066 “N66 c-6 .n < 0.0866 «.66 66n6 366 6v .. 5 ..
636 000N066 @666 um < 636 0QF6 62.6 .n ... on a
«36 0006.06 .N66 .N66 3. < 6.66 0036 .n66 3 I on n
686 000vg6 3N6 .n < 036 0006666 nvN6 .n I 3 n
”no.6 00.0.86 2.66 an < 686 0.0.2.6 2.6 an .. 6w 6.
6.6.6 636 626 8 < 566 000506 66.6 3. ... nu 6.
$66 000366 «6.6 686 cm. < 366 0.0.3.6 «666 3.6 0666 6n. .. nu 6
636 coouah6 3.6 686 6n. < .n66 0000666 3N6 6.66 on. ... 3 o
#86 ano6 :66 «v < 366 .36 .36 an .. 3 o
366 00062.6 «66.6 «8.6 .N. < 686 0002.66 6666 .....6 n86 an. .. 3 n
636 000366 n86 «:6 3. < 366 000.36 00.6 “666 .6. ... 3 n
636 000N86 ..66 5.6 0666 .n. < 366 0006.06 6666 6:6 «066 an. 2 ma N
366 n36 ....6 o < .86 0008.6 2.6 so ... 3 N
6n66 00066.6 6n~6 6N66 6n < 0660.6 66.6 6n .. nu .
«.66 00002.6 2.6 an < 366 000306 3.6 «n ... 3 .
n36 2.6 on < I I I I I I .. nu ..
0... 00.- a «VII?! .0 .50.. 0:. 00.. fillfiqlljlfinj. .

 

H.000.

0.00000 .000 50.0% 00......
000.500 0... 500.00 3.000.000 .~. .0 000030000 2. 0.000 .000 6... 0000.300 .0. .00.. 00.. 50.000002. 0.0..< ...N 0.00 ...

102

 

 

 

 

.8058? 930.03-00.00 52. 800.50 .5002...” a. 03-0 .80 mm ... 8.0 w u .. 8.0 w u _.
02.0 00.0.0 0.0.0 8 < 20.0 00.0.0 00 m 00 v
08.0 2.0.0 8 < 8.0 03.0 0.. = S h
08.0 000.0 000.0 0.0.0 00 < 03.0 000.0 08.0 .v a 00 0
08.0 ..0.0 000.0 ..0.0 0.. < 80.0 0.0.0 50.0 0.0.0 00 .. 8 n
00.0 08.0 08.0 08.0 0.0.0 .0 < 30.0 000.0 0.0.0 8 m 3 n
80.0 30.0 08.0 8 < 30.0 08.0 000.0 .0 .. 00 0.
3.0.0 «8.0 .000 00.0.0 00 < 9.0.0 3.0.0 000.0 3 m 3 0.
0.8.0 000.0 000.0 0.0.0 3.. < 80.0 000.0 0.0.0 000.0 0.0.0 02 m 3 0
50.0 8.0 0.00.0 000.0 00. < «00.0 000.0 «00.0 «8.0 00. m 3 0
02.0 08.0 0.0.0 0.0.0 08.0 .n < 20.0 000.0 08.0 0.0.0 0.0.0 vn .. 3 0
50.0 0.0.0 30.0 0.0.0 on. < 000.0 0.0.0 80.0 80.0 R. m 0.0 n
08.0 30.0 0.00.0 000.0 0.00.0 00. < 30.0 30.0 000.0 30.0 08.0 .0. m 3 0.
2.0.0 08.0 000.0 30.0 000.0 00. < 30.0 0.0.0 000.0 08.0 0.... m 3 N
000.0 3.0.0 0 < 08.0 000.0 000.0 . .0.0 8 m 3 a
000.0 08.0 20.0 .m < 08.0 0.0.0 ..0.0 .n m 0.0 .
00.0.0 0.0.0 80.0 8 < .80 08.0 N... m 00 .
8.0 3.0.0 00 < I I I I I I I m 3 ..
um. 6m. .... 00. 0a an 2 .00.. mm. 0w. .... 00. 3 an ... .00.. ufifi a...“
.00
0.00000 000 00.0.0
09.03 000.000 0... 000.00 000000.00 .x ..0 000000.000 .00 0.000 .000 $6 000.030.. 00.. .0... .0.. 8.0000000 0.000. .8 0.00...

103

 

 

 

 

.8858... 82583-88. as. 8.8.8 .5088... ... .82-.. .85 mm ...: .... w u .. 8... .v. m .

...... .8... 8.5 .85 ...... 8.5 .8 < ...... .8... 885 8... 8.5 .8 .. 8 v
.85 .85 ...... .85 .85 .85 8 < ..85 .85 8.... 8.... .85 885 ... .. 8 .
.85 .8... 8.... 8.5 .85 ... < 8... 8... 885 8.5 .85 .8 m 8 .
.85 8... 8.5 8... .85 ...5 8 < ...5 8.5 8... 8... .85 .85 .8 m 8 8
88... 8.... .8... .85 8 < .85 .85 8.... .85 .85 8 .. 8 8

...5 .85 88.5 .85 ...5 .85 8 < ..85 .8... .85 .85 .. 8. 8 ..
8.5 8.5 ...... 8.5 .85 8 < 8... .8... .85 8... 8... .85 8 .. 8. ..

.85 8.. 8.5 .85 .85 .85 8. < 885 8.5 .8... 88.5 88.. 88. .. 8. 8
885 2.... 8.... 8... .85 .85 .8. < .85 8.5 8.... .8... .85 88.. 8. .. 8 8
.85 .85 885 .8... .85 .85 8 < 88... .85 8.... 88... .85 8... 8 m 8. .
...... ...... E... 88.5 «85 «8.. 8. < 885 8... .85 2.... 88.. .... 88. .. 8. 8
8.5 88.. 8.... .85 ...5 8.. < ...5 .85 .8... 885 ...5 8.... 8. z 8 8
.85 88... .85 8.5 8... .85 8.. < 88.. 8.5 8.... 8... ...5 8... 8. .. 8. a
88.5 8.... 88... o < ...5 .8... .8... .85 8.... 88.. 8 .. 8 a
.85 88.5 .85 .85 .8 < .85 2.... .8... .85 .8 .. 8. .
...5 885 8.... 8.... ...5 .8... 8 < .85 ...... 8.... .85 .8... 8 .. 8 .
8.5 885 8... 88.. .8 < .. .. .. .. .. .. .. .. 8. ..
.... ... 1S. 8. 8 :j: ... 81 8. 8 3%..

5w.—
.§§U Us. .086

BED ...—283 o... 8896.. 8.8853 i no macaw—smog A3 039. ...... Q5 8053.... .5.. 5mm ...... 80.0553... 22?.

.am 038,—.

104

 

 

 

.8838... .2...3...5= 52. 88.8.. .58.... a. .85 .85 w ... ... .... v. u .. 8... w m .
.8... 8... 8:5 .85 .8 < 8.... .8... 8... .8 z 8 e
.85 8... .8... .8 < 8.... .85 8... ... = ... ..
.85 .8... 8.5 .85 9. < 8.5 .8... 8... .. .. 8 .
..........< --..........8.
8... .85 8.... .85 8 < ...... ...... ...... .85 8 .. 8 .
.85 .85 88.5 8 < .85 8.... E5 8 .. 8 ..
.85 .8... 8.5 .... < 88.. 8.... 8.5 .85 8 .. 8. ..
8.... ...... 8... .85 ... < 8... 8.... .8... 8.5 .8. .. 8. 8
.. .. .. .. .. < .. .. .. .. .. .. 8 8
.85 8.5 8.... .85 8 < .85 .85 .85 ...5 .8 = 8. .
8.5 .85 8.5 .85 ... < 8.5 .8... .85 8.... 8. z 8. 8
.. .. .. .. .. < .85 ...... .85 88 .. 8 8
88... 8.... .88... .85 .8. < 8.5 E... .85 8.... 8. .. 8. N
8.... .8... 8.... . < 885 .2... 2.... .85 8 m 8 a
.8... ...5 .8... ...5 .8 < .85 8.5 8.5 .8 .. 8. .
.8... 885 8.5 .85 .8 < .8... 885 ...... 8 .. 8 .
..........< ............8...
jflldlfiluzlfi: ... 8. 33

.84
3.50 .... 8....

355 .583 0... 880.6. 3.8853 .m .«o 8:38—38 2v 0...... a... at 85.33.... .6.. .0... 8.. 8.0532. 0.32.

.cm 03...“.

105

, .8898...» 93583.33: .5. 88...... .35.... .8. 3-0 .8. w u .... .... v. a : 8... w u .

 

n36 636 266

686
366

686
266

266
686

266

:66
686

n86

jglaJJJjqulaflldjngafi

>36
nnn6

v.36
_ 36

v36
636

636
~96

636

636
n86

$66
036

536
v86

n36

266
«n66

«36
366

686
686

n86
n86

«86

266
n 36

_ 86

266

686
«N66 266
v36

306
«2.6 686

6606
v2.6

306 266
306 «.66

«$6 686

036 «.66
«R6 366

6:6 366
u 56

«$6
636 686

2.06

8
an
we

3
an

_n
9.

an.
69

an

ha—
«2

mm—
a

3.
an

on

<<<<<<<<<<<<<<<<<

266
2.66
366

266
$66

866

366
366

266
366
«36
6.66

636
nun6
2&6

~26
n36

«36
686

9.9.6
n36

636
936
«36
«$6
636
636
636

266

— 86

366
266

$66

«36
«no.6
$66

356
#36

306

«N66 .636

366

n36
$06

$66

366
366
a 36
$66
«36
636

8
6‘
686 3.

we
an

«~66 6v
266 3.

:66 am—
866 cm—

866 an

6666 on.
:66 an

366 um—
266 No

686 3.
366 an

ZEIZZSEZZEIIZEZIZ

on
S
on

33
SSMMflhV

W
on

'5
fl

3
3

F'— NN run 0 00

—
—

 

:3

.3850 6... 8.8m
69...: 5883 o... 88.8. 3.8.6.82. .z .3 8.8.3308. 3.6 can. .8... 96 .553... .0.. :3 .6.. 88:33.. 0.0:...

. m m 038%

106

 

 

 

 

.8898... 28.58382... .2. 8...... .58.... a. .88.-.. .8.. mm ... .... mu .. 8... mm .
...... 8.... 88... .8 < :8. 8.... .... .8 .. 8. v
.8... 8..... ...... .... 8 < .... 8.... 8.... .8... ... z 8. .

8.... 8.... .8... 8.... 8. < .... .... 8... .. .. 8. .
.... 8.... 8.... 88 < .8... 88.. 88.. .... .. .. 8. 8
.8... 8.... .8.. .... .8 < .8... 88... ...... .... 8 .. 8 8
8.... .... 8.... .8.. 8 < 88... .8... ...... 8.... .8 .. 8. ..
.... 88... 8... .... 8.. < 8.... ...... 8.... ..8 m 8. ..
8.... .8... .... 88... 8.. 8. < .8.. 8.... .8... 88... ..8. 8. .. 8. 8
.... 8.... 88... .... 8... .8. < .... 8... 8.... .... 8... .8. .. 8 8

8... ...... 8.... 8 < .... .... .8.. .8 .. 8. 8
8... .... 8... 88.. 88. < 8.... 8.... .8... .... 8.... 8. .. 8. 8
8.. .... .8... 88... 8.. < .8... 88.. .8.. 8.... ... .. 8 8
8.... .8... .... 8... .... .8. < 8.... 8.... 88... 8.... 8.... 88. .. 8. .

8.... .... . < 88... .... 88... 8.... 8 .. 8 8
.... .8... .8. .... .8 < .8... .8.. .8... .... .8 z 8. .
.... 8.... ...... 8 < .... 88.. .8... 8 .. 8 .
8.... 2.... 8.... .8 < .. .. -- .. .. .. .. 8. ..
I... ... .... 3J8: ..fi ... ... at jlfinj.

....
.8850 ...... 88.88

62...: 80.88 0... 83.8.. £33528 .% ..o 8.3.338 A3 0...... a... E. 9853... 8.. 3. .0.. 80.8033 30:...

.Nm 03....

Chapter 4

Gene Frequency Clines for Rhagoletis pomonella
Host Races in the mid-Western United States

Introduction

Speciation, the rendering of one closed genetic system into two, is a
fundamental process responsible for the evolution of life on earth.
Unfortunately, speciation is a rare, time consuming event from the
perspective of the biologist. Consequently, much of our understanding of
the process is based on population genetic models and indirect analyses of
closely related species. Mathematical models of speciation tell us what is
theoretically possible but do not constitute proof that taxa diverged by a
particular mode of speciation. Post hoc studies of species differences are
also plagued by a major problem: It is extremely difficult to distinguish
differences which were actively responsible for restricting inter-
populational gene flow and initiating the divergence process from
differences which arose subsequent to speciation. We therefore have very
few examples directly documenting how populations become
reproductively isolated from one another. Consequently, questions have
been raised as to whether speciation occurs by "macro-evolutionary"

mechanisms which qualitatively differ from "micro-evolutionary"

107

108

processes responsible for gene frequency changes within populations
(Gould and Eldredge, 1977).

The apple maggot fly, Rhagoletis pomonella, has a biology and
natural history which makes it ideal for the study of speciation. R.
pomonella is part of a sibling species complex of true fruit flies whose
members are broadly sympatric in their geographic distributions across
North America (Bush, 1966). Rhagoletis larvae are obligate internal
parasites in the fruit of their hosts, with each species in the group being
monophagous or oligiophagous for a different set of host plants (Bush,
1966). Adult females lay their eggs directly into the host fruit which they
identify by specific visual, tactile and olfactory cues (Bush 1969a,b). Males
are attracted by the same cues and courtship and mating occur almost
exclusively on or near the fruits of the host plant (Prokopy et al., 1971,
1972). Because host recognition and mate selection are directly coupled in
Rhagoletis, variation for host preference and host associated survivorship
traits can act as genetically based, pre-mating barriers to gene flow. These
considerations led Bush (1966, 1969a,b, 1975) to propose that speciation in
the R. pomonella group occurs sympatrically and is initiated when flies
colonize and adapt to new host plants.

Verifying that sympatric speciation occurs in the R. pomonella group
is a two step process. First I must show that traits adapting fly populations
to different host plants are also responsible for reproductively isolating host
associated populations. Second, I must demonstrate that these "host
specific" traits can evolve within geographically continuous populations. In
this regard, the recent shift of the species R. pomonella from its native host
hawthorn (C rataegus spp.) to introduced, domestic apple (Malus pumila;
Ward, 1866; Walsh, 1867) provides a unique ecological timeframe to

109

determine whether host specialization is sufficient to differentiate
populations in the absence of geographic barriers to gene flow.
Agricultural records suggest a probable origin for the apple infesting
"race" in the Hudson Valley, New York, and document the apple fly's
subsequent spread across eastern North America over the last 120 years (see
Illingworth, 1912; 0' Kane, 1914; Bush, 1969a). Because domestic apples
were introduced into the United States by European settlers (Chapman and
Lienk, 1971), we have an upper limit of approximately 350 years for the
age of the apple infesting form of R. pomonella. Of course, apple
populations could conceivably have originally shifted from a host other
than hawthoms. However, the only other reasonable native hosts in North
America are five species of endemic crab apple (Pyrus angustzfolia, P. '
glabrata, P. ioensis, P. rivularis and P. coronaria), none of which support
populations of R. pomonella (O'Kane, 1914; Porter, 1928; Bush, 1966).
The reason for this is unclear but it has been suggested that the fruits of
native crabs are too acidic and ripen too late for R. pomonella larvae to
complete development before the onset of winter (O'Kane, 1914). It is
therefore highly likely that hawthoms were the original host for the apple
race.

In the preceding Chapters, I have shown that sympatric apple and
hawthorn populations of R. pomonella are genetically differentiated.
Consistently significant allele frequency differences exist for six allozyme
loci between sympatric populations of the two host races from across the
eastern United States (Chapter 3). These six loci map to three different
regions of the genome (Chapter 1) and linkage disequilibrium occurs in
natural populations between non-allelic genes within each of these three
regions (Chapters 2, 3). Allele frequencies for five of the six loci

1 10
displaying inter-host differentiation also co-vary significantly with latitude
among both apple and hawthorn populations (Chapter 3). This suggests that
north-south allele frequency clines exist for R. pomonella. However, the
geographic analysis of eastern North America involved widely separated
populations from a variety of different longitudes. Verification of clines
therefore requires more detailed examination of fly populations collected
along latitudinal transects.

The purpose of the current study is two-fold. First, I want to
describe the regional pattern of allozyme variation for R. pomonella in the
midwestem United States. For instance, how extensive is geographic
variation for R. pomonella in the Midwest? What levels of intra- and inter-
host variation exist on a county by county basis within a geographic region?
Do apple and hawthorn flies continue to show consistent genetic differences
when examined on a much finer geographic scale than previous studies? Do
allele frequency clines exist for R. pomonella? If so what is the nature of
these Clines? Do all loci co-vary similarly with latitude or do loci differ in
the shapes and configurations of their frequency clines? In short, the first
set of questions addresses how genetic variation is partitioned within and
between R. pomonella host races.

The second aim of this Chapter is to explore the causal basis for any
observed geographic patterns of variation. In other words, what factors are
responsible for differentiating apple and hawthorn populations? Does
natural selection act differentially between the two races? Do flies migrate
randomly between different species of host plants? Do the same processes
act unifomrly across the Midwest or do local conditions dictate the extent
and pattern of genetic differentiation? Ambient temperature has been

implicated as an important environmental factor influencing host-parasite

l l l
interactions for R. pomonella (Prokopy, 1968; Reissig et al., 1979). Is

there any indication of this in the regional pattern of genetic variation?
Finally, if gene frequency clines exist are they the result of differential
selection (perhaps related to ambient temperature) or is secondary contact a
possibility? The second set of questions therefore focuses on whether
reproductive isolation in R pomonella is the direct outcome of ecological
and behavioral adaptations resulting from host shifts.

Materials and Methods

Flies were collected from 34 different sites across the midwestem

United States from 1985 to 1987 (Figure 5, Table 33). In general, the sites
were organized into five north-south transects which ran from the states of
Wisconsin through Illinois or from Michigan through Indiana (see Figure 5
and Table 34 for a map and breakdown of sites comprising each of the five
latitudinal transects). At 28 of the sites, both hawthorn (C rataegus mollis)
, and feral apple (Malus pumila) infesting populations were sampled. Host
plants were sympatric (that is, separated by a distance of less than 100
meters) at 13 of the 28 paired locations. In only five instances were apple
and hawthorn trees situated more than 4 km apart with a maximum distance
of 8 km at site 8 near Parkerdale, Indiana. Six of the 28 paired sites (listed
as 9, 15, 19, 20, 26 and 28 in Table 33) were the same apple and hawthorn
populations that were analyzed in the earlier allozyme study of the eastern
United States (these sites were designated 7, 5, 1, 2, 3 and 4, respectively, in
Chapter 3).

Three different strategies were used to collect flies. At a majority of

sites larvae were dissected from fallen fruit in the field and immediately

Figure 5.

112

Collecting sites for R. pomonella in the midwestem United
States (see Table 33 for detailed descriptions of study sites).
Solid lines indicate isothermal clines for growing degree days
base 50°F (GDD). GDD values for Illinois and Indiana sites
are given in parentheses. Dashed lines connect sites
comprising each of the five latitudinal transects collected in
the study. 0: site where the frequency of Acon-275 was >
0.100 for either apple or hawthom flies. G = site where the
frequency of Acon-Z 75 was > 0.100 for both apple and
hawthorn flies. A = site where the frequency of Acon-289 was
> 0.200 for hawthorn flies and/or > 0.100 for apple flies.

 

 

113

2|00

 

 

 

 

20‘ (3506)

 

 

Figure 5.

114»

Table 33. Numerical designations and descriptions of field sites.
Sites were generally organized into five latidudinal
transects (see Figure 5 and Table 34). Latitude and growing

degree days base 50°F are also given for sites.

Distc
Site.Iear.anatinn Stage? agar? 1xm1.Larirl.Gnn
1 87 Rodgers Ave, Clare, MI. L(L) A,H 3.2 43°40' 2380
2 87 MI. 20, Mt. Pleasant, MI. L(L) A,H 3.0 43°35' 2660
3 87 Alger Rd., Alma, MI. L(L) A,H sym. 43°25' 2650
4 87 MI. 27, Eureka, MI. L(Ad,L) A,H sym 43°06' 2720
5 86 Jolly Rd., Okemos, MI. RtR) A,H 1.0 42°41' 2570
6 87 F Dr., Eckford, MI. L(Ad) A,H sym, 42°09' 2880
7 87 IN. 20, Lagrange, IN. (L) H - 41°38' 2986
8 87 Interstate 24, Parkdale, IN. L(L) A,H 8.0 41°11' 3376
9 87 Interstate 69, Gas City, IN. L(L) A,H 7.1 40°50' 3278
10 87 No. 45 Rd., Cadillac, MI. L(L) A,H 1.0 44°11' 2060
11 87 MI. 46, Lakeview, MI. L A - 43°25' 2550
12 87 Grand River Ave, Saranac, MI. L(Ad) A,H 4.9 42°52' 2870
13 87 Sprinkle Rd., Portage, MI. L(L) A,H 4.2 43°14' 3083
14 87 MI. 20, Woodville, MI. L(Ad) A,H sym 43°40' 2320
15 87 112th sr., Grant, MI. Ad,L(Ad,L) A,H sym. 43°21' 2550
16 87 24th sr., Harrington, MI. L(Ad) A,H 2.0 43°04' 2690
17 87 MI. 89, Fennville, MI. L(Ad,L) A,H 1.5 42°34' 2800
18 87 East Rd., Three Oaks, MI. Ad(Ad) A,H sym. 41°46' 2940
19 86 Hogan Fanm, Ephraim, WI. R(R) A,H sym, 45°07' 2060

20 85 Kuehn Farm, Carlsville, WI. R(R) A,H sym. 44°56' 2100
21 87 County MM Rd., Green Bay, WI. L(L) A,H sym. 41°28' 2310
22 87 Gass Lake Rd., Manitowoc, WI. L(Ad) A,H sym. 44°02' 2290
23 87 County EB Rd., Sheboygan, WI. L(Ad) A,H sym. 43°43' 2340
24 87 County C. Rd., Port Wash.,WI. L(Ad) A,H sym. 43°22' 2380

25 87 50th. St., N. Cape, WI. L(Ad) A,H 0.5 42°46' 2450
26 86 ILL. 132, Waukegan, ILL. R(R) A,H 3.3 42°20' 2733
27 87 River Park, Kankakee, ILL. (L) H - 41°09' 3211
28 86 Univ.ILL. Campus, Urbana, ILL. R(R) A,H sym 40°05' 3506
29 87 WI. 29, Hatley, WI. L(L) A,H 1.6 44°52' 2230
30 87 WI. 49, Berlin, WI. L(L) A,H 6.3 43°54' 2450
31 87 I 90 and WI. 12, Madison, WI. (L) H - 43°00' 2710
32 87 WI. 81, Beloit, WI. L(L) A,H 2.5 42°30' 3139
33 87 WI. 23, Greenbush, WI. (L) H - 43°46' 2490
34 87 WI. 67, Iron Ridge, WI. (L) H - 43°22' 2600

 

a Life history stage in which flies were collected. L=larvae, Ad=adult,
R-flies sampled as larvae and reared to adulthood in laboratory.
Stage designation given in parentheses is for hawthorn flies, without
parentheses for apple flies.

b Host plants from which flies were collected at a site. A=apple,
H=hawthorn.

c Distance (km) separating apple and hawthorn trees sampled at paired
sites. sympsympatric site where host trees were within 100 m.

115

frozen in liquid nitrogen (these sites are listed as L in Table 33). Only late
second and third instar larvae were electrophoretically analyzed from these
collections. At other locations adult flies were captured directly from off
of their respective host plants using a vacuum aspiration system and were
subsequently frozen in liquid nitrogen (these sites are designated Ad in
Table 33). For three hawthorn sites (4, 15, 17) and one apple site (15) both
larvae and adults were collected from host trees to allow comparisons of the
two life history stages which represent sequentially different fly
generations (i.e., parents and their immediate offspring). At sympatric site
15, adults and larvae were collected from both apple and hawthorn host
plants to determine whether inter-host genetic differentiation was consistent
across different life history stages and to qualitatively gauge the extent of
inter-host migration. Finally, fruits were collected from the ground
beneath host trees at the five sites designated R in Table 33 and the larvae
infesting these fruits were reared to adulthood in the laboratory.
Laboratory rearing included chilling pupae at 40C in a refrigerator for five
months to simulate winter diapause conditions.

Standard horizontal starch gel electrophoretic techniques were used
and are described fully in Chapter 1 and in Berlocher and Bush (1982). We
resolved and scored the six allozyme loci which previously displayed the
greatest amounts of inter-host and/or latitudinal variation in the earlier
survey of the eastern United States (these six allozymes being: Malic
enzyme [Me], Aconitase-Z [Acon-Z], Mannose phosphate isomerase [Mpi],
NADH-diaphorase-Z [Did-2] , Aspartate amino-tranjerase-Z [Ant-2] and
Hydroxyacid dehydrogenase [Had]). A Mendelian mode of inheritance has
been established for all six of these. allozymes (Chapter 1). Isozymes that
migrated the nearest to the cathode were designated system 1, the second

116

nearest system 2, etc. Alleles were numbered according to their relative
anodal mobilities with the most common allele for each locus designated
100 and used as a standard. Electromorphs resolved for each of the six loci
are given in Tables 39-43, which appear after the Discussion section on
pages 141-150. Only two alleles were scored for the loci Did-2 and Me.
Consequently, only the frequencies for the rarer of the two alleles for these
loci appear in Table 40. Three alleles were scored for Had, but one of these
alleles (Had9 7) was very rare and so for brevity only frequencies for the
second most common allele (HadI 22) are given in Table 40.
Electrophoretic alleles migrated identically in adult and larval samples.

Genotypic frequencies were examined for deviations from Hardy-
Weinberg expectation by G-tests. Alleles were pooled, as required, so that
all genotypic classes had expected numbers >1. The Levene conection
(Spiess, 197 7) was applied when sample sizes were <100 or when alleles
could not be pooled to make expected numbers >1. G-contingency tests
were used to test for allele frequency heterogeneity. Alleles were pooled,
when necessary, to ensure that each cell in the G-contingency test had an
observed number 25. F-statistics were calculated by the method of Weir
and Cockerham (1984) with standard deviations estimated by jackknifmg
over loci or populations. Corrections for unequal sample sizes in the
method of Weir and Cockerham can result in negative Fs'r values. In these
instances we report PM as 0.

Linear and curvilinear regression analyses were performed between
arcsine transformed allele frequencies and either latitude or growing
degree days (GDD) base 50°F. GDD is calculated by determining the
number of degrees that the average temperature for a day was above 50°F

and summing these daily totals throughout the course of the year (negative

117

daily values are considered to be 0). Fifty degrees Fahrenheit is not an
arbitrarily chosen temperature for calculating GDD but, rather, represents
the approximate lower threshold condition under which post-pupal
development occurs for R. pomonella (Reissig et al., 1979). Fahrenheit was
used as the unit of temperature in this study because weather stations in the
midwestem United States do not provide GDD values in centigrade base.
Unfortunately, converting GDD between the two temperature scales is not
exact without daily readings from stations, a task which imposed
insurmountable logistical and computational hurdles. GDD values from
Wisconsin were the yearly averages from 1950 to 1974, in Illinois from
1951-1980, in Michigan from 1931 to 1960, and in Indiana from 1975 to
1987. Regressions were performed separately for apple and hawthorn
populations from Michigan/Indiana and Wisconsin/Illinois transects. At
sites where both adults and larvae were collected, the allele frequencies used

in the regressions were the mean between the two life history stages.

Results

Consistently significant allele frequency differences were observed
for Me, Acon-Z, Mpi, Did-2, Ant-2, and Had between apple and hawthorn
fly populations from across the midwestem United States (Tables 34, 39-
43). Of the 28 paired apple and hawthorn sites, only site 4 (Eureka,
Michigan) and site 32 (Beloit, Wisconsin) did not show a significant
difference for at least one of the six loci analyzed. Me and Acon-Z
displayed the greatest amount of inter-host variation being significantly
different at 14 and 21 paired sites, respectively. Me and Acon-Z are, in fact,
tightly linked on chromosome II along with Mpi (Chapter 1) and significant

118

 

Table 34. G-contingency tests for allele frequency heterogeneity
between paired apple and hawthorn populations. Degrees of
freedom.for tests involving each locus are given in
parentheses following locus abbreviations. At site 15 both
adults (Ad) and larvae (L) were tested for inter-host
variation.

Site Mall). Amm- min. mm- mam.- mun.

Transect 1

1 *** *** ***

2 t t * *

3 ** ***

4

5 * **

6 t

8 'k

9 **

Transect 2 _

10 *** *** * ***
12 *** ***

13 *

Transect 3

14 *** *** *** * ***

15(Ad) *** *** *** * * ***
15(L) *** *** *** *** *** *
16 *** *** ***

17 ***

18 **
Transect 4

19 *** *** **

20 *** *** ** *** *** **
21 * **

22 * *
23 **

24 ** * **

25 ***

26 *** **

28 ** **

Transect 5

29 **

3o ** ***

32

* 2.3 0.05 ** E S 0.01 *** E S 0.001 G-contingency test

119

linkage disequilibrium has been found between non-allelic genes for Me and
Acon-Z in natural apple and hawthorn populations (Chapters 2, 3). Genetic
hitchhiking due to selection at a linked locus could therefore account for the
observed correlation in allele frequencies between Me and Acon-Z. The
same may also be true for Die-2 and Ant-2, which are separated by a map
distance of 3.2 centimorgans on chromosome I (Chapter 1). In this study, 5
of the 7 sites which showed significant frequency differences between apple
and hawthorn flies for Aat-Z were also significant for Dia-2 (Fable 34).
High levels of linkage disequilibrium have also been observed between Dia-
2 and Act-2 in natural fly populations (Chapters 2, 3).

Allele frequencies for Me, Acon-Z, Mpi, Did-2, Aat-Z and Had were
similar in intra-host comparisons between adults and larvae collected from
hawthoms at sites 4, 15, and 17 and from apples at site 15. Only Me for
hawthom flies at site 15 (Grant, Michigan) showed a significant frequency
difference between adult and larval life history stages out of a total of 24
tests (Me100 adults = 0593,11 = 345; Me100 larvae = 0.655, n = 467; G-
contingency test = 6.61, E _<_ 0.05). Larvae dissected from host fruits
represent the offspring of adults collected from off of apple and hawthorn
trees. The near constancy of gene frequencies observed between the two
generations suggests that intra-host selection is not intense between the two
life—history stages. The fact that hawthorn and apple populations were
generally in Hardy-Weinberg equilibrium lends further support for this
view (Tables 39-43). The 16 significant deviations out of 330 total tests for
apple populations and the 24 significant deviations out of 407 tests for
hawthorn populations do not differ appreciably from the number of
significant tests expected due to random type I error (16.5 and 20.3,
respectively, for apple and hawthorn populations).

120

The pattern and magnitude of inter-host differentiation between
hawthorn and apple populations was similar for both larval and adult life
history stages at site 15 (Tables 34, 3943). Gene frequencies were
significantly different between apple and hawthorn populations for both
larvae and field captured adults for Me, Acon-Z, Mpi, Die-2, Act-2, and
Had (Table 34). As mentioned before, only Me for hawthoms displayed
significant intra-host variation at site 15 between adults and larvae. Adults
collected directly from apple trees are, therefore, as genetically distinct
from hawthorn adults as apple larvae are from hawthorn larvae. This
finding indicates that adult flies do not migrate freely between apple and
hawthorn trees but instead tend to utilize the same species of host plant
which they infested as larvae. Because courtship and mating in R.
pomonella occur almost exclusively on or near the fruits of their host
plants, differences in host utilization between hawthom and apple flies at
site 15 produce pro-mating barriers to gene flow which help maintain host
associated polymorphism.

Geographic variation was apparent among both apple and hawthorn
populations across the rrridwestem United States. Substantial spatial
heterogeneity was observed across each of the five north-south transects
(Table 35) and the magnitude of geographic variation differed between the
two host races. Hawthorn flies displayed greater inter-populational
differentiation than apple flies as evidenced by the 5 x higher overall FST
value among hawthorn compared to apple populations (0.0626 and 0.0125,
respectively; Table 36). In addition, individual loci and alleles showed
variable levels of geographic variation. For instance, Me, Acon, and Had
had FST values of 0.1403, 0.0984 and 0.806, respectively, among hawthorn
populations and 0.218, 0.212 and 0.183 among apple populations. In

121

Table 35. Gwcontingency tests for allele frequency heterogeneity
among apple and hawthorn populations along five
latitudinal transects in the midwestern United States.

mwmmmmmm

Apple 1 *** t
2 *** ***
3 ** *** *** ***
4 *t* *** ** ***
5 ** ** ***
Hawthorn l *** *** ** *tw
2 *** *** ** *** * ***
3 *** *** * *** * ***
4 *** *** *** *** *** ***
5 ‘ * *** *** ***

 

* 2,5 0.05 ** 2,: 0.01 *** 2,3 0.001 G-contingency test

122

 

 

Table 36. PST values (Weir and Cockerham, 1984) across
the 28 paired apple and hawthorn populations
analyzed in the study. R refers to the PST
value calculated by pooling all alleles
remaining for a locus besides those listed
in the table.

Locus Allele Apple Hawthorn

Me 100 0.0218 1 0.0092a 0.1403 1 0.0314

Acon-Z 100 0.0214 i 0.0120 0.1033 i 0.0305
95 0.0312 3 0.0175 0.1346 1 0.0377
R 0.0132 1 0.0057 0.0380 i 0.0170

Total 0.0212 1 0.0099 0.0984 i 0.0257

Mpi 100 0.0123 i 0.0060 0.0341 i 0.0170

Dia-2 100 0.0059 1 0.0049 0.0379 1 0.0165

Aat-2 100 0.0023 1 0.0033 0.0025 1 0.0037
75 0.0050 1 0.0053 0.0126 i 0.0064
50 0.0020 1 0.0024 0.0266 1 0.0121
R 0.0031 i 0.0029 0.0061 1 0.0036

Total 0.0032 1 0.0025 0.0108 i 0.0042

Had 100 0.0183 1 0.0063 0.0806 1 0.0229

All loci 0.0125 :1; 0.0058b 0.0626 1 0.0293

 

Jackknife estimates of standard deviation calculated
over populationsa or over locib.

123

comparison, Aat-Z had a FST of 0.0108 for hawthorn flies and 0.0032 for
apple flies across the same set of populations. Furthermore, Fs'r among
hawthom populations for the alleles Acon-ZI 00 and Acon-295 (0.1033 and
0.1346, respectively; Table 36) were greater than the value for the
remaining pooled alleles at this locus (0.0380). A similar pattern existed
for allelic variants of Adz-2 among hawthorn populations (Table 36). The
observed heteogeneity of PST values suggests that differential selection is
occurring among alleles and loci, and that the effects of this selection are
more pronounced for hawthorn than apple flies.

Patterns of geographic variation were not random across the
Midwest. Linear and curvilinear regressions performed between arcsine
transformed allele frequencies and either latitude or growing degree days
(GDD base 50°F) indicated that frequency clines exist for both apple and
hawthorn flies. In only 6 of 52 cases did second order polynomials
significantly improve the fits of the regression lines compared to first order
linear equations (Tables 37, 38). Consequently, I will confine my
discussion to the results of the first order analysis. Apple flies did not
display nearly as many significant linear regressions for individual loci
with either GDD or latitude as hawthorn flies (8 compared to 22). Also, r2
values (the amount of genetic variation accounted for by the regression)
were always lower among apple than hawthorn populations (Tables 37, 38).
Apple populations therefore have more unexplained geographic variation
than hawthorn populations; a result which suggests that stochastic processes
(i.e., genetic drift or founder events) may play a more prominent role in
shaping the genetic architecture of local apple than hawthom populations.

The absolute values of the slopes of the regression lines were always

greater for hawthorn than apple populations (Tables 37, 38). This result is

124

..¢>oz¢. coaumzvo Hoouo umuwu o» conQEoo
mafia cownnoumou mo uHu vo>ouQEH haucmofimacmflm HmHEochaoa “mono Ucooom u

.ncowmmoumou HOOH
add on» cH ccmwnz How soon» we cmouncw cons mums noHocosvoum oHoHHm nmwnz n
.NroH mo nuHc: cH one ncOHnnoumou now How nuouuo cumccmun can modoam m
.unourm ¢>oz¢ ~o A Acownmoumou umocwa Hocuo
unuHm mo macaw. ucowowwuooo coHnnoumom Hoo.c W_N «Ha Ho.o w.m «« mo.o w M «

 

 

 

 

 

OHH.O H OOO.O- .OO.O OOH.O H OOO.H «lOm.O nHooH HH<

OOm.O H. OHO.O- .OO.O OO0.0 H NOO.~ .4NO.O OOH on:

OH~.O H Om0.0- HO.O OOO.O H. Hm~.O HO.O OOH «tuna

OO~.O H mOO.O- m~.O mOO.O H. OOO.H mm.O OOH mumHo

OH~.O .H HOO.O «OO.O OO0.0 H ONO.H- «HO.O OOH Her

OHO.O H OOm.O- mH.O OOO.H H -O.~ O~.O mO mucous

mam.O _H OOO.O- OH.O OOO.O H mmO.H OH.O OOH 6: .HHH\.Hz

OOH.O H mam.O- .4OO.O OOm.O H OOO.H «rOm.O nHoOH HHa

OO~.O H OmO.O- «O~.O OOO.H H mmO.H OH.O OOH our

OHm.O .H OOO.O- HO.O OOO.H H OO~.O OO.O OOH H-064

mmm.O H OOH.O- ~O.O OHH.H H «OO.O OO.O OOH NumHo

«O0.0 .H OOH.O HO.O OOm.H H. «OO.O- O0.0 OOH Ha:

OOm.O .H OOH.H- erm.O OOm.H H. OOO.O urrmm.O mO ~ueoo<

OOm.O H OOH.H- «.HO.O OOO.H H OOO.~ -.O OOH 6: .:H\.Hz
ilfll 3|? gag

mama .IlrrllluuquHuaqt
uaqHuanqquuuHaaa

.GOHnnoumou unocHH on» an cocwmamxo GOHuMHum>

owuocoo mo ucaoem I ma .nuooncmuu «Hocwaauxcanoonwz no McMHccH\cmmHnon Honuflo mcon

ocoquHsQom mamas How Anna. meow when when menace mcHzouo no ocsuHumH Hocuwo can
moHocosvouu oHoHHm coauouncmuu cHnoum cooauon nonaamcm cowmnouoou unocHH uocuo unuHm .nm magma

125

.A¢>oz<v :OHumsvo nopno umnwm ou

coanEoo ocHH cOHmnonmon uo uHu Um>onQEH aHucmOHuHcon HmHEocaHoa nocno vcooom u
.mGOHmmonmon HOOH

HHd 0:» GH OOHHQE now onocu mo Umouncw no»: one: moHocoonnu oHoHHm nmwnr n
.NIOH mo nuHc: cw one nGOHnnonuon now now nnonno Unmocmum can moQOHm m

.umourm ¢>oz< .o A Aconmonmon nmocHH noono
unnHu mo 332 23.8388 5338? HO0.0 w u 2:. H0.0 .v. a I. m0.0 w m ...

 

 

 

 

 

OhH.O H OOM.HI «HHO0.0 Hum.O H. ovO.v HHHH0.0 QHOOH HHQ

OvN.O H OOO.HI HHHH0.0 mvh.o H. th.m HHHOO.O OOH Com

OON.O H ObO.OI H0.0 hO0.0 H. MON.O H0.0 OOH Nlumd

me.O .H hOO.OI HON.O OO0.0 _H omm.MI HHO¢.O OOH NIMHQ

MON.O H. OOM.H HHHOO.O OHm.O H mv¢.m «Hom.O OOH HQ:

mmm.o .H OOO.NI «Humm.O mON.H H OOh.m HHHOm.O mm NIGOOG

OHm.O _H OOO.HI UHHOm.O mmm.H H. HOO.m usaHm.O OOH 02 .HHH\.H3

vhH.O H OON.NI HHHH0.0 mH0.0 H mcm.m «HHM0.0 QHOOH HHG

NMN.O H oom.Nl HHHO0.0 HOO.H H Omm.h HHHF0.0 OOH Cm:

NNH.O H www.cl «Hamm.o OO0.0 H OOO.H «Hum.O OOH Ntumd

bMN.O H OOM.HI HHHOO.O be.O H MON.v HHHHO.O OOH NIMHQ

HhN.O H OOO.H HHHOF.O Hb0.0 H mmv.mr UHHHN0.0 OOH HQ:

Ohm.O H. OOO.MI HHHOO.O ¢m¢.H H NvO.HH HHHNO.O mm Nicoom

mOm.O H. OOO.MI UHHMO.O th.H H Omm.OH uHHHNb.O OOH 02 .GH\.H2
lHHflJflHWJmlluflddfll _IIIMHIII duuflduuulmlluddAWl. IIIMHIII ududfld ddddfl.uuudddHH

.mmdu udddfludqlllllllll
a?

.GOHmnonoon noocHH can an vOGHMmem coquHnm>
owuocom mo unseen I Nn .nuooncmnu nHocHHHH\canoouH3 no mcMHucH\cmmHnowz nonuHo mGOHm
ncOHHMHsmoQ :nocusmc now Hana. boom anon name menace ocwzono no oosuHuMH nonuwo can

anocosvonm oHoHHm emanouncmnu cHnonm cowsumn nomaHmcm GOHnmonmon nmoCHH nmcno unnwm .mm mHQma

126

consistent with the higher levels of genetic variation found among hawthorn
populations and also helps to account for the observed pattern of inter-host
differentiation. Figure 6 graphically depicts the relationship between GDD
and the average arcsine transformed allele frequency for Me] 00, Acon-
295, Mpi37, Did-2100, Aat-2100 and Had100 for sites along
Michigan/Indiana and Wisconsin/Illinois transects (Note: Plotting allele
frequencies versus latitude gave roughly the same result as graphing allele
frequencies against GDD). Inter-host differentiation was greatest at the
colder, northem sites and was primarily due to the contributions of Me,
Acorn-2 and Mpi (Tables 34, 39-41). At centrally located sites with more
moderate temperatures, hawthorn and apple populations were less
genetically distinct, until at site 4 near Eureka, Michigan, and site 34 in
Beloit, Wisconsin, none of the six loci analyzed in the study showed a
significant inter-host frequency difference (Table 34). Hotter sites further
to the south once again began to show increased inter-host heterogeneity.
However, unlike the northern sites, Had displayed the greatest amount of
inter-host differentiation in Illinois and Indiana (Tables 34, 40). Genetic
differentiation between apple and hawthorn flies is, therefore, a function of
local thermal conditions (or latitude, as the two factors are correlated), with
inter-host divergence resulting from differences in the slopes of allele
frequency clines between the two races.

Me, Acon-Z, Mpi, Aat-Z, Die-2 and Had showed both similarities and
differences in their clinal patterns of variation across the Midwest. Because
allele frequency clines were more clearly defined for the hawthorn race,
my discussion will initially focus on hawthorn populations along transect l
(Michigan/Indiana) and transect 4 (W isconsin/ Indiana), the two most
intensively surveyed transects in the study (Note: Site 10 was considered to

Figure 6.

127

Average arcsin transformed allele frequency for M3100,
Acorn-295, Mpi-3 7, Act-2100, Did-2100 and Had] 00 for apple
and hawthorn populations along Michigan/Indiana and
Wisconsin/Illinois transects plotted against growing degree
days base 50°F. The first order linear regression equations
are given for apple and hawthorn flies along with r2 values
(the amount of genetic variation explained by the regression
equation) and the probability level (12) for the AN OVA F-test
of the null hypothesis that the slope of the regression line = 0.

 

 

 

 

 

128

Growing Degree Days versus Average Allele Frequency

 

 
  
 

   

Michigan I Indiana transects

    

 

 

 

60‘ OHewthompopulatlom OApple populations
3" .
t 55- + Hawthorn y-102.m.0.0220x 12.0.1113 a; 0.0001
5 1 + Apple y- 52033-00050: 3.0.404 a; 0.0016
O 50.
g .
g 45.
.- 1
.g 40.
a. 1
“T
a V I V I V I V r V I V t r I v, ‘1
2000 2200 2400 2600 2000 3000 3200 3400 3000
GrowingDeg'eeDaysmueso‘h
Wisconsin I Illinois transects
55‘ O Hawthorn populations OApple populations
3" . +
i;- so. Hawthorn y-7s.331-0.01301r 12.-0.302 3; 0.0001
2 . ‘ Apple y-50.694-0.0040x 3.0.443 as 0.0102
2 O
1 . +
g 4m .
,_ 1 (one 2 + .J\ n
E 35.1 9
.- 304
2 .
as v v V U V r V v V I V f ' V ' '
2000 2200 2400 2000 2000 3000 3200 3400 3000
Grewlngchecanswmso‘h

Figure 6.

129

be the northern most site along transect 1 in this discussion). For transect 4,
allele frequencies for Me100, Acon-295 and Mpi37 displayed sharp drops
between hawthorn populations 20 and 21 north of Green Bay in eastern
Wisconsin, leveled off and even showed somewhat of an increase between
populations 21 through 25, and then fell for a second time south of site 25
near the boarder between Wisconsin and Illinois (Figure 7, Tables 39-41).
The linked loci Me, Acon-Z and Mpi therefore behaved in a similar fashion
among transect 4 hawthorn populations. In contrast, Had] 00 frequencies
did not change appreciably between hawthorn sites 20 and 21 but did fall
along with Me] 00, Acon-295 and Mpi-3 7 south of population 25 (Figure 8,
Table 40). Did-2100 frequencies decreased steadily from hawthorn
populations 20 to 24 (Figure 8, Table 40). However, sites 25 and 26 formed
a plateau of elevated Bic-2100 frequencies that was more pronounced and
occurred slightly further south, than similar peaks for Me100, Acon-295,
Mpi37 and Had! 00. Allele frequencies for Did-2100, like those for

Mel 00, Acon-295 , Mpi3 7, and Had100, also declined sharply toward the
southern end of transect 4. Aat-2100 displayed a marked drop in allele
frequency at hawthom site 24 which was also seen for Did-2100 and to a
lesser extent for Had] 00, but was not observed for Acon-295, Me] 00 or
Mpi3 7 (Figures 7, 8). In general, Aat-ZIOO showed a very similar pattern
of geographic variation along transect 4 as Did-2100 (Figure 8, Tables 40,
42) as would be expected due to the close proximity of Act-2 and Did-2 on
linkage group I (Chapter 1).

Comparisons of allele frequency clines for hawthorn flies between
transects 1 and 4 revealed both similarities and differences. As was the case
for transect 4, Me] 00, Acon-Z 95 and Mpi3 7 all displayed large frequency
decreases at the northern end of transect 1 between sites 10 and 4 (Figure 7,

Figure 7.

130

Allele frequencies for M e1 00, Acon-295 and M1913 7 for apple
and hawthorn populations along transect l (Michigan/Indiana)
and transect 4 (W isconsin/Illinois). Sites are arranged by
decreasing latitude from left to right along the x-axis. Site 10

is included as the northern most site for transect 1.

131

Allele Frequencies tor Memo, [icon-295 and Mpi37
Along Transect 1 (ML/Ind.) and Transect 4 (Wl.llll.)

Allele ireq. Me 100
O
O

 

MelTreneectt

OW
+4001-

 

 

 

0-2 I I T I I Y
10 1 4 s 6 7 e 9
1reneecttenee

0.81-
, Aeonl'l’reneectt
c «that
8 .
<
304-:
4
000 I I I I I I
10 1 4 5 O 7 9
Tmnecttellee
0.3-
MplITreneecti
l
h Gm
202-1 '0'”
a
2
6r
2
3
.2
2

 

 

 

Allele ireq. A0011 1: Allele tree. Me 100

411.1. mg. Mpi 17

 

 

 

 

0.71
1 MeITreneecu
”-1
. 0- w
051 ...”
4
0.44
”-1
0.2-
0" f I r I I I I Ifi
19 20 21 22 23 24 25 28 27 20
Trencect4eltee
“q
AconITreneect4
05‘ 0W
*w
0.4-
”-1
4
“a
0.1-1
on i I I I I I T j 1
19 20 2122 23 24 25 26 27 2
Tmeecueltee
0.21
Mpi/Tmeecu
‘ O- W
'0' ~01-

       

 

 

I

19 20 21 22 23 24 25 26 27 20
Tnneect4eltee

Figure 8.

132

Allele frequencies for Aat-2100, Did-2100 and Had100 for
apple and hawthorn populations along transect 1 (Michigan/
Indiana) and transect 4 (Wisconsin/Illinois). Sites are
arranged by decreasing latitude from left to right along the x-

axis. SitelO is included as the northern most site for transect 1.

133

Allele Frequencies tor Ant-210°, Die-210° and Hadloo
Along Transect 1 (ML/Ind.) and Transect 4 (Wi./lll.)

 

 

 

 

 

 

 

 

 

 

 

  
 

 

 

 

 

0.0-1 0.0-
AetITreneect‘l
”-1 o W 05-1
§ ' *- Aeoh §
‘-’ a
‘ ‘ 0.4
0.4-1 ' a
g 1 g I
i i ‘
063‘ ‘ 0.3-1
0'2 I I I I I I I I I 0‘2 III I I I I I I I
10 1 2 3 4 5 3 7 3 9 12 20 21 22 23 24 25 23 27 23
Treneecneltee Tm4elee
0.33 10.
l DielTreneectl OieITreneect4
3 01-1 ° W 3 o- W
e- 7.’ W0 "
'a' a.
i .9
a
"<' <
0.5 I I I f I I I I I 0" I I I I I I 1 I I
10 1 2 3 4 5 3 7 3 9 12 20 21 22 23 24 25 23 27 2
Treneectieltee Treneeet4eitee
1.0! ‘OH
0,... °"‘ 0- W
8 o- w 3 «e- Apple
v. 1 + A”. " 0.3
g 01- E
1 0.7"
g or: g
.2 0.6-1
3 ‘ a .
3 one ‘ 031
1 - 1
05 v e T e I I 1 Vfi 0" V ' ' V ' I ' V ‘
10 1 2 3 4 s g 7 g g 13 20 21 22 23 24 23 26 27 2O
Truncation“ Treneect4el|ee

134

Tables 3941). However, allele frequencies for M e1 00 and Acon -295
fluctuated more sharply among centrally located populations in Michigan
than in Wisconsin (a possible cause for the frequency peaks observed for
hawthorn flies at site 5 will be discussed later). Had100 frequencies
changed roughly equivalently between the two transects (Figure 8, Table
40) but via-2100 and Ant-2100 clines were different, as both Did-2100 and
Aat-2100 displayed much more even and gradual reductions in allele
frequencies along transect 1 than transect 4 (Figure 8, Tables 40, 42).

Allele frequency clines for apple infesting populations, while not as
distinct or as statistically significant as those for hawthorn flies, were also
apparent. For instance, Me] 00 and Acorn-295 frequencies both dropped
between apple populations 20-21 and 1-4 just as they did for the
corresponding paired hawthorn populations (Figure 7, Tables 39, 40). In
addition, Mpi3 7 allele frequencies fell between apple sites 20 and 21 and
sites 10 and 1 (Figure 7, Table 41). Act-2100, Die-2100 and Had100 allele
frequencies also displayed a general decline in allele frequencies between
apple sites 10 and 2 along transect 1 just as they did for hawthorn flies
(Figure 8, Tables 42, 43). However, none of these three loci showed an
obvious decrease in frequency among apple populations at the northem end
of transect 4. Had! 00 did vary significantly with both latitude and GDD for
apple sites in Wisconsin and Illinois (Table 37). However, Hadloo
frequencies changed more gradually and uniformly among apple
populations along transect 4 than for hawthorn populations (Figure 8, Table
40).

Allele frequencies for Acon-Z suggest that at least some gene flow
may be occurring between local apple and hawthorn populations.
Frequencies for Acon-Z 75 were higher for both hawthorn and apple

135

populations in central and western Michigan than they were in any other
region of the Midwest (Figure 5, Table 39). The same phenomena also
occurred for Acon-289 at sites 30 and 32 in Wisconsin and Illinois (Figure
5, Table 39). In addition, the rare allele Acon-273 was present only in
apple and hawthorn populations along transect 4 (Table 39). Selection
could be responsible for the elevated frequencies of certain Acon-Z alleles
only at specific paired apple and hawthorn sites. However, Acon-Z allele
frequencies were usually significantly different between the two host races
at paired sites (Table 34); a result suggesting that selection pressures differ
between sympatric populations of hawthorn and apple flies. It is also
unlikely that stochastic processes alone could account for the independent
increases of rare Acon-Z alleles in apple and hawthorn populations at only
certain paired sites. Inter-host gene flow must therefore be at least partly
responsible for the increased frequencies of Acon-Z 75 and Acon-289 in
both host races in central Michigan and Wisconsin, respectively, and for the
occurrence of the rare allele Acon-Z 73 only along transect 4. Gene flow
may also account for the observation in Chapter 3 that polymorphic
allozymes which do not display host associated differentiation at sympatric
sites also show low levels of intra-host geographic variation across eastern
North America.

Although latitude and temperature are inter-related environmental
factors, their relationship is not perfect. Several irregularities exist in the
pattern of isothermal clines through Michigan and Wisconsin (Figure 5)
which are informative for determining whether latitude or GDD is the most
accurate predictor of gene frequencies for R. pomonella populations. For
Michigan/Indiana transects GDD proved to be the more reliable

determinant and explained approximately 10% more of the genetic

136

variation among both apple and hawthorn populations than latitude (see r2
values for all loci in Tables 37, 38). GDD and latitude accounted equally
well for genetic variation among hawthorn populations in Wisconsin and
Illinois (Table 38) but latitude was a better predictor among apple
populations (r2 latitude = 0.59, r2 GDD = 0.44; Table 37 ). Unfortunately,
four hawthorn sites in Wisconsin and Illinois (27, 31, 33, 34) did not have
paired apple collections. Three of these four unpaired sites (31, 33, 34)
were located within thermal warm areas in Wisconsin (Figure 5). The
addition of apple populations from these three sites would more than likely
make the r2 values for latitude and GDD among Wisconsin/Illinois apple
populations very similar to one another. If one also considers that GDD
estimates for sites are less accurate than latitude (GDD values being baSed
on multiple year averages obtained from the closest weather station to a
collecting site in years when temperature data were available), then ambient
temperature would appear to be a slightly better indicator of the genetic
composition of local R. pomonella populations, at the very least in Michigan
and Indiana, than latitude.

Hawthom populations displayed several allele frequency shifts for
Me100 and Acon-295 that were, in fact, correlated with irregularities in
local thermal conditions. First, site 5 (Okemos, Michigan) is located within
a pocket of colder GDD in central Michigan (Figure 5). Gene frequencies
for Memo and Acon-295 were significantly higher in hawthorn than in
apple populations at this site (Me1 00 hawthorn = 0.455 and Me] 00 apple =
0.298; Acon-295 hawthom= 0.295 and Acon-295 apple = 0.098; Tables 34,
39, 40, Figure 7). Site 4 (Eureka, Michigan) is located approximately 20
km north of site 5 but is warmer. Me] 00 and Acon-295 did not show inter-

host differentiation at site 4 (Table 34) and were present at respective

137

frequencies of 0.283 and 0.214 in hawthorn larvae and 0.333 and 0.200 in
hawthorn adults. Warmer temperature sites south of Okemos, Michigan
(sites 6-9; see Figure 5) showed the same trends for Me and Acon-Z allele
frequencies that were observed at site 4 (Figure 7, Tables 39, 40). A "blip"
therefore occurred for Me100 and Acon-295 frequencies that corresponded
with a cold bubble of GDD in central Michigan.

A second example involves fingers of higher GDD which extend up
into Michigan around the Kalamazoo area (see site 13 in Figure 5) and in
Wisconsin northeast of the city of Madison (see site 31). Gene frequencies
for Me1 00 and Acon-295 were almost always lower for hawthorn
populations within these warm areas (site 13, Portage, Michigan, and sites
33 and 34 near Greenbush and Iron Ridge, Wisconsin) compared to sites of
equivalent latitude in parallel transects (sites 6, l7 and 18 in the case of site
13, sites 23 and 30 for site 33, and site 24 for site 34; see Tables 39 and 40
for Acon-Z and Me allele frequencies, respectively). The only exception to
this general trend was for Acon-295 in comparisons between sites 23 and
33, for which the fiequency of the allele was lower at the colder site 23
(0.221) than at the warmer site 33 (0.306).

Finally, in a related study of hawthorn populations along a transect
running from north of Milwaukee, Wisconsin to south of Chicago, Illinois
(Feder et al., manuscript in prep.), gene frequencies for Me] 00 and Acon-
295 dropped dramatically around Milwaukee corresponding to an increase
in GDD presumably due to a heating effect of the city.

138

Discussion

The results from this study strongly suggest that partially
reproductively isolated host races can evolve in sympatry. However, to
prove that the apple race originated via a sympatric host shift I must still
show that 1.) the apple infesting form of R. pomonella was not
reproductively isolated from hawthorn flies prior to its colonization of
apples and 2.) that the apple race could not have evolved in allopatry from
the hawthorn race and then been secondarily introduced into the
northeastern United States.

All available evidence indicates that the apple race is directly
descendant from a hawthom infesting form of R. pomonella and was not
derived from an unrecognized sibling species. For instance, experimental
crosses give no indication of any postrnating or etlrological premating
reproductive isolation between hawthorn and apple flies (Reissig and Smith,
1978; Smith, 1988; R. Prokopy, pers. comm.). Also, no single
morphological or genetic character exists which diagnostically
distinguishes apple from hawthorn flies, as the host races are currently
known to differ only in allele frequencies for six specific allozyme loci. In
fact, because allele frequency clines are steeper among hawthom than apple
populations, hawthorn populations at southern sites in the Midwest are
actually more genetically similar to northern apple populations than they
are to northern hawthorn populations. Consequently, different apple
populations from across the Midwest do not cluster together genetically as a
discrete subdivision from hawthorn populations as might be expected if the
apple race was formed from a sister taxon to the hawthorn race.

Furthermore, the non-random pattern of Acon-Z 73, Acon-Z 75 and

139
Acon-289 allele frequencies in Michigan and Wisconsin (Figure 1) and the
similar pattern of linkage disequilibrium in apple and hawthorn populations
(Chapters 2, 3) suggest that at least some gene flow is occurring between
local apple and hawthorn populations. It is therefore very unlikely that
apple and hawthorn populations of R. pomonella are completely
reproductively isolated.

It is still possible, of course, that different races of R. pomonella exist
on different species of hawthoms and that the apple race was derived from
just one of these "hawthorn races". As many as 1,000 species of Crataegus
comprising 19 different species groups may, in fact, be endemic to North
America (Femald, 1950; Correll and Johnston, 1970). However, the
taxonomic status of a majority of these Crataegus species is questionable as
hybridization, polyploidy and apomixis is common in the genus (Phipps,
pers. comm). Many of the trees that we identified as C. mollis in this study
may, therefore, have been of hybrid origin. Nevertheless, R. pomonella
attacks only a restricted set of Crataegus "species", with infestations
confirmed for only 14 endemic hawthoms (Bush, 1966; Wasbauer, 1972;
Berlocher, 1976; McPheron et al., 1988). In the northeastern United States,
only C. mollis or hawthorn species which are phenotypically similar to C.
mollis in the Brainerdianae, Coccinnea , Macracantha, Pruinosa, Punctata
and Tenuifoliae series have been reported to be parasitized by the fly
(O'Kane, 1914; Wasbauer, 1952; Bush, 1966). It therefore appears that R.
pomonella utilizes only a small number of potential hawthorn hosts in the
Northeast, attacking only those species that share a similar fruiting
phenology and fruit characteristics; A trend not conducive to the formation
of different hawthorn races. In addition, electrophoretic analysis of R.

pomonella populations infesting C. punctata, C. brachyacantha, C. douglasii

140

and C. monogyna from the southern and western United States has given no
evidence for genetically differentiated hawthorn races (Berlocher, 1976;
McPheron, 1987). Although we cannot completely discount the possibility
of host specific hawthorn races, all available data suggest that hawthorn
populations of R. pomonella represent a single race which displays
latitudinal variation.

Several points also make it unlikely that the apple race formed in
allopatly from the hawthorn race and was subsequently introduced into the
Hudson Valley region of New York. First, genetic bottlenecks due to
reductions in population size frequently occur when a species colonizes a
new area. Reduced amounts of genetic variation have, in fact, been
observed for newly introduced populations of R. pomonella and R.
completa in the western United States (McPheron et al., 1988c; Berlocher,
1984b). However, apple populations in the eastem United States have as
much or more allelic diversity as eastern hawthorn populations (Chapter 3).
There is, therefore, no genetic evidence indicating that the apple race was
introduced into the New England area.

Second, the R. pomonella species group is endemic to North America
(Bush, 1966), thereby ruling out the possibility that the apple race
originated in Europe and was brought to the United States along with
cultivated apples. Therefore, if the apple race formed in allopatry, it must
have originated somewhere in North America. The range of domestic
apples is, however, contained entirely within that of hawthoms in North
America. Consequently, even if R. pomonella did not originally shift onto
apples in the Hudson Valley, the apple race could not have been
geographically isolated from hawthorn infesting populations at the time of

its inception.

141

Third, B.D. Walsh (1867) noted at the time of the initial discovery of
the apple race that although R. pomonella "exists both in the East and in the
West, it attacks the cultivated apple only in a certain limited region, even in
the East, for this new and formidable enemy of the apple is found in the
Hudson-river valley, but has not yet reached New Jersey". Therefore, up
until 1867, no apple infesting population of R. pomonella had been reported
outside of the New England area. We now know, however, that an isolated
population of R. pomonella does infest both hawthoms and domestic apples
in the highlands of Mexico (Bush, 1966). But Mexican flies are unique in
that they possess a hyaline spot at the base of the apical band on their wings
which makes them morphologically distinguishable from all other taxa in
the R. pomonella group (Bush, 1966). The apple race in the United States
could not, therefore, have been derived from the Mexican population.

Although the Mexican population does not clarify the issue of the
geographic origin of the apple race, it does lend further indirect support for
hawthoms as the original host for the apple race. In Mexico, R. pomonella
has independently shifted from hawthoms to domestic apples within
historical times (apples were introduced into Mexico from Spain in 1522;
Standley, 1922). However, Mexico does not have any endemic species of
crab apple (Standley, 1922; J. Beaman, pers. comm.). Therefore, in
Mexico at least, R. pomonella shifted directly from hawthoms to domestic
apples.

Two factors are primarily responsible‘for the genetic differences
observed between hawthorn and apple populations of R. pomonella.

First, adult flies do not migrate randomly between hawthorn and apple
trees. This was seen from the fact that adult flies collected off of hawthorn

and apple trees at site 15 showed the same pattern and magnitude of genetic

142

divergence as their larval offspring. Host fidelity therefore exists for these
flies with adults tending to utilize the same species of host plant which they
infested as larvae.

Second, differential selection occurs for flies infesting apples and
hawthoms and appears to be related in some way with ambient temperature.
The latitudinal allele frequency clines among both apple and hawthorn
populations suggest that selection pressures are not constant across the
ranges of both host races, however. Allele frequency Clines are also steeper
for hawthorn flies than for apple flies and intersect between the two host
races in the central portion of R. pomonella's range (Figures 6-8). This
accounts for the greater amount of geographic differentiation in the
hawthorn race as well as for why northern and southern sites show higher
levels of inter-host divergence than centrally located areas in the Midwest.

Further research is needed in three areas to clarify specific details of
the host specialization process. These include: 1.) establishing the extent of
and mechanisms responsible for host fidelity and host related selection; 2.)
determining how temperature acts as a differential selective factor and 3.)
estimating effective levels of gene flow among and between apple and
hawthom populations.

Mark-recapture and controlled rearing experiments would help
address questions concerning host fidelity and gene flow. Collecting larvae
from infested fruit and marking and releasing the resulting "naive" adults
the following summer would provide estimates of intra- and inter-host
migration. Hardwired behavioral differences in host recognition and
conditioning in adults are two likely mechanisms responsible for host
fidelity (larval conditioning has never been found for Rhagoletis flies but
adult conditioning has; Prokopy et al., 1982). Releasing naive apple flies

143

under hawthorn trees and vice-versa would establish the relative
importance of genetically based differences in host preference. Adult flies
could also be exposed to alternative host fnrits in the laboratory prior to
their release. Analyses of the subsequent host acceptance behavior of these
"trained" flies under field conditions would indicate the extent to which
adult conditioning affects host choice. Furthermore, by combining the
results of mark-recapture studies with those from reciprocal egg transplant
experiments measuring survivorship levels of progeny derived from
hawthorn x hawthorn, hawthorn x apple and apple x apple crosses, I will be
able to directly estimate effective gene flow between the host races.
Ambient temperature has always been regarded as an important
environmental factor influencing host plant-parasite interactions for
Rhagoletis. Rhagoletis pomonella is, in general, univoltine across its range
and adult life expectancy has been estimated at from 3 to 6 weeks in the field
(see review by Boller and Prokopy, 1976). Adults must therefore eclose at
times paralleling the fruiting phenologies of their host plants. "Temporal
windows" for infesting apples and hawthoms are different with the fruit of
domestic apple varieties favored by R. pomonella generally ripening and
falling one month earlier than fruit of native hawthorn species. In addition,
flies must also develop as larvae and pupae at rates closely matching local
environmental conditions. If development proceeds too quickly flies may
not diapause and emerge as a second generation in the fall when suitable
host fruit is no longer available. Alternatively, if development occurs too
slowly flies are exposed to potential predators and parasites for a longer
period of time and also run the risk of freezing to death during the first
frost. Ambient temperature does, of course, correlate with latitude and is

also a major determinant of the phenology of fruit maturation.

144

Temperature may therefore have a profound effect on development rates
for hawthorn and apple flies across their respective ranges. Consequently,
latitudinal variation in temperature could account for the allele frequency
clines in the two races. Furthermore, differences in the fruiting
phenologies of apple and hawthorn trees could explain why allele frequency
clines differ between the two host races.

The temperature hypothesis is contingent, of course, on Me, Acon-Z,
Mpi, Did-2 and Had being related in some way with developmental timing
in R. pomonella. Elucidating the genetic basis for diapause related traits in
insects has proven to be a difficult task, however. Although diapause
characters do show latitudinal variation associated with temperature and
photoperiod (see review by Tauber and Tauber, 1976), few examples exist
documenting the genetic basis for this phenotypic variation. One possible
exception may be the pitcher plant mosquito, Wyeomyia smithii, in which a
weak association has been found between Phosphoglucomutase and Phenyl-
alanyl- leucine amino-peptidase (two loci showing latitudinal frequency
clines within the species) and phenotypic variation for diapause and
developmental timing (Istock and Weisberg, 1987). Genetic and/or
phenotypic markers associated with diapause characters have also been
found for the green lacewings, Chrysopa carnea and C. downesi (Tauber et
al., 1977); the European corn borer, Ostrinea nubilalr's (Shower, 1981); and
the eastern tiger swallowtail butterfly, Papilio glaucus (Rockey et al.,
1987). R. pomonella appears to represent another example as preliminary
studies indicate that allele frequencies for Me, Acon-Z and Had correlate
with the timing of diapause termination and adult eclosion for hawthom

flies (Feder et al., manuscript in preparation)

145

A potentially interesting test of the temperature/development
hypothesis for R. pomonella would be to sample hawthorn and apple fly
populations along altitudinal transects in the Appalachian mountains. The
prediction would be that if temperature is important and migration rates are
low or moderate, allele frequency clines similar to those observed in the
Midwest should also be found for the two host races along the slopes of
these mountains. Larval transplant experiments performed under
controlled rearing conditions would also provide valuable information on
how temperature influences host plant interactions for R. pomonella.

It is not uncommon for allozyme surveys to uncover one or more loci
whose alleles display a clinal pattern of variation correlated with some
aspect of the physical environment (Johnson, 1971; Clegg and Allard," 1972;
Harnrick and Allard, 1972; Christiansen and Frydenberg, 1974; Koehn et
al., 1976; see review by Hedrick et al., 1976). Latitudinal, seasonal and
altitudinal variation in temperature have been implicated as causal factors
for a number of allozyme clines including those for the freshwater fish,
Catostomus C larkii (Koehn, 1969); the marine fish, F undulus heteroclitus
(Mitton and Koehn, 1975); the fathead minnow, Pimephales promelas
(Merritt, 1972); the fruit fly, Drosophila melanogaster (Berger, 1971;
Johnson and Schaffer, 1973; Miller, Pearcey and Berger, 1975); and the
butterfly, Colias meadii (Johnson, 1976). What is unusual about R.
pomonella is that ambient temperature apparently has different
consequences for conspecific populations infesting different host plants.
Host plant associations therefore add an additional level of environmental
heterogeneity for R. pomonella which helps maintain increased amounts of

intraspecific genetic variation; a result consistent with predictions of niche

146

theory (Levene, 1953; Maynard-Smith, 1966, 1970; Levins, 1968,
Christiansen and Feldman, 1975; Felsenstein, 1976).

Allele frequency Clines for hawthorn and apple flies are not
completely discordant, however, and regions where the two races show
similar patterns of genetic variation have important evolutionary
implications. The three areas where the largest allele frequency changes
occur for apple and/or hawthom populations are located in or near major
ecological transition zones in the Midwest. Sites 14 in Michigan and 20-21
in Wisconsin are situated at the boundary between deciduous eastem forest
and arboreal lake forest habitats. Likewise, sites 25-26 near the
Illinois/Wisconsin border lie where the central prairie meets deciduous
eastern forest. The ecological consequences of these zones are significant
and mark cutoff points in the distribution of numerous plant species.
Changes in voltinism and/or host plant associations occur across these
regions for several different phytophagous insects including the European
com-borer, (Showers, 1981) and the swallowtail butterfly, Papilio glaucus
(for review see Scriber and Hainze, 1987). The finding of a similar pattern
for R. pomonella supports the hypothesis that these zones have a general
impact on the evolution of life history traits for phytophagous insects.

The observed frequency clines in R. pomonella could, of course, be
due to secondary contact rather than selection. However, there is no
historical or biogeographic evidence indicating that geographically isolated
northern and southern populations of R. pomonella ever existed. It is
therefore much more likely that R. pomonella recolonized northern
latitudes along with hawthoms following the retreat of ice sheets in North
America some 10,000 to 20,000 years ago than for populations of the fly to

have expanded from pleistocene refugia. Also, although apple and

147
hawthorn flies both display allele frequency clines, the configurations of

these clines do show some differences between the two host races and
among different loci within the races. Furthermore, Memo and Acon-295
frequencies for hawthorn populations co-vary with irregularities in local
thermal conditions. These latter two observations are not consistent with
the secondary contact hypothesis and argue for differential selection as the
causal basis for allele frequency clines in R. pomonella.

Another important question is whether apple and hawthorn host races
are incipient species. A host race is a parasitic population which, by
adapting to a preferred host, has become partially reproductively isolated
from other conspecific populations specialized on alternative hosts (Diehl
and Bush, 1984). Host races may therefore be at various stages of
divergence ranging from populations which almost freely interbreed to
those that rarely exchange genes. Based on available data, it is difficult to
judge whether apple and hawthom flies have reached or will ever reach the
evolutionary point of no return when they should be considered distinct
species. In essence, the results of this and the preceding chapters have
shown that host specific traits can evolve within geographically continuous
populations. The key question then is whether these host associated
adaptations are sufficient, in and of themselves, to eventually cause the
complete reproductive isolation of populations. Indirect support that they
are has come from additional population genetic studies involving sibling
species in the R. pomonella complex. For instance, R. mendax is a closely
related sibling species to R. pomonella which infests blueberries
(Vaccinium sp.) and huckleberries (Gaylussacia sp.) in the eastern United
States and Canada (Bush, 1966). Although field captured R. pomonella and
R. mendax adults will hybridize in the laboratory and produce fertile and

148

viable F1 progeny (J. Frey pers. comm.; Feder and Bush, submitted),
sympatric populations of the sibling species are genetically distinct in nature
and to not interbreed (Berlocher and Bush, 1982; Feder et al., submitted).
Genetic analysis of reproductively mature adults collected from
interdigitated blueberry bushes and apple trees indicate that R. mendax and
R. pomonella maintain 100% complete host fidelity in the field (Feder and
Bush, submitted). Differential host plant recognition is therefore a very
effective pre-mating isolating barrier between at least two sympatrically
distributed sibling species in the R. pomonella group.

The allele frequency clines indicate that both hawthorn and apple
races are not evolving uniformly across the Midwest, however. Primary
frequency clines have been implicated in speciation (Fisher, 1930; Murray,
1972; Endler, 1977) and it is possible that R. pomonella populations are
"isolated by distance" as a consequence of being differentially adapted to-
local thermal conditions. Both hawthorn and apple races did display sharp
allele frequency "steps" across environmental transition zones in the
Midwest (Figures 7, 8). In addition, hawthorn populations at the northern
and southern ends of R. pomonella’s range are essentially fixed for
altemative alleles at the loci Had, Did-2 and Me (McPheron, 1987). R.
pomonella populations at different latitudes may therefore be "genetically
incompatible" with one another even though they are inter-connected by a
series of populations exchanging genes. It will be interesting to see whether
allele frequency clines for R. pomonella can become steep enough to result
in "hybrid zones" given the fly's apparent ability to disperse over relatively
long distances.

Whether or not R. pomonella host races are "cohesive evolutionary

units" is not central to the issue of sympatric speciation, however. The

149

important question is whether host recognition behaviors are constant
across the range of a race. For example, if flies infesting hawthoms
continue to prefer hawthoms when experimentally transplanted into
different portions of their distribution, then host shifts ensure the
divergence of populations regardless of whether selection pressures vary
from locality to locality within races. The results from site 15 confirm that
some host fidelity exists for R. pomonella in at least one portion of the fly's
range. However, apple and hawthorn populations at sites 4 and 32 showed
no significant genetic differentiation (Table 34). Is the lack of genetic
divergence between apple and hawthorn populations at these two sites due to
high levels of inter-host gene flow? Or, alternatively, is host fidelity still
strong at sites 4 and 32 but selection pressures on Acon-Z, Mpi, Dia-2, Act-
2 and Had the same for apple and hawthom flies? The shapes of the allele
frequency clines among hawthorn and apple populations suggest that the
latter possibility is true. However, detailed studies measuring inter-host
gene flow across the range of R. pomonella are needed to resolve this point
and establish the extent to which apple and hawthorn races are discrete
populations on separate evolutionary pathways.

In conclusion, R. pomonella is rapidly becoming a model organism
for the study of speciation. Genetic differences exist between recently
formed host races of the species and the geographic pattern of allozyme
variation suggests that differences in host plant utilization and host
associated development (in some way related to ambient temperature) are
responsible for reducing gene flow between apple and hawthorn races. My
next task is to characterize exactly how these factors operate at the
population level and to translate their mode of action to the speciation

process. In this regard, the ability to hybridize several different R.

150

pomonella species in the laboratory will help us elucidate the genetic basis
for host associated traits, such as differential host recognition, which appear

to play a major role in reproductively isolating taxa in the R. pomonella
group.

151

 

 

 

 

.eOHunHHHavo OnoncheroHoz eonu coHocH>0o ocnoHuHcoHn now encore HOO.O w m use HO.O w u a. OO.O w m s
NON.O OO0.0 .OHH.O OO0.0 OO0.0 HO 4 NO0.0 NO0.0 OON.O NO0.0 OOO.O NO O OH
om0.0 ON0.0 OmO.o OO0.0 Om0.0 On a NO0.0 OO0.0 OO0.0 OH0.0 OOO.O NO x HHOOH

r r r r r t r r a OH0.0 HHH.O OO0.0 HO0.0 ON0.0 OOH.O on O .oOOOH
OOH.O OO0.0 OOH.O OO0.0 OOH.O on < OOO.O OO0.0 HO0.0 cH0.0 NOH.O NO O OH

mO0.0 OOH.O Oom.O OOH.O ON0.0 OO0.0 moo r OO0.0 OO0.0 OO0.0 ON0.0 NO0.0 ON0.0 ONO O AHOOH
coo.O OOH.O NOm.O OOH.O OO0.0 OO0.0 Own a NO0.0 OOO.O OOO.O mO0.0 NO0.0 OO0.0 OHO O loOOOH
OOH.O OOm.OrrooH.O OO0.0 OOH.O on r NH0.0 OON.O OH0.0 OOH.O NH0.0 NO O OH
H. HZ. m uUOmcmuB
mH0.0 NOH.O OOm.O OOH.O OO0.0 OO0.0 Om c OH0.0 HN0.0 NO0.0 OOH.O mHH.O OOH.O OO O OH
NNH.O HHm.O OO0.0 OO0.0 OOH.O me < OOO.O ON0.0 NHO.O OOO.O OOH.O OO O NH
OOH.O ON0.0 omN.O OO0.0 ON0.0 HO e r r r r r r r r O HH
OOH.O NOO.O NO0.0 OO0.0 OO0.0 Om < HHO.O omN.O OOO.O OOO.O On O OH
..HZON oooncnna
OH0.0 mNH.O OOH.O OOO.O OH0.0 Nm 4 ON0.0 OO0.0 OOO.O OO0.0 ON0.0 HO O O
OON.O Nmm.o OOH.O ON 4 HOH.O OOH.O OO0.0 ON0.0 OH0.0 mm O O

r r r r r r r r c OO0.0 Om0.0 OHo.O OHH.O OO0.0 OOH.O on r O
OHH.O HHm.O OHH.O NO0.0 ONN.O OO c OO0.0 OON.O .NNN.O OO0.0 HO0.0 on m o
ONN.O OO0.0 OO0.0 NH0.0 ON0.0 HO r ON0.0 OOO.O OO0.0 mON.O ONO.O OOO.O OO O m
OOH.O OOc.O OOH.O OH0.0 ON a OH0.0 NOO.O.OOO.O..OHN.O ON O .HOO

r r r t r r r r 4 OOH.O OO0.0 OON.O OH0.0 OO O HoOOO
«O0.0 NO0.0 mmN.O Om0.0 mON.O NO 2 OO0.0 ONN.O ON0.0 ON0.0 NOH.O OO O O

OH0.0 ONH.O ON0.0 NOH.O ON0.0 NOH.O On c ONH.O OO0.0 NOO.O HH0.0 HH0.0 OO O N
OOH.O OHm.O OON.O omO.o ON0.0 on c ON0.0 Om0.0 OOH.O oOo.O Om0.0 ONO.O HO O H
..cH\.Hz. H commence
OHH OOH OOH mO OO HO ON zlqwqu OOH .OOH OOH «O .OO OH OOrtHHZIHHOHIHHHu
Ntflco<
£58 9 5.5: 80h unawa—

3 «588 288.5 26 05 £55 Bung 2O moan .233 8 HBO 93 onto— 9 232 Arc
628:8 203 233 can 832 53 823 2 e5 2 .v 8% 2 .OoHSm 355 5283-38 05

388 £2853 .m .3 Ozone—smog AS came 93 $5 F8532 Bu N13,; .Hou 865:3 222

.3 033—.

152

 

 

 

.eOHunHHHOuo OumncHoznanuOO scuu OoHuOH>oo ucmuHOHOOHO you Hmouuu HOO.O w m ... HO.O w m .. OO.O w m .

u u u u n u - n O HHH.O ON0.0 ONH.O OON.O OO O OO

u u u u u u u n O OO0.0 ON0.0 OOO.O OHH.O OH O OO
..Hz. .oOHz

OOH.O OO0.0 OOH.O OOH.O OH0.0 OO O HHN.O OO0.0 OO0.0 OON.O HH0.0 OO O NO

u u u u n u u n O OO0.0 OO0.0 OO0.0 OOH.O OHO.O ON O HO

OO0.0 OO0.0 OOH.O OHH.O OO0.0 OO O HH0.0 OOH.O ON0.0 OON.O OO0.0 OO O OO

NH0.0 OOH.O OO0.0 OON.O NO0.0 NO O OO0.0 OO0.0 OO0.0 ONH.O NO O ON
A. :3 m uowmcpum.

OH0.0 OOH.O ON0.0 OHH.O OO0.0 OH0.0 NO0.0 OO O ON0.0 NO0.0 NO0.0 OO0.0 ONH.O NHO.O OHO.O OO O ON

a : u u u u u a O OH0.0 ONH.O OO0.0 .OOH.O OOH.O ON0.0 HO O ON

ON0.0 HOH.O OO0.0 HOH.O OO0.0 ON0.0 OH0.0 OO O OO0.0 OHH.O OO0.0 OOH.O HO0.0 NO O ON

OO0.0 OOH.O OO0.0 OOH.O OO0.0 NO0.0 ON O HO0.0 OO0.0 OO0.0 OOH.O OH0.0 OO O ON

OOH.O ON0.0 NOH.O HO0.0 OH0.0 OO O OO0.0 OO0.0 OH0.0 OO0.0 OH0.0 ON O ON

OHH.O OH0.0 OOH.O ON0.0 OO0.0 OO O HOH.O ON0.0 .HNN.O OOo.O OO0.0 OO O ON

NN0.0 OOH.O OO0.0 OOH.O OO0.0 NN0.0 OO O ON0.0 OOO.O OOO.O ON0.0 OO0.0 ON0.0 OO O NN

NH0.0 OHN.O NO0.0 OO0.0 OO0.0 ON0.0 ON0.0 OO O OOH.O HH0.0 .OON.O HO0.0 ON0.0 OO O HN

«O0.0 OOH.O OO0.0 OO0.0 OO0.0 «O0.0 NOH O OO0.0 OO0.0 OON.O NO0.0 OOO.O OO0.0 OO0.0 OOH O ON

OO0.0 OOH.O OH0.0 OON.O ON0.0 «H0.0 OO0.0 OOH O OO0.0 OO0.0 OON.O OO0.0 NNH.O «O0.0 HHH O OH
..HHH\.H3. O Oommcmua

IIOHHIIIOOH OOH OO Oa‘ ON OO z HOOO OHH OOH OOH OO Om ON OO alqquluqau

Nn=¢9<
3.83 .OO 2%»

.eOHunHHHOvm OumnOszuOquO sown OOHOOH>OO OOOoHuHOOHO OoO ammuuu HOO.O w N .«O H0.0 w m .« OO.O w N O

 

153

vmv.o an C mmm.o Nm : NNm.o Hm C m~m.o NM 2 oom.o Hm C omh.o mm n ma
pwv.o on C an.o mN a mma.o on C m~n.o mm : vmm.o on C bmw.o mm m Aqvha
I I C NOH.O on m I I C Nwm.o mm m I I C omw.o mm = AvahH
0mm.o mm C wma.o mm m NmN.o mm C NOH.O mm : won.o mm C Hon.o mm a ma
OHH.O wa C oo~.o Pov m no~.o omw C OOH.O vwv m mmw.o wow C mvm.o va m Aqvmfl
omn.o Pom C va.o man a oNN.o 0mm C va.o Own : omw.o hmm C rev.o mom 2 AvamH
mam.o he C umva.o an : omH.o om C OOH.O N¢ m omv.o on C wNN.o Nv m vH
A.Hz.m powmcmue
vhN.o Hm C vov.o hv m mmN.o mm C «amv.o ov a $05.0 vm C HO0.0 mv : ma
th.o mv C mh~.o 0v : oon.o mv C omN.o ov : bmm.o mv C ham.o ov : NH
mmm.o mm C I I m mmw.o N¢ C I I : hmm.o av C I I a HH
mmN.o mm C ooa.o 5v = NmN.o mm C owo.o on m mmm.o mm C oma.o on m 0H
..HZVN uomOCMHB
va.o Hv C ummm.0 ov a mhm.o «m C mhv.o ow m ¢mh.o mm C Ooom.o ov a m
omn.o on C m~n.o mm C mmv.o on C oom.o mm a 095.0 on C hm>.o mm = o
I I C nnm.o 5N m I I C omv.o an m I I C oo~.o on m h
~m~.o av C QhN.o an m mvN.o av C NmN.o on m mmn.o av C wmw.o mm m w
mam.o mv C mnm.o we a th.o me C me.o mv : Now.o mv C mcm.o co m m
mmw.o on C Nmm.o mm : OsomN.o on C moH.o on m OHN.O on C hab.o on m Aqv¢
I I C annm.o on = I I C va.o aw m I I C www.o on m .nva
hom.o vv C «OHQN.O mm m mnn.o mv C w>~.o on m wmm.o mv C Hmv.o mm m m
«awee.o on C uuhhw.o hv m wvm.o mm C mvm.o hv m hmm.o ov C Nmm.o hv m N
vom.o Hm C «vNN.o av m wow.o Hm C mmH.o mv m oww.o on C mmm.o ov a H

..OH\.Hz. H uooOOOOO
333%

«-39 as: a:

 

£58 9 58: Bob 0365 9 «5988 3085 55? Bang

pa 8.5 .933 9 as O5 85H 9 52 3O 38:8 203 8:2 OOO O0.0.“ 53

82.3 2 9a 2 .v O23 OC .OoHSm 62:5 5283.28 05 388 322.38% .M «o
88238 3O 03% O5 SO 59.52 .8 N65 93 E: O: 8. $83.3 222 .OO ”OE.

 

154

.eOHEHHHOOm OHBOHOOIOEOO so: 5333.0 28223“ SO ummqu HOO.O w u t: HO.O w m : OO.O w u .

I I O .OH0.0 OO O I I O OO0.0 OO O I I O OH0.0 NO O OO

I I O ONN.O ON O I I O OON.O ON O I I O OON.O ON O OO
..Hz. .ome

Oom.o NO O OON.O on O HO0.0 OO O NNm.o OO O OOO.o OO O NOO.o OO O NO

I I O OON.O NN O I I C mOn.o ON O I I C omm.o ON O HO

HNn.o NO O OON.o Om O omN.o «O C cmON.o OO O OOO.o OO O OOO.o OO O on

OOH.O NO O OOH.O Nn O OO0.0 NO O ONH.o Nm O OOm.o NO O ON0.0 Nm O ON
..Hz. O uoomcuua

ooO.o ow C oom.o on O nom.o ow O Nom.o om O NON.o om C ON0.0 om O ON

I I O OO0.0 OO O I I O OO0.0 NO O I I O ONN.O OO O NN

..HO0.0 ON O OON.O NO O NO0.0 OO O NO0.0 OO O OO0.0 NO O OON.O OO O ON

OHm.o OO O OON.O OO O OON.O mm C OON.O OO O OOHN.o om O ONO.o OO O ON

NO0.0 OO O ON0.0 ON O OON.O NO O OO0.0 OH O ..OO0.0 OO O NO0.0 ON O ON

OO0.0 HO O OO0.0 OO O ONN.O HO O OOH.O OO O OO0.0 HO O OO0.0 OO O ON

NNm.o OO O OON.O Nn O ONN.O OO O ONH.o ON O OOO.o OO O oOO.o OO O NN

OON.O HO O NON.o HO O ONN.o NO O OOH.O OO O OHN.O NO O OOO.o OO O HN

ON0.0 NOH O NOH.O OOH O ONN.O OOH O NOH.O NOH O OO0.0 OOH O .OO0.0 OOH O ON

ON0.0 OOH O OHN.o OOH O OON.O OOH O oNN.o OoH O NOm.o OoH O Oom.o oHH O OH

jju. 7.3.2.2.: v uommcmue
NNH a anO NNH z ..HOOO jnjOIIJOJOqO ufld

NO:— ca: «2

 

H.388 .3 033.

155

.EsHunHHOOUm

 

 

 

muoncfiszOUumO Eouu noduuH>ou ucmoauacoOm new ummqu Hoo.o W N «cu Ho.o w N «u O0.0 w M.:
OOO.o 0Oo.o OO0.0 HO O OOO.o OH0.0 ON0.0 OH0.0 NO O OH
ON0.0 ON0.0 NO0.0 OH0.0 NN O OOO.o HO0.0 OOo.o OH0.0 OO O AAONH
I I I I I I O NOO.o OOo.o OH0.0 OOo.o HO O .UOONH
NOO.o ON0.0 OO0.0 OO O OON.O OOo.o OOH.O OO0.0 ON O OH
NOO.o OO0.0 OO0.0 HH0.0 COO O moo.oac«OHO.o NO0.0 OO0.0 OOo.o NHO O .OOOH
Ooo.o OOO.o OO0.0 OOo.o OH0.0 OOO O oHo.o OHO.o NO0.0 NHH.o OHo.o OOO O .nOOOH
OHO.o HH0.0 NO0.0 NO0.0 NO O NHo.o OHN.O OOo.o OHN.O NHo.o NO O OH
1sz m Howmcmufi
OH0.0 ON0.0 OO0.0 OO0.0 OH0.0 OO O NH0.0 COO.o OOo.o OO0.0 on O OH
NN0.0 OOO.o OO0.0 NO0.0 NN0.0 OO O OOO.o OOo.o omo.o ON0.0 oO O NH
OOO.o OO0.0 NH0.0 NO O I I I I I I O HH
OON.O HO0.0 OOH.O OH0.0 OO O NHo.o NNO.o OO0.0 HON.o OOo.o ON O oH
A.Hz. N pommcmua
OOO.o OO0.0 OO0.0 OO0.0 on O ON0.0 omo.o ON0.0 oO O O
omO.o omo.o on O OOO.o ON0.0 NO0.0 OH0.0 OO O O
I I I I I I O OO0.0 OOO.° OHo.o HO O N
«OO.O «OO.O NOH.O ON O HOO.O NO0.0 moo.o ONO.O on O O
OOO.o ON0.0 OO0.0 OHo.o OO O NN0.0 NOO.o OOo.o NO0.0 OO O m
OOO.o OO0.0 NH0.0 OO O OOO.o NO0.0 OO0.0 OO O AAOO
I I I I I I O OOO.o NO0.0 omo.o OO O AUOOO
OON.O ON0.0 oOH.o OO0.0 OO O OON.O oOo.o OOH.O oOo.o NO O O
oOO.o omo.o omo.o OO O OHO.o OO0.0 OOH.O NO O N
OH0.0 HO0.0 OH0.0 ON0.0 OO O OON.O OO0.0 OOH.O OHo.o NO O H
..OH\.OOO H uommcoua
ONO ooa oN NO HO Iz umOO ONO ooa OH NO HO 2 unan Mann
2. 2
£58 9 5.5: 8.5 832: 9 25:88 388.3 £55
BOSEO 2O «BB .333 9 85 O5 095: 8 OBOE 3V 628:8 203 823 Ea
933 53 823 NH 93 2 .O «3% .< .338 BED E2836? 2: 888
SOSBEQO .m mo gag—smog A5 03% O5 at E0532 .5. .32 O8 86:33qu 322 .HO 2%?

156

.Ezaunwaaacm

 

 

 

muwncamsu>numm Eouu coauma>ou unmoauwcoam new ummuuo HO0.0 W N uua H0.0 w N «a no.0 w M I

I I I I I I < 050.0 mH0.0 000.0 0>0.0 mm x vm

I I I I I I t mN0.0 mNm.0 0m0.0 mN0.0 m>0.0 ON I mm
..axv .omHz

«v0.0 0N0.0 0N0.0 en 4 wmm.0 vv0.0 mv m Nm

I I I I I I C 000.0 vmo.0 mN a an

Na0.0 nHH.0 0m0.0 0N0.0 0v c 000.0 an.0 h0H.0 n00.0 NH0.0 Nv m on

~m0.0 mm0.0 hm0.0 hm0.0 hm0.0 av < 0H0.0 000.0 N00.0 «00.0 mm a mm
..azv m uommcmua

mmm.0 vco.0 NN0.0 mv < mmm.0 mm0.0 hH0.0 5H0.0 00 x am

I I I I I I < Mvm.0 mH0.0 0H0.0 0~0.0 mm 3 FN

000.0 0>0.0 000.0 mN0.0 «F < 000.0 0v0.o 000.0 Nv0.0 mm x mm

000.0 mN0.0 vv0.0 vv0.0 00 fl N00.0 MHH.0 m~0.0 0v : mm

mh0.0 000.0 n00.0 NH0.0 0v C 000.0 MN0.0 mvH.0 am m cm

noa.o ~mo.o wno.o «no.0 an t mon.o boo.o mwa.o 0q 0 mN

NH0.0 mmo.0 mm0.0 NN0.0 mv < th.0 0N0.0 NOH.O mm x mm

000.0 000.0 0H0.0 mH0.0 on t 0N0.0 N00.0 hv0.0 m0H.0 NH0.0 mc m HN

mm0.0 005.0 mh0.0 H00.0 v00.0 NMH C Hv0.0 000.0 «00.0 mmH.0 mma = 0N

0H0.0 vm0.0 050.0 0v0.0 ON0.0 mm 4 m00.0 00h.0 vm0.0 avH.0 vN0.0 vofl : mfi
A.HHH\.H30 v uUOMCMHH

n2 2: S R an jam mad ooa 2 2” am 3

a:
@9880 .3 035.

157

.Esqunqaazvo oumncamza>oumm Eouu ceaumq>mu new ummuIU Hoo.o w N a«* H0.0 w M ;. no.0 w N a

 

 

 

0H0.0 mmm.0 0m~.0 vmw.0 eva.o NM : 0H
0H0.0 omm.o HvN.0 ovH.0 ”no.0 hmH.0 5N z .Avfia
hm0.0 mom.0 Hma.o >0H.0 vH0.0 mmH.0 mm m Av<05H
mH0.0 mum.0 mon.0 0>0.o “00.0 mm m 0H
000.0 ov0.0 va.0 H>N.0 hma.o HO0.0 ~00.o Hm0.0 HO0.0 vac : .gvmfi
000.0 mmo.0 mHv.0 m-.0 ma0.0 mmfl.o 000.0 moH.0 000.0 ahw : AU<VmH
vmo.o 05v.0 v>~.o hmfi.o mmo.0 «v : vH
A32: m UUchmuB
amn.0 mH~.0 HH0.0 mom.0 00H.0 HH0.0 pv 2 Ma
vvo.o Nav.0 m0~.o ~0a.0 mHH.o cm 2 NH
I I I I I I I I I I a HH
w~0.0 nam.0 wh~.0 oHH.0 000.0 mm m 0H
..Hz. N pummcmua
mHo.o mnn.o ww~.0 om0.o OON.O oo~.o 0c m m
mmo.o nmn.0 whH.0 amn~.0 wo~.o cm a m
ov0.0 00n.0 oom.o ON0.0 0vH.0 0v~.o mw : h
vHo.o Nv0.o pav.0 mnH.0 ~v0.0 -~.0 mmo.0 m00.0 0N0.o 0m 0 w
5N0.0 mwn.0 om~.0 om~.o mva.o pm a m
mHo.0 wwv.0 mm~.0 50H.0 HHH.O 0H0.0 hm z Aavv
0mo.0 00N.0 nnv.0 FH0.0 vma.0 OOH.O on a Ancvv
vv0.0 wwv.0 om~.0 «mmm.0 «v0.0 cm a m
v00.0 nan.0 nv~.0 n-.0 vho.0 he a N
NH0.0 omv.0 hmm.0 Hma.o 050.0 nv a H
..:H\.Hz. H uommcmua
cad MNH NHH I100H MN mm on NM HN H0 aludddluqflfl
«$3
£58 9 53: :50 0082: 9 9:288 9085.: 30 05 £53, comes—S
2d 33m .833 9 0.3 93 332 9 E32 30 628:8 203 333 05
outs 53 22.3 S 05 n— .v new. .< £28m 08:5 E2836? 2: 388
33.853 .x .3 285.58 0.0 c.8532 he «-8.». 50 38593.0 222 .me 033.

158

.Ezaunaafisum oumncdozu>vuom Eouu coauma>mu uOu

um0UIU a00.0 .W N «a; H0.0 W N «a no.0 W N a

 

 

 

Hmo.o .van.o ~¢~.o ~m~.o HNH.o mm = on
omo.o mpv.o m-.o mpd.o m~o.o om = mm
..fiz. .oqu
mno.o pov.o ~m~.o H-.o moH.o no x mm
omo.o owm.o nmo.o pw~.o poH.o ma a an
mmo.o mpn.o pom.o HHo.o oma.o ema.o flao.o «v m on
«no.0 amm.o mnn.o oHo.o HoH.o mmo.o aw a on
..sz m uommcmue
”Ho.o mNo.o mo~.o oo~.o pan.o NvH.o nao.o oo = ow
moo.o mm..o «mH.o .an.o moH.o mm a pm
«vo.o mpv.o nH~.o moo.o mma.o woa.o cm a ow
Hmo.o pwq.o .mmH.o mmH.o mmfi.o oHo.o we x mm
.Hw~.o ~m~.o wm¢.o omH.o MN a «N
~Ho.o ~Ho.o ppv.o oHH.o ~Ho.o H-.o mmfl.o nmo.o me x mm
mao.o puo.o ~an.o .mH~.o wpa.o wpfl.o pm a -
avo.o Nov.o «Ho.o om~.o NH0.0 «mH.o «Ho.o mmo.o He a Hm
moo.o mqo.o moo.o o~¢.o mmm.o moo.o omo.o moo.o mmo.o mmfl m om
moo.o pwo.o mom.o mmn.o Hv~.o moo.o pno.o moo.o ma a aH
..HHH\.03. q uommcmua
cma mwa cod an Ian omI Nm Hm Ha zlquauluuau
«.23
0353 .NV 2%...

159

.EaaunHHszo oumncHo3I>Uum= eouu coHumH>mn “on ummqu Hoo.o w u ... Ho.o w.m .. mo.o w u .

 

 

 

mHo.0 mmo.0 va.o NvN.0 mom.o HOH.0 0H0.o Hm c mH
00v.0 0vH.0 oom.0 00H.0 mN c AAVFH
I I I I I I I I I I I < .05 p H
mHm.0 HNH.0 NvN.o mmN.0 mm < 0H
Noo.o wNo.o «vvm.o sammN.o Hoo.0«wwH.o Hoo.o Noo.0 omH.o Noo.o N00 a .qva
0N0.o >mm.0 000.0 mbH.0 moo.0 mmH.0 Noo.o th < Avcva
HmH.o th.o HH0.0 mvH.0 Nmo.o HmH.0 nv a vH
..Hz. m uoomcmue
m0H.0 mwn.o HNN.0 omH.0 mHH.0 en t NH
omo.o va.o Hom.o omH.o mHH.o Nv c NH
Nv0.0 mov.0 NmN.0 mmH.0 hm0.0 «H0.0 mm 4 HH
nH0.0 Hon.0 mmN.0 mHo.oahnN.o NNH.O on 4 0H
..Hz. m powmcmua
ovo.o 0mm.o oHo.o oom.0 oHo.o ONN.O omo.0 on c m
mmN.o nmN.o mmN.0 omH.0 on 4 m
I I I I I I I I I I I t h
vmo.0 wNn.0 HH0.0.«N»N.0 HH0.0 me.o OMH.0 0v a 0
«No.0 vmn.o mwN.0 vMH.0 ONN.O He a m
pHo.o NH0.0 mNn.o on.o mmN.0 moo.o mN c .qvv
I I I I I I I I I I I t Andvv
nN0.o «mmN.0 hNN.0 th.0 HH0.0 05H.0 we t m
mN0.0 omN.0 mH0.0 th.0 «0MN.0 uHmH.0 en t N
0q0.0 0mm.0 omn.0 00H.0 0vH.0 om < H
A.GH\.HZV H #00MCMHB
0nd. MNH 00H\ «m mu 0m 0w NM AN H0 2 unam uufim
~93
.538 9 5.5: :50 00303 9 9:288 8085.: 55? Bang
Ea 83m .933 9 A30 93 85. o. .808 30 88:8 203 833
05 out: 53 82.3 S 05 n— .v 8.6 ~< 628m BED 5383-28 05
88a §§§R .m .«o means—anon A5 03% he .33.. 8.0 88838.0 03.2 .mV 035.

160

.EaHunHHszw ouancH03I>Uumm Eouu coHuuH>mu now

mammalo .noo.o W N «4... H0.0 M. N a; mo.o W N I

 

 

 

I I I I I I I I I I I t v0

I I I I I I I I I I I < mm
A.H30 .UmHZ

HN0.0 00v.0 hma.o NQH.0 mH0.0 NoH.0 mm < Nm

I I I I I I I I I I I < Hm

mmm.0 NH0.0 mmN.0 NH0.0 NON.0 mHH.0 Nc 4 0m

vN0.0 mnv.0 mmN.0 0H0.0 NNH.O m00.0 He 4 mN
..«3. m uowmcmua

NH0.0 000.0 hHm.0 mmN.0 MMN.0 an.0 00 ( mN

I I I I I I I I I I I 4 FN

0N0.0 vH0.0 moN.0 mmN.0 0>H.0 h00.0 00H.0 Hp C 0N

wv0.0 mvn.0 ONN.O HmH.0 va.0 05 < mN

0H0.0 000.0 mvN.0 0H0.0 nhH.0 0N0.0 mmH.0 mv c VN

0N0.0 0M0.0 va.0 v0n.0 h0H.0 th.0 Hm 4 MN

HH0.0 vvn.0 0hN.0 «00N.0 h0H.0 mo 4 NN

nmn.0 th.0 MH0.0 HvH.0 0N0.0 mHH.0 an t HN

000.0 0N0.0 mmN.0 whN.0 Hv0.0 NNH.O 000.0 00H.0 000.0 MMH 4 0N

m00.0 0H0.0 hNn.0 Nmm.0 00H.0 0H0.0 avH.0 voH t 0H
A.HHH\.N3. v wommcmua

dflfl MNH MHH 00H v0 Q. &n 0n 0“ NH HN H0 33

Néa<
3.83 .9. use.

Chapter 5

Local and Microgeographic Genetic Variation
between Host Races of Rhagoletis pomonella

Introduction

The niche-variation hypothesis (Levene, 1953; Van Valen, 1965;
Levins, 1968) predicts that genetic polymorphism is correlated with the
ability of a population to utilize different environmental resources. If true,
then differential selection in heterogeneous environments should be an
important factor maintaining genetic variation (Levene, 1953; Levins,
1968; Maynard-Smith, 1970) and promoting speciation (Maynard-Smith,
1966) in natural populations.

Niche theory has found wide application in the study of
phytophagous insects due in part to the existence of numerous pest
"biotypes" which attack economically important plant cultivars (see review
by Diehl and Bush, 1984). Differential habitat use is also central to a long-
standing debate concerning modes of speciation. Speciation in animals is
generally thought to require the geographic isolation of populations to
eliminate the homogenizing effects of gene flow (Mayr,l963). However,
Bush (1966, 1969a,b, 1975) has proposed that speciation in certain

phytophagous and parasitic insects occurs sympatrically and is initiated

161

162
when insects shift and become affiliated with new hosts.

Although a number of examples demonsu'ate associations between
particular phenotypic or genotypic attributes of a population and some
aspect of the environment (see reviews by Hedrick et al., 1976 and
McDonald, 1983), documenting the existence of host associated
polymorphism in sexually reproducing insects has been difficult (Futuyma
and Mayer, 1980; Futuyma and Peterson, 1985). First, it must be
established that populations utilizing different hosts are conspecific and not,
in fact, distinct species. Second, characters varying between habitats must
have a genetic basis and not be environmentally induced modifications (this
is not a particularly strict requirement for allozyme or molecular genetic
studies). Third, resource related polymorphism has to be consistent on
both fine and broad spatial scales, as well as being temporally stable, to
distinguish it from variation due to stochastic processes such as founder
effects and genetic drift. Examples of resource associated variation do exist
for sympatric populations (for reviews see Diehl and Bush, 1984 and
Futuyma and Peterson, 1985) but, in almost all of these cases, the
evolutionary and/or geographic origin of the host-insect relationship are
not known.

True fruit flies in the Rhagoletis pomonella sibling species complex
have a biology and natural history which make them ideal candidates for
studies of host associated polymorphism and speciation. Rhagoletis larvae
are obligate internal parasites in the fruit of their host plants and each
member of the R. pomonella group is monophagous or oligiophagous for a
different set of host plants (Bush, 1966). Females lay their eggs directly
into the fruit of the host plant which they identify by specific visual, tactile
and olfactory cues (Bush 1969a,b). Males are attracted by the same cues,

163

and courtship and mating occur almost exclusively on or near the fruits of
the host plant (Prokopy et al., 1971, 1972). Because host recognition and
mate selection are directly coupled in Rhagoletis, variation for host
preference traits can produce genetically based, pre-mating barriers to
gene flow. This reduction in gene flow can result in the evolution of
partially reproductively isolated host races, and eventually species, adapted
to different host plants (Bush, 1966, 1969a,b, 1975).

R. pomonella group species and their host plants overlap broadly
with one another in their distributions across eastern North America (Bush,
1966) and at no time during their evolutionary history was it likely for
geographically isolated fly populations to have existed. This point is most
evident in the recent shift of the species R. pomonella from its native host
hawthorn (Crataegus spp.) to introduced domestic apple (Malus pumila).
Agricultural records indicate a probable origin for the apple infesting fly in
the Hudson Valley, New York and document the subsequent spread of the
apple fly across the eastern United States and Canada over the last 120 years
(Ward, 1866; Walsh, 1867; Bush, 1969a). Hawthorn and apple races of R.
pomonella therefore offer a unique opportunity to study the initial stages of
the divergence process. The historical timeframe of R. pomonella's shift(s)
to apples provides a direct test of whether partially reproductively isolated
"host races" can form sympatrically in the absence of prior periods of
geographic isolation.

The previous studies in this dissertation have shown that sympatric
apple and hawthorn populations of R. pomonella are genetically
differentiated. Significant gene frequency differences exist for six
allozyme loci between sympatric apple and hawthorn populations from
across the eastern United States (Chapters 2, 3, 4). These six loci map to

164

three different regions of the genome (Chapter 1) and linkage
disequilibrium occurs in natural populations between non-allelic genes
within each of these time regions (Chapter 2, 3). The pattern of linkage
disequilibrium is similar for apple and hawthorn flies suggesting that
inversions are present in different frequencies in the two races (Chapters 2,
3). However, gene frequencies for five of the six loci displaying inter-host
differentiation also show significant linear relationship with latitude
(Chapters 3, 4). The slopes of these latitudinal allele frequency clines are
different, however, between hawthorn and apple flies implying that
differential selection occurs both within and between host races across their
ranges (Chapters 3, 4).

Gene frequency clines among hawthorn and apple populations
suggest that host races may not evolve uniformly across their respective
ranges. Differential selection caused by vagaries in environmental
conditions could dramatically shape the local pattern of intra- and inter—
host differentiation. In addition, hawthorn and apple fruits are not
necessarily constant resources in either time or space. Host trees can mast
and it is not unusual for entire sites to be virtually devoid of fruit in certain
years. Mark and recapture studies indicate that adult flies will migrate over
distances as great as 2 km when host fruit is not readily available (Phipps
and Dirks, 1933a, Bourne et al., 1934; Maxwell, 1968). As a result,
founder events and differential migration could drastically affect patterns
of host associated polymorphism for R. pomonella. McPheron et al.,
(1988b) have, in fact, found significant microgeographic allozyme
variation among fly populations infesting different hawthorn trees within
210 meters of one another at a field site in Urbana, Illinois. It is, therefore,

imperative to characterize local and microgeographic levels of

165

differentiation among host populations as a baseline for assessing processes
occurring on broader geographic scales.

The current study addresses the issue of local and microgeographic
genetic differentiation by examining allozyme variation for apple and
hawthorn populations within a 60 km2 area near the town of Grant, in
western Michigan (Newaygo Co.). The results indicate that genetic
differentiation among apple and hawthorn populations are a microcosm of
patterns observed across the eastern United States with inter-host
divergence being a consequence of differences in north-south allele
frequency clines between the two races. Genetic variation is, therefore,
finely subdivided in R. pomonella according to host plant utilization and to
local environmental conditions, a result supporting predictions of the niche

variation and sympatric speciation hypotheses.
Materials and Methods

All flies analyzed in this study were collected as larvae from infested
ground fruit. At three sites (A, E and H; see Figure 9a; Note: Site E is the
original sympatric site analyzed in Chapter 2) in 1987, larvae were
dissected from infested fruit and immediately frozen in liquid nitrogen for
later electrophoretic analysis. In all other cases, infested fruit was
transported back to the laboratory and placed in trays of moist vermiculite.
Fruit from individual trees and from different sites were kept in separate
trays and covered with fine mosquito netting to prevent larvae from
moving between trays. All trays were stored in a greenhouse under
constant temperature (210 C) and light (15:9 hr., lightzdark cycle)

conditions. After two months pupae were sifted from the vermiculite and

Figure 9.

166

Collecting sites for R. pomonella in western Michigan. Panel b
is a diagram of local populations sampled within a 60 km2 area
near the town of Grant, Michigan. 0 = apple tree,

A = hawthorn tree. Panel c gives the distribution of apple and
hawthorn trees at site B. A=apple H=hawthom.

 

 

167

.o 2:5

 

 

 

 

a“. ¢<
I ...8. L a a
n: z. 2.
..m 5N:
o NO<
04 .(o
0:1
4.... u
0(0
‘
a: 2
h(

63 arm 28:; a

 

 

 

 

_a .235

_

[III

 

’«l

 

 

LEN

¢u>_¢ 2009.031

..fl 3N:

 

.. L

 

 

 

. ..m 3'0.

        

 

 

 

 

... .J.

z 2

 

 

168

kept at 4° C for five months to simulate winter diapause conditions. Pupae
were removed from the cold and held at 21° C, 15 :9 L:D to terminate
diapause. Adults were collected when they emerged after 1-3 months of
heating and were frozen at -80° C until they could be analyzed
electrophoretically.

Sympatric hawthorn and apple trees (sympatric being defined as host
plants separated by a distance of less than 0.4 km) were individually
sampled over a four year period at a centrally located site designated E
(Figure 9b,c). Two apple trees and one hawthorn tree at site B were
monitored across all four years of the study. Collections were made from
other trees in years when enough fruit was available to ensure against
population extinction. The year 1987 was exceptional with virtually every
tree fruiting in large quantity at site B, except for hawthorn trees 2 and 3.
Analysis of microgeographic variation was therefore most extensive in
1987.

Local geographic variation was assessed in 1986. Flies from five
hawthorn and five apple trees were sampled from within a 60 km2 area
surrounding site E (Figure 9a,b). Local hawthom and apple sites were not
sympatric but were distributed as uniformly as possible in all directions
around site B. In addition, sympatric hawthorn and apple pOpulations were
collected in 1987 from sites A and H located approximately 30 km north
and south, respectively, of site E (Figure 9a).

Standard horizontal starch gel electrophoretic techniques were used
and are described fully in Chapter 1 and in Berlocher and Bush (1982). We
resolved and scored six different allozyme systems: Malic enzyme (Me),
Aconitase-Z (Acon-Z), Mannose phosphate isomerase (Mpi), NADH-
diaphorase-Z (Did-2), Aspartate amino-transferase-Z (Aat-Z) and

169

H ydroxyacid dehydrogenase (Had). These were the six allozymes which
showed the highest levels of inter-host differentiation in the previous
geographic surveys of genetic variation across eastern North America
(Chapters 3, 4). Isozymes that migrated the nearest to the cathode were
designated 1, the second nearest 2, etc. Alleles were numbered according to
their relative anodal mobility to the common allele (100) at a locus. Tables
50-55, which follow the Discussion section on pages 183-188, list a
majority of the electrophoretic alleles which were resolved in this study.
Only two alleles were scored for the loci Me and Did-2 and so only the
relative mobilities and frequencies for the rarer of the two alleles for these
allozymes are given in Tables 50-55. Three alleles were scored for the
locus Had, but one of these alleles (Had9 7) was very rare and so only
frequencies for the second most common allele (Hadlzz) are presented in
Tables 52, 53 and 55. A Mendelian mode of inheritance has been
established for all six of the loci analyzed in the study (Chapter 1) and
electromorphs migrated identically in larval and adult samples.

Genotype frequencies were examined for deviations from Hardy-
Weinberg expectation by G-tests. Alleles were pooled, if required, so that
all genotypic classes had expected frequencies >1. The Levene correction
(Spiess, 1977) was applied when sample sizes were < 100 or when it was not
possible to pool alleles to make the expected frequency for all genotypic
class >1. Allele frequencies were analyzed for significant heterogeneity
between host plants and across time and space using G-contingency tests.
Allele frequencies were pooled, when necessary, so that each cell in the G-
contingency test had a frequency 2 5. F-statistics were calculated by the
method of Weir and Cockerham (1984) with standard deviations estimated

by jackknifing over populations, years or loci. Corrections for unequal

170

sample sizes in the method of Weir and Cockerham can result in negative
FST values. In these instances we report FST as 0. One of the F-statistics, f
(the correlation of genes within individuals within populations 2 F13 of
Wright, 1951) was determined independently for each population from
allele frequency data (see Barker et al., 1986). Multi-allele f values were
tested against the null hypothesis of f=0 by a Chi-square, single-degree of
freedom test (Barker et al., 1986).

Results

Results are presented sequentially for either one locus or two loci at a
time. We first discuss the microgeographic pattern of differentiation for a
gene(s) on a tree by tree basis at site B and then follow with a description of
genetic variation among local R. pomonella populations in the surrounding

area.

m2

Significant allele frequency differences between apple and hawthorn
flies were observed for Acon-Z on both local and microgeographic scales.
At site B, Acon-Z displayed significant inter-host differentiation in each of
the four years of the study but showed no significant intra-host variation
within any year (Tables 44, 50, 51). Gene frequencies for Acon-Z were
heterogeneous for hawthorn flies across years at site B (Table 45).
However, intra-host temporal variation neither overlapped nor was of
similar magnitude as inter-host differentiation (Tables 50, 51). F-statistics
quantify this point. The FST value for Acon-Z calculated for both host
populations over years at site B (0.0604 _-l; 0.0136, jackknife estimate of
standard deviation calculated over years; Table 46) was 6 x greater than the

171

3...:me— H
N: .0
303.

200.0
*AHv0.v

......AHVHfi
add

.m 2%. 8 088388: 88:80 22? $8-82: 08 ..EE é 98. 88388-0

5mm
.63

......ACHd
2.32013

.934.
N84

...:vadm
8:.
:0:

3900.2
...:AVVNHN
$.va

800.2
«and—

.....:..€: .5
AvaN

Nana

......wAandH
.....SANVm.nH
LNVHS

*9.va H
000;

.....SANVVNM
N: .H

.33

-- ....A :3 SEA 8; 2&3.

-- ......8...$ 93:; use 2%...

3:88 -- 38-2%? season

53 38 ea... 5: $82.
61o .104.” $2... 29:. mace... £2

I I 882$ out. 203.

I I mBaIN RE. 2%?

I I 28:; 8F 2%?

I I m 8:. Ba:-mo_&<

.. I N 8.0. 3am-8_&<

I I H 08,—. Bam-8_&<

......a: .3. ......czu 8 33-2%... $028

65.2 a; .o ”8:. 2.: $8....
602 a: .m “8:. 2&3. «:82 22

I I m3amN out. 20.?

I I 33:; 02... 29?.

Inseam .....c:..£ 33-2%... 5253
58 38 see 2%... ass... 8%
.3.qu 42. 8H 8H
.3. 29¢

172

.3 ooh 32-33% .06 .8:

85¢ 2.. Ba 39 3.8g 2: 35%.“ £89 3: .3 3: a .5 30:38 0.83 was won-Hefi .38. a BM 3388

83 bmoaow820= .85-g5 .505:me =25, ABE-3&5 c823 rod 38. 32.-BE He now: 293 £89

@289 805 9a mug “no: $an 3.08 0.83 was “8..-8.5 .§o£&_m-=o: “8 «omega—gob o_o=< 385539
E :03» 0.3 $8. ..8 880on no 39500 . .8. .8.—$3.890 Sod .v- m .3... :3 w M ...... 36 .v- m ...

 

83 -- 82. -- -- ...-.2: as was: 3:. ea?
32. -- ES -- -- ...-ism? aim-c 3:- 2%?
33 -- 83 -- -- ...-2:3 WEE-m. 8:. 2%?
82 -- 8c... -- -- ...-€08 aim-v 8:. ea?-
83 -- :82- -- -- ...-Let:- mgm-m 8: 293
Ste -- 32 -- -- 3.2: .9. mafia 3: 293
1.35.2 I :...CK.£ I I trucvedm 95534 02H 039‘
I ...:Amvcfim I :LCQVN ...:AmV—dom I Balbia< 503.05

63“ 8:2 53 53 $3 52 m3:- 3: ”85¢-
...365.m~ 8:2" ......va.8 65: £83 ...-Lem? m3; 2%? ”:82 23.

3.258 .3 03$.

173

:2 55358-0 :5de ...: 85 mm ...... 2:. w m ...

 

.230: .33 non $.56 AnvnS ...Aevoda $33. I
55 A252 8:3 59w ......8:_.n~ .....va9 < 8% .82 «:82

L00.” ANS~.3 ...:Amvvdfi ...:GV—AN «......vadm .138:de I
3va ...:ANGmfim 3va 3:6 33.2 356 < m 03mins?» mmob<
add Maud Mag .33 «Maud d2 add duH

“8055.3.“ 5 :03» 2a 33 Be 803 ..o 82on .mcousaom Ev F8532 98 A3 03%
.82 macaw can m 3% a who.» 38% £05388: >So=§ 30:“ .8.—-555 Bu v.33 aocowaucoaé .3 03¢

174

datum—g 060:0» 12882 65 605883 .8.. :00an 860825 05 .6 36853262 28205
8a 83? .8.. 506 gens—=28 26 E 68205 93 meg Egg“: 65 03% 53 5:3 82? ,5”. 05 Ba 98: :86 o

68— .85 03868 0.630% - 863306 6.3655 n

.30» ~05 25:68 0.630% - 59656 6.36536 «

 

6366 H 2.86
6866 H 69666
6~666 H 66666
2666 H $.86
356 H «~86
6966 H 3666
mm~66 H 66.66
final

6366 H #6666
3666 H «866
6866 H 3666
66666 H 6866
«6666 H 9266
3666 H 66666
6366 H 6366
3

6866 H 66666
6866 H 3666
n 566 H n 566
3666 H 66666
5366 H 556
3666 H 9666
5266 H 8666

«6666 H 2666
N366 H N866
6N666 H 3666
6866 H 3666
3666 H 6666.6
3666 H 2666
6866 H 2666

6866 H $666 66866 H 3666

6266 H $666
3666 H 3666
6266 H 66666
66666 H 6666.6
3666 H 2666
2666 H 6 6666

g “2un :6 N mud” unamd

6866 H 6666.6
N866 H 3666
66666 H 66666
3666 H 6666.6
2666 H :666

«me—66 H :86
_ uflH uamd

.m 27.

8 moo: E9632 6.3 03% .56 E3» 828 9.66: 8568.80 65 60>» ..o 3508 05 .3 Esau—3 32? pm”—

_8_ =<
Ex
$2
35
a:
N-:Ou<
0:

good

9 23.

175

value for hawthorn flies alone (0.0099 _-t_- 0.0095) and over 50 x greater than
the PST for apple flies (0.0015 i 0.0025).

The allele frequency differences found between apple and hawthorn
flies at site B suggest that differential selection may be occurring between
the host populations. However, if selection occurs for Acon-Z, its effects
were not reflected in genotypic deviations from Hardy-Weinberg
expectation within host populations. Only 3 out of a total of 96 tests showed
significant deviations for Acon-Z when individual host trees were analyzed
within years (Tables 50, 51). In addition, individual tree f values at site B
showed no trend for either an excess or deficiency of heterozygotes (Table
47).

Inter-host differentiation for Acon-Z was also consistent across the
12 local apple and hawthorn populations analyzed in 1986. The highest
frequency for the allele Acon-295 in any apple population was 0.268 at site
D (Table 54). In comparison, the lowest frequency for Acon-295 for a
hawthorn population was 0.392 at site B (0.392 represents the pooled Acon-
295 frequency among hawthorn trees at site B in 1986 and also happens to
be the lowest frequency observed for hawthorn flies at site B for all four
years of the study, the next lowest frequency being 0.479 in 1987; Table
54). Nevertheless, allele frequencies for Acon-Z were still significantly
different in 1986 between apple flies at site D and hawthorn flies at site B
(G-contingency test = 6.3, B 5 0.05, 2 df).

Acon-295 allele frequencies among local hawthom populations fit
within a broader geographic pattern of latitudinal clines observed across
the Midwest (Chapter 4). Local hawthorn populations I through M had
Acon-295 frequencies that were bracketed within the range of values found
for hawthorn flies at site A (0.619) and site H (0.392) located

176

Table 47 . Disu'ibution of f values (: Fm of Wright, 1951) for
individual populations calculated by the method of
Barker, East and Weir (1986). The number of f values

significantly greaterorlessthanOaregivenin

parentheses.

 

119.911.513.929

Me 16
Acon-2 10(1)
Mpi 19(2)
Dia-2 12(1)
Aat-2 9(1)
Had 10

All loci 76(5)

8 8(3) 4
14(1) 8 4

5 9(1) 3
12 9 3
15(2) 3 9(1)

14(2) 8 4(1)
68(5) 45(4) 27(2)

24(3)
18(1)
28(3)
21(1)
12(1)
18
120(9)

125.0111 1.3.0ng

12

18(1)
8

15

24(3)
18(3)

96(7)

 

177

approximately 30 km north and south, respectively, of site B (Table 54,
Figure 9). However, Acon-295 allele frequencies in 1986 still fluctuated
between 0.571 at site L and 0.392 at site B. Latitudinal clines for Acon-295
were, therefore, not perfect across local hawthorn populations.

Acon-Z also showed significant intra-host variation among local
apple and hawthorn populations (Table 45 ). For example, the frequency of
Acon-275 was 0.250 in hawthorn population K compared to 0.038 at
hawthorn site I (Table 54). The same was also true for local apple
populations where Acon-Z 75 varied in frequency from 0.205 to 0.026
(Table 54). Despite the significant amount of local variation for Acon-Z,
only Amt-295 at site I showed a significant deviation from Hardy-
Weinberg equilibrium for this locus out of a total of 88 tests (Table 54).
Also, local intra-host heterogeneity for Acon-Z did not overshadow inter-
host differentiation. The F31 values among local apple (0.0103 i 0.0074;
Table 48) and hawthorn populations (0.0127 :1; 0.0098) were approximately
4 x lower than the value across all host populations (0.0472 i 0.0184).

11:311de

Me and Mpi also displayed consistent microgeographic
differentiation between apple and hawthom flies at site E (Table 44).
However, unlike Acon-Z, significant intra-host variation was observed for
Me and Mpi within years. Me varied significantly among apple trees at site
B in both 1986 and 1987 (Table 44). However, inter-host allele frequency

differences for M6” were never less than 0.100 and were usually greater

than 0.200 between any chosen pair of hawthom and apple trees within a
given year (T ables 50, 51). PST values for Me of 0.1233 :1: 0.0313 across

 

Table 48.

Locus

Me
Acon-2
Mpi
Dia-2
Aat—2
Had
All loci

178

F31- values for local apple and hawthorn populations calculated
by the method of Weir and Cockerham (1984).

_ADRL._
0.0185 1 0.01598

0.0103 1- 0.0074
0.0093 1 0.0086
0.0084 1: 0.0067
0.0018 3 0.0060
0.0000 x 0.0036

0.0071 i 0.0032b

_flamhm—

0.0000 i 0.0043
0.0127 :1; 0.0098
0.0000 i 0.0018
0.0202 1; 0.0180
0.0801 i 0.0598
0.0092 :9; 0.0147
0.0321 i 0.0230

m

0.0406 i 0.0141
0.0472 1 0.0184
0.0214 i 0.0169
0.0384 i 0.0177
0.0427 1 0.0354
0.0038 : 0.0049
0.0395 3 0.0046

 

a Standard deviation- jackknife estimate over sites.

b Standard deviation- jackknife estimate over loci.

9 Both hosts are F31 values when both local apple and hawthorn populations
are included in the calculations. Both host values therefore represent the
interaction between host associated and spatial genetic variation.

179

.8» 555 m 3% 8 5635.058 088» “8:83 .«o 83:68 0338 0255 38: 56m 6
.28» 5.23 count? 063% 0268838? 3.2.8.5 .«e .08. 05 ©6536 83? 9mm E6533 65 3%? o

.82 8>e 895.8 06503 - 56386 Essa a
.8» 8>o 3258 £65.82 - 56386 68653 a

 

 

an 56 H 63.66 2666 H 2666 2866 H 3666
856 H 6266 3666 H 66666 6366 H 6686
«.366 H 3666 3666 H 3666 9666 H 6866
$86 H 3.86 3.86 H 3666 6266 H 886
66666 H 9266 #866 H 6866 6866 H 9666
«~56 H 936 ”N666 H 66666 3666 H 386
6686 H 3.86 @366 H 8666 66666 H 3.666
Handmade! Inga: landfill
gm—

 

mNN66 H 6636 9.666 H 9666 63666 H «6666
2.56 H 386 @866 H 66666 6366 H on 86
2.86 H «N56 3666 H 8666 6266 H @6666
386 H 386 6966 H 3666 386 H N366
:86 H 6686 @366 H 3666 2.666 H 66666
386 H 63.66 356 H $666 6866 H 66666
286 H mug—6 6866 H 8666 «6986 H N686
35%| lasagna Ieauaqolt
www—
.Avwg 6

eafioxooo e5 no; do 858. o... .3 Beanie 5% cs 22 5 m 2e. a noon 2%: one .8532 ace 8.? a".

3 5.
on:
22
25
a:
«-83.
oz

3&4

6v 035.

180
all host trees at site B in 1985 (Table 49) and 0.0844 :1; 0.0269 in 1987

document the substantial amount of inter-host differentiation for this locus.

Mpi displayed significant heterogeneity among apple trees at site B in
1986 and among hawthorn trees in 1985 (Table 44). However, uner
Acon-Z and Me, apple and hawthorn populations overlapped slightly in
their gene frequencies for Mpi in 1987. Hawthorn tree #4 had a frequency
for Mpi100 of 0.860 in 1987 while apple tree #4 had a frequency of 0.858
(Tables 50, 51). Within no other year was the frequency of Mpi100 greater
for a hawthom tree than for an apple tree. In fact, the pooled Mpi] 00 gene
frequency for hawthorn trees at site B was still significantly different from
that for pooled apple trees even ill 1987 (G-test = 24.2, 13 5 0.001, 1 df;
Table 44). However, the magnitude of inter-host variation was much less
for Mpi than for either Acon-Z or Me as evidenced by the lower FST values
calculated for Mpi across both hosts in 1985 and 1987 (0.0390 -_l-_ 0.0171 and
0.0145 1; 0.0089, respectively; Table 49).

Gene frequencies for Me and Mpi at site B varied significantly across
years for hawthorn flies but not for apple flies (Table 45). The result was
indicative of a general trend in the partitioning of intra-host differentiation
at site B. Hawthorn flies showed greater temporal heterogeneity than apple
flies. The overall Fs'r value across years for all six loci was almost 5 x
higher for hawthorn populations (0.0084 i. 0.0040; Table 46) than apple
populations (0.0017 i 0.0002). In contrast, apple flies displayed higher
levels of spatial differentiation within years at site B. This was exemplified
by the 5 x larger Fs'r values for all loci among apple trees in 1985 and 1987
(0.0062 i 0.0051 and 0.0064 :1; 0.0031, respectively; Table 49) compared
to hawthorn trees (0.0015 i 0.0043 and 0.0013 _-l_- 0.0013).

181
Local differentiation for Me and Mpi was similar to the pattern of

microgeographic variation observed for these loci at site B. Hawthorn
populations I through M showed limited spatial heterogeneity for Mpi and
Me in 1986 (Fs'r = 0.0000 for both loci across hawthorn sites; Table 48).
In comparison, apple populations displayed higher levels of local spatial
diversity (FST Me = 0.0185 1: 0.0159, Fs-r Mpi = 0.0093 i 0.0086). In fact,
significant allele frequency heterogeneity existed among local apple
populations for Me (G-contingency test = 16.4, E g 0.01, 5 df; Table 45).
Inter-host differentiation among local host populations, while not as
great as that observed for Acon-Z, was also apparent for both Me and Mpi.
Fs'r values for Me and Mpi across all 12 local host populations were over 2
x greater than the values calculated for apple populations alone (Table 48).
Mpi allele frequencies were homogeneous among both local apple and
hawthorn populations (Table 45), while pooled hawthorn and apple Mpi
frequencies were significantly different from each other (G = 33.4, B _<_
0.001, 1 df). Local apple and hawthorn populations overlapped in Mpi100
allele frequencies, however. Apple sites B and C both had Mpi100
frequencies below that of the highest hawthorn population at site I (Table
54). Mpi100 frequencies for local sites therefore formed a continuum,
with apple populations tending to have higher frequencies than hawthorn
populations. A similar pattern was also observed for Me”. However,
unlike Mpi, local differentiation for Me can be partly explained by gene
frequency clines. All apple populations north of site B had M e80 allele
frequencies 5 0.560 while all sites south of E had Me80 frequencies _>_ 0.630
(Table 54). Apple site B fluctuated in frequency between 0.578 and 0.671
from 1984 to 1987. Local hawthorn sites displayed similar clinal
tendencies with northern populations generally having lower Me80

182
frequencies than southern populations (Table 54). Inter-host differences

for Me were therefore superimposed on clinal patterns of variation, with
apple flies having higher Me80 allele frequencies than hawthorn flies at
every site where sympatric populations were sampled (sites A, E and H;
Table 54).

Genotypic proportions for Me and Mpi showed general trends for an
excess of heterozygotes at both the local and microgeographic levels.
Twenty four and 28 out of a total of 36 f values were negative (indicating an
excess of heterozygotes) for Me and Mpi , respectively (Table 47 ). Three f
values out of 36 were significantly less than 0 for both Me and Mpi.

Aat-Z and Die-2 had roughly equivalent patterns of genetic
differentiation as Mpi. Both loci showed significant microgeographic
variation among apple trees at site B within 2 of the 4 years surveyed (Table
44). For the two years when intra-host gene frequencies were
homogeneous for Die-2 or Act-2, significant inter-host differences were
observed (Table 44). However, frequencies for Did-285 and Act-2100
overlapped between apple and hawthorn trees in both 1985 and 1987
(Tables 52, 53). A spectrum of gene frequencies was therefore also seen
for Dia-Z and Aer-2 at site B. Hawthorn trees tended to reside on one end of
the frequency distribution and apple trees on the other with certain trees
(apple trees # 2, 7 and hawthorn tree # 1, 5) overlapping in the middle of
the spectrum (Tables 52, 53). Aat-Z was unusual in that it was the only
locus which varied significantly across years for apple flies at site B (Table
45). Did-2 displayed the normal temporal pattern showing heterogeneity
across years for hawthorn but not for apple populations (Table 45).

183

Die-2 and Aat-Z had the highest Fs'r among the six local hawthorn
populations (0.0202 _+_ 0.0180 and 0.0801 :1; 0.0598 respectively, Table 48).
Intra-host variation for Aat-Z among local hawthorn populations was
actually greater than inter-host levels observed for Me and Acon-Z (Table
48). For example, Ant-250 varied in frequency from 0.256 to 0.083
between hawthorn sites I and M while Ant-221 changed from 0.0191 to
0.050 between sites L and M (Table 55).

Inter-host differentiation still existed for both Aat-Z and Dia-Z,
however, as rim-2100 and Die-2100 gene frequencies were generally
higher in local hawthorn than apple populations (Table 55). In fact, pooled
Did-2 allele frequencies across hawthorn and apple sites were significantly
different from one another in 1986 (G-contingency test = 23.4, E _<_ 0.001, 1
df). However, just like Mpi] 00, inter-host frequency differences for Aar-
2100 and Did-2100 were not discreet and overlap was found among local
hawthom and apple populations. Die-2100 did not display a local pattern of
variation within either host race that was indicative of a latitudinal
frequency cline. Aat—Z 75 did show a hint of a frequency cline among local
apple populations (Table 55), yet the allele did not co-vary significantly
with latitude in a broader geographic survey of apple populations from
Michigan through Indiana (Chapter 4).

Aat—Z was unusual in that it was the only locus which displayed a
trend for an excess of homozygotes (positive f values; Table 47 ). Four f
values for Act-2 were significantly different than 0 out of a total of 36 tests
(Table 47). Genotype frequencies for Aat-Z therefore deviated from

equilibrium proportions more often than expected by chance.

184

lied
Had showed intra-host variation at both the local and

microgeographic levels but only the slightest trace of inter-host
differentiation. Had varied significantly among apple trees at site B in 1984
and 1987 as well as between hosts in 1985 (Table 44). Allele fiequencies
for the locus were heterogeneous for hawthorn flies across years at site B as
well as among local hawthorn populations in 1986 (Table 45). However,
Had displayed lower levels of inter-host differentiation between local
hawthorn and apple populations than any of the other five loci analyzed in
the study (Fs'r Had = 0.0038 i 0.0046; Table 48). In addition, no
geographic pattern of variation was discemable for Had allele frequencies.
Had f values were equally divided between positive and negative estimates
(Table 47).

Discussion

The results of this and the preceding studies confirm that partially
reproductively isolated host races of phytophagous insects can evolve in
sympatry by shifting and adapting to new host plants. R. pomonella
' populations are genetically finely subdivided according to the species of
host plant they infest; a result which agrees with predictions of the niche
variation hypothesis. Even for different species of host trees within 100
meters of one another at site B, Memo and Accra-295 allele frequencies are
significantly higher for hawthorn than apple flies. Inter-host
differentiation is also stable over time. Although allele frequencies for
Me100 and Acon-295 did vary significantly for hawthorn flies at site B

over a four year period, intra-host variation did not compare ill magnitude

185

or overlap in frequency with inter-host differences. Me and Acon-Z are in
fact, closely linked on chromosome II in R. pomonella (Chapter 1) and
disequilibrium has been found between non-allelic genes for these two loci
in natural fly populations (Chapters 2, 3). The similar pattern of
microgeographic variation for Me and Acon-Z is therefore not surprising
and suggests that selection in this region of the genome is causing genetic
hitchhiking effects among linked loci.

Genetic studies of R. pomonella suggest that differences in host
recognition and host associated survivorship are two important factors
differentiating fly populations infesting hawthoms and apples. Mark-
recapture studies and larval transplant experiments would help determine
the extent to which these factors reduce gene flow between the two host
races. The fine level of inter-host divergence revealed in this Chapter
indicates that reproductive isolating mechanisms can be reliably examined
even for local R. pomonella populations. Evidence already exists indicating
that adult flies do not randomly migrate between different species of host
plants. Adults eclosing from under and collected directly from off of
hawthorn and apple trees at site E showed the same pattern and magnitude
of inter-host divergence as larvae sampled from ground fruit (Chapters 2,
4). R. pomonella therefore tend to utilize the same species of host plant as
adults which they infested as larvae. We do not know whether inherent,
genetically based differences in host recognition and/or conditioning in
adults (Prokopy et al., 1982) cause this bias ill host choice but mark-
recapture and larval transplant experiments could be designed to
specifically address this issue.

The association between mate and habitat selection for R. pomonella

host races has important implications for the modeling of non—allopatric

186
speciation. The theoretical work of Felsenstein (1981) has often been cited

to support the view that sympatric speciation is a rare phenomenon in
nature (Futuyma and Mayer, 1980; Butlin, 1987). However, in
Felsenstein's model individuals randomly select habitats, an assumption
which is not valid for R. pomonella group flies and potentially many other
phytophagous and parasitic insects (Price, 1980; Diehl and Bush, 1984).
More recent theoretical studies (Bush and Diehl, 1982; Rice, 1984; Diehl
and Bush, 1989) indicate that sympatric populations can differentiate
rapidly when mating is confined to specific habitats chosen by the animals.
These models suggest that non-allopatric divergence is as much, or more
dependent, on the fidelity of habitat related assortative mating than on the
sheer intensity of differential selection within habitats. This consideration
has not been adequately appreciated in our understanding of how
heterogeneous environments maintain genetic diversity and promote
speciation.

An interesting implication of my work is that host races may not
evolve uniformly across their ranges. Gene frequency clines, although not
perfect, were observed for Memo and Acon-295 even across the limited
geographic area surveyed in this study. Inter-host differentiation for Me
and Acon-Z is, therefore, superimposed on north-south clinal patterns of
intra-host variation. Besides Me and Acon-Z, however, none of the other
four loci analyzed, with the possible exception of Aat-Z for apple flies,
showed a distinct pattern of clinal variation among local populations.
However, Mpi, Dia-Z and Had do display latitudinal allele frequency clines
when analyzed on a broader geographic scale across the midwestem United
States (Chapter 4). For a number of reasons discussed in Chapters 3 and 4,

secondary contact is an unlikely explanation for the observed frequency

187

Clines. Instead, environmental factors affecting R. pomonella appear to be
experienced ill a graded fashion from north to south, with the two races
being differentially affected at a given latitude. Environmental gradients
are rarely exact, however, and vagaries in local environmental conditions
can obscure broader geographic patterns of genetic variation. This could
account for why allele frequencies for Mpi, Did-2 and Had were not
correlated with latitude among local populations despite showing clinal
patterns of variation across the Midwest.

Differences in micro-environmental conditions could also explain
the pattern of intra-host variation at site B. Although apple flies at site B
had higher levels of intra-host variation within years than hawthorn flies,
hawthorn populations were more variable across years. This suggests that
differences exist in the environmental "buffering capacities" or "selective
constraints" imposed by the two species of host plants, with flies attacking
hawthoms being more acutely affected by environmental fluctuations than
flies infesting apples. Buffering capacities may vary more among
individual apple trees than hawthorn trees, however, accounting for the
higher levels of intra-host variation for apples within years. Indirect
support for this hypothesis comes from the fact that apple tree #1 at site B,
the most variable apple tree across years, also deviated the most in allele
frequencies from the other apple trees within years (Tables 50-53). In
addition, egg to pupal survivorship is known to fluctuate widely among
different varieties of apples (Phipps and Dirks, 1933b; Reissig, 1979), with
flies surviving better on earlier, sweeter varieties. Reciprocal egg
transplant experiments, among different apple and hawthorn trees at site B,

are needed to directly test the environmental buffering hypothesis.

188

The discussion has thus far concentrated on the importance of
deterministic factors in influencing the genetic structure of fly populations.
However, stochastic processes undoubtedly interact with selection to shape
local and micro geographic pattems of genetic differentiation. Aat—Z
represents the best example of this. High levels of intra-host variation exist
for Ant-221, Aat-250 and Act-2 75 among local hawthorn populations. The
pattern of differentiation for these alleles appears to be random among sites
which suggests the action of genetic drift.

Aat~2 gene frequencies are not totally dictated by chance, however.
Allele frequencies for Act-2100 are generally higher in hawthom than
apple populations (Tables 52, 5 3, 55). In addition, genotypic frequencies
for Aer-2 reveal a significant trend for an excess of homozygotes (Table
47) and deviate from Hardy-Weinberg equilibrium more often than would
be expected by random chance (Table 47). The excess of homozygotes for
Aer-2 could be due to selection, errors in enzyme scoring caused by a silent
alle1e(s), positive assortative mating within host populations or a Wahlund
effect resulting from population subdivision on a fine scale. The latter
possibility seems unlikely because none of the other five loci analyzed had a
surplus of homozygotes (Table 47). Population sub-structuring should
affect neutral genes in an equivalent manner. Consequently, the lack of a
significant excess of homozygotes at loci besides Aat-Z argues against a
Wahlund effect.

Positive assortative mating within apple and hawthonl populations
based on Aat-Z genotypes is also unlikely. Did-2 is closely linked to Aat—Z
(Chapter 1) and non-allelic genes for these two loci show substantial
amounts of linkage disequilibrium in natural populations (Chapters 2, 3).

If non-random mating were occurring relative to Aat-Z, its effects should

189

also be apparent for Dia-Z. However, Die-2 had an excess of heterozygotes
thereby making assortative mating an improbable scenario.

Little information is currently available on the frequencies of null
alleles in Rhagoletis populations. Null alleles appear to be relatively rare in
natural populations, however (Langley et al., 1981). In addition, no
individual was observed to be homozygous for a null allele out of a total of
2,729 flies scored in the study. Although we cannot completely rule out the
possibility of null alleles, selection at or near the Aat-Z locus is the most
probable explanation for the excess of homozygotes.

The loci analyzed in this study therefore appear to represent a
continuum in terms of the inter-action of evolutionary processes. Me and
Acon-Z form one side of the spectrum with allele frequencies for these two
loci primarily determined by selection. In contrast, genetic drift may play
a larger role in influencing local patterns of variation for Aer-2 and Had.
There is no reason to assume, however, that selection acts directly on M e,
Acon-Z, Mpi, Die-2, Act-2 or Had. The map distances of these six loci to
genes actually under differential selection must, therefore, also be
considered when interpreting the relative importance of stochastic versus
deterministic processes.

To quantify the interplay between natural selection and genetic drift
requires detailed analyses of the distribution and movement of adult flies.
First, estimates of effective population sizes are needed for local hawthorn
and apple populations. Second, we must determine whether adults confine
their reproductive activities to a single tree (perhaps the same tree they
infested as larvae) or several different trees at a site. Finally, neighborhood
sizes must be established for R. pomonella which require estimates of

migration rates among local populations. Previous work indicates that R.

190

pomonella adults can travel over distances of up to 2 km (Phipps and Dirks,
1933b; Boume et al., 1934; Maxwell, 1968) but that most flies disperse less
than 60 m from release plants (Neilson, 1971; Buriff, 1973; Reissig, 1977).
Maxwell (1968) and Reissig (1977) both report the proportion of apple
flies remaining on a release tree as approximately 10%. This implies that
adult flies distribute their reproductive capacities among a number of
different apple or hawthorn trees. If true, then either differential selection
on larvae or differential fecundity among adults must be responsible for the
observed microgeographic genetic variation among apple trees at site B.

In addition to being spatially patchy, host resources are also
temporally variable. Partial or complete fruit failure can dramatically
affect population structure. For example, in mast years entire local sites
can be virtually devoid of host fruit. On a microgeographic scale,
hawthorn trees #2 and #3 at site B usually had large fruit loads. However,
these two trees bore no fruit in 1987 . Adult flies emerging from under
hawthorn trees #2 and #3 in 1987 were, therefore, forced to disperse in
search of suitable host fruit. Such forced migrations would greatly increase
mixing within and among local fly populations and could account for the
general trend for an excess of heterozygotes for loci in this study (Table
47). Ill addition, adult flies may not randomly recolonize hawthorn trees
#2 and #3 in subsequent years. Founder events could therefore play an
important role in affecting the pattern of genetic differentiation at site E.

The diapause characteristics of R. pomonella may ameliorate some of
the effects of mast years, however. In both the field and the laboratory, a
certain proportion of pupae (ca. 10%) require exposure to more than one
cycle of chilling and heating before eclosing as adults (Phipps and Dirks,
1933b). Adults emerging after two winters would tend to homogenize gene

191

frequencies over years for trees experiencing fruit failure provided, of
course, that the loci analyzed are not involved either directly or indirectly
with diapause termination. This later assumption may not be true,
however, as preliminary studies indicate a relationship between allele
frequencies for Me, Acon-Z and Had and the timing of adult eclosion for the
hawthorn race (Feder et al., manuscript in preparation). Continued
monitoring of hawthorn trees #2 and #3 would help tease apart how
diapause and founder effects influence gene frequency distributions within
and among fly populations.

The results of this study indicate that local and microgeographic
genetic variation for hawthorn and apple flies are a microcosm of the
patterns observed across eastern North America. Detailed examination of
local demes will therefore provide useful information concerning the
processes involved in host specialization and population divergence.
Further studies are needed to discern exactly how selection, drift and
migration interact to produce the fine level of genetic subdivision observed
for R. pomonella. Perhaps foremost, we must identify the nature and mode
of action of selective factors differentially affecting apple and hawthorn
populations.

Previous studies indicate that ambient temperature has an important
influence on the phenology of host-parasite interactions for R. pomonella
(Prokopy, 1968; Reissig et al., 1979; also see Chapter 4). Temperature is a
key factor in the initiation, maintenance and termination of diapause in
these flies (Prokopy, 1968; Reissig et al., 1979) and, of course, has a direct
effect on the phenology of fruit maturation for host plants. Apple and
hawthorn flies do, in fact, differ in the timing of adult eclosion with apple
flies emerging consistently earlier than hawthom flies in both the

192

laboratory and the field (Diehl, 1984; Smith, 1986, 1988; Feder er al.,
manuscript in preparation). Temperature also correlates with latitude,
which could account for the observed north-south frequency clines in R.
pomonella. The temperature/development hypothesis therefore requires
serious consideration and further empirical testing.

Positive correlations between genetic diversity and environmental
variation have been documented for populations both in the laboratory
(Powell, 1971; McDonald and Ayala, 1974; Powell and Wistland, 1978)
and in the field (Clarke, 1968; Berger, 1971; Dobzhansky and Ayala, 1973;
Kettlewell, 1973; Taylor and Powell, 1976; Saul et al., 1978; Mully et al.,
1979; Steiner,l979; Franklin, 1981; Barker et al., 1986; Lynch, 1987). The
finding of significant allozyme variation between host associated races of R.
pomonella indicates that in addition to environmental heterogeneity in time
and space, biotic considerations related to host-parasite interactions can also
promote and maintain genetic polymorphism in populations. Furthermore,
because R. pomonella is an economically important pest species, much is
known about the fly's ecology, behavior and natural history (see review by
Boller and Prokopy, 1976). This information, combined with the recent
formation of several new host races in the species (Herrick, 1920; Bush,
1966; Shervis er al., 1970; Prokopy and Bush, 1972; Prokopy and
Berlocher, 1980) makes R. pomonella a unique organismal system for
understanding evolutionary processes within a real time framework. In
addition, many of the sibling species in the R. pomonella group, although
distinct in nature, can be hybridized in the laboratory to produce viable and
fertile F1 progeny. Experiments can therefore be performed to elucidate
the genetic basis of traits responsible for restricting gene flow among R.

pomonella flies. Comparisons can then be made to determine whether

193

factors affecting gene frequencies within R. pomonella host races are
directly involved in the development of reproductive isolation between R.
pomonella group species.

194

.coaunuoomxo muoncfiozthvumm Scum COHumw>ov unmoflmwcmflm

How umeIO HO0.0 W Nee; H0.0 .W Nee no.6 tvt Ne

 

 

 

 

ummm.o meo.o oho.o rao.o HHH oNH.o mmm.o hmH.o mao.o omo.o mod bmw.o HHH pfl

Ham.o hmo.o moo.o mm hmo.o mmH.o mmm.o ema.o omo.o «mo.o mm mmm.o Hp m4

mmm.o muo.o pmo.o ow pmo.o Hmm.o hmo.o omo.o rmo.o mm mpo.o ov mm

omo.o mho.o mmo.o omo.o FNH voo.o mva.o omm.o mHN.o «vo.o abo.o mNH mum.o «ma ¢¢

oom.o mco.o omo.o hoo.o or vma.o «Hm.o >ma.o omo.o pwo.o pm mvm.o or md

mom.o aeo.o vmo.o mmo.o flfla moo.o cma.o mpm.o mo~.o wflo.o omo.o HHH mhm.o ONH N4
mmm.c ~mo.o mmo.o >oo.o Hva moo.o hhH.o nmam.o oH~.o mmo.o mwo.o mma mqm.o «ma H4 no

~mm.o Hmo.o meo.o «Ho.o mma wma.o mmm.o mpa.o Hmo.o NoH.o voo.o mma wmm.o MNH m<
omm.c mmo.o emo.o ooo.o mpa moo.o Nea.o amm.o mmH.o mmo.c pmH.o NPH arm.o ova Ha mm

omc.o ovo.o cwo.o oao.o on ova.o cam.o o-.o omo.o omo.o oHo.o om mam.o Hm m4

aoo.o mom.o omo.c moo.c mm moo.o maa.ceu¢ov.o mma.o pmo.o maa.o mm pom.o mm N4
oom.o oeo.o omc.o mm ooH.o oov.c o¢~.o omo.o oma.o mm omv.o mm ad mm

moo.o vom.o omo.o mvo.o moo.o mm bao.o -H.o Nam.o H-.o omo.o cho.o mm Hmm.o em ma
moo.o mmm.o mmo.c mmo.o moo.o mm oao.o NmH.o mcm.o ~ma.o moo.o mmo.o mm bob.o mm Ha «a
mud cod on mm Hm 6H ml «Ha mod cod Ha mm mp mp Z on 2 uuHH H»

a: «.83. a:
3% 0388 u 2 8665.386 8b 93% 86305 523 m an me 899% a “8 co 0.5mm”—

oom .5: 9 £2 82.. m 2e. a 828:8 82. 3 2...: .3333 he a: e8 28% no: he essence 223. .on osfi

195

.cOHumuoomxo muoncwoslavunx Bonn coaunfiwon 96603.25?

How unmulo .866 W Ne: .86 W New no.6 V: Ni

 

 

 

 

000.0 e000.0 000.0 0H..0 000.0 mm. 000.0 000.0 000.0 m00.0 000.0 0N0.0 med 00m.0 N0. 00
0m0.0 000.0 H00.0 m~0.0 00 ~00.0 00m.0 H00.0 000.0 000.0 00 mem.0 00 mm
000.0 000.0 .00.0 000.0 on 0.0.0 000.0 00~.0 0.0.0 0H..0 0.0.0 .0 000.0 em 0:
000.0se0N0.0 000.0 00H.0 0N0.0 mm. 000.0 000.0 000.0 000.0 H00.0 hm0.0 0m. 00m.0 00H .m 00
m00.0 HH0.0 000.0 H00.0 0N0.0 on. 000.0 000.0 000.0 000.0 000.0 000.0 00. 000.0 00. N:
000.0 «00.0 000.0 -H.0 m~0.0 00. 000.0 000.0 H00.0 000.0 000.0 00. m~0.0 00. a: 00
000.0 000.0 000.0 m0..0 «0 0.0.0 000.0 0m~.0 000.0 0.0.0 0.0.0 um mmm.0 an 0:
0.0.0 0.0.0 000.0 0.~.0 mm 0.0.0 0N0.0 m0~.0 000.0 000.0 000.0 0.0.0 mm 00~.0 mm «x
0m0.0 m00.0 000.0 000.0 0.0.0 000.0 mm 000.0 e0m~.0 000.0 000.0 000.0 00 00~.0 mm H: mm
000.0 000.0 000.0 00..0 000.0 .0. 000.0 0.6.0 00~.0 000.0 0m0.0 000.0 .0. ~0~.0 .0. H: 00
mud 00H 00 em Hm ma 21 «AH 00H 00. mm 00 we 00 z. 00 21 mama 4»
.02 2.80. as
.83 2683 u 2 86606286 8... E6532 86365 666.3 m 6.8 ..6 80.9% 0 .6.. 6... Beam
8m .52 6. $3 86.... m 6.6 .0 66.62.66 82. :6 E6530: 306.266. 666 RE 6:0 ”:66... .0: .66 8.666062”. 222. ..n 630,—.

196

.cowueuoomxo muonewmzlaoumx Scum cowu0H>o6

0083.030 no. 83-6 .000 w a...

H6.6 W N00 66.6 W N0

 

 

 

80.0 6.6800 .1. 2 86606686 8... 2660 86265 6023 m o... 66 80660.6 0 .6.. 66 2066.
86 .066. 6. 066— 82.. m 3.0 .0 66.8666 82. 606 6.660 3.6566. .6.. 66: 6:0 N030 .~.6..Q 666 8.080096 0.05.

b<
6d
m<
66
md
N<
ad

N<
HG

Md
N4
ad

NC
ad

66

66

66

v6

N6H.6 baa 666.6 hav.6 66N.6 66H.6 666.6 66H.6 vHH 6>N.6 HHH
«mmH.6 H6 666.6 666.6 th.6 666.6 666.6 66H.6 66 666.6 66
66H.6 6m 6H6.6 666.6 6N6.6 6Nm.6 «HN.6 H66.6 mm dam.6 hm
66N.6 66H 666.6 0666.6 00H6N.6 666.6 66H.6 vma.6 666.6 66H 66N.6 66H
66N.6 66 6H6.6 66N.6 666.6 H6N.6 66H.6 66 666.6 66
66H.6 NHH 666.6 6H6.6 666.6 0006Hm.6 «66.6 666.6 H6H.6 666.6 6HH 6m~.6 NNH
666.6 66H 6H6.6 66N.6 66N.6 th.6 666.6 6hH.6 66H 666.6 66H
66H.6 MNH 666.6 «No.6 666.6 NDN.6 666.6 66H.6 666.6 666.6 66H.6 666.6 MNH hH6.6 MNH
6HN.6 66H 6N6.6 Hbm.6 «066N.6 666.6 66H.6 66H.6 666.6 66H 666.6 66H
vhN.6 H6 6N6.6 666.6 v6N.6 6H6.6 mvN.6 6H6.6 66H.6 6H6.6 6N6.6 66H.6 6H6.6 H6 666.6 66
N6H.6 66 6H6.6 6H6.6 6H6.6 666.6 6H6.6 66N.6 666.6 666.6 6H6.6 66H.6 N6 N6H.6 66
66N.6 6N 6N6.6 6N6.6 666.6 666.6 66N.6 6NH.6 6N HFN.6 6N
666.6 66 666.6 6H6.6 6N6.6 666.6 666.6 . 6HH.6 666.6 666.6 66H.6 66 666.6 66
666.6 66 6N6.6 6h~.6 666.6 666.6 6H6.6 66H.6 6H6.6 66H.6 666.6 66 Nam.6 66
NNHIIIIZI 66H MNH MHH 66H Mm 66 66 NM .HN. H6 ZI IfldlllZl uuHH MN
60: «.36. «.05

.N6 630,—.

197

.cowunuoaxm mugcwmznauumm Scum 60333.6 unmoflmfigfin now .3976 36.6 w an: 3.6 w m: 66.6 w m...

 

N6H.6
NHN.6
N6H.6
H6H.6

66H.6
66H.6

6NH.6
MNH.6
6HH.6

vNH.6

 

N6H 666.6 666.6 H6v.6 66N.6 ova.6 666.6 666.6 66H
66 H66.6 666.6 «N6m.6 6Ha.6 666.6 66H.6 66
66 666.6 «va.6 «66N.6 6HH.6 666.6 H6
66H HH6.6 N66.6 666.6 666.6 66N.6 6HH.6 666.6 666.6 and
66H 6N6.6 666.6 v6N.6 6NH.6 6H6.6 666.6 66H
66H HN6.6 666.6 666.6 66a.6 666.6 66H.6 66H
N6 6H6.6 666.6 6H6.6 666.6 666.6 66H.6 6N6.6 666.6 66
66 6H6.6 6Hv.6 666.6 H6H.6 666.6 66
66 6H6.6 666.6 th.6 6N6.6 6N6.6 bma.6 6H6.6 666.6 H6
H6H 666.6 666.6 N6N.6 666.6 666.6 666.6 66H.6 666.6 Hoa
MNH 66H 66 66 Nm HN

NM" 2 GMH

6HN.6
H6N.6
66H.6
66H.6

moN.6
NHN.6

66H.6
MNH.6
moN.6

66H.6

N6H 6m
66 6:
pm 63
66H H:

66H Nm
66H Hm

N6 mm
mm Na
66 Hz

66H Hm

 

66:

Néa<

.86.“ 2689. u 2 25636606 8.6 6.5532 8632: 606.3 m 86.“ 66 8.26.66 6 .86 06 956$ 36
.662 Q 662 826 m 3a a 683—30 82. $6 59632 366565 .56 66: 65 N466. 6-6.5 .56 86:33.6 20.2

66

66

66

66

2| 66 z uuHH Hfi

N55

.66 035.

198

.cowumuommxm oumnc603|>6umm Scum

coauow>mv ucmofiwficmfln Mom ammunw H00.0 w N «*«

H0.0 .W M a; 600.0 W N «

 

 

 

 

665.6 666.6 66H.6 666.6 6N 550.6 666.6 N66.6 6Ho.o N5H.6 Nm H66.o mm
666.6 6N6.6 6NH.6 6H6.6 66 656.6 666.6 H66.6 666.6 666.6 H6 666.6 66
665.6 ooa.o oNH.o oNo.o 6N mvo.o 666.6 H56.o 666.6 666.6 66 666.6 Nv
666.6 5H6.6 5vo.o 666.6 666.6 5H6 666.6 656.6 6cm.o 656.6 N6o.o 6No.o 6N6 666.6 566
666.6«666.6 566.6 666.6 6No.o 56N 666.6 N5o.o 666.6 Nmm.o 666.6 N6o.o 56N H¢¢.o 66N
666.6 N65.6 666.6 66H.o 666.6 65H 666.6 666.6 56N.o 666.6 N6o.6 666.6 666.6 66H 66N.o 66H
066.6 665.6 656.6 65H.o 666.6 Hoa 666.6 «HH.6 56N.6 06v.6 666.6 666.6 HoH N5N.o Hoa
665.6 666.6 v6a.o mHo.o 6m 66N.o H¢v.6 moa.o 66N.o 66 666.6 6v
666.6 6No.o N6H.6 NH6.6 06 6No.6 6Nm.6«666.6 6HH.6 666.6 66 «666.6 N6
666.6 Nao.o NNH.6 NH6.6 av 6No.6 656.6 666.6 N6o.6 566.6 66 6N¢.o oc
NHO.6 665.6 666.6 vHN.6 NH6.6 N6 NH6.6 66N.6 6H6.6 506.6 NH6.6 N6 6NN.6 No
666.6 666.6 666.6 5N 56H.o mv6.o ooa.o 666.6 56H.o mm 665.6 66
HH6.6 avo.o 666.6 6H6.o 6m 66N.6 666.6 6NH.6 ”66.6 66N.6 mm 6N5.o N6
666.6 N66.o NH «No.6 6HH.6 666.6 N6N.6 NHo.o 666.6 H6 666.6 N6
N66.o 666.6 666.6 mHo.o 666 666.6 66H.6 666.6 66H.6 6No.o 666.6 666 6N6.o 665
«HH6.6 H¢o.6 666.6 oao.6 Nom N66.6 66H.6 H66.6 66H.6 ”66.6 NNa.o Noo.o 66N HN6.6 mmN
566.6 656.6 666.6 656.6 «66.6 66H «66.6 oNH.o 666.6 6oN.o 666.6 mHH.o 666.6 56H 656.6 6NH
666.6 666.6 666.6 H66.6 666.6 66H 6H6.6 666.6 666.6 v6a.o 666.6 666.6 66H H56.o 66H
666.6 666.6 56H.6 NH6.6 Nv NH6.6 66H.o 666.6 66N.6 ¢No.o H66.o NH6.6 av 666.6 N6
666.6 NHH.6 6 6No.o 66H.6 656.6 5mN.6 N66.6 6No.o 6H «a666.6 6N
6N6.o ¢Ho.o vaa.c 666.6 66 66H.o H6¢.6 66N.o NHo.o 666.6 «6 NH6.o 56
6H6.o 666.6 N¢o.o Nmo.o 5c 66H.6 666.6 66H.o 666.6 ooa.o 66 666.6 66
6NH 66H 65 56 N6 2! «NH 66H 66H 66 66 65 65
a: «-53.. a:

56
66
66
56
66
66
v6
66
66
66
56

56
66
66
56
66
66
v6
66
66
66
56

.05 2683 n 2 286666806 2% 563m
56 6.66 056$ 36 £566.32— 50 52632. 65 A3 2666 682 86 .3: 65 Né8< .6: 86 memo—8:626 22?

6666666662231
CHUXMNMMr-‘sz

fid‘fitficd‘dfiidd
flmUQfxlf-dldlflkawm

2| 66 z HduN.uudn.uHfiw

.66 0366.

 

 

 

.606umuomaxw muwnc6mzn>Uum6 Scum :06um6>ov u660666cm6n uom unmuuo 666.6 M N«*« 66.6 w_m«« 66.6 w_m«
666.6 66 666.6 666.6 666.6 656.6 666.6 66 666.6 66 56 6 6
666.6 66 566.6 666.6 «566.6 666.6 666.6 66 66N.6 66 66 6 2
666.6 N6 6N6.6 666.6 66N.6 666.6 666.6 66 66N.6 66 66 6 A
656.6 666 666.6 666.6 666.6 666.6 65N.6 5N6.6 N66.6 666.6 666.6 666 66N.6 566 56 6 6
656.6 66N 666.6 6N6.6 566.6 66N.6 666.6 666.6 666.6 N66.6 56N 66N.6 66N 66 6 m
6N6.6 656 566.6 6N6.6 666.6 666.6 666.6 566.6 666.6 666.6 666.6 666 566.6 666 66 6 6
666.6 666 666.6 666.6 N6N.6 666.6 666.6 666.6 666.6 666.6 666 666.6 666 66 6 6
566.6 66 N66.6 566.6 566.6 556.6 556.6 66 N66.6 66 66 6 6
66N.6 66 6N6.6 5N6.6 «666.6 N66.6 666.6 666.6 66 566.6 66 66 6 6
666.6 N6 666.6 666.6 656.6 666.6 66N.6 6N6.6 66 6N6.6 66 66 6 6
566.6 N6 6N6.6 656.6 65N.6 566.6 666.6 N6 666.6 66 56 6 a
N6N.6 66 666.6 6N6.6 N6N.6 66N.6 66 666.6 66 56 6 6
N6N.6 N6 66N.6 66N.6 666.6 N66.6 56N.6 66 N66.6 N6 66 6 w
m”«6NN.6 N6 6N6.6 «666.6 666.6 65N.6 N6 656.6 56 66 6 m
11 66N.6 666 N66.6 6N6.6 666.6 66N.6 666.6 666.6 666.6 N66.6 666.6 N66.6 N66 666.6 N56 56 6 6
666.6 666 666.6 5N6.6 666.6 «56N.6 666.6 666.6 666.6 666.6 666.6 666.6 66N 666.6 66N 66 4 m
66N.6 666 666.6 5N6.6 666.6 666.6 N66.6 N6N.6 666.6 666.6 666.6 N66.6 666.6 666.6 6N6 66N.6 566 66 6 6
666.6 666 666.6 NN6.6 666.6 666.6 666.6 666.6 5N6.6 666.6 666.6 656.6 666.6 666 65N.6 N66 66 6 6
666.6 N6 N66.6 66N.6 666.6 666.6 666.6 N66.6 66 66N.6 66 66 6 a
66N.6 6N NNN.6 666.6 566.6 566.6 6 666.6 66 66 6 U
«NNN.6 66 5N6.6 66N.6 666.6 «66N.6 «66N.6 56 666.6 66 66 6 6
666.6 66 666.6 566.6 666.6 666.6 N66.6 666.6 56 666.6 56 56 6 6

N66 2 666 6N6.\\666 666 66 65 66‘ 66 66 N6 6N 66 2| 66 6 MN Adan «Add

86 $3 «.66

.366 26866 n 2 6666666686 3?.

.8.: E 36 2:66 86 2822.8 66 .5532 65 36 26% 6.8. a. 36 65 «.36 6.6.5 .8 “382.606 222 .66 226.

Chapter 6

Concluding Summary and Remarks

"One would think that it should no longer be necessary to devote
much time to this topic, but past experience permits one to predict
with confidence that the issue will be raised again at regular
intervals. Sympatric speciation is like the Lemaean Hydra which
grew two new heads whenever one of its old heads was cut off.
There is only one way in which final agreement can be reached and
that is to clarify the whole relevant complex of questions to such an
extent that disagreement is no longer possible".

13. Mayr (1963, pg. 451)

In this dissertation, I have shown that apple and hawthorn populations
of R. pomonella are genetically differentiated host races which arose in
sympatry. I first reported the observation of significant inter-host allele
frequency variation for six allozymes (Me, Acon-Z, Mpi, Dia-Z, Aat-Z and
Had) at a field site near the town of Grant, in western Michigan. A number
of alternative, but not mutually exclusive, causal explanations were present
which could account for the genetic difference. Most importantly,
however, adults captured directly from off of host trees showed a similar
pattern of inter-host variation as larvae reared from host fruits. Adult flies

are therefore not randomly migrating between different species of host

200

201

plants and, as a consequence, hawthorn and apple flies are not randomly
mating in nature. Differential host utilization is therefore a key factor
restricting gene flow between the host races; a finding consistent with
predictions of sympatric speciation models in which host race formation
represents the initial stage of the divergence process.

I next verified a Mendelian mode of inheritance and determined map
positions for the six allozymes displaying allele frequency differences
between the host races. Did-2 and Aat-Z were shown to be separated by a
distance of 3.2 centimorgans on linkage group I, while Me, Acon-Z, and
Mpi were found to be tightly linked on chromosome 11. Had was localized
on linkage group III. The six loci showing differentiation between apple
and hawthorn populations therefore mapped to only three specific regions
of the genome. A genetic basis of inheritance and linkage associations were
also determined for three other previously unresolved allozymes. One of
these three loci, Aspartate amino-tran.sferase-I (Aat-I ), was found to be sex
linked and indicated for the first time that males are the heterogametic (XY)
sex in R. pomonella. In addition, significant gametic disequilibrium was
observed between pairs of linked, non-allelic genes in the laboratory strain
used for the genetic mapping experiments. The pattem of linkage
disequilibrium for non-diapausing flies was different from that seen in
natural populations, however, suggesting that the laboratory line has
become genetically modified during culture.

A geographic study of allozyme variation was then conducted for
hawthorn and apple populations from across the eastern United States and
Canada. The results showed that host associated genetic divergence is
widespread across the range of R. pomonella and is primarily due to the

same six loci which varied at the Grant, Michigan site. Although genetic

202

variation was much less among apple than among hawthorn populations
across eastern North America (consistent with the apple race having gone
through a recent bottleneck), the pattern of geographic differentiation gave
no genetic clue as to the original source of the apple race. In addition, high
levels of linkage disequilibrium were observed between loci within, but not
between, each of the three genomic regions displaying host related
differences. The patterns of disequilibrium among loci were similar,
however, between apple and hawthorn flies. Therefore, although at least
three specific regions of the genome of R. pomonella are co-evolving , the
relationships among genes within these regions are the same for the two
host races. Consequently, it is just the frequencies of certain gene or
chromosomal "blockS" which differs between apple and hawthorn
populations; a result which suggests that inversion polymorphisms exist for
these regions and are promoting hitchhiking effects by reducing
recombination among linked loci. Furthermore, epistatic interactions were
not found between the different chromosomal regions displaying racial
variation. Taken together, these data indicate that host race formation and,
by implication, speciation, in the R. pomonella group involves genetic

' changes at only a few key loci, rather than a "revolution" involving a
restructuring of genetic interactions throughout the genome.

Latitudinal allele frequency clines were also found among both apple
and hawthorn populations for a majority of the allozymes displaying inter-
host variation. The slopes of the allele frequency clines were different for
apple and hawthom races, however, being much steeper among hawthorn
than apple populations. Inter-host genetic differentiation is therefore
superimposed on clinal patterns of variation within the host races such that

the magnitude of host associated divergence is a function of latitude. The

203
allele frequency clines suggest that the host races are not experiencing

uniform selection pressures across their ranges but instead may be strongly
influenced by local environmental conditions. The configurations of the
allele frequency clines do, in fact, implicate ambient temperature as an
important environmental parameter affecting host-parasite interactions for
R. pomonella. Allele frequencies for two loci (Aconitase-Z and Malic
enzyme) in the hawthorn race correlated with irregularities in isothermal
clines for growing degree days (base 500 F) in the states of Wisconsin and
Michigan. Furthermore, allele frequencies changed most dramatically for
both apple and hawthorn populations across three areas in the Midwest that
mark major ecological transition zones. Variation in voltinism and host
plant associations are now known for a number of different phytophagous
insects across these three regions indicating that environmental changes
within these areas have a general impact on the evolution of life history
traits for insects.

Finally, I examined local and microgeographic genetic variation
among and between apple and hawthorn populations within a 60 km2 area
surroundinggthe original field site near Grant, Michigan. The results
indicated that local and microgeo graphic genetic differentiation for these
flies are a microcosm of the patterns seen across eastern North America.
Even for trees separated by distances of less than 100 meters at the Grant
site, Malia Enzyme] 00 and Aconitase-295 allele frequencies were always
significantly higher for hawthom than apple flies every year over a four
year period from 1984 to 1987. The two loci also showed a similar pattern
of inter-host differentiation across 12 local populations analyzed in 1986.
However, the local pattern of genetic variation was not identical to that seen

at the Grant site as allele frequencies for Me and Acon-Z showed a general

204

correlation with latitude even across the limited area surveyed in the study.
Although genetic drift and founder effects undoubtedly shape the fine scale
spatial and temporal pattern of genetic variation for R. pomonella,
differences in host recognition and host related survivorship had to be
invoked to account for the maintenance of genetic differentiation between
apple and hawthorn races. The data therefore agreed with the niche
variation hypothesis that differential selection in heterogeneous
environments promotes genetic polymorphism and lent further support to a
sympatric mode of divergence for R. pomonella.

Several important questions still need to be answered concerning
speciation in the R. pomonella group, however. First, we need to have a
better understanding of the factors involved in differentiating R. pomonella
populations. This requires establishing the extent of and mechanisms
responsible for host plant fidelity and host associated survivorship,
determining exactly how temperature acts as a differential selective factor,
and estimating levels of effective gene flow among and between apple and
hawthorn populations across the range of R. pomonella in North America.
Second, we must more accurately quantify the extent and pattern of genetic
divergence between host races and sibling species in the group. At least
three genomic regions show host associated allele frequency differences.
Are these the only regions of the genome containing genes responsible for
the formation of the apple race or are more, as yet unidentified, regions
involved? Do sibling species in the group show greater amounts of genetic
variation in the same regions of the genome which display racial
differences or are species specific mutations randomly scattered throughout
the genome? What is the nature of genetic change within genomic regions

differing among the host races and sibling species (do base pair

205

substitutions, additions or deletions predominate?) and what phenotypic
traits do these regions encode? High levels of linkage disequilibrium were
found between non-allelic genes within the three regions of the genome
displaying allele frequency differences between apple and hawthorn flies.
Is linkage disequilibrium common throughout the genome of these flies or
is disequilibrium confined only to those specific regions diverging between
the host races and sibling species? Are inversion polymorphisms more
prevalent in those regions of the genome displaying linkage disequilibrium?
Third and finally, we have to demonstrate that host associated traits, such as
differential host recognition, are directly responsible for the development
of reproductive isolation among higher taxa in group. In other words, we
must show that host races are, in fact, incipient species which may have
reached the evolutionary point of no return. An important point mentioned
in this regard is whether host fidelity behaviors are constant across the
distribution of a host race. If they are and barring extinction, then host
races should be considered to be potential species.

As a post-doctoral fellow at Princeton University and later as a
faculty member at a major research University, I plan to continue work on
the R. pomonella group through a series of experiments addressing the
issues raised above. Field experiments will be performed to measure levels
of intra- and inter-host gene flow and to identify and characterize specific
phenotypes related to host recognition and host associated development that
are restricting gene flow between apple and hawthorn populations.
Molecular genetic studies are currently underway with Dr. Martin
Kreitrnan of Princeton University to construct an extensive restriction site
polymorphism linkage map for R. pomonella. The development of such a

map will let us comprehensively screen the genome of these flies for genetic

206

differentiation. In addition, the linkage map will provide a large number of
genetic markers evenly distributed across the genome thereby making it
possible to resolve potentially polygenic traits related to host preference
and development, into discrete and mappable Mendelian factors. The
feasibility and power of this strategy has recently been demonstrated by
Paterson et al. (1988) working on the domestic tomato, Lycopersicon
esculentum. Dr. Eric Lander of the Whitehead Institute and Harvard
University, a co-researcher on the tomato project, has agreed to lend his
expertise and computer programs in a collaborative effort with Dr.
Kreitman and myself to accomplish similar phenotypic mapping
experiments for Rhagoletis. What makes the R. pomonella group
particularly amenable to genetic studies of speciation is that several pairs of
sibling species in the complex, although genetically distinct in nature, can be
hybridized in the laboratory to produce viable and fertile F1 progeny which
can be backcrossed to either parental type. It will therefore be possible to
determine whether genetic changes affecting phenotypes responsible for
host shifts also form the underlying basis for reproductive isolation among
higher taxa in the group.

By combining a variety of different experimental strategies,
Rhagoletis workers will eventually reach their long term goal of
understanding the ecological, behavioral and genetic basis for host race
formation and speciation in the R. pomonella group. The unique biology
and natural history of the R. pomonella group therefore make the flies an
ideal model system for bridging the supposed gap between "micro-
evolutionary" changes occurring at the population level and the "macro-

evolutionary" process of speciation.

List of References

List of References

Barker, J.S.F., P.D. East and BS. Weir. 1986. Temporal and
microgeo graphic variation in allozyme frequencies in a natural
population of Drosophila buzzatii. Genetics 112:577-611.

Bateson, W. 1922. Evolutionary faith and modern doubts. Science 55 :55-
61.

Bazykin, AD. 1965. On the possibility of sympatric species formation.
Bull. Moscow Soc. Nat. Biol. Div. 70:161-165.

Bazykin, AD. 1969. Hypothetical mechanisms of speciation. Evolution
23:685-687.

Berger, EM. 1971. A temporal survey of allelic variation in natural and
laboratory populations of Drosophila melanogaster. Genetics
67:121-136.

Berlocher, S. H. 1976. The genetics of speciation in Rhagoletis (Diptera:
Tephritidae). Ph. D. Thesis. University of Texas at Austin.

Berlocher, S. H. 1980. Genetics of allozymes of the apple maggot. J.
Heredity 71: 63-67.

Berlocher, S. H. 1984a. Segregation and linkage of allozymes of the walnut
husk fly. J. Heredity 75 :392-396.

207

208

Berlocher, S.H. 1984b. Genetic changes coinciding with the colonization
of California by the walnut husk fly, Rhagoletis completa. Evolution
38: 906-918.

Berlocher, SH. and G.L. Bush. 1982. An electrophoretic analysis of
(Diptera:Tephritidae) phylogeny. Syst. 2001. 31:136-155.

Berlocher, SH. and DC. Smith. 1983. Segregation and mapping of
allozymes of the apple maggot fly. J. Heredity 74:337-340.

Boume, A.I., W.H. Thies and RR. Shaw. 1934. Some observations on
long distance dispersal of apple maggot flies. J. Econ. Ent. 27:352-
355.

Boller, BF and R.J. Prokopy. 1976. Bionomics and management of
Rhagoletis. Ann. Rev. Ent. 21:223-246.

Brues, GT. 1924. The specificity of food plants in the evolution of
phytophagous insects. Am. Nat. 58:127-144.

Buriff, CR. 1973. Recapture of released adult apple maggot flies on
sticky-board traps. Environ. Entomol. 2:757-758.

Bush, G.L. 1966. The taxonomy, cytology and evolution of the genus
Rhagoletis in North America (Diptera: Tephritidae). MCZ,
Cambridge, Mass.

Bush, G.L. 1969a. Sympatric host race formation and speciation in
frugivorous flies of the genus Rhagoletis (Diptera: Tephritidae).
Evolution 23:237-251.

Bush, G.L. 1969b. Mating behavior, host specificity, and the ecological
significance of sibling species in frugivorous flies of the genus
Rhagoletis (Diptera: Tephritidae). Am. Nat. 103:669-672.

Bush, G.L. 1975a. Modes of animal speciation. Ann. Rev. Ecol. Syst.
6:339-364.

209

Bush, G.L. 1975b. Genetic variation in natural insect populations and its
bearing on mass-rearing programmes. pp.9-18. In: Controlling Fruit
Flies by the S terile-Insect Technique. International Atomic Energy
Agency. Vienna, Austria.

Bush, G.L. and R.W. Neck, Jr. 1976. Ecological genetics of the
screwworrn fly, Cochliomyia hominivorax (DipterazCalliphoridae)
and its bearing on the quality control of mass reared insects.
Environ. Entomol. 5:821-826.

Bush, G.L., R.W. Neck Jr., and GB. Kitto. 1976. Screwworm eradication:
Inadvertent selection for noncompetitive ecotypes during mass
rearing. Science 193:491-493.

Bush, G.L. and GB. Kitto. 1979. Research on the genetic structure of
wild and laboratory strains of olive fly. In: F .A.0. Report,
Development of Pest Management Systems for Olive Culture
Programs. F.A.O. of the United Nations. Rome. 27pp.

Bush, G.L. and SR. Diehl. 1982. Host shifts, genetic models of sympatric
speciation and the origin of parasitic insect species. pp. 297-306.
In: Insect-Plant Relationships. Proceedings of the 5th Intemational

symposium. (J .H. Visser and A.K. Minks, eds.) Pudoc, Wageningen,
Holland.

Butlin, R. 1987. A new approach to sympatric speciation. Tree 2:310-311.

Caesar, L. and WA. Ross. 1919. The apple maggot. Ontario Dept. Agr.
Bull. 275.

Carson, H. 1959. Genetic conditions which promote or retard the

formation of species. Cold Spring Harbor Symp. Quant. Biol. 24:87-
105.

Carson, H. 1965. Chromosomal morphism in geographically widespread
species of Drosophila. pp. 503-531. In: Genetics of Colonizing

species. (H.G. Baker and G.L. Stebbins, eds.) Academic Press, New
York.

210
Carson, H. 1968. The population flush and its genetic consequences. pp.
123-137. In: Population Biology and Evolution. (R.C. Lewontin,
ed.) Syracuse University Press, Syracuse, New York.

Carson, H. 1971. Speciation and the founder principle. Stadler Genetics
Symposia 3:51-70.

Carson, H. and A.T. Ohta. 1981. Origin of the genetic basis of colonizing
ability. pp. 365-370. In: Evolution Today, The Proceedings of the
Second International Congress of Systematic and Evolutionary
Biology. (G.G.E. Scudder and J .L. Reveal, eds.). Hunt Institut. for
Botanical Documentation, Camegie-Mellon Univ., Pittsburgh, Pa.

Chapman, PJ. and SE. Lienk. 1971. Tortricid Fauna of Apple in New
York. Cornell Univ., Ithaca, New York. 122 pp.

Christiansen, RB. and O. Frydenberg. 1974. Geographic patterns of four
polymorphisms in Zoarces viviporus as evidence of selection.

Genetics 77:765-770.

Christiansen, EB. and M.W. Feldrnan. 1975. Subdivided populations: A
review of the one- and two-locus deterministic theory. Theor. Pop.
Biol. 7:13-38. -

Clarke, B. 1968. Balanced polymorphism and regional differentiation in
land snails. In: Evolution and Environment. (E.T. Drake, ed.) Yale
Univ. Press, New Haven, Conn.

Clegg, M.T. and R.W. Allard. 1972. Patterns of genetic differentiation in
the slender wild cat species, Avena barbata. PNAS 69:1820-1824.

Cockerham, CC. and B.S. Weir. 1977. Digenic descent measures for
finite populations. Genetical Res. 30:12-147.

Correll, D.S. and MC. Johnston. 1970. Manual of the Vascular Plants of
Texas. 1881 pp. Texas Research Foundation, Renner, Texas.

211

Darwin, C. 1859. On the origin of species by means of natural selection, or
the preservation of favoured races in the struggle for life. John
Murray, London, England.

Dean, R.W. and PJ. Chapman. 1973. Bionomics of the apple maggot in
eastern New York. Search Agric. Entomol. Geneva No. 3.

Diehl, SR. 1984. The role of host shifts in the ecology and speciation of
Rhagoletis flies (Diptera:Tephritidae). Ph. D. Thesis. Univ. of
Texas at Austin.

Diehl, SR. and G.L. Bush. 1984. An evolutionary and applied perspective
of insect biotypes. Ann. Rev. Entomol. 29:471-504.

Diehl, SR. and G.L. Bush. 1989. The role of habitat preference in
adaptation and speciation. In: Speciation and I ts Consequences. (D.
Otte and J. Endler, eds.) Sinauer Assoc.

Dobzhansky, Th. and FJ. Ayala. 197 3. Temporal frequency changes of
enzyme and chromosomal polymorphisms in natural populations of
Drosophila. PNAS 70:680-683.

Economopoulous, AP. 1980. SIRM against the olive fruit fly: Differences
between wild and lab-reared (normal or sterilized) insects. pp. 17-
26. In: Proceedings of XVI. Int. Congress of Entomology, Symp. on
Fruit Fly Problems. Kyoto and Naha, Japan.

Economopolous, AP. and GA. Zervas. The quality problem in olive flies
produced for SIT experiments. pp. 357-368. In: Proceedings of Int.

Symp. on Fruit Flies of Economic Importance. Athens, Greece.
1982.

Endler, J .A. 1977. Geographic variation, speciation, and Clines. Princeton
University Press, Princeton, New Jersey.

212

Feder, J .L. and G.L. Bush. Differential host utilization is an effective pre-
mating barrier between sibling species in the Rhagoletis pomonella
group. Evolution (submitted).

Feder, J .L., C.A. Chilcote and G.L. Bush. Are the apple maggot,
Rhagoletis pomonella and the blueberry maggot, R. mendax, distinct
species? Implications for sympatric speciation. Entomol. exp. appl.
(In press).

Felsenstein, J. 1976. The theoretical population genetics of variable
selection and migration. Ann. Rev. Genet. 10:253-280.

Felsenstein, J. 1981. Skepticism towards Santa Rosalia, or why are there so
few kinds of animals. Evolution 35:124-138.

Felsenstein, J. 1988. Phylogenies from molecular sequences: Inference and
reliability. Ann. Rev. Genet. 22:521-565.

Femald, ML. 1950. Gray's Manual of Botany. Eigth Edition. 1632 pp.
American Book Company, New York.

Fisher, R.A.F. 1930. The Genetical Theory of Natural Selection.
Clarendon, Press. Oxford, England.

Franklin, LR. 1981. An analysis of temporal variation at isozyme loci in
Drosophila melanogaster. pp. 217-236. In: Genetic Studies of
Drosophila Populations. Proceedings of 1979 K ioloa Conference.
(J .B. Gibson and J .G. Oakeshott, eds.). Australian Natl. Univ.,
Canberra.

Futuyma, DJ. and G.C. Mayer. 1980. Non-allopatric speciation in
animals. Syst. 2001. 29:254-271.

Futuyma, DJ. and SC. Peterson. 1985. Genetic variation in the use of
resources by insects. Ann. Rev. Entomol. 30:217-238.

213

Gasperi, G., A. Malacrida, M. Tosetti and R. Milani. 1986. Enzyme
variability: A tool for investigating the genome organization and the
population structure of Ceratitus capitata. pp. 153-161. In: Fruit
Flies of Economic Importance 84. (R. Cavalloro, ed.) A.A.
Balkema, Rotterdam, Netherlands.

Gould, SJ. and N. Eldredge. 1977 . Punctuated equilibrium: The tempo and
mode of evolution reconsidered. Paleobiology 3:115-151.

Harnrick, J .L. and R.W. Allard. 1972. Microgeographic variation in
allozyme frequencies in Avena barbata. PNAS 69:2100-2104.

Harris, H. 1966. Enzyme polymorphisms in man. Proc. Roy. Soc. Ser. B
164:298-310.

Hedrick, P.W., M.E. Ginevan and ER Ewing. 1976. Genetic
polymorphism in heterogeneous environments. Ann. Rev. Ecol.
Syst. 7:1-32.

Herrick, G.W. 1920. The apple maggot in New York. Cornell
Univ. Agr. Exp. Sta. Bull. 402:89-101.

Huettel, MD. and G.L. Bush. 1972. The genetics of host selection
and its bearing on sympatric speciation in Procecidochares
(Diptera: Tephritidae). Entomol. Exp. Appl. 15:465-480.

Illingworth, J .F. 1912. A study of the biology of the apple maggot
(Rhagoletis pomonella), together with an investigation of
methods of control. Cornell Univ. Agr. Expt. Stat. Bull.
324:129-187.

Istock, C.A. and W.G. Weisberg. 1987. Strong habitat selection and the

development of population structure in a mosquito. Evol. Ecol. 1:
348-362.

J aenike, J. 1987. Genetics of oviposition-site preference in
Drosophila tripunctata. Heredity 59:363-369.

214

Johnson, RM. and HE. Schaffer. 197 3. Isozyme variability in species of
the genus Drosophila. VII. Genotype-environmental relationships in
populations of Drosophila melanogaster from the eastern United
States. Biochem. Genet. 10:149-163.

Johnson, GB. 1976. Polymorphism and predictability of the alpha-
glycerophosphate dehydrogenase locus in Colias butterflies.
Biochem. Genet. 14:403-425.

Johnson, MS. 1971. Adaptive lactate dehydrogenase variation in the
crested blenny, Anoplorchus. Heredity 27:205-226.

Jordan, K. 1896. On mechanical selection and other problems. Novit.
2001. 3:426-525.

Kettlewell, H.B.D. 1973. The Evolution of Melanism. Clarendon, Oxford,
England.

Koehn, R.K. 1969. Esterase heterogeneity: Dynamics of a polymorphism.
Science 163:943-944.

Koehn, R.K., R. Milkman and J .B. Mitton. 1976. Population genetics of
marine pelecypods. IV. Selection, migration and genetic
differentiation in the blue mussel, Mytilus edulis. Evolution 30:2-32.

Langley, C.H., R.A. Voelker, A.J. Leigh Brown, S. Ohnishi, B. Dickson
and E. Montgomery. 1981. Null allele frequencies at allozyme loci in

natural populations of Drosophila melanogaster. Genetics 99: 151-
l 5 6.

Laurie-Ahlberg, C. C. and B. S. Weir. 1979. Allozymic variation and
linkage disequilibrium in some laboratory populations of Drosophila
melanogaster. Genetics 92: 1295-1314.

Levene, H. 1953. Genetic equilibrium when more than one ecological
niche is available. Am. Nat. 87:331-333.

215

Levins, R. 1968. Evolution in changing environments. Princeton Univ.
Press, Princeton, NJ.

Lewis, H. 1962. Catastrophic selection as a factor in speciation. Evolution
16:257-27 1.

Lewis, H. 1966. Speciation in flowering plants. Science 152:167-172.

Lewontin, R.C. 1974. The Genetic Basis of Evolutionary Change.
Columbia University Press, New York, New York. pp. 346.

Lewontin, R.C. and Hubby. 1966. A molecular approach to the study of
genetic heterozygosity in natural populations. II. Amount of
variation and degree of heterozygosity in natural populations of
Drosophila psuedoobscura. Genetics 54:595-609.

Loukas, M.G., A.P. Economopoulos, E. Zouros and Y. Verini. 1985.
Genetic changes in artificially reared colonies of the Olive Fruit Fly.
Ann. Ent. Soc. Amer. 78:159-165. 1985.

Lynch, M. 1987. The consequences of fluctuating selection for isozyme
polymorphisms in Daphnia. Genetics 115:657-669.

Maruyama, T. and P.A. Fuerst. 1985. Population bottlenecks and
nonequilibrium models in population genetics. 11. Number of alleles
in a small population that was formed by a recent bottleneck.
Genetics 1 1 1:675-689.

Maxwell, CW. 1968. Apple maggot adult dispersion in a New Brunswick
apple orchard. J. Econ. Ent. 61:103-106.

Maynard Smith, J. 1966. Sympatric speciation. Am. Nat. 100:637-650.

Maynard Smith, J. 1970. Genetic polymorphism in a varied environment.
Am. Nat. 104:487-490.

216

Mayr, E. 1942. Systematics and the origin of species. Columbia
University Press, New York.

Mayr, E. 1954. Change of genetic environment and evolution. pp. 157-
180. In: Evolution as a Process. (J. Huxley, A.C. Hardy and EB.
Ford, eds.) Allen and Unwin, London, England.

Mayr, E. 1963. Animal Species and Evolution. Harvard Univ. Press,
Cambridge, Mass. 797 pp.

McDonald, J .F. 1983. The molecular basis of adaptation: A critical review
of relevant ideas and observations. Ann. Rev. Ecol. Syst. 14:77-102.

McDonald, J .F . and FJ. Ayala. 1974. Genetic response to environmental
heterogeneity. Nature 250:572-574.

Mcleron, BA. 1987. The population genetics of the colonization of the
western United States by the apple maggot, Rhagoletis pomonella
(Walsh) (Diptera: Tephritidae). Ph. D. Dissertation, Univ. of
Illinois, Urbana, Illinois.

McPheron, B. A. and S. H. Berlocher. 1985. Segregation and linkage of
allozymes of Rhagoletis tabellaria. J. Heredity 76: 218-219.

McPheron, B.A., D.C. Smith and SH. Berlocher. 1988a. Genetic

differences between host races of Rhagoletis pomonella. Nature
336:64-66.

McPheron, B.A., D.C. Smith and SH. Berlocher. 1988b.
Microgeographic genetic variation in the apple maggot,
Rhagoletis pomonella. Genetics 119:445-451.

McPheron, B.A., C.D. J orgensen and SH. Berlocher. 1988c. Low
genetic variability in a Utah cherry-infesting population of the

apple maggot, Rhagoletis pomonella. Entomol. Exp. Appl.
46:155-160.

217

Menken, S.B .J. and SA. Ulenberg. 1987. Biochemical characters in
agricultural entomology. pp. 305-360. In: Agricultural Zoological
Reviews. Vol. 2. Intercept, Ponteland, Newcastle upon Tyne.

Merritt, RB. 1972. Geographic distribution and enzymatic properties of
lactate dehydrogenase allozymes in the fat head minnow, Pimephales
promelas. Am. Nat. 196:173-184.

Miller, S., R.W. Pearcy and E. Berger. 1975. Polymorphism at the alpha-
glycerophosphate dehydrogenase locus in Drosophila melanogaster.
1. Properties of adult allozymes. Biochem. Genet. 13:175-188.

Mitton, J .B. and R.K. Koehn. 197 5. Genetic organization and adaptive
response of allozymes to ecological variables in F undulus
heteroclitus. Genetics 79:97-111.

Mully, J .C., J .W. James and J .S.F. Barker. 1979. Allozyme genotype-
environment relationships in natural populations of Drosophila
buzzatii. Biochem. Genet. 17:105-126.

Murray, J. 1972. Genetic Diversity and Natural Selection. Oliver and
Boyd, Edinburg, Scotland.

Nei, M. 1977. F-statistics and analysis of gene diversity in subdivided
populations. Ann. Hum. Genet. 41:225-233.

Nei, M., T. Maruyama, and R. Chakraborty. 1975. The bottleneck effect
and genetic variability in populations. Evolution 29:1-10.

Neilson, W.T.A. 1971. Dispersal studies of a natural population of apple
maggot adults. J. Econ. Ent. 64(3):648-653.

O'Kane, W.C. 1914. The apple maggot. New Hampshire Expt. Sta. Bull.
No. 171., 120 pp.

218

Paterson, A.H., E.S. Lander, J .D. Hewitt, S. Peterson, S.E. Lincoln and
SD. Tanksley. 1988. Resolution of quantitative traits into
Mendelian factors by using a complete linkage map of restriction
fragment length polymorphisms. Nature 335:721-726.

Phipps, CR. and GO. Dirks. 1933a. Dispersal of the apple maggot fly-
1932 studies. J. Econ. Ent. 5:576-582.

Phipps, CR. and CO. Dirks. 1933b. Notes on the biology of the apple
maggot fly. J. Econ. Ent. 26:349-358.

Porter, BA. 1928. The Apple Maggot. US. Dept. Agr. Tech. Bull. 66.,
48 pp.

Powell, J .R. 1971. Genetic polymorphism in varied environments. Science
174:1035-1036.

Powell, J .R. and H. Wistrand. 1978. The effect of heterogeneous
environments and a competitor on genetic variation in Drosophila.
Am. Nat. 112:935-947.

Price, P.W. 1980. Evolutionary biology of parasites. Princeton Univ.
Press, Princeton, NJ.

Prokopy, R.J. 1968. Influence of photoperiod, temperature and food on
initiation of diapause in the apple maggot. Canad. Ent. 100:318-329.

Prokopy, R.J., E.W. Bennett and G.L. Bush. 1971. Mating behavior in
Rhagoletis pomonella (Diptera: Tephritidae). I. Site of assembly.
Canad. Ent. 103:1405-1409.

Prokopy, R.J., E.W. Bennett and G.L. Bush. 1972. Mating behavior in
Rhagoletis pomonella (Diptera: Tephritidae). II. Temporal
organization. Canad. Ent. 104:97-104.

219

Prokopy, R.J. and G.L. Bush. 1972. Apple maggot infestation of pear. J.
Econ. Ent. 65:597.

Prokopy, R.J., G.E. Haniotakis and AP. Economopoulos. 1975.
Comparative behavior of lab-cultured and wild-type Dacus oleae
flies in the field. pp. 101-108. In: Controlling Insects by the Sterile
Insect Technique. International Atomic Energy Agency,Vienna.

Prokopy, R.J. and SH. Berlocher. 1980. Establishment of
Rhagoletis pomonella (Diptera: Tephritidae) on rose hips in
southern New England. Can. Ent. 112:1319-1320.

Prokopy, R.J., A.L. Averill, S.S. Cooley and C.A. Roitberg. 1982.
Associative learning in egglaying site selection by apple maggot flies.
Science 218:76-77.

Prokopy, R.J., S.R. Diehl, and SS. Cooley. 1988. Behavioral evidence for
host races in Rhagoletis pomonella flies. Oecologia 76: 138-147.

Ramshaw, J .A.M., J .A. Coyne and R.C. Lewontin. 1979. The sensitivity of
gel electrophoresis as a detector of genetic variation. Genetics
93:1019-1037.

Reissig, W.H. 1977. Capture of released apple maggot flies, Rhagoletis
pomonella (Diptera:Tephritidae), on sticky panels. Can. Ent.
109:631-634.

Reissig, W.H. 1979. Survival of apple maggot larvae, Rhagoletis
pomonella (Diptera: Tephritidae), in picked and unpicked apple.
Can. Ent. 111:181-187.

Reissig, W. H. and DC. Smith. 1978. Bionomics of Rhagoletis pomonella
in Crataegus. Ann. Ent. Soc. Amer. 71: 155-159.

Reissig, W.H., J. Barnard, R.W. Weires, E.H. Glass and R.W. Dean. 1979.
Prediction of apple maggot fly emergence from thermal unit
accumulation. Environ. Entomol. 8:51-54.

220

Rice, WR. 1984. Disruptive selection on habitat preference and the
evolution of reproductive isolation: a simulation study. Evolution
38:1251-1260.

Rockey, SJ ., J .H. Hainze and J. M. Scriber. 1987. Evidence of a sex-linked
diapause response in Papilio glaucus subspecies and their hybrids.
Physiol. Entomol. 12:181-184.

Saul, SH. 1986. Genetics of the Mediterranean fruit fly (Ceritatis
capitata) (W iedemann). pp. 73-108. In: Agricultural Zoological
Reviews. Vol I . Intercept, Ponteland, Newcastle upon Tyne.

Saul, S.H., M.J. Sinsko, P.R. Grimstad and GB. Craig, Jr. 1978. P0pulation
genetics of the mosquito Aedes triseriatus: Genetic-ecological
correlations at the esterase locus. Am. Nat. 112:333-339.

Scriber, J .M. and J .H. Hainze. 1987. Geographic invasion and abundance as
facilitated by differential host plant utilization abilities. pp. 433-
468. In: Insect Outbreaks: Ecology and Evolutionary Processes. (P.
Barbarosa and J .C. Shultz, eds.) Academic Press, N.Y.

Shervis, L.J., G.M. Boush and CF. Koval. 1970. Infestation of sour
cherries by the apple maggot: Confirmation of a previously uncertain
host status. J. Econ. Ent. 63:294-295.

Showers, W.B. 1981. Geographic variation of the diapause response in the
European corn borer. In: Insect Life History Patterns (R.F. Demo
and H. Dingle, eds.) pp. 97-111. Springer-Verlag, Berlin and New
York.

Singh, P. 1984. Insect diets: Historical developments, recent advances and
future prospects. pp. 32-44. In: Advances and Challenges in Insect
Rearing. (E.C. King and NC. Leppla, eds.). Agricultural Research
Series. USDA. Washington, DC.

Smith, DC. 1986. Genetics and Reproductive Isolation of Rhagoletis flies.
Ph. D. dissertation. Univ. of Illinois, Urbana, Illinois.

221

Smith, D.C. 1988. Heritable divergence of Rhagoletis pomonella host races
by seasonal asynchrony. Nature 336:66-67 .

Smith, HS. 1941. Racial segregation in insect populations and its
significance in applied entomology. J. Entomol. 34:1-12.

Spiess, E.B.1977.Genes in Populations. Wiley and Sons, New York. 780pp.

Standley, RC. 1922. Trees and Shrubs of Mexico. 1721 pp. Smithsonian
Press, Washington, D.C.

Steiner, WM. 1979. Genetic variation in Hawaiian Drosophila. VI.

Seasonally dependent gene changes in Drosophila mimica. Evolution
33:543-563.

Tauber, C.A., M.J. Tauber and J .R. Nechols. 1977. Two genes control
seasonal isolation in sibling species. Science: 197: 592-593.

Tauber, M.J. and C.A. Tauber. 1976. Insect seasonality: Diapause
maintenance, termination, and postdiapause development. Ann. Rev.
Entomol. 21:81-107.

Taylor, CE. and J .R. Powell. 1976. Microgeographic differentiation of

chromosomal and enzyme polymorphisms in Drosophila
pseudoobscura. Genetics 85:681-695.

Templeton, AR. 1981 . Mechanisms of speciation- A population genetic
approach. Ann. Rev. Ecol. Syst. 12:23-48.

Thompson, J. 1988. Evolutionary genetics of oviposition preference in
Swallowtail Butterflies. Evolution 42: 1223-1234.

Thorpe, W.H. 1930. Biological races in insects and allied groups. Biol.
Rev. 5:177-212.

222

Van Valen, L. 1965. Morphological variation and the width of the niche.
Am. Nat. 99:377-390.

Wagner, M. 1868. Die Darwin’sche Theorie und das Migrationsgesetz der
Organismen. Duncker and Humblot, Leipzig, Austria.

Wagner, M. 1889. Die Entstehung der Arten durch raumliche Sonderung
Benno Schwalbe, Basel, Switzerland.

Walsh, B.D. 1864. On phytophagic varieties and phytophagic species.
Proc. Ent. Soc. of Philadelphia 3:403-430.

Walsh, B.D. 1867. The apple-worm and the apple maggot. J. Hort. 2:338-
343.

Ward, C. 1866. Answers to correspondents. Practical Entomologist 2:20-
21.

Wasbauer, MS. 1972. An annotated host catalog of the fruit flies of
America north of Mexico (Diptera: Tephritidae). Bureau of

Entomol., Dept. of Agric., Sacarrnento , Calif. Occasional papers
no. 19.

Weir, B. S. 1979. Inferences about linkage disequilibrium. Biometrics
35: 235-254.

Weir, B.S. and QC. Cockerham. 1984. Estimating F-statistics for the
analysis of population structure. Evolution. 38:1358-1370.

White, M.J.D. 1968. Models of speciation. Science 159:1065-1070.

White, M.J.D. 1978. Modes of Speciation. W.H. Freeman and Co., San
Francisco. pp. 455.

223

Williams, S.M., J .L. Feder, B.A. McPheron, S.H. Berlocher and G.L.
Bush. The effects of recent colonization on hybridization and
introgression between two sibling species in the Rhagoletis
pomonella group (Diptera: Tephritidae). (Manuscript in
preparation).

Wright, S. 1931. Evolution in Mendelian populations. Genetics 16: 97-
159.

Wright, S. 1940a. Breeding structure of populations in relation to
speciation. Am Nat. 74:232-248.

Wright, S. 1940b. The statistical consequences of Mendelian heredity in
relation to speciation. pp. 161-183. In: The New Systematics. (J.
Huxley, ed.) Oxford Univ. Press, Oxford, England.

Wright, S. 1943. Isolation by distance. Genetics 28:114-138.

Wright, S. 1951. The genetical structure of populations. Ann. Eugen.
(Lond.) 15:323-354.