—
s E .
{$35. .
n . .1 v 1.

 

,.

o . n
.0 W . .
23?"? . . . . .
r :1 \1. . . . . . . , ..
.e . t... . , . . . . , . , , r
‘ , A v , . . , . . . c _
. . . . ,, . z. 3.: . AM! . 2 .n... x
. , . . fwduwumwntsrfi and

I;

vb .
.35.. z.
.e ..

r J . .
$4.4...
.3 .

 

 

 

 

 

Lek} .
. . 1. 11: .
in 7,453. .1

3?

if: .
r. . a

. 3.}...

. . \.
5.1%....
r...

)

L f :

mawfitzat
rift

 

fiifififi? . , ,. 9% $33

 

 

THESIS
w")
‘-

900i

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the

dissertation entitled

THE NATURE OF INFERTILITY AND RESPONSE TO INBREEDING IN
VACCINIUM (BLUEBERRY) SPECIES

presented by

Gary William Schott

has been accepted towards fulfillment
of the requirements for

Ph.D. Botany and Plant Pathology

degree in

 

 

8101‘ pro es

Date /y‘@~2¢ 00

 

MS U is an Affirmative Action/Equal Opportunity Institution 0- 12771

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

moo c/CIRCIDaieDuopss-p.“

mafia? .‘., . __

THE NATURE OF INFERTILITY AND RESPONSE TO INBREEDING IN
VACCINIUM (BLUEBERRY) SPECIES

By
Gary William Schott

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Botany and Plant Pathology

2000

ABSTRACT

THE NATURE OF INFERTILITY AND RESPONSE TO INBREEDING IN
VACCINIUM (BLUEBERRY) SPECIES

By
Gary William Schott

The study of plant mating systems is a very active area of research. In
particular, determining the nature of infertility is important. The two
predominant schools of thought on why plants have low or no self-fertility are
self-incompatibility (SI) and inbreeding depression (ID). With SI, self pollinated
flowers set no seeds. Arrest of the reproductive process can occur at any stage
between pollen germination and ovule development. In late-acting SI,
otherwise known as ovular self-incompatibility (OSI), self-fertilized ovules are
aborted at a relatively uniform stage of development. With ID, inbred progeny
have lower fitness, which can manifest itself at any stage of development.
Whether the abortion of self-pollinated ovules is due to OSI or early-acting ID
will impact the mating system, which will impact genetic variation within and
between populations, and influence the evolution of plant populations and
species. I examined self-infertility and the response to inbreeding in several
blueberry species (Vaccinium).

The response to inbreeding was studied by performing self and
outcross pollinations on 7-13 plants in each of two populations of Vaccinium
myrtilloides (diploid) and LI. cmmbosum (tetraploid). l monitored fruit

development, seed set, seed germination and seedling growth for evidence of

ID, which is calculated as 5 = 1 - ws/wo, with ws and W0 representing the fitness
of selfed and outcross progeny respectively. The effect of outcrossing distance
on fertility was studied in the same two species by performing outcross
pollinations of increasing distance from experimental plants. The pollination
treatments consisted of self, close (w/in 10 m), distant (>100 m) and other
(other population) pollinations. The nature of infertility was studied in V.
corymbosum and \_l. angustifolium (tetraploid) by monitoring the development
of self and outcross ovules. Self and outcross fruits were sampled at 1, 2, 3,
and 4 weeks following pollination and at maturity. Seed length was measured
and germinability was tested.

I found that both y. myrtilloides and \_l. corymbosum suffer ID at most life-
history stages examined with ID being greater at early stages in the diploid and
consistent across stage in the tetraploid. Cumulative ID was comparable in
the two species. The diploid shows evidence of outbreeding depression while
the tetraploid does not. Seed development and size was variable with V.
corymbosum being more self-fertile than V. angustifolium. The data suggest
that seed abortion and low self-fertility in blueberries is due to early-acting ID,
not OSI. The phenomena of pseudo-seIf-fertility is invoked to explain how
blueberries respond to inbreeding. It is thought that relatively high self-fertility

in V. corymbosum is an artifact of polyploidy formation.

DEDICATION

I would like to dedicate this dissertation to my wife Carolyn and my children
Sara, Cole and Jacob. They help me keep everything in the proper perspective.

ACKNOWLEDGEMENTS

I would like to thank the many people who made this dissertation
possible. I am indebted to Jim Hancock for accepting to be my graduate
advisor after my first advisor left the university. His generosity has been above
and beyond the call of duty and I am very grateful. The people I have worked
with in Jim's laboratory are the best. Pete Callow is the worlds best technician.
Pete, Karen Hokanson, Jamie Houghton, Sedat Serge, Chris Owens, Rebecca
Keesler, Kym Washo and Tom Harpstead helped with field work, laboratory
work and offered advice on interpreting data and presentations.

I would like to thank the faculty, staff and students of the W. K. Kellogg
Biological Station. In particular I would like to thank Jeff Conner, Kay Gross,
Heather Reynolds, Brian Black, Kevin Kosola and Tara Darcy for assisting with
data analysis, document preparation and laboratory and greenhouse work.
Nina Consolatti was indispensable. I think she dreaded seeing me come
down the hall, knowing that there was something else I needed. Char Adams,
Alice Gillespie and Sally Shaw all helped to keep assistantships in order and
are usually patient with the demands of graduate students.

For financial assistance, I would like to thank Peter Murphy and the
Department of Botany and Plant Pathology for the many teaching
assistantships they provided. The Kellogg Biological Station also provided
many summer teaching assistantships and support for one year through the
Research Training Grant. The graduate school and George Lauff provided

generous support in aid of my research.

I would like to thank Mike Green, the head ranger at Cheyboygan State
Park. Mike provided assistance in my field work at the state park and would
notify me when the plants were doing their thing, saving me the time and effort
of going up to find out myself.

Finally, I would like to thank my family. My wife Carolyn and children
Sara, Cole and Jacob were very patient during these many years. They offered
unwavering support and kept my spirits up during those times when I didn’t
think all of this effort would amount to a hill of beans. Carolyn's ability to
support the family and stroke my fragile ego has helped tremendously. My
parents, Bill and Jeannine Schott, and Carolyn’s parents Fred and Marilyn
Wenzel, have been indispensable. They have been very generous with
financial support and took care of the kids and our home during those times
when l was very busy. Extended family member often wondered what I was up
to these many years. Although I am sure they often wondered about me at
times, they were always supportive and kept a straight face when I told them

that I studied reproduction in blueberries.

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES
INTRODUCTION

CHAPTER 1: THE RESPONSE OF DIPLOID AND TETRAPLOID
VACCINIUM SPECIES TO INBREEDING UNDER NATURAL
CONDITIONS

Abstract

Introduction

Materials and Methods
Results

Discussion

Literature Cited

CHAPTER 2: THE EFFECT OF CROSSING DISTANCE ON FERTILITY
IN VACCINIUM SPECIES

Abstract

Introduction

Materials and Methods
Results

Discussion

Literature Cited

CHAPTER 3: THE NATURE OF INFERTILITY IN VACCINIUM
SPECIES

Abstract

Introduction

Materials and Methods
Results

Discussion

Literature Cited

CONCLUSIONS

vii

Page

viii

10

11
12
16
25
36
43

49

50
51
53
58
71
75

78

79
80
86
91
1 00
1 04

109

Table

LIST OF TABLES
CHAPTER 1
Page

Table 1. Mean fitness of selfed (S) and outcrossed (OC)
offspring in the blueberry species Vaccinium myrtilloides
and \_/. corymbosum. The abbreviations represent percent
fruit set (°/o Frt. Set), mean number of fertilized ovules per
fruit (# Seeds), percentage of fertilized ovules that develop
into mature seeds (Seed Set), percent germination (°/o
Germ), growth rate of seedlings (Growth) and cumulative
fitness of the product of the five life-history stages (Cum.
Fitness). Cumulative inbreeding depression (ID) is 1-
cumulative self/cumulative OC. Values in parentheses are
standard errors. F values are for a one-way ANOVA
comparing the means of the four populations. 26

Effects of year, species and pollination treatment on fitness

at five life-history stages. Pop[Species], and all interactions

involving it, are treated as a random effect. Significance tests

for all transformed data were conducted on log-transformed

data (Johnston and Schoen 1994). 28

A comparison of self and outcross performance between

Hokanson and Hancock (2000) and the current study. Relative
survival is the percent mature seed per self pollination relative

to outcross pollination (standard error). Percent self is the

percentage of plants that set a self pollinated fruit. Lethal

equivalents is an estimate of the genetic load . See the text for

a detailed description of lethal equivalents. 36

CHAPTER 2

Effects of species and pollination treatment on fitness at five life-
history stages in Vaccinium myrtilloides and \_/. corymbosum
blueberries. Pop[Species], and all interactions involving it, is

treated as a random effect. The Pollination x Species interaction

was tested on log transformed data for all stages as per

Johnston and Schoen 1994. 59

viii

CHAPTER 3

1. Criteria for differentiating between pseudo-seIf-fertility and
true self-fertility as specified by Levin (1996). The statements
apply to a species that is subject to pseudo-seIf-fertility.

2. Criteria for differentiating between inbreeding depression (ID)
and ovular self-incompatibility (OSI) as specified by Seavey
and Bawa (1986).

3. Seed germination data for y. angustifolium (V. a.) and \_/.
corymbosum (V. c.). The cross types are self pollen, outcross
pollen (00) or open pollinated (OP). Also included are the
maximum (Max) and minimum (Min) size of germinating seeds,
the number of seeds germinating (n), the mean size of
germinating seeds, and the standard error of the mean (SEM).
All seed sizes are in micrometers.

4. Effect of species and pollination treatment on the size of
germinating seeds.

5. Data for fruits that reached maturity. Plant represents the 10
experimental plants. AAW represents the y. angustifolium
population and CCL represents the y. corymbosum population.
Cross represents self, outcross (OC) and open (OP) pollinated
flowers/fruits. # fruits = the number of mature fruit sampled, #

ovules is the total number of fertilized ovules in all of the mature

fruits, # germ is the number of ovules that germinated, % germ
is the percent of ovules that germinated, #>1100 is the number
of seeds that were larger than 1100 um, #>820 is the number
of seeds that were larger than 820 um, #>11/fruit is the number
of seeds larger than 1100 um in each fruit, #>820/fruit is the
number of seeds larger than 820 um in each fruit.

LIST OF FIGURES
Figure
CHAPTER 1
1. Range in seed sizes from a representative Vaccinium

corymbosum fruit. Size and appearance in \_l. myrtilloides is
comparable. The text in the picture is 12 point courier.

80

82

97

97

99

Page

14

The locations of study populations. COL: _\_l. commbosum

at Otis Lake, CCL= _\_l. corymbosum at Crooked Lake, MFR= \_/.
myrtilloides at Fox River, MPR= \_l. myrtilloides at Poe Reef.

See the text for detailed descriptions of the location and

habitat of each population. 20
Relative performance of the blueberries Vaccinium

myrtilloides and \_l. corymbosum at several life history stages

with % Frt. Set = fruit set, # Seeds = the total number of fertilized
ovules, Seed Set = the fraction of fertilized ovules developing

into mature fruits, °/o Germ. = the percentage of mature seeds

that germinated, Growth = the growth rate of seedlings. Fruit

set is significantly different at p=0.05 (t-test). 30

Fruit set of selfed and outcrossed (OC) pollinations in two
blueberry species in 1997 and 1998. 30

Total number of ovules fertilized in selfed and outcrossed
(OC) pollinations in two blueberry species in 1997 and 1998. 32

Average seed set in selfed and outcrossed (OC) blueberries
in the years 1997 and 1998. 32

Percent germination in two blueberry species for self and
outcross (OC) pollinations. 34

Growth rate in two blueberry species for self and outcross
(OC) pollinations. 34

CHAPTER 2

The locations of study populations. Vaccinium corymbosum

at Otis Lake (COL), _\_l. corymbosum at Crooked Lake (CCL), \_l.
myrtilloides at Fox River (MFR), \_l. myrtilloides at Poe Reef

(MPR). The Black River population served as the distant pollen

donor for the MPR population. See the text for detailed

descriptions of the location and habitat of each population.

Pollen was pooled for all the outcrossed pollinations. 55

The percent fruit set in _\_l. myrtilloides and \_/. corymbosum

blueberries for self, close (pollen source within 10 m), distant

(pollen source 100-200 m distant) and other (reciprocal

pollination between populations within species) pollinations. 61

The relative performance (RP) for fruit set in y. myrtilloides and
\_l. corymbosum blueberries for self, close (pollen source within
10 m), distant (pollen source 100—200 m distant) and other
(reciprocal pollination between populations within species)
pollinations. [RP is 1-ws/wo when we a ws and wo/ws - 1 when
ws > we, ws and we are the fitness of outcross and selfed
individuals respectively] An RP value > 0 represents ID while

a value < than 0 represents OD.

The total number of ovules fertilized (aborted & mature) in

\_/. myrtilloides and V. corymbosum blueberries for self, close
(pollen source within 10 m), distant (pollen source 100—200 m
distant) and other (reciprocal pollination between populations
within species) pollinations.

The relative performance (RP) for total number of ovules
fertilized in y. myrtilloides and \_l. corymbosum blueberries for
self, close (pollen source within 10 m), distant (pollen source
100-200 m distant) and other (reciprocal pollination between
populations within species) pollinations. [RP is 1-ws/wo when
WC 2 ws and wo/ws - 1 when ws > wo, ws and wo are the fitness
of outcross and selfed individuals respectively] An RP value > 0
represents ID while a value < than 0 represents OD.

The percentage of fertilized ovules developing into mature
seeds (seed set) in \_l. myrtilloides and \_/. corymbosum
blueberries for self, close (pollen source within 10 m), distant
(pollen source 100-200 m distant) and other (reciprocal
pollination between populations within species) pollinations.

The relative performance (RP) for seed set in y. myrtilloides
and \_l. commbosum blueberries for self, close (pollen source
within 10 m), distant (pollen source 100-200 m distant) and
other (reciprocal pollination between populations within
species) pollinations. [RP is 1-ws/wo when wo 2 ws and

wo/ws - 1 when ws > we, ws and we are the fitness of outcross
and selfed individuals respectively] An RP value > 0 represents
ID while a value < than 0 represents OD.

The percentage of mature seed germinating in y. myrtilloides
and y. corymbosum blueberries for self, close (pollen source
within 10 m), distant (pollen source 100-200 m distant) and
other (reciprocal pollination between populations, within
species) pollinations.

xi

10.

11.

The relative performance (RP) for seed germination in V.
myrtilloides and \_l. corymbosum blueberries for self, close
(pollen source within 10 m), distant (pollen source 100-200 m
distant) and other (reciprocal pollination between populations
within species) pollinations. [RP is 1-ws/wo when WC 2 ws and
wo/ws - 1 when ws > wo, ws and WC are the fitness of outcross
and selfed individuals respectively] An RP value > 0 represents
ID while a value < than 0 represents OD.

Seedling growth rate in \_l. myrtilloides and \_l. corymbosum
blueberries for self, close (pollen source within 10 m), distant
(pollen source 100-200 m distant) and other (reciprocal
pollination between populations, within species) pollinations.
The value for seedling growth rate represents the height of the
seedling, in mm, divided by the number of days since the
seedling had emerged from the soil.

The relative performance (RP) for growth rate in y. myrtilloides
and \_l. corymbosum blueberries for self, close (pollen source
within 10 m), distant (pollen source 100-200 m distant) and
other (reciprocal pollination between populations within species)
pollinations. [RP is 1-ws/wo when WO a WS and wo/wS - 1 when
ws > we, ws and we are the fitness of outcross and selfed
individuals respectively] An RP value > 0 represents ID while a
value < than 0 represents OD.

CHAPTER 3

Map of the Allegan State Game Area, Allegan county Michigan,
showing the locations of the study populations. AAW
represents the Vaccinium angustifolium population and CCL
represents the V. coLymbosum population.

Picture of mature and aborted seeds from a Vaccinium
corymbosum fruit. For purposes of scale, the text is 12
point courier.

"Images in this thesis/dissertation are presented in color."

xii

67

69

69

88

90

Ovule/seed size for self and outcross pollinations in five

Vaccinium angustifolium plants. Rows represent the five

plants. The first column is the mean ovule size for selfed

and outcrossed fruits sampled at several dates before

maturity, and at maturity (day 60). In addition, ovule size was
quantified in unopened flowers, that were not pollinated, and

in flowers that were naturally pollinated (OP). The second

and third columns represents the distribution of ovule sizes

for self and outcross pollinated fruits, respectively, at different

sample dates. 93

Ovule/seed size for self and outcross pollinations in five

Vaccinium commbgsum plants. Rows represent the five

plants. The first column is the mean ovule size for selfed

and outcrossed fruits sampled at several dates before

maturity, and at maturity (day 60). In addition, ovule size was
quantified in unopened flowers, that were not pollinated, and

in flowers that were naturally pollinated (OP). The second

and third columns represents the distribution of ovule sizes

for self and outcross pollinated fruits, respectively, at

different sample dates. 95

xiii

INTRODUCTION

Understanding the various aspects of plant reproduction is very
important for conservation biology and agriculture. Human population growth
contributes to the loss of natural habitats and arable land. Maintaining viable
wild populations requires knowledge of the reproductive dynamics of plant
populations. Maintaining adequate food supplies requires maximizing the
reproductive output of crops. A study that examines the nature of infertility and
the performance of plants under different pollination regimes will add to our
knowledge base and aid in conservation efforts and agriculture. This study is
unique because it tests theories on the nature of infertility, and response of
different ploidy levels to inbreeding, in a wild species that is the direct
progenitor, and same species, as a widely cultivated species, the highbush
blueberry (Vaccinium corymbosum). This species is of increasing importance
due to its widely recognized health benefits (Prior et al., 1998). Studying wild
populations of \_l. corymbosum species, and closely related species, will help
us to develop better cultivation practices and assist in the conservation of wild
populations that can serve as reservoirs of genetic information.
In reedin Infertili and O timal Outcrossin

Inbreeding is, of course, the mating of closely related individuals. It often
results in inbreeding depression, a decrease in the fitness of resulting
progeny. There are several theories that can explain the phenomenon of ID
(Charlesworth and Charlesworth, 1987). In the partial dominance model,

deleterious recessive alleles are preferentially expressed in the homozygous

state after inbreeding. In the overdominance model, heterozygotes have higher
fitness than either homozygote, and their frequency decreases with inbreeding.
The higher fitness of heterozygotes presumably comes from positive
interactions between different alleles. Overdominance can also result from the
heterozygote having more diverse gene products (enzymes) which provides a
multiplicity of metabolic pathways for enzymatic reactions (Richards, 1997).
Empirical data suggests that partial dominance plays a greater role in
inbreeding depression (Charlesworth and Charlesworth, 1987).

Since the advent of plant cultivation and animal husbandry, humans
have recognized the negative consequences of inbreeding. Darwin (1876)
was the first person to systematically study inbreeding in plants. He found that
24 out of the 40 species he studied exhibited inbreeding depression. He also
found that some species thought to be outbreeders showed no evidence of
inbreeding depression and other species thought to be inbreeders suffered
from inbreeding depression (Darwin, 1876). Our knowledge of the impact of
inbreeding depression has increased a great deal over the past century and
this has led to the recognition that inbreeding depression can occur at many
stages in plant development. Inbreeding is particularly detrimental in small
populations. However, inbreeding can be a valuable tool for the plant breeder
by allowing them to create true breeding lines, that are then used to create
hybrids that are often of superior quality. This phenomenon of hybrid vigor, or

heterosis, has been used extensively in agriculture.

Many mechanisms have evolved in natural populations to reduce or
avoid inbreeding. In animals, sex-biased dispersal and kin recognition reduce
the likelihood that closely related individuals will mate. Plants have many
morphological and biochemical mechanisms to avoid inbreeding. Dioecy
assures outcrossing while monoecy can favor outcrossing. Plants with
herrnaphroditic flowers often spatially separate stigma(s) and anthers
(herkogamy) or mature the stigma(s) and anthers at different times
(dichogamy). However, dioecy cannot prevent biparental inbreeding and
herkogamy, dichogamy and spatial separation of male and female flowers
cannot completely prevent the pollination of one flower, by another, on the
same plant (geitonogamy). As a result, many species are self-incompatible to
prevent the production of inbred progeny. The introduction of chapter 3
provides a detailed description of the forms of self-incompatibility found in
plants.

Outbreeding depression can also be detrimental in plant populations.
Progeny of a mating between individuals that are too distantly related and/or
widely spaced may have decreased fitness, relative to progeny resulting from a
mating of closer relatedness and/or distance. When a species exhibits both
inbreeding depression and outbreeding depression, there may be an
intermediate, optimal outcrossing distance that results in progeny with the
highest fitness. Outbreeding depression may arise in situations where mating
with a too dissimilar individual breaks up epistatic, coadapted gene

complexes. An intermediate optimal outcrossing distance can help explain the

phenomenon of mixed mating systems, where species or populations have
crossing rates that are somewhere between predominate selfing and
predominate outcrossing (Waser, 1993).
Blue erri s

The true blueberries are in the family Ericaceae, section Cyanococcus,
genus Vaccinium. Other species in this genus that are not considered true
blueberries include cranberries. The term huckleberries is also used when
referring to blueberries. While huckleberries are Ericaceous, they are in the
genus Gaylussacia. The genus Vaccinium occurs primarily in North America,
but there are a few European and Asiatic species (Camp, 1945). Three
species of Vaccinium are included in this study and are described below.

Vaccinium corymbosum is known as the highbush blueberry. It is a
North American species found from Florida to southern Canada, across to
eastern Texas in the south and Michigan in the north. It is found in open
swamps, bogs and along the margins of lakes and ponds in acidic sandy soil.
Vander Kloet (1978) provides a detailed description of the morphology and
ecology of y. corymbosum. Camp (1945) recognized 12 highbush blueberry
species. He divided these species between northern and southern complexes
with different diploid progenitors. The putative diploid progenitors of the
northern complex are y. pallidum,\_/. ceasariense,\_l. atrococcum and \_L
angustifolium. Camp mistakenly thought that V. angustifolium was a diploid, it
is tetraploid. For the southern complex they are y. ceasariense,\j. atrococcum,

y. darrowii and \_l. tenellum. Vander Kloet (1980, 1988) combined Camp's 12

 

species into one species, Vaccinium corymbosum, with infraspecific taxa at the
diploid, tetraploid and hexaploid levels. Vander Kloet (1980) believes that V.

tenellum V. darrowii and V. pallidum are the diploid progenitors of all highbush

 

 

blueberries. This argument suggests that V. corymbosum has multiple origins
and a diverse genetic background. Hybridization among local Vaccinium taxa
is creating genetically distinct groups. In the southeastern US, tetraploid (4n)
V. corymbosum hybridizes with 4n V. myrisinites (Cockerham and Galletta

1976) and diploid (2n) V. corymbosum hybridizes with V. darrowii,V. tenellum

 

and V. pallidum (Darrow and Camp, 1945; Vander Kloet, 1977; Lyrene and
Sherman, 1981; Ballington and Galletta. 1978; Ballington et al., 1979; 1982;
1985). In northern areas, 4n V. corymbosum commonly hybridizes with 4n V.
angustifolium (Vander Kloet, 1976).

Vaccinium angustifolium, commonly called the low sweet blueberry, is a
North American species partial to dry upland habitats. It is found in oak and
pine barrens, rock outcrops and burned areas. It requires acidic soil. It is
found primarily in the north eastern US. and eastern Canada, as far south as
Virginia and west to Minnesota. Camp (1945) recognized V. angustifolium as a
unique species but mistakenly classified it as a diploid. More recent studies
clearly demonstrate that this species is a tetraploid (Hall and Aalders, 1961;
Vander Kloet, 1978). The putative progenitors ofy. angustifolium are the

diploids V. boreale and V. pallidum (Vander Kloet, 1977).

 

Vaccinium myrtilloides, commonly called the velvet-leaf blueberry, is a diploid

that was also recognized by Camp (1945) as a unique species. Vander Kloet

and Hall (1981) maintain that status and provide a detailed description of the
species. Preferred habitats are boreal forest, bogs, barrens and rocky
outcrops. It is found across North America as far south as Michigan and
Pennsylvania and as far north as the mid-latitudes of Hudson Bay. In Michigan,
V. mVrtilloides is often found in sympatry with V. angustifolium, and is
sometimes sympatric with V. corymbosum.

Blueberries are monoecious with bisexual flowers. lnflorescences
consist of approximately 5-12 flowers and are usually clustered towards the
ends of year-old branches. There is spatial separation of stamens and the
stigma. The stamens are attached to the base of the fused corolla and extend
half way up the corolla. The style extends to the opening of the corolla and the
stigma surface is located just inside the corolla or extends beyond the corolla.
The anther sacks are poricidal and pollen is released with the ultrasonic
vibrations of buzz pollinators or from physical manipulation. They are generally
acknowledged to be highly outcrossed. However, many geitenogamous
pollinations (within plant) do occur (Vander Kloet and Lyrene, 1987).

The broad geographic ranges of blueberries in North America, and
variation in ploidy among species, make them ideal for studying the
relationship between inbreeding and polyploidy. Not only do individuals from
different regions have divergent evolutionary histories, but the species also
have variant life-history strategies. The lowbush types, such as V.
angustifolium and V. myrtilloides, are predominantly rhizomatous and form

large colonies of single genotypes. In contrast, the highbush types are crown

formers that exist in more finely-grained patches. The rhizomatous growth
pattern in the lowbush blueberry potentially favors more self pollinations. As
clones expand in lateral extent, pollinators move between shoots of plants in
the same clone. Relatively speaking, a greater proportion of outcross
pollinations should occur in the highbush blueberry because of the closer
proximity of different clones. However, it is also possible that the extensive
lowbush clones are interdigitated, and a greater number of outcross
pollinations are occurring than otherwise might be expected.

The experiments described in this dissertation address questions
concerning the reproductive biology of blueberries focusing on the responses
to inbreeding, the nature of infertility and the effect of outbreeding. In Chapter 1,
I examine how diploid (V. myrtilloides) and tetraploid (V. corymbosum)
blueberries respond to inbreeding. I examine ID at several life history stages
and reevaluate notions of inbreeding from previous work. In Chapter 2, I
examine outbreeding depression in V. rrLLrtilloides and V. corymbosum. In
Chapter 3, I address the nature of self-infertility in blueberries, following seed
development and viability in the tetraploid species V. angustifolium and V.
corymbosum. I review previous studies (Krebs and Hancock, 1990; Hokanson
and Hancock, 2000) to conclude that low self-fertility in blueberries is due to
inbreeding depression, not self-incompatibility as previously suggested

(Vander Kloet and Lyrene, 1987).

LITERATURE CITED
CAMP, W. H. 1945. The North American blueberries with notes on other

groups of Vacciniaceae. Brittonia 5:203-275.

 

CHARLESWORTH, D. and B. CHARLESWORTH. 1987. Inbreeding depression and
its evolutionary consequences. Annual Review of Systematics and
My 18:237-268.

DARWIN, C. 1876. The effects of cross and self ferilisation in the vegetable
kingdom. Murry, London.

HALL, l. V. and L. E. AALDERS. 1961. Cytotaxonomy of lowbush blueberries in
eastern Canada. American Journal 91m 48:199-201.

HOKANSON, K. E. and J. F. HANCOCK. 2000.

KREBS, S. L. and J. F. HANCOCK. 1990. Early acting inbreeding depression and
reproductive success in the highbush blueberry, Vaccinium
corymbosum L. Theoretical and Applied Genetics 79:825-832.

PRIOR, R. L., G. CAO, A. MARTIN, E. SOFIC, J. MCEWEN, C. O'BRIEN, N. LISCHNER,
M. EHLENFELDT, W. KALT, G. KREWER AND C.M. MAINLAND. 1998.
Antioxidant capacity as influence by total phenolic and anthocyanin

content, maturity, and variety of Vaccinium species. Journal 9_f

 

Agriculture and Food Chemistry 46:2686-2693.

RICHARDS, A. J. 1997. Plant breeding systems. Chapman & Hall, London, p.
386.

VANDER KLOET, S. P. 1977. The taxonomic status of Vaccinium boreale.

Canadian Journal gt Botany 55: 281-288.

. 1978. Systematics, distribution, and nomenclature of the polymorphic
Vaccinium angustifolium. Rhodora 80:358-376.

_. 1980. The taxonomy of the highbush blueberry, Vaccinium
corymbosum. Canadian Journal of B_otany 58: 1187-1201 .

. 1988. The genus Vaccinium in North America. Research Branch
Agriculture Canada, Publication 1828.

_, and l. V. HALL. 1981. The biological flora of Canada: 2. Vaccinium
myrtilloides Michx., velvet-leaf blueberry. Canadian Field Naturalist 95:
329-345.

_, and P. M. Lyrene. 1987. Self-incompatibility in diploid, tetraploid, and
hexaploid Vaccinium corymbosum. Canadian Journal of Botany 65:660-
665.

WASER, N. M. 1993. Population structure, optimal outbreeding, and
assortative mating in angiosperms. II_I_ N. W. Thornhill [ed.], The natural
history of inbreeding and outbreeding, 173-199. University of Chicago

Press, Chicago.

CHAPTER 1

THE RESPONSE OF THE BLUEBERRIES

VACCINIUM MYRTILLOIDES AND V. CORYMBOSUM TO INBREEDING UNDER

NATURAL CONDITIONS

10

ABSTRACT

Low self fertility can result from various forms of self incompatibility or
inbreeding depression (ID). The nature of self-fertility in blueberries
(Vaccinium) is in question with some studies suggesting self-incompatibility
and others suggesting ID. This study examined the response of Vaccinium
corymbosum (tetraploid) and V. myrtilloides (diploid) blueberries to inbreeding.
In both species, there was significant ID for fruit set, seed set and growth rate.
V. myrtilloides had significantly greater ID for fruit set relative to V. corymbosum.
There was significant ID for germination rate in V. corymbosum but not V.
myrtilloides. The ability of self and outcross pollen to fertilize ovules was
comparable, especially in V. corymbosum. Vaccinium cormbosum had a
more consistent response to inbreeding across life history stage than V.
myrtilloides. Inbreeding depression in V. myrtilloides was more severe for early
traits and decline more dramatically at later stages. Estimates of lethal
equivalents were higher in V. corymbosum. These data suggest that ID in the
diploid, V. myrtilloides, is caused by strong deleterious alleles that are readily
purged in early reproductive stages, while self-infertility in the tetraploid, V.
corymbosum. results from the more gradual loss of higher order allelic
interactions. The results are consistent with earlier studies that have shown

that decreased self fertility in blueberries is caused by early-acting inbreeding

depression, not late-acting self incompatibility.

11

INTRODUCTION

In plants, a reduction in fitness due to inbreeding results from one of
three phenomena: pre-zygotic self-incompatibility (Proctor et al., 1996), post-
zygotic self-incompatibility (ovular self-incompatibility or OSI) (Seavey and
Bawa, 1986; Sage et al., 1994) or inbreeding depression (Charlesworth and
Charlesworth, 1987). The theoretical nature of these phenomena has been
discussed in detail in the literature and the citations above provide recent
reviews. Elucidating the nature of self-sterility is important when one is
interested in the evolution of mating systems as genetic system and ecological
conditions interact to optimize the level of outcrossing in successful species.

Inbreeding depression (ID) is defined as a reduction in fitness from self
pollination or a cross between closely related individuals relative to outcrossed
individuals. Inbreeding depression is a quantitative trait, and individuals
express differing levels of self-fertility depending on the number of deleterious
alleles they carry. Inbreeding depression can manifest itself at any stage of the
life cycle, affecting early to later traits such as fruit set, seed set, seed weight,
germination rate, seedling development, growth and survival, flower production
and fruit production.

The degree to which ID is manifest can depend on environmental
conditions (Pray et al., 1994). Poor or unpredictable conditions can favor the
survival of outcrossed, highly heterozygous individuals. In contrast, variation
between habitats can favor local adaptation and the maintenance of coadapted

gene complexes, thus favoring inbred individuals (Schmitt and Gamble, 1990).

12

Ideally, one should carry out a study of ID under field conditions where resource
levels, herbivory and other environmental factors may be of importance, relative
to the greenhouse.

Many studies have documented the presence of inbreeding depression
(Rathcke and Real, 1993; Mayer et al., 1996; Daehler, 1999). Hokanson and
Hancock (2000) found evidence of inbreeding depression in native blueberry
species of the genus Vaccinium. In a greenhouse study, they demonstrated
that selfing reduced fruit and seed set. For both traits, the lowered fertility of
self crosses was attributed to early acting inbreeding depression, as opposed
to self-incompatibility. Evidence for this assertion included a range in self-
fertility among parents and the presence of fertilized, yet apparently aborted
ovules in fruits (Figure 1). These aborted ovules varied in size and shape,
suggesting that they were aborted at different stages. However, this
assumption was not directly tested, except to show that self-pollen can
sometimes accomplish fertilization (Krebs and Hancock, 1990; Hokanson and
Hancock, 2000). Hokanson and Hancock (2000) suggested that differences in
ploidy level among Vaccinium species may influence the expression of ID, with
selfing accelerating the purging of deleterious alleles in diploids and polyploidy
slowing this process down.

Polyploidy is a common phenomenon in plants which can influence
patterns of genetic variation and the response to inbreeding (Grant, 1981;
Bever and Felber, 1992). In general, polyploids are more heterozygous than

diploids and therefore can potentially harbor more deleterious alleles than

13

diploids. A tetraploid can have up to three deleterious recessive alleles without
reducing its fitness. if only one "good" allele is needed. For a given level of
heterozygosity, progeny of diploids are more likely to have been the result of
outcrossing than those of tetraploids, because diploids carry only two alleles at
a locus and have a higher likelihood of fixing deleterious alleles through

selfing.

 

Figure 1. Range in seed sizes from a representativemmm
fruit. Size and appearance in V. myrtilloides is comparable. The text in the

picture is 12 point courier.

Given the same amount of outcrossing and heterozygosity, diploid species are
expected to initially suffer more inbreeding depression than polyploids, but will

more rapidly purge deleterious alleles after one or more generations of selfing.

14

Among polyploids, the mode of inheritance affects how genetic variation
is distributed within and between individuals. Autopolyploids have what is
termed tetrasomic inheritance and form multivalents, or a random association
of bivalents during meiosis; therefore, an allele can segregate with any of the
other three alleles in gametes. Allopolyploids have disomic inheritance with no
pairing between homeologous loci. This can result in "fixed heterozygosity"
when the alleles on homeologous loci are different. When inbred, a diploid
population will approach homozygosity more rapidly than an alloploid
population which will approach homozygosity quicker than an autoploid.
However, an alloploid with fixed heterozygosity will never be homozygous
across homeologous loci (Hancock, 1992). Thus, knowing the mode of
inheritance in tetraploids is important because of the differences in response
of allo- and autopolyploids to selection and inbreeding depression. Krebs and
Hancock (1989) demonstrated, using isozyme segregation data, that the
tetraploid V. corymbosum is an autopolyploid.

Until recently, most of the information on self-fertility in blueberries came
from studies of fruit set in cultivars of narrow genetic background. These
cultivars were selected for their ability to set fruit when planted in blocks of one
or a few clones and were likely to be self-fertile (Morrow, 1943; Krebs and
Hancock, 1989; Lang et al., 1990; Lang and Danka, 1991) . However, the
reproductive success of a wild species is much more dependent on genotypic
and environmental variability. Krebs and Hancock (1990) and Hokanson and

Hancock (2000) have measured seed set in native clones in greenhouse

15

studies and found considerable inbreeding depression. However, no one has
studied the reproductive success of blueberries under field conditions.

One of the goals of this study was to determine the degree to which the
results of Hokanson and Hancock (2000) are reproducible in the field. The
number of fitness traits examined was also expanded by including germination
rate and seedling growth rate, in addition to the fruit set and seed set included
in their studies. These traits are expressed later in development and the
expectation was that ID would be more severe under natural conditions in
these presumably highly outcrossing species. In addition, the influence of
ploidy level on inbreeding depression was examined with the hypothesis that
diploid V. myrtilloides would have a more negative response to inbreeding than
the autotetraploid V. cowbosum.

MATERIALS AND METHODS

StudySystemuThe true blueberries in the genus Vaccinium, section
Cyanococcus, are ideal for studying the relationship between inbreeding
depression and ploidy. The section consists of closely related species at
different ploidy levels with presumably high outcrossing rates. Camp (1942)
referred to the entire section as "interlocking parts of a common polyploid
complex". The diploid and tetraploid species of Vaccinium are described in
detail in Vander Kloet (1980, 1981). The species used in this study were the
diploid V. myrtilloides and tetraploid V. corymbosum.

Study Sites-~Two populations of each species were studied (Figure 2).

The two V. corymbosum populations are found in southwest Michigan. The

16

COL population is located in the Barry State Game Area at Otis Lake. The CCL
population is located in the Allegan State Game Area at Crooked Lake. Both of
these populations are found along the margins of shallow, acidic lakes with
organic soils. The surrounding vegetation is primarily deciduous forest on
sandy soils, with Acer rubrum, Quercus alba and figu_s grandifolia as the
dominant overstory trees.

The V. myrtilloides populations are located further north. The MPR
population is located at Cheboygan State Park in a mixed forest dominated by
Thuja occidentalis and Acer rubrum. This habitat is characterized by low, wet
areas and dry uplands which contain the bulk of the blueberry plants.
Vaccinium angustifolium is sympatric with V. myrtilloides at MPR, but is much
less common. The MFR population is located in Michigan's Upper Peninsula,
in the Superior State Forest. The habitat here is a matrix of clear-cut and
secondary growth with V. angustifolium common in cut areas and V.
myrtilloides more often found under forest canopy. The forests are dominated
byflug bagksiana, Pinus resinosa, Populus tremuloides and BetuLa
papyrifera.

Procedures--Pollinations were performed on a total of 76 experimental
plants over two years, including 44 tetraploid and 32 diploid plants. There
were 7-13 experimental plants in each population. Unique plants of V.
myrtilloides were used each year, while several of the same V. corymbosum
plants were used both years, one in the COL population and seven in the CCL

population.

17

To examine the response to inbreeding, I performed a series of outcross
and self-pollinations in the field. Pollen was collected from five randomly
chosen individuals in a population and pooled for all the outcrossed
pollinations. These individuals were termed "pollen donors". Pollen was
collected by removing several recently opened flowers from a donor plant and
rolling each of them between the thumb and forefinger, which caused the
pollen to dehisce from the anthers. Pollen was collected on a cover slip and
kept in a small petri dish. 30-40 flowers were emasculated on each
experimental plant, divided between 2-3 areas of the plant. Approximately 2/3
of the emasculated flowers were pollinated with self-pollen while the remaining
third were pollinated with the outcross pollen. More flowers were self-
pollinated because of the anticipated lower set of selfed fruit. Crosses were
performed in mid-May on V. corymbosum and in mid-June on V. myrtilloides in
1997. Pollinations were performed approximately three weeks earlier in 1998
due, presumably, to an earlier, warmer spring which sped up flower
maturation. Each bunch of emasculated, hand-pollinated flowers was covered
with a nylon mesh bag to exclude pollinators and herbivores, and to prevent
removal of the developing fruits. Fruit from the artificial crosses were harvested

when ripe, approximately 11 weeks after pollination.

18

Fig. 2. The locations of study populations. COL= V. corymbosum at Otis Lake,
CCL= V. corymbosum at Crooked Lake, MFR= V. myrtilloides at Fox River,
MPR= V. myrtilloides at Poe Reef. See the text for detailed descriptions of the

location and habitat of each population.

19

 

20

Several fitness traits were monitored including fruit set, total ovules
fertilized, seed set, percent germination and seedling size. Fruit set was
determined as the number of mature fruit/flowers pollinated. Total ovules
fertilized was the number of seeds, either aborted or mature, in a mature fruit.
Fertilized seeds were identified by sorting through the contents of a fruit with a
spatula, seeing the larger seeds and feeling for the hard seed coats of smaller
seeds. Both of these species can have anywhere from 100-122 total ovules in
a fruit, but in no fruit are all of the ovules fertilized and fully developed (Palser,
1961). Seed set was determined by the number of plump, mature seeds/total
number of fertilized ovules. Percent germination was the fraction of mature
seeds that germinated. Seedling growth rate was the height of a seedling
when it initiated its fourth leaf/number of days to reach that height.

Individual fruits were sorted by plant and cross type, weighed and sliced
through the middle with a razor blade and the seeds were spread on a piece of
filter paper. Mature seeds were identified as those that were plump and brown,
while ovules that were not filled or were white or small nubbins were
considered aborted (Hokanson and Hancock 2000; Figure 1). Mature seeds
were transferred to filter paper in a petri dish and moistened with a water-
phosphoric acid solution of pH=5.0. The seeds were placed in a growth
chamber under 16 hour/25° C days and 15° C nights, and seed germination
was monitored every other day from August through December taking note of

the day of germination.

21

To compare seedling growth rate, a random subset of the germinated
seeds were removed from the petri dishes and transplanted into a 50:50
mixture of Scotts Metro Mix (Scotts-Sierra Horticultural Products Co.) and peat
moss. The height of seedlings at the four leaf stage, in relation to time since
emergence, was quantified in the fall of 1997. It was necessary to factor in the
age of a seedling when it was measured, as not all seeds germinated on the
same day.

Data AnaysisnMean population and species level values were
calculated for fruit set, number of seeds, seed set, percent germination and
growth for self and outcross pollinations. Cumulative fitness was calculated
using these mean values for each trait, from each population as follows: the
mean value for each trait, for each population was divided by 100, these new
values were summed for each trait, within population and this sum was
multiplied by 100 to arrive at cumulative fitness for each population for self and
outcross pollinations. Cumulative inbreeding depression for each population
and species was calculated as 1 - ws/wo where ws and wo are cumulative
fitness for self and outcross pollinations respectively.

Fruit set, total ovules and seed set were analyzed using a three-factor
split-plot ANOVA with nesting. The sources of variation were Year, Species,
Population [Species] (random), Pollination, Pollination x Species, Pollination x
Population[Species] (random), Year x Species, Year x Population[Species]
(random), Year x Pollination, Year x Pollination x Species and Year x Pollination

x Population[Species] (random). Percent germination was analyzed using a

22

two-factor split-plot ANOVA with year nested within species because
germination data was not available for the MFR population in 1997 (Table 1).
Seedling size was analyzed with a two-factor ANOVA with Species, Pollination
and Pollination x Species as sources of variation. Seed set and seedling size
were log transformed to normalize the data. The other three variables did not
require transformation. The significance of the Pollination x Species interaction
term was tested using log transformed data. This was necessary in order to
compare the relative rather than absolute difference between selfing and
outcrossing (Johnston and Schoen, 1994).

Inbreeding depression was also measured at the family level using

relative performance (RP), where RP is 1-ws/wo when wo > ws and RP =
wo/ws - 1 when ws > wo. we is the mean fitness for a particular trait when
outcrossed and ws is the mean fitness when selfed. This estimator is more
informative than the traditional estimator of inbreeding depression (1 - ws/wo)

when one has cases where selfed pollinations perform better than outcross
pollinations (Agren and Schemske, 1993). Relative performance values that
are significantly > 0 indicate that outcrossed individuals were superior to inbred
ones, signaling inbreeding depression. Relative performance values
significantly < 0 indicate better performance of selfed individuals and
outbreeding depression. For each species, cumulative RP was calculated as
the sum of individual family RP values for fruit set, number of seeds, seed set

and % germination. The data for growth rate were not quantified on a per family

23

basis and RP values are simply means for all of the self and outcross
pollination data.

Relative survival was calculated for each species and compared to the
relative survival values of Hokanson and Hancock (2000). Relative survival is
the percent mature seed per self pollination relative to outcross pollinations. A
value is calculated for each plant, and then a mean is calculated for each
species.

Lethal equivalents (LE) were also calculated. Lethal equivalents
estimate the mean number of deleterious genes (genetic load) carried by each
member of a population multiplied by the mean probability that each gene will
cause genetic death when homozygous (Morton et al., 1956). Lethal
equivalents were estimated as:

LE = 2(-1/F)lnRS F = Fixation Index

F = 1/2 in a Diploid

F = 1/6 in a Tetraploid

RS = Relative Survivorship (°/o mature seed

per self pollination relative to outcross pollinations)
Statistical analyses were conducted using JMP 3.1.6 software (SAS Institute,
1994)

For all the traits, except seedling size, standard errors for relative
performance were calculated using family means The family level analysis
necessitated excluding some of the data because some plants did not set fruit

when self pollinated (V. myrtilloides - 14 of 32 plants, V. corymbosum - 5 of 44

24

plants). A fruit set of zero was considered relevant; however, families that set
no fruit when selfed were not considered for subsequent traits because it is not
possible to follow the fate of a seed that never existed.

RESULTS

Inbreeding depression was apparent for all the fitness traits measured
in both species, except percent germination in the diploid species, V.
myrtilloides (Table 1). However, there was a great deal of plant to plant
variation, with some individuals showing little evidence of ID, especially several
V. corymbosum clones in the CCL population. Cumulative ID was comparable
between the two species (Table 1). The estimated number of lethal
equivalents (LE) was 17.4 and 3.3 in the V. corymbosum and V. myrtilloides,
respectively (p=0.01, t-test).

Fruit set was the only trait in which there was a significant Pollination x
Species interaction (Table 2). Vaccinium myrtilloides had a more negative
response to inbreeding than V. corymbosum, particularly in 1997 (Figure 4). V.
corymbosum had a comparable response to selfing in both years, with
reductions of 42% (1997) and 40% (1998), while V. myrtilloides was reduced
by 77% (1997) and 43% (1998). Inbreeding depression measured at the family

level was significantly greater in V. myrtilloides (Figure 3; p<0.05, Mesh.

25

Soda .dea .. .modva ..

 

CCCw.MN CCCV.MF CCCN.N CCCm.N CCQ." COCV. FF CCio.Nm CGCF.m? N.N ilQ.‘ m
E. 22 66. 6.86.6 6.86.? 6. :36 6.8 F. 6 6.3.3 6. :3: 6.80.8 6. :36 6.46.8 6.36.3 :82
vom. Bow 8: 6.93» 6.336 6.316 6.0.6.6 6. :03 6. :02 6.. 3.8 6. Co? 66:66 6.36.? do

so. 3.; 63 6.65.9. 6.65.8 6.30.8 6.3.3 6.3.8 6.3.? 6.52% 6.56.2 6.66.8 6.66.8 uoo
Eamonano Eagoolmfl

 

6:. 83 62. 6N6? 6.96.3” 6.4336 6.66.9. 6.56.3 6.56.9 6.50.3 6.36.3 6.66». 66:5 :82

6%. 86. 2 F. 6.36.9. 6.3.66 6.86% 6.56.2 6.32: 6.56.8 6.36.3 2.65.6” 6.66.2 Es.
66. 66: NS. 6.82.? 6.86.3” 3.66.9. 656.8 SONS 6.83 6.58 6.66.3 666$ 6.96.2 E2
322:: E E:_c_oom>

oo 6 co 6 co 6 co 6 co m 00 w

o. 685: .55 ciao saw as am 88 886 a saw .3. as

 

 

.mcozmfiaoa 58 9: ho mcmoE 05 9:59:00 <>Oz< 636:0 m .2 9m. mm:_m> n. .995

98:86 0:... 6085:6ch E mm:_m> .00 626383206 02.6.:an -F m. E: 86858 9.66655 026.:an
685$. .838 66966 E06362. 02» 65 Go 6:86 65 Co 365; 9:3:an 6cm 3:599 35.6006 .6 SE 5306
.AELmO oi cosmEEcmm Emeoa .omm 3me «comm 2:ng 95 8.656 85 8:50 63:.th ho mmflcmocma .68mm

3 :3: Log 9235 3.2.2.9 .6 59:5: :69: dew .t..._ oi “mm ES Emocoa “commie. 2263653 on... .3.»
new 2 ___t E 833386on 3.30:3 m5 5 mctawto GOV 66669230 new Amy 3:3 .6 865: cows. .6 2an

 

26

Table 2. Effects of year, species and pollination treatment on fitness at five life-
history stages. Pop[Species], and all interactions involving it, are treated as a
random effect. Significance tests for all transformed data were conducted on
log-transformed data (Johnston and Schoen 1994).

27

 

Life-History Stage

Fruit Set

Total Ovules

Seed Set1

% Germination

Seedling Growth Rate‘

Source

Year

Species

Pop[Species]

Pomnafion

Pollination x Species
Pollination x Pop[Species]
Year x Species

Year x Pop[Species]

Year x Pollination

Year x Species x Pollination
Year x Pollination x Pop[Species}
error

Year

Species

Pop[Species]

Pollination

Pollination x Species
Pollination x Pop[Species]
Year x Species

Year x Pop[Species]

Year x Pollination

Year x Species x Pollination
Year x Pollination x Pop[Species}
error

Year

Species

Pop[Species]

Pollination

Pollination x Spades
Pollination x Pop[Species}
Year x Species

Year x Pop[Species]

Year x Pollination

Year x Species x Pollination
Year x Pollination x Pop[Species}
error

Species

YearlSpecies]

Pop[Species]

Pollination

Pollination x Species

Pollination x Pop[Species]

Year x Pop[Species]

Year x Pollination

Year x Pollination x Pop[Species]
error

Species

Pollination

Pollination x Species
BflOf

Q
qr.

NNHHNHNr—‘HNHl—J

p...
H

ONHHNHNHr—‘Nl—‘H

100

\O
(I)
mNHr—‘NHNHF—‘NH H

MIJF‘F‘NDHFJKJHrA

m
U

H

198

SS

2528.
102.
503.

5513.

1696.

2344.
769.
214.

1522.

1546.
483.

61120.

5785.
23158.
355.
4705.
3096.
1846.
186.
1823.
3.

53.
840.
220484.

688.

2008.
3087.
3134.
291.
2018.
2321.
5829.
486.
2.
1591.
310979.

17937.
5956.
3461.

162.
1007.

14228.
3579.
3372.
1947.

1133612.

15736.

r—A

6 ‘33) .
667.
73591.

~J

HOGO‘QQ‘OWOOQi—J mmNm‘OCOkOHOI-I'

(DUNONNO‘OWOWH O

OGWQNHQQUICD

L»)

11

p...

owNOthCJOOA

H

p.) b.)
.‘\
.1

b.)

HOOHONhr—IOQ

NOO\wHHOO‘.OU N
U10

HiAhah<3<>c>H»»
émewD-‘N‘JL

H .L.
m (3\

c-aaaihi»raL.L:MtoEn

c>a>a-b;slbc> N

mnoc>u>biuandhafiifi
‘

.

fix

\_I
I

m

‘

‘
n.

I>

fit‘

it,

 

*P<0.05; "P<0.01; *"P<0.001
1 Log transformed data

28

Fig. 3. Relative performance of the blueberries Vaccinium myrtilloides and \_l.
corymbosum at several life history stages with "/0 Frt. Set = fruit set, # Seeds =
the total number of fertilized ovules, Seed Set = the fraction of fertilized ovules
developing into mature fruits, °/o Germ. = the percentage of mature seeds that
germinated, Growth = the growth rate of seedlings. Fruit set is significantly
different at p=0.05 (t-test).

Figs. 4. Fruit set of selfed and outcrossed (OC) pollinations in two blueberry
species in 1997 and 1998.

29

70

 

I V. corymbosum n V. myrtilloides

 

 

% Frt. Set # Seeds Seed Set

 

 

% Germ. Growth

 

 

8

rn'llll'llllIlI'II'llllIIIIUIIIII

Fruit Set
8 8

20

 

 

 

 

—I— V.cory97
+ V.cory98
+ V.myrt97
.4. V.myrt98

 

 

Fig. 5. Total number of ovules fertilized in selfed and outcrossed (0C)
pollinations in two blueberry species in 1997 and 1998.

Fig. 6. Average seed set in selfed and outcrossed (OC) blueberries in the
years 1997 and 1998.

31

 

um 97
+ V. corymbosum 98

' oides 97

 

V. m 'lloides 9
97
98

V. o
V.m

 

 

l-I—

—I—
+
+

 

 

 

 

 

 

2.

Self

 

 

 

 

—--th- —bbFF
pP-br-nP—ppnp nun-pnpppb nub-nnph-n -b-p hb- - pupph Pb. p-

. . 661%.... .. .. .. . . ........... m .. ..

 

 

32

Self

Fig. 7. Percent germination in two blueberry species for self and outcross (0C)
pollinations.

Fig. 8. Growth rate in two blueberry species for self and outcross (OC)
pollinations.

33

 

 

me mm
.W ....
H

mm. mm
_: _:

 

 

 

 

 

 

 

T
Self

 

 

hp-h—n...—.-..—P--—p-..—..- —-.. --—--—--—--.—--—--—-.-

o o 0 o 0 o 0
m w w 5%?ng as. 2 1 7 w mum”— fiWEHU Wan—comm 1

 

 

 

 

Self

The two species had comparable reductions in the total number of
ovules fertilized when selfed. For both species, the percent reduction was
greater in 1997 than 1998, when more ovules were fertilized (Figure 5). The
number of fertilized ovules was reduced 13% (1997) and 7% (1998) in v.
corymbosum, and 28% (1997) and 14% (1998) in \_l. myrtilloides. Inbreeding
depression measured at the family level was not significantly different in the
two species (Figure 3).

There was significant inbreeding depression in both species for seed
set, with average reductions of 50% and 32% in y. myrtilloides and \_l.
mrymbosum respectively (Figure 6). \_/. myrtilloides fitness was reduced 64% in
1997 vs. 39% in 1998, while y. corymbosum seed set was reduced 34% and
30% in 1997 and 1998, respectively. The degree of inbreeding depression
measured at the family level was greater in \_l. myrtilloides (Figure 3) at the p<
0.10 significance level .

For percent germination, species was the only significant source of
variation in the ANOVA, although there was only one year of data for \_l.
myrtilloides (Table 2). Outcrossed progeny of LI. corymbosum germinated 30%
greater than y. rmrtilloides; however, the decrease in germination after selfing
(14% vs. 10% decrease) was not significantly different between y. corymbosum
and \_l. myrtilloides (Figure 7).

Inbreeding depression was also apparent in seedling growth rate
(Figure 8). Species was significant in the ANOVA, because y. corymbosum

seedlings were, in general, taller than those of \_/. myrtilloides. However, the

35

relative performance of inbred \_l. myrtilloides progeny was higher (Figure 3). V.
corymbosum seedlings had a 22% decrease in size when selfed, while \_l.
myrtilloides decreased 15%.

Absolute values of relative survival were higher in our study than in
Hokanson and Hancock and were significantly different for y. myrtilloides
(Table 3).

Table 3. A comparison of self and outcross performance between Hokanson
and Hancock (2000) and the current study. Relative survival is the percent
mature seed per self pollination relative to outcross pollination (standard error).
Percent self is the percentage of plants that set a self pollinated fruit. Lethal

equivalents is an estimate of the genetic load . See the text for a detailed
description of lethal equivalents.

 

 

Relative n Percent Lethal
Survival Self Equivalents

Hokanson

V. myrtilloides 0.02 (.01) 4 27 15.7

V. corymbosum 0.26 (.06) 7 47 18.6

Schott

V. myrtilloides 0.38 (.10) 14 60 3.3

V. corymbosum 0.28 (.09) 18 90 17.4

DISCUSSION

Overall, inbreeding depression was quite severe in both _V. corymbosum
and \_l. myrtilloides as was shown previously by Hokanson and Hancock
(2000). This is expected in species that are presumed to have high
outcrossing rates. While there have been no direct estimates of outcrossing
rates in blueberries, their negative response to selfing, floral architecture and

wide ranging bee pollinators would suggest that they have high realized

36

outcrossing rates. Still, these two species of blueberries are hardly self-
infertile or self—incompatible, particularly \_l. corymbosum. There is a great deal
of intraspecific variation in selfing ability, with some genotypes performing as
well or better when selfed and others that are relatively self-infertile (Krebs and
Hancock, 1991; Hokanson and Hancock, 2000).

Vander Kloet and Lyrene (1987) suggested that geitonogamous
pollinations are probably very common in blueberries and this may be one
reason why a small fraction of their total ovules reach maturity. The low
number of fertilized ovules and low seed set observed in this study are
consistent with previous studies and support the idea that natural pollination
involves many geitonogamous pollinations.

A comparison of Hokanson and Hancock (2000) to this study, indicates
that relative survival and genetic load were comparable in y. corymbosum but
differed considerably for y. myrtilloides. This field study observed much higher,
and more diverse levels of self fertility in _\_l. myrtilloides than Hokanson and
Hancock's greenhouse study. As a result, estimates of genetic load are much
higher in Hokanson and Hancock. My data may more accurately reflect the
nature of fertility in y. myrtilloides, as my study had a considerably larger
sample size, covered two years instead of one, and was performed in the field.

It is likely that ploidy level influences the self-fertility of diploid V.
myrtilloides and tetraploid V. corymbosum. The response of diploid and
polyploid species to ID can have a complex genetic basis and lead to

considerable variation in response to inbreeding. Inbreeding is known to

37

purge diploid populations of genetic load (Barrett and Charlesworth 1991), but
it is less effective in polyploids. Relatively poor performance of selfed diploids
is usually attributed to the expression of deleterious alleles in what is known as
the partial dominance model of ID. Relatively poor performance of selfed
polyploids is usually attributed to the loss of higher order allelic interactions in
what is known as the overdominance model of ID (Bever and Felber, 1992).
Serial inbreeding will increase levels of homozygosity and purge deleterious
recessive alleles relatively quickly in diploids. However, polyploids reach
homozygosity much slower than diploids, and if ID in polyploids is due to the
loss of higher order allelic interactions, inbreeding and purging are less likely
to yield plants and populations free of ID.

In this study, the diploid y. myrtilloides had fewer lethal equivalents than
tetraploid \_/. corymbosum, but \_l. myrtilloides still carried sufficient numbers of
deleterious alleles to significantly reduce its self-fertility. The highly outcrossed
mating system of \_l. myrtilloides, and high reproductive output, may have
slowed its approach to homozygosity.

The results of other studies examining the response to inbreeding in
related diploid and polyploid species have varied. Polyploid complexes with
lower seed and/or fruit set in diploids include wheatgrass (Dewey, 1969),
maize (Alexander, 1960), clover (Townsend and Remmenga, 1968), ferns
(Hedrick, 1987), and Epilobium angustifolium (Husband and Schemske,
1997). Polyploid complexes with less ID in diploids include orchardgrass

(Kalton et al., 1952) and Amsinckia sp. (Johnston and Schoen, 1996).

38

Husband and Schemske (1997) examined several life history stages in
Epilobium angustifolium and found consistently higher ID in diploids across
stage, although selfing rates in the two ploidy levels were nearly identical, and
relatively high (r=0.45). They attributed greater ID in the diploid to the
expression of deleterious alleles in the homozygous state. The diploid studied
by Johnston and Schoen (1996) was highly selfed and suffered significantly
lower lD than the outcrossing tetraploid. Vander Kloet and Lyrene (1987)
demonstrated significant, and comparable, ID in diploid, tetraploid and
hexaploid y. corymbosum for seed set and seedling survival. Hokanson and
Hancock (2000) found significantly greater ID in diploid y. myrtilloides, relative
to tetraploid y. corymbosum for seed set, but the level of ID in tetraploid \_/.
angustifolium was comparable to \_l. myrtilloides. The variation among these
studies in the response of diploids and polyploids to inbreeding is not
surprising, as the ability of a species to purge deleterious alleles depends on
many factors including not only its ploidy, but also its mating system, selection
pressure and population size.

The degree of ID varied across developmental stage in this study, with
the highest values occurring for fruit set and seed set in both species, although
the difference in the magnitude of ID between early and late stages was much
greater in the diploid. These results suggest that the diploid may be more
effective in purging, genetic load at these early stages than the polyploid, and
the progeny that make it through the selective sieve consist of generally highly

fit genotypes. Beneficial allelic combinations are probably not broken up as

39

quickly in the tetraploid and while there is ID, there are fewer low quality
individuals at the early reproductive stages amongst the tetraploids. In fact, the
results for total fertilized ovules (Figure 3) show relatively low ID, suggesting
that self pollen is only somewhat poorer at fertilizing ovules than outcrossed
pollen.

Husband and Schemske (1996) found that in primarily outcrossing
species most expressed ID at early (seed set) or late (growth and
development) life history stages, while primarily selfing species expressed the
majority of their ID at later stages. Their explanation was that outcrossers carry
many strong recessive lethals, that are readily purged in early stages while the
selfers carry only weakly deleterious mutations that are hard to purge until late
stages. Husband and Schemske (1996) predicted that if the nature of ID is the
same in two species, there should be consistent levels of ID across stages,
but they did not consider ploidy level. The data described here My indicate that
the nature of ID may differ between the diploids and tetraploids. The diploid
may be expressing strong deleterious alleles in the homozygous state at early
stages, while the more gradual decline in ID across stages in the tetraploid
may be due to the loss of higher order allelic interactions as per the
overdominance model of ID (Bever and Felber, 1992).

The higher degree of self-fertility in the polyploid may also have a simple
historical basis. Vaccinium corymbosum presumably arose from diploid
progenitors as a minority cytotype in a predominately diploid population.

Consequently, most matings involving the new polyploid would be diploid x

40

polyploid crosses resulting in inviable or infertile triploid hybrids. The ability of
the new polyploid to persist requires a certain degree of self fertility. Tetraploid
blueberries in the middle latitudes of North America are relatively young
species that may have multiple origins (Vander Kloet, 1988). Relatively high
self fertility in current tetraploid populations may be an historical artifact of their
creation and establishment.

The difficulty Camp (1945) and Vander Kloet (1988) have had in
classifying North American Vaccinium species and the intraspecific fertility that
exists (Ballington and Galletta, 1978) suggests this is a young, rapidly evolving
phyletic group. Indeed, the fact that morphologically indistinguishable \1.
corymbosum populations exist at the diploid, tetraploid and hexaploid levels
(Vander Kloet, 1988) indicates that this group is in the early stages of
speciation. The mating system of the group is probably evolving as well. It may
be moving from a state of low self-fertility, due to ID, to self-incompatibility at the
postzygotic stage. The variation in levels of self-fertility that exists in the
populations studied suggest past purging of deleterious alleles in certain
genotypes or segregation of self-compatibility and self-incompatibility alleles.

The response to inbreeding and nature of self-infertility is quite variable
among many plant species. The response in blueberries depends on cytotype,
with diploids easily purging deleterious alleles and tetraploids purging with
greater difficulty. Blueberries have, in general, low self-fertility and are probably

highly outcrossed. However, they are clearly not self-incompatible, especially

41

the tetraploid _\_/. corymbosum. Vaccinium corymbosum may have relatively high

self fertility due to its recent hybrid origin from diploid progenitors.

42

LITERATURE CITED

AGREN, J. AND D. W. SCHEMSKE. 1993. Outcrossing rate and inbreeding
depression in two annual monoecious herbs, Begonia _hfis_uta and _B.
semiovata. Evolution 47: 125-135.

ALEXANDER, D. E. 1960. Performance of genetically induced corn tetraploids.
American Seed Trade Association 15: 68-74.

BALLINGTON, J. R. AND G. J. GALLETA. 1978. Comparative crossability of 4
diploid Vaccinium species. Journa_l of the American Society of
Horticultural Science 103: 554-560.

BARRETT, S. C. H. AND D. CHARLESWORTH. 1991. Effects of a change in the

level of inbreeding on the genetic load. Nature 352:522-524.

 

BEVER, J. D. AND F. FELBER. 1992. The theoretical population genetics of
autopolyploidy. Ln D. Futuyma and J. Antonovics [eds.], Oxford Surveys in
Evolutionary Biology, 185-217. Oxford University Press.

CAMP, W. H. 1942. The structure of populations in the genus Vaccinium.

Brittonia 4:189-204.

 

. 1945. The North American blueberries with notes on other groups of

Vacciniaceae. Brittonia 52203-275.

 

CHARLESWORTH, D. and B. CHARLESWORTH. 1989. Inbreeding depression and
its evolutionary consequences. Annual Review of Systematics and

Ecology 18: 237-268.

43

DAEHLER, C. C. 1999. Inbreeding depression in smooth cordgrass (Spartina

alterniflora Poaceae) invading San Francisco Bay. American Journal of

 

_B_ot_apy. 86: 131-139.

DEWEY, D. R. 1966. Inbreeding depression in diploid, tetraploid, and
hexaploid crested wheatgrass. Crop Science 6: 144-147.

_. 1969. Inbreeding depression in diploid and induced autotetraploid
crested wheatgrass. Crop Science 9: 592-595.

HANCOCK, J. F. 1992. Plant evolution and the origin of crop species. Prentice
Hall, Englewood Cliffs, NJ.

HEDRICK, P. W. 1987. Genetic load and the mating system in homosporous
ferns. Evolution 41: 1282-1289.

Hokanson, K. E. 1995. The consequences of polyploidy on inbreeding
depression in Vaccinium (blueberry) species. Ph. D. dissertation,
Michigan State University.

_, and J. F. Hancock. 2000. Early-acting inbreeding depression in three

species of Vaccinium (Ericaceae). Sexual Plant Reproduction (In press).

 

HUSBAND, B. C. AND D. W. SCHEMSKE. 1996. Evolution of the magnitude and
timing of inbreeding depression in plants. Evolution 50: 54-70.

_. 1997. The effect of inbreeding in diploid and tetraploid populations of
Epilobium angustifolium (Onagraceae): implications for the genetic
basis of inbreeding depression. Evolution 51: 737-746.

JOHNSTON, M. O. AND D. J. SCHOEN. 1994. On the measurement of inbreeding

depression. Evolution 48: 1735-1741.

44

_. 1996. Correlated evolution of self-fertilization and inbreeding
depression: an experimental study of nine populations of Amsinckia
(Boraginaceae). Evolution 50:1478-1491.

KALTON, R. R., A. G. SMIT and R. C. LEFFEL. 1952. Parent-inbred progeny
relationships of selected orchardgrass clones. Agronomy Journal 44:
481-486.

KREBS, S. L. AND J. F. HANCOCK. 1989. Tetrasomic inheritance of isoenzyme
markers in the highbush blueberry, Vaccinium corymbosum. Helm
63: 11-18.

_. 1990. Early-acting inbreeding depression in reproductive success in
the highbush blueberry, Vaccinium corymbosum L. Theoretical and
Applied Genetics 79: 825-832.

_. 1991. Embryonic genetic load in the highbush blueberry Vaccinium
corymbosum (Ericaceae). American Journal of Botany 78:1427-1437.

LANG, G. A., R. G. DANKA AND E. J. PARRIE. 1990. Pollen-stigma interactions
and relationship to fruit development in low-chill southern highbush
blueberries. Proceedings of the Sixth North American Blueberry
Research and Extension Workers Conference. Portland, Oregon. pp.
36-44.

LANG, G. A. AND R. G. DANKA. 1991. Honey-bee-mediated cross- versus self-
pollination of 'Sharpblue' blueberry increases fruit size and hastens
ripening. Journal of the American Society of Horticultural Science 116:

770-773.

45

MAYER, S. 8., D. CHARLESWORTH AND B. MEYERS. 1996. Inbreeding
depression in four populations of Collinsia heterophylla Nutt
(Scrophulariaceae). Evolution 50: 879-891.

MORTON, N. E., J. F. CROW AND H. J. MULLER. 1956. An estimate of mutational
damage in man from consanquineous marriages. Proceeding of the
National Academy of Sciences, USA 42: 855-863.

PALSER, B. F. 1961. Studies of floral morphology in the Ericales. V.
Organography and vascular anatomy in several United States species of
the Vacciniaceae. Botanical Gazette 123: 79-111.

PRAY, L. A., J. M. SCHWARTZ, C. J. GOODNIGHT and L. STEVENS. 1994.
Environmental dependency of inbreeding depression: implications for
conservation biology. Conservation Biology 8: 562-568.

PROCTOR, M., P. YEO, AND A. LACK. 1996. The natural history of pollination,
pp. 323-330. Timber Press, Portland, Oregon.

RATHCKE, B. AND L. REAL. 1993. Autogamy and inbreeding depression in

mountain-laurel, Kalmia latifolia (Ericaceae). American Journal of

 

 

_Bgta_ny 80: 143-146.

SAS INSTITUTE. 1994. Statistical software for the Macintosh. JMP, statistics
and graphics guide. SAS Institute, Inc., Cary, NC.

SAGE, T. L., R. I. BERTIN, AND E. G. WILLIAMS. 1994. Ovarian and other late -
acting self-incompatibility systems. In_ E. G. Williams, R. B. Knox, and A.

E. Clarke [eds.], Genetic control of self-incompatibility and reproductive

46

development in flowering plants, 116-140. Kluwer Academic,
Amsterdam.

SCHMITT, J. AND S. E. GAMBLE. 1990. The effect of distance from the parental
site on offspring performance and inbreeding depression in Impatiens
capensis: a test of the local adaptation hypothesis. Evolution 44: 2022-
2030.

SEAVEY, S. F. AND K. S. BAWA. 1986. Late-acting self-incompatibility in
angiosperms. Botanical Review 52: 195-218.

SOLTIS, D. E. AND P. S. SOLTIS. 1999. Polyploidy: recurrent formation and
genome evolution. Trends Ecology and Evolution 14: 348-352.
TOWNSEND, C. E. AND E. E. REMMENGA. 1968. Inbreeding in tetraploid alsike

clover, Trifolium hybridum L. Crop Science 8: 213-217.

VANDER KLOET, S. P. 1980. The taxonomy of the highbush blueberry,
Vaccinium corymbosum. Canadian Journal of Botany 58: 1187-1201.

_. 1988. The genus Vaccinium in North America. Research Branch
Agriculture Canada, Publication 1828.

______. 1991. The consequences of mixed pollination on seed set in
Vaccinium corymbosum. Canadian Journal of Botany 69: 2448-2454.

VANDER KLOET, S. P. and l. V. HALL. 1981. The biological flora of Canada: 2.
Vaccinium myrtilloides Michx., velvet-leaf blueberry. Canadian Field

Naturalist 95: 329-345.

47

VANDER KLOET S. P. and P. M. LYRENE. 1987. Self-incompatibility in diploid,
tetraploid, and hexaploid Vaccinium corymbosum. Canadian Journal of

Botany 65: 660-665.

48

CHAPTER 2
THE EFFECT OF CROSSING DISTANCE ON FERTILITY

IN VACCINIUM SPECIES

49

ABSTRACT

Outbreeding depression (OD) is defined as a reduction in fitness when
distantly related or distantly spaced individuals are mated, compared to when
more closely spaced individuals are crossed. This reduction is thought to
occur when locally coadapted gene complexes are broken up in areas where
there is spatial variation in selective regimes. I examined whether or not OD is
important in diploid (Vaccinium myrtilloides) and tetraploid (V. corymbosum)
blueberries, by performing self and outcross pollinations at differing distances
in four populations (two for each species). I corroborated earlier studies
demonstrating significant inbreeding depression in both of these species, but
found that parental distance had no effect on outcross performance in V.
cogmbosum. In \_/. myrtilloides, there was higher fruit set and percent
germination at intermediate outcrossing distances. Differences in ploidy level,
and clonal growth patterns, may contribute to the differing responses of these

species to crossing distance.

50

INTRODUCTION

Mating systems in plants evolve depending on how different types of
crosses effect the fitness of individuals and progeny. Inbreeding depression
(ID), outbreeding depression (OD), decreased resources and insufficient
pollinators can all influence the reproductive output of individuals. In the
previous chapter, I demonstrated that there is significant ID for fruit set, ovule
fertilization, seed set, germination and seedling growth in two species of wild
blueberries \_l. corymbosum and \_/. myrtilloides. This study was designed to
examine the importance of OD in these same two species.

Outbreeding depression is defined as a decrease in any fitness trait
when mating occurs between distantly related, and possibly distantly spaced
individuals. It arises when there is spatial variation in selection regimes
resulting in the evolution of specific locally adapted genes or gene complexes.
An intermediate, optimal outcrossing distance may emerge when populations
are highly differentiated and maximum fitness is associated with intermediate
genetic similarities between parents (Campbell and Waser, 1987).

Studies searching for optimal outcrossing distances have both
supported and refuted its existence. Vander Kloet and Lyrene (1987) found that
crosses between unrelated, presumably more distantly spaced, individuals of
y. corymbosum performed better than self-pollinated individuals. Species with
intermediate, optimal outcrossing distance include Zfl mgys (Moll et al.,
1965), Picea canadensis (Coles and Fowler, 1976), Delphinium nelsonii

(Waser and Price, 1979 & 1994), lpomopsis aggregata (Price and Waser,

51

1979), Amphicarpaea bracteata (Parker, 1992) and Agave schottii (Trame et al.,

 

1995). Species without an optimal outcrossing distance include gaija
tembloriensis (Holtsford, 1996), Gentiana pneumonanthe (Oostermeijer et al.,
1995), and Chamaecrista fasciculata (Fenster and Sork, 1988). Clearly,
numerous factors interact to determine which types of pollinations are
successful in a species including: how closely related is the pollen received by
a plant, levels of genetic differentiation, the genetic system that accepts or
rejects pollen, the number of deleterious alleles carried, patterns of
segregation, and environmental effects.

The evolution of mating systems in hermaphroditic plants is complex.
The degree of pollen diversity accepted by a plant is dependent on the
importance of genetic differentiation in regulating fertility. The plant is
constrained by what pollinators bring to the plant and what arrives via wind and
water. Several mechanisms have evolved to modify the degree to which a plant
receives self or outcross pollen. Physiologically, plants can accept or reject
self pollen through incompatibility systems. Morphologically, inflorescence
design can influence rates of geitonogamy. Dispersal distance and plant
spacing can also regulate the degree to which biparental inbreeding occurs.

The coevolution of plants and pollinators is also complex, with fitness
being maximized in each to the point that does not compromise the fitness of
the other. The genetic system of the plant will constrain the coevolutionary
process and determine the extent to which it will accept or reject self or

outcross pollen. lf pollination between close related and/or spaced plants

52

results in biparental inbreeding, and inbreeding is deleterious, then plant
reproductive success will be highest with pollen from distant sources. If local
adaptation is more important, then distant pollen will reduce reproductive
success. Examining the effect of pollen source on fitness characters can lend
insight into the evolution of a mating system and whether or not genetic
differentiation is important.

Herein, a crossing experiment is described using Vaccinium myrtilloides
and \_l. corymbosum that tests whether or not there is an optimal outcrossing
distance in these diploid and tetraploid species. The data support previous
work demonstrating that ID has a negative impact on the reproductive success
of both these species (Hokanson and Hancock, 2000; Schott, 2000), but OD
was found to be only important in y. myrtilloides.

MATERIALS AND METHODS

Four different types of crosses were made in the field: 1) self - self
pollination, 2) close - pollen from five local, randomly selected donors (<10 m
apart, 3) distant — pollen from five distant, randomly selected donors from the
same population (100-200 m apart), 4) other — pollen from five, randomly

selected donors from another population.

53

Fig. 1. The locations of study populations. Vaccinium corymbosum at Otis
Lake (COL), y. commbosum at Crooked Lake (CCL), \_I. myrtilloides at Fox River
(MFR),y. myrtilloides at Poe Reef (MPR). The Black River population served as
the distant pollen donor for the MPR population. See the text for detailed
descriptions of the location and habitat of each population.

Pollen was pooled for all the outcrossed pollinations.

54

OU‘fliflfl

 

55

Pollen was collected by removing several recently opened flowers from a
donor plant and rolling each of them between the thumb and forefinger, which
caused the pollen to dehisce from the anthers. Pollen was collected on a cover
slip and kept in a small petri dish. Outcross pollen was applied
intraspecifically using the CCL population as the 'Other" donor for the COL
population and vice versa. The MPR population served as the 'Other' donor for
the MFR population; however, the reverse was not possible. The MFR
population was not releasing pollen when the MPR population was ready for
pollination. Instead, pollen was obtained from a more southern Upper
Peninsula population as the source of 'Other' pollen for MPR. This population
was located at the Black River Campground in the Lake Superior State Forest
(Fig. 1). There were nine experimental plants in each of the y. corymbosum
populations and four and five experimental plants in the MPR and MFR
populations, respectively. Vaccinium myrtilloides populations are, in general,
less robust than \_l. corymbosum and there are fewer plants with enough
flowers on which to perform the pollinations. On each _\_/. corymbosum plant, an
average of 23 self-pollinations were performed, along with 12 of each of the
three outcross pollinations. On each \_/. myrtilloides plant, an average of 19
self-pollinations were performed, along with 13 of each of the three outcross
pollinations.

Five fitness traits were analyzed including fruit set, total number of
fertilized ovules, seed set, percent germination and seedling size. Fruit set

was calculated as the proportion of pollinated flowers that developed into

56

mature fruit. Total number of fertilized ovules was averaged across fruit in each
treatment combination. Blueberries have 100-122 ovules per fruit (Palser
1961), and only a subset of these ovules are fertilized (approx. 44%) and
develop into mature seeds (approx. 24%) (Schott, 2000). Seed set is defined
as the proportion of fertilized ovules that developed into mature, plump seed
(Schott, 2000). To calculate percent germination, seeds were extracted from
ripe fruit and kept moist in a pH=5 solution of water and phosphoric acid. The
seeds were isolated by fruit in a petri dish and kept in a growth chamber set at
20°C, 15 hour day and 15°C night. Germination was monitored twice weekly
from late August until the middle of January, when it had ceased. A subset of
the germinated seeds were transferred to a 50:50 mixture of peat moss:Scots
Metro Mix 360 (Scotts-Sierra Horticultural Products Co.). Seedling height was
measured when a seedling had initiated its fourth set of leaves, which was, on
average, 76 days after germination and 65 days after emergence from the soil
surface. The seedlings were grown in the growth chamber set at 20°C, 15 hour
day and 15°C night.

The data were analyzed in a two-factor split-plot ANOVA with nesting.
The sources of variation were Species, Populations[Species] (random),
Pollination, Pollination x Species and Pollination x Population[SpecieS]
(random). The seed set data was log transformed and the seedling size data
was arcsine-square root transformed to improve normality. The Pollination by
Species interaction term indicates whether the response to pollination

treatment varies between species. The significance of this interaction was

57

tested on log transformed data for all traits, in order to determine the relative
difference between selfing and outcrossing treatments (Johnston and Schoen
1994). All analyses were performed using JMP 3.1.6 software for the MAC (SAS
Institute 1994).

Relative performance (RP) was calculated across families as an
alternative to the traditional measure of ID (1 - ws/wo). RP is 1-ws/wo when we
2 ws and wo/ws - 1 when ws > wo. we is the mean fitness for a particular trait
when outcrossed and ws is the mean fitness when selfed. RP is a more
appropriate estimator of ID when you have individuals performing better when
selfed (Agren and Schemske 1993). Relative performance values significantly
> 0 indicate better performance of outcrossed than selfed individuals and ID.
Relative performance values significantly < 0 indicate better performance of
selfed individuals and outbreeding depression. RP was calculated on all the
data for fruit set, but only those plants producing selfed seed were analyzed for
the other parameters.

RESULTS

Pollination treatment had a significant affect on fruit set (Table 1), but the
performance of the two species and their populations were not significantly
different. Selfs performed relatively poorly, while the various outcross
treatments performed equally well (Figure 2). Outcross performance in \_l.
myrtilloides peaked at the intra-populational 'Distant' treatment, suggesting an
optimal outcrossing distance, but this difference was not significant. In \_I.

corymbosum, fruit set was lower in selfed than out-crossed pollen, but the

58

Table 1: Effects of species and pollination treatment on fitness at five life-
history stages in Vaccinium myrtilloides and \_l. corymbosum blueberries.
Pop[Species], and all interactions involving it, is treated as a random effect.
The Pollination x Species interaction was tested on log transformed data for all
stages as per Johnston and Schoen 1994.

 

Stage
Fruit Set

Total Seeds/

Fruit

Seed Set1

"/0 Germination

Seedfing
Growth Rate2

Source

Species

Pop[Species]

Pollination

Pollination x Species
Pollination x Pop[Species]
Error

Species

Pop[Species]

Pollination

Pollination x Species
Pollination x Pop[Species]
Error

Species

Pop[Species]

Pollination

Pollination x Species
Pollination x Pop[Species]
Error

Species

Pop[Species]

Pollination

Pollination x Species
Pollination x Pop[Species]
Error

Species

Pop[Species]

Pollination

Pollination x Species
Pollination x Pop[Species]
Error

df

mOwwN-t ‘waNA "OQQN-l meUN-e

gawwN—s

88
84.7
3967.0
6754.2
2198.7
4855.7
66310.6

24457.9

1994.4
22713.1
22695.1
13984.7

222284.3

6960.4
9360.3
35477.9
1 5599.6
1 1778.9

3073858

14683.3

4625.8
29096.6
15583.3
13736.8

7771472

0.08
0.1
0.3
0.2
0.1
9.7

F
0.07
2.9
3.7*
0.2
1.1

30.6
0.4

23.9***
4.4*
7.4***

2.0
2.4
32.0***
3.5
6.3***

7.7

1 .0
9.7”
2.6
2.3*

1.1
2.6
4.8“
3.5*
1.0

 

1: log transformed

2: arcsine-square root transformed
*p<.05, **p<.01, ***p<.001

59

Fig. 2. The percent fruit set in V. myrtilloides and V. cogmbosum blueberries
for self, close (pollen source within 10 m), distant (pollen source 100-200 m
distant) and other (reciprocal pollination between populations within species)

pollinations.

Fig. 3. The relative performance (RP) for fruit set in y. myrtilloides and V.
corymbosum blueberries for self, close (pollen source within 10 m), distant
(pollen source 100-200 m distant) and other (reciprocal pollination between
populations within species) pollinations. [RP is 1-ws/wo when WC 2 ws and
wo/ws - 1 when ws > wo, ws and wo are the fitness of outcross and selfed
individuals respectively] An RP value > 0 represents ID while a value < than 0

represents OD.

60

 

Fruit Set

888838

I'lllllllllllllI‘T11T'1'l'llilIllIlfl—II

20

 

 

—I- may.
-O- V. mm' Oidfi

 

 

 

Self Close Dist

 

I V. corymbosum n V. myrtilloides

 

€04:
s --=

m n
0- 0.3-5
(3 -

 

.2 I
E 0.2-E
0 .
(I 0.1-5

0.:

Close Dist

 
 
    

 

Other

 

 

Fig. 4. The total number of ovules fertilized (aborted & mature) in V. myrtilloides
and V. corymbosum blueberries for self, close (pollen source within 10 m),
distant (pollen source 100-200 m distant) and other (reciprocal pollination

between populations within species) pollinations.

Fig. 5. The relative performance (RP) for total number of ovules fertilized in V.
myrtilloides and V. corymbosum blueberries for self, close (pollen source
within 10 m), distant (pollen source 100-200 m distant) and other (reciprocal
pollination between populations within species) pollinations. [RP is 1-ws/wo
when WO a WS and wo/ws - 1 when ws > wo, ws and wo are the fitness of
outcross and selfed individuals respectively] An RP value > 0 represents ID

while a value < than 0 represents OD.

62

8

 

\l
O

Total#FertilizedOvules
8 8 8 8

N
O

p—l
O

 

‘I'IlIIIIII'II'IIIIIIIII'IIIIIIIII'IIII

 

 

O

 

 

+ CCL
+ COL
+ MFR

+MPR

 

 

I V. commbosum I V. myrtilloides

 

Relative Performance - Total
N

—L
lllllllllllllll

 

O

 

 

Close Dist Other

63

 

Fig. 6. The percentage of fertilized ovules developing into mature seeds (seed
set) in V. myrtilloides and V. corymbosum blueberries for self, close (pollen
source within 10 m), distant (pollen source 100-200 m distant) and other

(reciprocal pollination between populations within species) pollinations.

Fig. 7. The relative performance (RP) for seed set in V. myrtilloides and V.
corymbosum blueberries for self, close (pollen source within 10 m), distant
(pollen source 100-200 m distant) and other (reciprocal pollination between
populations within species) pollinations. [RP is 1-ws/wo when WC 2 ws and
wo/ws - 1 when ws > wo, ws and WC are the fitness of outcross and selfed
individuals respectively] An RP value > 0 represents ID while a value < than 0

represents OD.

64

 

VIII'IIIIIIllllllfilfiir1iiii

 

 

 

 

—.- CCL
+ COL
+ MFR
+ MPR

 

 

 

Self Close Dist

Other

 

 

I V. cogmbosum I V. myrtilloides

 

 

 

Close Dist

65

 

 

Other

 

 

Fig. 8. The percentage of mature seed germinating in V. myrtilloides and V.
corymbosum blueberries for self, close (pollen source within 10 m), distant
(pollen source 100-200 m distant) and other (reciprocal pollination between

populations, within species) pollinations.

Fig. 9. The relative performance (RP) for seed germination in V. mflilloides
and V. corymbosum blueberries for self, close (pollen source within 10 m),
distant (pollen source 100-200 m distant) and other (reciprocal pollination
between populations within species) pollinations. [RP is 1-ws/wo when wo 2
ws and wo/ws - 1 when ws > wo, ws and wo are the fitness of outcross and
selfed individuals respectively] An RP value > 0 represents ID while a value <

than 0 represents OD.

66

0.9

 

0.8

0.7

%Germmabon
. . 9
Ln
IIIl'llil'lllllIIIT‘III‘IIIII'IIIUIIIIV‘IIUI

 

 

 

Self Close

Other

 

 

—I— CCL
—O— COL
+ MFR
+ MPR

 

 

.0
co

- V. cogmbosum

n v. myrtilloides

 

.0 .0 .0 9 P P .0 .0
M 0) «h 01
llllllllllllllllll

Relative Performance - % Germination

 

O

 

Close

Dist

67

 

Other

 

 

Fig. 10. Seedling growth rate in V. myrtilloides and V. corymbosum blueberries
for self, close (pollen source within 10 m), distant (pollen source 100-200 m
distant) and other (reciprocal pollination between populations, within species)
pollinations. The value for seedling growth rate represents the height of the
seedling, in mm, divided by the number of days since the seedling had

emerged from the soil.

Fig. 11. The relative performance (RP) for growth rate in V. myrtilloides and V.
corymbosum blueberries for self, close (pollen source within 10 m), distant
(pollen source 100-200 rn distant) and other (reciprocal pollination between
populations within species) pollinations. [RP is 1-ws/wo when wo 2 WS and
wo/ws - 1 when wS > wo, ws and WC are the fitness of outcross and selfed
individuals respectively] An RP value > 0 represents ID while a value < than 0

represents OD.

68

 

0.7 _
0.6 L

0.5 L

 

—I— V. coMum
+ V. mygtjlloidg

 

 

 

 

 

Self Close

Other

 

.0
on

I V. commbosum

n V. myrtilloides

 

.0
O)
IIIII

.0
1‘

.0
N
JJJIJJALJI

02-]
-0.4-‘

1
-0.6-:

Relative Performance - Growth Rate
0

 

.
-0.8 ‘

 

 

 

Close

Dist

Other

69

source distance was not important. Relative performance for fruit set was
comparable between pollination treatments and species (Figure 3), although
the absolute values were lower in V. myrtilloides suggesting less ID.

The total number of ovules fertilized also differed significantly between
pollination treatments. The response varied depending on the population
(Table 1). The COL and MFR populations showed significant ID for most
pollination treatments, while in the CCL and MPR populations, selfed and
outcrossed progeny had a comparable performance for all but the most distant
cross in the MPR population (Figure 4). Only in the MFR population was there
evidence of an optimal, intermediate, outcrossing distance (Figure 4).
Inbreeding depression was consistently higher in V. myrtilloides than V.
corymbosum. but only significantly for the 'Distant' treatment (Figure 5).

Pollination treatment had a significant effect on seed set. The effect
depended on species and population (Table 1). The CCL, COL and MFR
populations had a significant boost in performance when outcrossed. The
COL population showed a significant intermediate, optimal outcrossing
distance for this trait, while the MPR population again had the highest
performance with the most distant pollination treatment (Figure 6). Inbreeding
depression was highest for this trait in V. myrtilloides (Figure 7). 3

Pollination treatment and population had significant effects on
germination (Table 1). Germination rate was higher in V. corymbosum
populations for all treatments except 'Distant' and there was less ID in V.

corymbosum. The MFR population showed a steady increase in germination

7O

as pollen distance increased while the MPR population showed an
intermediate optimal outcrossing distance (Figure 8). Again, ID was greater in
V. myrtilloides than V. corymbosum, and selfed and outcrossed progeny of V.
corymbosum had similar germination rates, except for the most distant
outcross treatment (Figure 9).

For seedling growth rate, the performance of V. corymbosum progeny
was similar across treatments, while in V. myrtilloides there was considerable
outbreeding depression across all treatments (Figure 11). While few self-
pollinated progeny made it through to the seedling stage in V. myrtilloides,
those that did had significantly higher growth rates than outcrossed seed.
Many of the germinated seeds emerged from the soil, but died before they
could be measured. Over 70% of the transferred seeds from the MPR
population emerged from the soil surface, but none of them survived.

DISCUSSION

There was not a uniform response to crossing distance between V.
myrtilloides and V. commbosum, and between populations within species,
especially for the V. myrtilloides populations. This result is comparable to that
of Waser et al. (2000) who found variable expression of outbreeding
depression among cohorts of Ipomopsis aggregata. Vaccinium myrtilloides
showed evidence of an optimal outcrossing distance, or outbreeding
depression for several traits; however, V. corymbosum had a relatively uniform
response to outcrossing distance. The presence of outbreeding depression in

V. myrtilloides suggests that there is more genetic structuring within these

71

populations than V. corymbosum. This may be related to ploidy differences, as
the tetraploid V. corymbosum has been shown to be much more heterozygous
than the diploid V. myrtilloides (Hokanson and Hancock, 2000) and this
increased heterozygosity may adapt it to a broader range of environmental
conditions. There also may be selection within the polyploid genome to
maintain higher order allelic associations, as per the overdominance model of
inbreeding depression (Charlesworth and Charlesworth, 1987). In previous
inbreeding studies done on cultivated blueberries, Krebs and Hancock (1988)
found that all cultivars suffered reductions in fertility when selfed, but the use of
any outcrossed parent restored fertility, regardless of relatedness. The diploid
species may be more finely tuned to its environment, than the tetraploid, as the
diploid condition would encourage more homozygosity and the formation of
unique coadapted complexes.

The data from this study corroborates earlier reports of inbreeding
depression in both species (Hokanson and Hancock. 2000; Schott, 2000). The
more clonal growth pattern of V. myrtilloides may explain, at least a part of, the
difference in the degree of outbreeding depression we observed between the
two species. Vaccinium myrtilloides populations are composed of large
clones often extending several meters, while V. corymbosum clones are crown
formers covering much smaller, discrete surface areas. As a result,
pollinations within a 10 m area have a higher probability of involving self-
pollination in V. myrtilloides, than V. corymbosum. This effect may have been

most pronounced in the MPR population where clones were large and

72

continuous with undefined boundaries, compared to the MFR population which
consisted of smaller, well defined clones

In this study, the pollinations were performed in the field but germination
and seedling growth were monitored in environmental chambers. This may
have overestimated the fitness of outcrossed progeny, as plants resulting from
local pollination might perform better under local conditions. However,
seedling survival was very low in the environmental chamber and greenhouse,
and my anecdotal observations in the field suggest that seedling survival in the
field is low as well. Any study attempting to assess seedling survival under
natural conditions will likely suffer from small sample size given high putative
seedling mortality.

In summary, there is no evidence for OD, or an optimal outcrossing
distance, in the tetraploid V. corymbosum. However, in the diploid species V.
myrtilloides, there does appear to be an optimal outcrossing distance of
approximately 100 m for fruit set and percent germination. An interplay
between growth pattern and ploidy probably contributed to these differential
patterns. The tetrasomic inheritance of V. corymbosum may moderate the
negative effects of inbreeding or outbreeding, and the clonal growth pattern of
this particular species will favor a higher percentage of outcross pollinations.
The growth pattern of the diploid V. myrtilloides will favor a higher proportion of
self pollinations, relative to the tetraploid, and its disomic inheritance pattern

will more readily manifest the negative effects of inbreeding or outbreeding. It

73

appears that levels of inbreeding depression in natural species are regulated

by a complex interaction of factors.

74

LITERATURE CITED

Agren, J. and D. W. Schemske. 1993. Outcrossing rate and inbreeding
depression in two annual monoecious herbs, Begonia hirsuta and B.
semiovata. Evolution 472125-135.

Campbell, D. R. and N. M. Waser. 1987. The evolution of plant mating
systems: multilocus simulations of pollen dispersal. The American
Naturalist 129:593-609.

Charles Worth, D. and B. Charlesworth. 1987. Inbreeding depression and its
evolutionary consequences. Annual Review gf Ecolggy and Systematics
18:237-268.

Coles, J. F. and D. P. Fowler. 1976. Inbreeding in neighboring trees in two
white spruce populations. Silvae Genetica 25:29-34.

Fenster, C. B. and V. L. Sork. 1988. Effect of crossing distance and male
parent on in vivo pollen tube growth in Chamaecrista fasciculata.
American Journal of Botany 75:1898-1903.

Hokanson, K. E. and J. F. Hancock. 2000. Early-acting inbreeding depression
in three species of Vaccinium (Ericaceae). Sexual Plant Reproduction
(In press).

Holtsford T. P. 1996. Variation in inbreeding depression among families and
populations of Clarkia tembloriensis (Onagraceae). mm 76:83-91.

Johnston, M. O. and D. J. Schoen. 1994. On the measurement of inbreeding

depression. Evolution 48:1735-1741.

75

Kalisz, S., D. Vogler, B. Fails, M. Finer, E. Shepard, T. Herman, and R.
Gonzales. 1999. The mechanism of delayed selfing in Collinsia verna
(Scrophulariaceae). American Journal Botany 86:1239-1247.

Krebs, S. L. and J. F. Hancock. 1988. The consequences of inbreeding on
fertility in Vaccinium corymbosum. Journal of the American Society for
Horticultural Science 85:302-306.

Moll, R. H., J. H. Lonnquist, J. V. Fortuno and E. C. Johnson. 1965. The
relationship of heterosis and genetic divergence in maize. Genetics
52:139-144.

Oostermeijer, J. G. B., R G. M. Altenburg and H. C. M. den Nijs. 1995. Effects
of outcrossing distance and selfing on fitness components in the rare
Gentiana pneumonanthe (Gentianaceae). Acta. Botanica Neerlandica
44:257-268.

Palser, B. F. 1961. Studies of floral morphology in the Ericales. V.
Organography and vascular anatomy in several United States species of

the Vacciniaceae. Botanical Gazette 123279-111.

 

Parker, M. A. 1992. Outbreeding depression in a selfing annual. Evolution
46:837-841.

Price, M. V. and N. M. Waser. 1979. Pollen dispersal and optimal outcrossing
in Delphinium nelsoni. Nature 277:294-297.

SAS Institute. 1994. Statistical software for the Macintosh. JMP, statistics and

graphics guide. SAS Institute, Inc., Cary, NC.

76

SCHOTT, G. W. 2000. The response of the blueberries Vaccinium myrtilloides
and V. corymbosum to inbreeding under natural conditions. (In prep.)

Trame, A. M., A. J. Coddington and K. N. Paige. 1995. Field and genetic
studies testing optimal outcrossing in Agave schottii, a long-lived clonal
plant. Oecologia 104:93-100.

Vander Kloet, S. P. and P. Cabilio. 1984. Annual variation in seed production
in a population of Vaccinium corymbosum L. Bulletin of the Torrey
Botanical Club 111:483-488.

_, and P. M. Lyrene. 1987. Self-incompatibility in diploid, tetraploid, and
hexaploid Vaccinium corymbosum. Canadian Journal of Botany 65:660-
665.

Waser, N. M. and M. V. Price. 1979. Pollen dispersal and optimal outcrossing
in Delphinium nelsoni. Nature 277:294-297.

_. 1994. Crossing-distance effects in Delphinium nelsonii: outbreeding
and inbreeding depression in progeny fitness. Evolution 48:842-852.

___, M. V. Price and R. G. Shaw. 2000. Outbreeding depression varies
among cohorts of lpomopsis aggregata planted in nature. Evolution

542485-491 .

77

CHAPTER 3

THE NATURE OF INFERTILITY IN VACCINIUM SPECIES

78

ABSTRACT

The nature of self-infertility is an active area of research in plant biology.
The mode of self-infertility, inbreeding depression (ID) or post-zygotic
incompatibility, othenlvise known as ovular self-incompatibility (OSI), can have
important consequences for the evolution of mating systems. Previous studies
attempting to determine whether low self-fertility in blueberries is due to ID or
OSI have been equivocal. Criteria by which one can distinguish between ID
and OSI include measuring levels of inter-plant variation in self-fertility and
seed development. In this study seed development was monitored in self- and
cross-pollinated fruits in two species of tetraploid blueberries, Vaccinium
angustifolium and V. corymbosum. It was discovered that many self-fertilized
ovules begin to develop, but abort at different times. In addition, there was
considerable variation in self-fertility among individual plants. These data, in
conjunction with previous work, support the hypothesis that blueberries suffer

from ID, not OSI.

79

INTRODUCTION

An active area of research in plant biology is the nature of self-fertility or,
alternatively, self-infertility. Darwin (1876) was first to systematically examine
self-fertility in plants. He self-pollinated 40 species for 10 consecutive
generations and observed that some species showed a decline in fitness
when selfed, relative to outcrossed lines. This phenomenon of inbreeding
depression has subsequently been well documented.

Despite the known negative consequences of inbreeding, a common
pathway in the evolution of plant mating systems is a shift from self-
incompatibility (obligate outcrossing) to predominant self-fertilization (Stebbins,
1957). This shift is believed to have occurred in many plant families (Stebbins,
1974; Wyatt, 1988), including many cultivated species (Rick, 1988). The shift to
self-compatibility is believed to be due to genes that modify the functioning of
self-incompatibility alleles at so called S-loci (Lewis, 1979). This shift to self-
compatibility, due to modifier alleles, is known as pseudo-self-fertility (de
Nettancourt, 1977). Pseudo-self-fertility is differentiated from true self-fertility
based on the criteria outlined in Table 1.

Table 1: Criteria for differentiating between pseudo-self—fertility and true self-

fertility as specified by Levin (1996). The statements apply to a species that is
subject to pseudo-self-fertility.

 

Greater seed production with cross pollination relative to self pollination.
Earlier germination and more rapid pollen tube growth with cross pollen.
A continuous distribution of self-fertility levels in progeny.

Self versus outcross performance depends on environmental conditions
and plant age.

Variable levels of seed set are found in crosses between plants that share
both S alleles.

PP’N.‘

9‘

 

80

The inability of self-pollinated flowers to set seed can be grouped into
many classes according to the time at which the pollination/fertilization process
is arrested. Self-infertility occurs most immediately when self pollen fails to
germinate on the stigma. This is usually attributed to sporophytic self-
incompatibility (SSI) where gene products of the diploid sporophyte, deposited
on the pollen grain coat, are recognized and rejected. The next class arises
when pollen tube growth is arrested in the style. This is usually attributed to
gametophytic self-incompatibility (GSI) where the haploid genotype of the
pollen grain nuclei is recognized and rejected (Proctor et al. 1996). The
existence of SSI and GSI systems has been known for quite some time, and
recognition of self is attributed to alleles at one or a few of the S-Ioci (Lewis,
1979). The appropriateness of using only two classes to describe pre-zygotic
SI has been recently questioned by Sage et al. (1999) who believe that there
are a wide array of pollen-pistil interactions. They hypothesize that in their
system (Narcissus triandrus) embryo sacs degenerate due to the lack of a
proper stimulus, even though pollen tubes appear to develop normally. Indeed,
pre-zygotic incompatibility probably occurs on a continuum from pollen
recognition, and rejection, to gamete rejection after the pollen tube penetrates
the micropyle.

The next stage of reproductive arrest is post-zygotic, and can occur due
to inbreeding depression (ID) or ovarian self-incompatibility (OSI). With ID,
developing embryos abort at different stages of development due to genetic

load or the loss of heterotic interactions, and selfing results in variable seed

81

production among genotypes (Charlesworth and Charlesworth 1987). In OSI
systems, there is a uniform rejection of self fertilized ovules in all individuals at
an early stage of embryo development, and species segregate within families
for self- incompatibility alleles.

Seavey and Bawa (1986) first spelled out the criteria for differentiating
between ID and OSI (Table 2). A number of recent studies have documented
the presence of OSI. Naaborgh and Willemse (1992) observed a uniform
rejection of self-pollinated ovules approximately two days after fertilization in
Gasteria verrucosa. Gibbs and Bianchi (1999) observed zero self-seed set in
Dolichandria cynanchoides and Tabebuia nodosa despite penetration of the
micropyle by pollen tubes. Lipow and Wyatt (2000) determined that Asclepias

exaltata exhibits OSI and that the occasional self—fertile plant is due to pseudo-

 

self-fertility alleles that alter S-Iocus functioning.

Table 2: Criteria for differentiating between inbreeding depression (ID) and
ovular self-incompatibility (OSI) as specified by Seavey and Bawa (1986).

 

1. Uniform abortion of embryos at the same stage of development suggests
OSI while abortion at various stages suggests ID.

2. Variability among plants in self seed set suggests ID.

3. With ID, those selfed seeds that do mature will likely express the effects of
genetic load at following stages of development.

4. Embryo rescue should be possible on aborted embryos subject to OSI, not
ID.

 

82

Numerous studies have documented ID at various stages of plant
development including seed set in alfalfa (Holland and Bingham 1994), maize
(Pinnisch and Stucker 1998) and blueberries (Harrison et al. 1994, Hokanson
and Hancock 2000; Schott, 2000), forage production in alfalfa (Holland and
Bingham 1994, Sain et al. 1994) and leeks (Smith and Crowther 1995 ), and
disease resistance in winter rye (Miedaner et al. 1995). In addition, selfing has
been shown to result in variable seed production among genotypes in maize
(Vasal et al. 1995) and blueberries (Krebs and Hancock, 1991; Schott, 2000 ).

Traditionally, ID was considered the primary factor in post-zygotic
infertility. However, OSI has gained credence and is considered a widespread
phenomena (Sage et al. 1999). Since the nature of the infertility system can
impact the evolution of the mating system, distinguishing between ID and OSI
is important. Where self-fertility is minimal, early recognition of self pollen
minimizes the wastage of resources into partially developed seeds, and the
fouling of the style with self pollen tubes that inhibit subsequent outcross
pollinations. Because fertilized ovules are rejected early and uniformly in OSI,
resource wastage is decreased with OSI relative to abortion due to ID. Strict
self-incompatibility allows for little or no leakage of selfed progeny through the
system. However, some leakage can be beneficial when other pollen sources
are unavailable or unreliable, or if there is a reasonably high probability that
beneficial recessive alleles are found in the population. Such an allele has a
higher probability of being expressed with self pollination, as a higher

proportion of loci will be homozygous. Where ID is severe, and a premium is

83

placed on outcrossing, one can envision a scenario where a species evolves
from later to early stage self recognition and rejection. This would depend on
there being a selective advantage to individuals who reject self pollen earlier in
the reproductive process. Pollen discounting, the decrease in paternal fitness
due to self pollination, can promote this trend. However, pollinator behavior will
also need to evolve where pollen discounting is a factor. Self-compatibility is
favored under the following conditions: when the fitness of selfed progeny is at
least half that of outcross progeny (Fisher 1941); when autogamy can assure
successful reproduction under variable pollinator activity (Kalisz et al. 1999);
when pollinator behavior promotes geitonogamy, which is the movement of
pollen between flowers on the same plant.

The nature of inbreeding depression in blueberries is still unfolding.
Some researchers have claimed that reduced seed set is due to self-
incompatibility (Vander Kloet and Lyrene 1987), while others maintain that
blueberries suffer from early acting ID (Krebs and Hancock, 1990). Both of
these studies demonstrated that self pollen tubes penetrate ovules in
Vaccinium corymbosum. Thus, pre-zygotic incompatibility is not likely in
blueberries and decreased self-fertility is due to either ID or OSI. Vander Kloet
(1991) has examined ovule development in diploid V. corymbosum. He
observed relatively uniform rejection of ovules and suggested that these
species are subject to OSI. Several other observations on blueberry species
suggest that Vaccinium suffer from inbreeding depression due to the

expression of deleterious alleles. First, there exists a range of self fertility in

84

different genotypes, thus addressing criteria #2 of Seavey and Bawa (Table 1).
There is also evidence for a positive correlation between self and outcross
seed set (Krebs and Hancock 1990) and a range of "seed" sizes from fully
developed seeds to nubbins that have presumably developed little beyond
fertilization, partially addressing criteria #1 of Seavey and Bawa (Table 1)
(Hokanson and Hancock 2000; Schott, 2000). In addition, Krebs and Hancock
(1990) found that V. corymbosum has an inbreeding threshold of F > 0.3 (F =
Wright's inbreeding coefficient). Individuals below the threshold were
facultative selfers and those above it were obligate outcrossers. In a separate
study, Krebs and Hancock (1991) found that ID was more severe in those
individuals with the greatest number of alleles per locus. They suggest that the
polyploid nature of tetraploid V. cogmbosum allows the harboring of a large
genetic load resulting in low self-fertility.

Comparing fruit and seed set between self- and out-cross pollinations
alone is inadequate in differentiating between inbreeding depression and self-
incompatibility (Gibbs and Bianchi 1999). One needs to examine the
development of fertilized ovules to determine if self pollinated ovules are
preferentially aborted and if these abortions occur at various stages over a
relatively long period of time. Self-incompatibility will result in uniform abortion
at an early stage of development. Abortion due to the expression of deleterious
alleles (ID) will occur at different stages over an extended period of time

(Seavey and Bawa 1986).

85

In this study, the question of ID versus OSI in blueberries is addressed
by taking the work of Krebs and Hancock (1990) and Hokanson and Hancock
(2000) a step further. Seed abortion is shown to occur at various stages of
seed development which implicates ID. Blueberries appear to suffer from
reduced self-seed set due to the manifestation of genetic load upon
inbreeding.

MATERIALS AND METHODS

my mnWe used the wild, tetraploid blueberries Vaccinium
angustifolium and V. corymbosum in this study. These are primarily bee-
pollinated species that are subject to many geitonogamous pollinations
(Vander Kloet and Lyrene, 1987). These species readily reproduce vegetatively
with V. angustifolium spreading laterally, via rhizomes, and V. corymbosum
producing many stems, or canes, from a central underground stem. The study
populations are located in the Allegan State Game Area in Allegan County,
Michigan (Figure 1). The V. angustifolium population is in a dry, upland habitat
dominated by white oak (Quercus alba). The V. corymbosum population is on
the margin of a shallow, acidic lake with red oak (Quercus rubra), red maple
(Acer rubrum), beech (Fagu_s grandifolia) and white pine Qinus strobus) as
dominant tree species. Vander Kloet (1988) provides detailed descriptions of

these blueberry species.

86

Figure 1. Map of the Allegan State Game Area, Allegan county Michigan,
showing the locations of the study populations. AAW represents the Vaccinium
aerstifolium population and CCL represents the V. corymbosum population.

87

luau-ran-Ingur... .nurug._‘rfla...»u.n. “shah-5;:
.H194. .l‘ni.lnt~laudqini‘-hP-nnn a .;:‘4..-::..

.1554: ... stink .h»I-.u..F-I..I.Du..tl).iadt-A-
mr‘fl ”.5"... 5.. :35! .QISS‘E‘DIIHIJ

.In!!«. » in: u

. i=5

y§luép.3§rgilfiit.l dirt.
in in Pin-3: EH3:

 

88

Procedures--In each population, five plants were chosen for the crossing
experiment. In the _\_l. angustifolium population, these plants were chosen
based on whether they produced enough flowers to perform the crosses. In
the V. corymbosum population, the plants were chosen to represent a variety of
selfing abilities as determined from crosses in previous years (Schott, 2000).
A sixth plant was chosen in each population to serve as the outcross donor for
the other five plants.

On each of the five experimental plants, pollen was collected from
opened flowers, and 30-40 unopened flowers were emasculated.
Approximately 2/3 of these emasculated flowers were self-pollinated and the
remainder were pollinated with the out-cross pollen from the same species.
The inflorescence with pollinated flowers were covered with a fine mesh nylon
bag to prevent fruit loss and to exclude pollinators.

To get a temporal picture of the range in sizes of developing ovules, self-
and out-cross fruits were sampled from each plant at one, two, three and 4 1/2
week intervals following pollination, and the remaining fruit were allowed to
completely mature. In some cases, there were insufficient self pollinated fruit
to sample at all four dates on an individual plant. When this occurred, the
sampling was aborted when there was only one fruit remaining, and it was
allowed to mature. The harvesting of mature fruits began approximately 60
days after pollination. Two open pollinated fruits were also collected at maturity

from each plant.

89

To process the immature and mature fruits, an approximately 1-2 mm
slice of the top of the fruit was removed with a razor blade, and the contents of
the fruit were removed with a spatula. The number of maturing ovaries was
quantified and all of the ovaries were spread out on a glass slide using a VVlld
M5 observation scope. The 'size' of each ovary was measured using the ocular
micrometer of a Zeiss Axioskop. Blueberry ovules and seeds have a generally
oblong shape with rounded ends (Figure 2), and the size was measured as the

longest length of each ovule/seed.

 

Figure 2. Picture of mature and aborted seeds from aVaanigmmumpgsum
fruit.

90

The viability of aborted and mature seeds from ripe fruits was tested by
transferring them to moistened filter paper in petri dishes. The seeds were
kept moist with a solution of tap water and phosphoric acid (pH = 5.0). Low pH
water was used to mimic field conditions. Seeds were than scored for their
ability to germinate. Using this procedure, I was able to quantify the
approximate size a seed must reach in order to germinate.

Dag analysis-~Data were compiled on a species, cross and individual
plant basis. The mean size of seed that germinated was calculated and
species and crosses were compared with a nested analysis of variance.
Sources of variation were species and cross within species. We also
quantified the number of fertilized ovules and germinating seeds for each
species and cross, and calculated the percent of fertilized ovules that
germinated. All data analyses were performed using JMP 3.1.6 software for the
MAC (SAS Institute, 1994).

RESULTS

Self-fertility varied greatly among individuals of both species. In general,
self fertility was lower in Vaccinium angustifolium than V. corymbosum, with
four of the five plants failing to mature any self-pollinated fruit (Figure 3). The
fertilized ovules began to develop in all the V. angustifolium plants, but ceased
developing at variable times depending on genotype. Seed development
ceased by day 23 in AAW 1 and AAW 5, and by day 33 in AAW 4. No ripe fruit
were collected from AAW 2, and most of its seeds had aborted by day 23, but a

few of its developing seeds had already gotten rather large by that time.

91

Figure 3. OvuIe/seed size for self and outcross pollinations in five Vaccinium
angustifolium plants. Rows represent the five plants. The first column is the
mean ovule size for selfed and outcrossed fruits sampled at several dates
before maturity, and at maturity (day 60). In addition, ovule size was quantified
in unopened flowers, that were not pollinated, and in flowers that were naturally
pollinated (OP). The second and third columns represents the distribution of
ovule sizes for self and outcross pollinated fruits, respectively, at different
sample dates.

92

”MI-(IN

“who”

  

Whit-'0
mun-(an

     

AAW 4

“Mum

 

Hod-lam
MIDI!!!“

   

93

Figure 4. Ovule/seed size for self and outcross pollinations in five Vaccinium
corymbosum plants. Rows represent the five plants. The first column is the
mean ovule size for selfed and outcrossed fruits sampled at several dates
before maturity, and at maturity (day 60). In addition, ovule size was quantified
in unopened flowers, that were not pollinated, and in flowers that were naturally
pollinated (OP). The second and third columns represents the distribution of
ovule sizes for self and outcross pollinated fruits, respectively, at different
sample dates.

94

CCLZ

...-$33.1

 

5‘3!

CCL14

 

azllia

 

ralllail

 

Ev}:

 

as)":

 

 

Aivlflg

 

fuel-3.6!

95

AAW 3 was the onlyV. angustifolium that had large seeds and produced fully
mature fruits. In four of the five plants, cross pollinated fruits reached maturity.
Only in plant AAW 2 did the remaining outcross fruits not reach maturity. Out-
crossed ovules were fertilized and began to develop on all plants. Open
pollinated fruits performed comparably to the outcrossed fruits, except for plant
AAW 4, where the open pollinated fruits had significantly smaller seeds.

In V. corymbosum, four of the five plants matured both self and outcross
fruit while plant CCL 2 matured neither (Figure 4). In the four self-fertile plants,
it is clear that self-pollinated ovules begin to develop at an early stage. Again,
seed size was comparable in open-pollinated and cross-pollinated fruits,
except in plant CCL 14, which had significantly larger seeds in the OP fruit. No
data are available for days 22 and 32 for plant CCL 53 because there were too
few self- and out-cross fruits remaining by day 22. Collection of fruit on day 22
would have left nothing to reach maturity.

In general, seeds of both species had to reach a size of 1100 um before
they would germinate (Table 3), although an outlier was found that germinated
at 820 um, is an outlier, falling well outside 95% confidence intervals (1433 <->
1579) in the bottom 0.5% quantile of selfed seeds in V. corymbosum. The
widest range in size of germinable seeds was found in self pollinated V.
corymbosum and only one of the selfed seeds germinated in V. angustifolium.
An analysis of variance revealed no differences between species or crosses

in the mean size of germinating seeds (Table 4).

96

Table 3: Seed germination data for V. angustifolium (V. a.) and V. cogmbosum
(V. c.). The cross types are self pollen, outcross pollen (00) or open pollinated
(OP). Also included are the maximum (Max) and minimum (Min) size of
germinating seeds, the number of seeds germinating (n), the mean size of
germinating seeds, and the standard error of the mean (SEM). All seed sizes
are in micrometers.

 

 

Cross Max Min n Mean SEM
Self 1320 1320 1 1320
V. a. QC 2000 1140 96 1556 17.4
OP 1860 1140 10 1594 61.9
Self 2080 820 44 1506 36.2
V. 0. OC 1800 1280 19 1547 30.1
OP 1980 1100 57 1579 28.4

 

Table 4. Effect of species and pollination treatment on the size of germinating

 

 

seeds.

Source of variation d.f. MS F
Species 1 1435 .03ns
Cross(Species) 4 50899.6 1.32“3
Error 221 38428.6

 

97

Table 5. Data for fruits that reached maturity. Plant represents the 10
experimental plants. AAW represents the V. angustifolium population and CCL
represents the V. corymbosum population. Cross represents self, outcross
(OC) and open (OP) pollinated flowers/fruits. # fruits = the number of mature
fruit sampled, # ovules is the total number of fertilized ovules in all of the
mature fruits, # germ is the number of ovules that germinated, % germ is the
percent of ovules that germinated, #>1100 is the number of seeds that were
larger than 1100 um, #>820 is the number of seeds that were larger than 820
um, #>11/fruit is the number of seeds larger than 1100 um in each fruit,
#>820/fruit is the number of seeds larger than 820 um in each fruit.

98

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Plant Cross # # # % # 2 # 2 #21 10 #2820
fruits ovules germ erm 1 100 820 0/fruit lfruit
AAW1 Self 0
OC 4 210 21 10.0 26 32 6.5 8.0
OP 2 89 2 2.2 9 20 4.5 10.0
AAW 2 Self 0
OC 0
OP 2 114 5 4.4 67 94 33.5 47.0
AAW 3 Self 2 143 1 0.7 22 59 11.0 29.5
OC 2 72 3 4.2 30 56 15.0 28.0
OP 2 75 0 0.0 45 57 22.5 28.5
AAW4 Self 0
OC 5 347 62 17.9 93 145 18.6 29.0
OP 2 104 2 1.9 9 19 4.5 9.5
AAW 5 Self 0
OC 3 153 10 6.5 16 21 5.3 7.0
OP 2 94 1 1.1 15 26 7.5 13.0
CCL 2 Self 0
OC 0
OP 2 124 4 3.2 63 95 31.5 47.5
CCL 9 Self 5 316 10 3.2 33 55 6.6 11.0
OC 5 227 2 0.9 32 57 6.4 11.4
OP 2 117 1 0.8 26 43 13.0 21.5
CCL 14 Self 2 145 5 3.4 32 46 16.0 23.0
OC 3 190 2 1.0 43 84 14.3 28.0
OP 2 135 13 9.6 86 105 43.0 52.5
CCL 46 Self 6 323 26 8.0 71 87 11.8 14.5
00 2 127 1 1 8.7 13 17 6.5 8.5
OP 2 128 15 11.7 41 53 20.5 26.5
CCL 53 Self 1 67 3 4.5 15 43 15.0 43.0
OC 1 50 4 8.0 25 38 25.0 38.0
OP 2 123 24 19.5 68 96 34.0 48.0

 

 

 

 

 

 

 

 

 

 

99

 

Average germination rates for fertilized ovules were quite variable across
plants and ranged from 0.0-19.5% (Table 5). For most fruits, a majority of
seeds fall below the germination threshold of 1100 um. Thus, it is not
surprising that such a small percentage of fertilized ovules germinate. The
number of seeds exceeding the germination thresholds of 1100 um and 820
um is comparable between self and outcross pollinations in the majority of
plants. The exceptions are selfed seeds in plant CCL 46, and outcross seeds
in plant CCL 53, which had greater number of seeds above the germination
threshold Figure 4.

DISCUSSION

The data presented in this study address the first two criteria for
differentiating between inbreeding depression (ID) and ovular self-
incompatibility (OSI), as outlined by Seavey and Bawa (1986). First, there is
evidence for abortion occurring at various stages of development. Second, the
response to inbreeding varied between plants. These data, in conjunction with
the ID data of Hokanson and Hancock (2000) and Schott (2000) which show a
range in self-fertility across plants (criteria #3 of Seavey and Bawa), clearly
demonstrate that the blueberry species suffer from ID, not OSI. While some
plants appear to be self-incompatible, or to have low self-fertility (AAW 1, AAW 4,
AAW 5, CCL 2), this is by no means a universal trait in eitherV. angustifolium or
V. corymbosum. In fact, plant CCL 2 has demonstrated high self and outcross

fertility in previous years (Schott, 2000), relative to these data from 2000.

100

Seavey and Bawa (1986) suggest that uniform failure of a wide sample
of ovules, at a comparable stage of development, suggests an OSI system.
Naaborgh and Willemse (1992) demonstrated uniform rejection of ovules two
days after fertilization in Gasteria verrucosa. They believe it is a 2+ loci
incompatibility system that results in abortion due to the termination of
maternal resources. While the majority of self-pollinated blueberry ovules are
aborted early on, a number of them continue to develop and a small
percentage do become mature seed.

Species suffering from ID will show individual variation in self-fertility
(Seavey and Bawa, 1996). Species that are truly self-incompatible show little or
no variation in self-fertility among plants. Plants with only a few sublethal
recessives can have close to normal seed production when selfed. It is
possible that a few self-fertile plants might arise through mutation in an
otherwise self-infertile population, but the frequency of self-fertile plants found
in our blueberry populations, and their variability in self-fertility from year to year
(Schott, 2000), suggests that we are dealing with ID rather than rare self-fertile
mutants in otherwise self-incompatible populations.

Where detrimental recessives lead to ID, mature selfed seeds are likely
to express ID at later stages of development. Examples include: Seed weight,
days to germination, plant height and number of flowers in Collinsia
heterophylla (Mayer et al., 1996); germination rate, seedling survival and
seedling mass in Epilobium angustifolium (Husband and Schemske, 1997);

and growth and survival of progeny in Spartina alterniflora (Daehler 1999). In

101

blueberries, inbreeding depression was observed for germination rate in V.
myrtilloides and V. corymbosum and growth rate in V. corymbosum (Schott,
2000)

The criteria outlined above eliminates SI as operating in blueberries;
however, they are not completely self-fertile and may suffer from pseudo-self-
fertility. According to Levin's definition of true self-fertility, similar seed sets
follow self and outcross pollination (Table 1), which is not the case in
blueberries. In addition, self pollen is less vigorous than outcross pollen in
blueberry (Hokanson and Hancock, 2000), a further criteria for pseudo-self-
fertility according to Levin (1996). Anecdotal observations also suggest that
self-fertility varies from year to year, and is probably environmentally dependent
(Schott, 2000).

Seavey and Bawa (1986) discuss species (Lilium sp.) for which self-

 

incompatibility is not always complete. The mechanism for this leaky SI likely
involves signaling, or lack there of, between embryonic and/or endosperm
tissue and maternal tissue. The self-infertility modifier alleles conferring
pseudo-self-fertility in Asclepias exaltata (Lipow and Wyatt, 2000) and other
species (Levin, 1996), are possibly the basis for leaky SI. Smaller, more
remote milkweed populations have higher frequencies of self-fertile individuals
(Lipow and Wyatt, 2000), fitting with the notion

Why would long-lived, vegetatively reproducing species evolve the ability
to self-fertilize? It is possible that the evolution from diploidy to polyploidy is a

causal factor. A rare polyploid appearing in a diploid population will increase in

102

numbers only if it is self-fertile (Thompson and Lumaret, 1992). Being long-
lived and clonal will help it persist in the short-terrn, but only the tolerance of
extreme inbreeding will allow it to develop into a distinct species that has
evolutionary potential. Jain (1976) and Stebbins (1974) maintain that self-fertile
lines that develop from self-infertile progenitors are almost always evolutionary
dead ends. However, the disruption in mating system associated with
polyploidy may allow a newly derived species to initially persist by selfing,
before ultimately devel0ping into a primarily outcrossed species. Such species
would maintain the ability to self, and might maintain intraspecific variation in

selfing ability.

103

LITERATURE CITED
CHARLESWORTH, D. and B. CHARLESWORTH. 1987. Inbreeding depression and
its evolutionary consequences. Annual Review g_fE_co_logy _rm
Systematics 18:237-268.
DAEHLER, C. C. 1999. Inbreeding depression in smooth cordgrass (Spartina

alterniflora Poaceae) invading San Francisco Bay. American Journal 9_f

 

Bgta_ny 86:131-139.

DARWIN, C. 1876. The effects of cross and self ferlilisation in the vegetable
kingdom, Murray, London.

DE NETTANCOURT, D. 1977. Incompatibility in angiosperms. Springer-Verlag,
New York, NY.

ENTANI, T., S. TAKAYAMA, M. IWANO, H. SHIBA, F. S. CHE and A. ISOGAI. 1999.
Relationship between polyploidy and pollen self-incompatibility

phenotype in Petunia hybrida Vilm. Bioscience Biotechnology mg

 

Biochemistry 63:1882-1888.

FISHER, R. A. 1941. Average excess and average effect of a gene substitution.
Aflal_s gf Eugenics 11:53-64.

GIBBS, P. E. and M. B. Bianchi. 1999. Does late-acting self incompatibility (LSI)
show family clustering? Two more species of Bignoniaceae with LSI:
Dolichandra cynanchoides and Tabebuia nodosa. Annals gf Bgtfly
84:449-457.

HANCOCK, J. F. 1992. The origin and evolution of crop plants. Prentice Hall,

Englewood Cliffs, NJ.

104

HARRISON, R. E., J. J. LUBY and P. D. ASCHER. 1994. Pollen source affects yield
components and reproductive fertility of 4 half-high blueberry cultivars.

Journal f th American Society fg’ Horticultural Science 119284-89.

 

HOKANSON, K. E. and J. F. HANCOCK. 2000. Early-acting inbreeding
depression in three species of Vaccinium (Ericaceae). Sexual Plant
Reproduction (In press).

HOLLAND, J. B. and E. T. BINGHAM. 1994. Genetic improvement for yield and
fertility of alfalfa cultivars representing different eras of breeding. Cm];

Science 34:953-957.

 

HUSBAND, B. C. and D. W. SCHEMSKE. 1997. The effect of inbreeding in diploid
and tetraploid populations of Epilobium angustifolium (Onagraceae):
Implications for the genetic basis of inbreeding depression. Evolution
51:737-746.

JAIN, S. K. 1976. The evolution of inbreeding in plants. Annual Review gf
_E_c_glggy and Systematics 7:469-495.

KREBS, S. L. and J. F. HANCOCK. 1990. Early acting inbreeding depression and
reproductive success in the highbush blueberry, Vaccinium
corymbosum L. Theoretical and Applied Genetics 79:825-832.

, and . 1991. Embryonic genetic load in the highbush
blueberry, Vaccinium corymbosum (Ericaceae). American Journal gf
_B_ota_ny 78:1427-1437.

LEVIN, D. A. 1996. The evolutionary significance of pseudo-seIf-fertility. Illa

American Naturalist 148:321-332.

105

LIPOW, S. R. and R. WYATT. 2000. Single gene control of postzygotic self-
incompatibility in Poke Milkweed, Asclepias exaltata L. Genetics
154:893-907.

MAYER, S. S., D. CHARLESWORTH and B. MEYERS. 1996. Inbreeding depression
in four populations of Collinsia heterophylla Nutt (Scrophulariaceae).
Evolution 50:879-891.

MIEDANER, T., D. E. ZIEGLER and H. H. GEIGER. Variation and covariation for
quantitative resistance to head blight (Fusarium culmorum) in 2

testcross series of S-2 lines in winter rye. Plant Breeding 114:155-159.

 

NAABORGH, A. T. and M. T. M. WILLEMSE. 1992. The ovular incompatibility
system in Gasteria verrucosa (Mill.) H. Duval. Euphfiica 58:231-240.

PINNISCH, R. and R. E. STUCKER. 1998. Inbreeding depression in the selfed
progeny of F-1 hybrids between Northern Flint and corn belt dent
populations of maize (Q m L.). MILE 43:283-289.

RICHARDS, A. J. 1997. Plant breeding systems. Chapman & Hall, London. p.
220.

RICK, C. M. 1988. Evolution of mating systems in cultivated plants. l_r1 L. D.
Gottlieb and S. K. Jain [eds.], Plant evolutionary biology, 133-147.
Chapman & Hall, London.

SAGE, T. L. , F. STRUMAS, W. W. COLE and S. C. H. BARRETT. 1999. Differential
ovule development following self- and cross-pollination: The basis of
self-sterility in Narcissus triandrus (Amaryllidaceae). American Journal

91‘ Botany 86:855-870.

106

SAIN, S. L., D. K. BARNES and D. D. BIESBOER. 1994. Hydroponic and tissue-
culture evaluation of alfalfa (Medicago satiya L.) subpopulations selected
for phosphorus efficiency. Plant Science 99:17-26.

SAS Institute. 1994. Statistical software for the Macintosh. JMP statistics and
graphics guide. SAS Institute Inc., Cary, NC.

SCHOTT, G. W. 2000. The response of the blueberries Vaccinium myrtilloides
and V. corymbosum to inbreeding under natural conditions. (In prep.)

SEAVEY, S. R. and K. S. BAWA. 1986. Late-acting self-incompatibility in
angiosperms. The Botanical Review 52:195-219.

SMITH, B. M. and T. C. CROWTHER. 1995. Inbreeding depression and single
cross hybrids in leeks (Allium ampeloprasum ssp. mm). Euphflica
86:87-94.

STEBBINS, G. L. 1957. Self-fertilization and population variability in the higher
plants. American Naturalist 91 :337-354.

1974. Flowering plants: evolution above the species level. Belknap,
Cambridge, MA.
THOMPSON, J. D. and R. LUMARET. 1992. The evolutionary dynamics of

polyploid plants: Origins, establishment and persistence. Trends i_n

 

E§o_logy and Evolution 71302-307.
VANDER KLOET, S. P. 1988. The genus Vaccinium in North America. Research
Branch Agriculture Canada, publication 1828, Ottawa.
. 1991. The consequences of mixed pollination on seed set in

Vaccinium corymbosum. Canadian Journal _o_f Botany 69:2448-2454.

107

, and P. M. Lyrene. 1987. Self-incompatibility in diploid, tetraploid, and
hexaploid Vaccinium corymbosum. Canadian Journal gf Bgt_agy 65:660-
665.

VASAL, S. K., B. S. DHILLON, G. SRINIVASAN, S. H. ZHANG and S. D. MCLEAN.
1995. Recurrent selection for inbreeding-stress tolerance in 4
intermediate-maturity maize populations. IVE-gm 40:159-164.

WYATT, R. 1988. Phylogenetic aspects of the evolution of self-pollination. I_ L.

D. Gottlieb and S. K. Jain [eds], Plant evolutionary biology, 109-131.

Chapman & Hall, London.

108

CONCLUSIONS

These data support previous work showing that there is significant
inbreeding depression in wild blueberries (Hokanson and Hancock, 2000;
Schott, 2000a). I am not aware of any published estimates of outcrossing rates
in blueberries, but bee pollination is thought to promote many geitonogamous
pollinations (Vander Kloet and Lyrene, 1987) and high cumulative inbreeding
depression would suggest that successful progeny are, in general,
outcrossed. The source of outcross pollen is irrelevant in the reproductive
success of tetraploid V. corymbosum, although the performance of
outcrossed progeny is superior to selfed ones when progeny are outcrossed.
As a result, I suggest that selection will move these populations towards
decreased self-fertility, which is counter to the common direction in which plant
mating systems are thought to evolve (Stebbins, 1974). However, there are a
few individuals that are relatively self-fertile, and even though cumulative
inbreeding depression across life-history stage can be high (Schott, 2000a),
self-pollinated progeny do emerge that are comparable or superior to
outcrossed progeny (Schott, 2000a). Whether or not the existence of these few
self-fertile plants is enough to maintain self fertility in the population as a whole
is still an open question.

The genetic system of most blueberry plants and pollinator behavior
seem to be at odds. Self-fertile plants would benefit from pollinator behavior
that encouraged geitonogamous pollinations; however, relatively self-infertile

plants would not benefit. Bee pollination in blueberries is thought to promote

109

many geitonogamous pollinations, and this is the main reason suggested for
why blueberries mature so few of their available ovules (Vander Kloet and
Lyrene, 1987). The resulting situation is an evolutionary tug-of-war between
two genetic systems in the plants, self-fertility versus self-infertility, and a
pollinator foraging method that favors geitonogamous pollinations.
Superimposed on this system are the survival rates of self and outcross
progeny under natural conditions. It is suggested that variable pollinator
service, due perhaps to environmental variation, will favor mixed mating (Kalisz
et al., 1999). Whether or not such a force can mold the mating system of a
long-lived perennial, like blueberries, is debatable. Blueberries can, and often
do, tolerate years of complete reproductive failure (Vander Kloet and Cabilio,
1984). Likewise, they can persist for long periods of time whether or not they
produce outcross progeny. It seems reasonable to surmise that the mating
system of an annual must be more flexible than that of a perennial.

There are several questions about blueberry mating systems that
remain to be answered: 1) What are outcrossing rates in natural populations
of blueberries? It is presumed that their outcrossing rates are relatively high;
however, evidence for outbreeding depression in V. myrtilloides and the relative
indifference to source of outcross pollen in V. corymbosum (Schott, 2000b)
suggest that a wide range of intermediate crossing distances are possible in
these species. High inbreeding depression in V. angustifolium, relative to V.
corymbosum (Hokanson and Hancock, 2000) suggest that the former species

will have relatively high outcrossing rates. 2) What is the exact nature of self-

110

infertility in blueberries? Interpreting previous studies, in light of my work and
others, strongly suggests that blueberries are not self-incompatible, but suffer
low self-fertility due possibly to pseudo-seIf—fertility. Only a detailed crossing
experiment using individuals of various levels of self-fertility, as per Lipow and
Wyatt (2000), and examining embryonic development, will fully address this
question. 3) In what direction are blueberry mating systems evolving? I believe
that they are evolving in the direction of decreased self—fertility. However,
obtaining a definitive answer in such a long-lived perennial would be difficult.
We can gain insight into this question throUgh a series of field and greenhouse
experiments. First, you would want to estimate outcrossing rates including the
intrapopulational variation in outcrossing rate. You would then generate an
array of progeny with of various levels of outcrossing, in the same proportion as
found in the natural population. You would then establish a common garden of
these progeny in the field and monitor their progress over several years taking
note of survival, growth rates, and various aspects of fertility. Comparing the
progress of self and outcross progeny would give us insights into the question
of mating system evolution. If we find that selfed progeny are at a
disadvantage, compared to outcross progeny, we might suspect that the self-
fertility found in blueberries is left-over from the time polyploid species first

arose, and is of little ecological and evolutionary importance.

111

LITERATURE CITED
HOKANSON, K. E. and J. F. HANCOCK. 2000. Early-acting inbreeding
depression in three species of Vaccinium (Ericaceae). Sexual Plant
Reproduction (In press).
KALISZ, 8., D. VOGLER, B. FAILS, M. FINER, E. SHEPARD, T. HERMAN, and R.
GONZALES. 1999. The mechanism of delayed selfing in Collinsia

verna (Scrophulariaceae). American Journal gf Botany 86:1239-1247.

 

LIPOW, S. R. and R. WYATT. 2000. Single gene control of postzygotic self-
incompatibility in Poke Milkweed, Asclepias exaltata L. Genetics
154:893-907.

SCHOTT, G. W. 2000a. The response of the blueberries Vaccinium
myrtilloides and V. corymbosum to inbreeding under natural conditions.
(In prep.)

_. 2000b. The effect of crossing distance on fertility in Vaccinium
species. (In prep.)

STEBBINS, G. L. 1974. Flowering plants: evolution above the species level.
Belknap, Cambridge, MA.

VANDER KLOET S. P. and P. CABILIO. 1984. Annual variation in seed

production in a population of Vaccinium corymbosum L. Bulletin f th

 

Torrey Botanical Club 111:483-488.
VANDER KLOET S. P. and P. M. LYRENE. 1987. Self-incompatibility in diploid,
tetraploid, and hexaploid Vaccinium corymbosum. Canadian Journal gt

Botany 65: 660-665.

112

 

llllllllllllllllllllll