 

 

 

 

 
 
 

 

 

 

 
 
 
 
   
 

 
   
 

 

 
 
 

 
 

 
 

 

 

 
 
 
 

 

 

 

 

    

 

 
 
 

. . x .n ..
2.... 3.
, .r?

 

 

 

 

 

 

 

 

 

 

 

 
 

.. : .5... ... . 43" u
13...: . £ . .. .. .wx ,. .3. 3.955;}; taxi...

  

 

This is to certify that the
thesis entitled

Disease, pollinator, and resource limitation influences on the
reproductive biology and growing season of Arisaema
triphyl/um, Jack-in-the-pulpit

presented by
Jessica L. Cook

has been accepted towards fulﬁllment
of the requirements for the

degree in Plant Biology

Major Pro ssysiénature
7 [2/ (9% ’
I

/ {ate

MSU is an Afﬁrmative Action/Equal Opportunity Institution

 

 

 

LIBRARY

Michigan State

University

 

 

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

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/DateDue.p65-p.15

 

DISEASE, POLLINATOR, AND RESOURCE LIMITATION
INFLUENCES ON THE REPRODUCTIVE BIOLOGY AND
GROWING SEASON OF ARISAEMA TRIPHYLLUM,

J ACK-IN-THE-PULPIT

By

Jessica L. Cook

A THESIS

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

MASTER OF SCIENCE
Department of Plant Biology

2004

ABSTRACT
DISEASE, POLLINATOR, AND RESOURCE LIMITATION INFLUENCES ON THE
REPRODUCTIVE BIOLOGY AND GROWING SEASON OF
ARISAEMA TRIPHYLLUM, JACK-IN-THE-PULPIT
By

Jessica L. Cook

Arisaema triphyllum, Jack-in-the-pulpit, is a common spring ephemeral in eastern
forests of North America. Female reproductive success in A. triphyllum has been shown
to be limited by the amount of available pollen and resources. Disease has also been
shown to limit female reproductive success in A. triphyllum; populations frequently
contain plants infected with the systemic rust Uromyces ari-triphylli. Previous research
demonstrated that disease reduces seed set in infected females, but results came from a
single population over a single growing season. This study investigated the relative
importance of pollen limitation, resource limitation, and disease in determining female
reproductive success as well as the effects of disease on the emergence and growing
season of vegetative plants at four populations in Michigan over two growing seasons.
The results of the two year study indicated that although pollen limitation and disease
both limit female reproductive success, disease has a much greater effect. Disease limited
flowering period length, seed set frequency, and seed weight. The amount of available
resources, i.e. the proximate environment, did not appear to affect female reproductive
success. Infected vegetative individuals emerged earlier than healthy plants and had a
shorter growing season. Plants growing in more alkaline soils had a longer growing

season than plants growing in more acidic soils.

“I am still learning ........ ” Michelangelo

This thesis is dedicated to loved ones and those still seeking
the “rainbow connection” (Thanks Kermit)

iii

ACKNOWLEDGEMENTS

The author gratefully acknowledges the members of her graduate advisory
committee: Dr. Andy J arosz, Dr. Jeff Conner, and Dr. Alan Prather for their positive
feedback, guidance, constructive criticism, and overall support for this project. The
author also wishes to especially thank Dr. Fernando Cardoso for conducting the majority
of the statistical analyses. Drs. Pat and Mukta Webber generously allowed the author
access to the J ack-in-the—pulpits on their property and tolerated gaudy pin ﬂags in their
backyard. Rob Schafer generously donated his time for fungicide preparation and
application. Funding for this research was provided in 2003 by a stipend from the Hanes
Trust. Monies to attend the 2003 Ecological Society of America meeting were provided
by the MSU Plant Biology Department and the MSU Graduate School. Special thanks to
my many ﬁeld assistants: Chuck Hughes, Frances Knapczyk, Chris Gaukler, Teddi
Bearman, Emily Cohen, Erica McConnell, Kyle Kirkby, and Erin Mason. The author also
wishes to thank her immediate & extended family, Carrie Scheele, Sara Kaiser, and Holly

Nieusma for their enthusiasm and support.

iv

TABLE OF CONTENTS

LIST OF TABLES .................................................................................. vii

LIST OF FIGURES ................................................................................... x

CHAPTER 1

EFFECTS OF POLLEN LIMITATION, RESOURCE LIMITATION, AND DISEASE

ON FEMALE REPRODUCTIVE SUCCESS .................................................... 1
Pollen limitation .............................................................................. 2
Resource limitation ........................................................................... 6
Disease ........................................................................................ 8
Literature Cited ............................................................................. 16

CHAPTER 2

FEMALE REPRODUCTIVE SUCCESS IN ARISAEMA TRIPHYLLUM: EFFECTS
OF POLLEN LIMITATION, RESOURCE LIMITATION, AND DISEASE. . . . . .......21

 

Introduction .................................................................................. 21
Materials & Methods ...................................................................... 24
Results ....................................................................................... 29
Inﬂorescence height ............................................................... 29
Flowering period length .......................................................... 31
Seed set frequency ................................................................. 32
Seed production ................................................................... 34
Discussion ................................................................................... 36
Literature Cited ............................................................................. 42
CHAPTER 3
EFFECTS OF DISEASE AND RESOURCE LIMITATION ON THE GROWING
SEASON OF VEGETATIVE ARISAEMA TRIPHYLLUM. . ............................... 46
Introduction ................................................................................. 46
Materials & Methods ...................................................................... 48
Results ....................................................................................... 5 1
Plant emergence .................................................................. 51
Growing season length ........................................................... 53
Discussion .................................................................................. 57
Literature Cited ............................................................................ 62
CHAPTER 4
CONCLUSIONS ................................................................................... 64
Literature Cited ............................................................................ 67

TABLE OF CONTENTS cont’d
APPENDICES

Appendix A: Analysis of variance of environmental data at Hudson and Lott
populations for the a) 2002 and b) 2003 ﬂowering seasons. .........68

Appendix B: Mean values of light availability, soil pH, soil moisture, and soil
temperature for Hudson (HU) and Lott (LS) at the beginning,
middle, and end of the 2002 and 2003 ﬂowering seasons ............70

Appendix C: Correlations among environmental variables in 2002 (above
diagonal) and 2003 (below diagonal) .................................... 71

Appendix D: Mean values of light availability, soil pH, soil moisture, and soil

temperature for all populations at the beginning, middle, and end of
the 2002 and 2003 growing seasons ...................................... 72

vi

Table 2.1.

Table 2.2.

Table 2.3.

Table 2.4.

Table 2.5.

Table 2.6.

Table 2.7.

Table 2.8.

Table 2.9.

LIST OF TABLES

Least square mean inﬂorescence height (cm) and standard error values for
healthy & diseased male and female A. triphyllum ................................ 30

Analysis of variance for the effects of population, sex, disease status and year
on inﬂorescence height ................................................................. 30

Least square mean inﬂorescence height (cm) and standard error values for A.
triphyllum at Hudson and Lott populations in 2002 & 2003. . 30

Analysis of variance for the effects of population, sex, disease status and year
on ﬂowering period length of healthy & diseased A. triphyllum at Hudson and
Lott populations in 2002 & 2003 ..................................................... 31

Least squares mean ﬂowering period length (days) and standard error values
for male and female A. triphyllum in 2002 & 2003 ................................. 31

Least square mean ﬂowering period length (days) and standard error values
for healthy and diseased A. triphyllum at Hudson and Lott populations in
2002 & 2003 ............................................................................ 32

Seed set frequency of unmanipulated controls and hand-pollinated healthy &
diseased A. triphyllum females ........................................................ 33

Logistic regression to test the effects of environmental variables (light, soil
pH, soil moisture, soil temperature) on seed set frequency in 2003 ............ 34

Analysis of variance for effects of population, disease status and year on the
number of seeds produced by healthy & diseased A. triphyllum setting seed at
Hudson and Lott populations in 2002 & 2003 ...................................... 35

Table 2.10. Least square mean number of seeds produced by hand-pollinated and

unmanipulated healthy and diseased A. triphyllum females setting seed in
Hudson and Lott populations ......................................................... 35

Table 2.11. Least square mean seed weight (g) for hand-pollinated & unmanipulated

healthy & diseased A. triphyllum females setting seed in Hudson & Lott
populations .............................................................................. 35

Table 2.12. Analysis of variance for effects of population, disease status and year on the

average weight of seeds produced by healthy & diseased A. triphyllum at
Hudson and Lott populations in 2002 & 2003 ................................... 36

vii

LIST OF TABLES cont’d

Table 2.13. Least square mean seed weight and standard error for A. triphyllum females
setting seed in Hudson & Lott populations in 2002 & 2003 ................... 36

Table 3.1. Analysis of variance of Julian emergence dates for the effects of population
& disease for vegetative A. triphyllum at all four populations in a)2002 &
b)2003 .................................................................................... 5 1

Table 3.2. Least square mean Julian emergence dates and standard errors for healthy
and diseased A. triphyllum in 2003 .................................................. 52

Table 3.3. Least square mean Julian emergence dates and standard errors for A.
triphyllum in Hudson, Webber, Elsesser, and Lott populations in 2002 and
2003 .................................................................................... 52

Table 3.4. Analysis of variance for the effects of population, disease, and environ-
mental variables early in the season on plant emergence in 2002 .............. 53

Table 3.5. Analysis of variance for the effects of population, disease, and environ-
mental variables early in the season on plant emergence in 2003 .............. 53

Table 3.6. Analysis of variance of growing period length for the effects of population
and disease on healthy & diseased vegetative A. triphyllum at all four
populations in 2002 ................................................................... 54

Table 3.7. Least square mean growing period length (days) and standard errors for
healthy & diseased vegetative A. triphyllum at all four populations in 2002.
........................................................................................... 54

Table 3.8. Analysis of variance of growing period length for the effects of population
and disease on healthy & diseased vegetative A. triphyllum at all four

populations in 2003 ................................................................... 55

Table 3.9. Least square mean growing period length (days) and standard errors for
vegetative A. triphyllum at all four populations in 2003 ......................... 55

Table 3.10. Least square mean growing period length (days) and standard errors for
healthy & diseased vegetative A. triphyllum in 2003 ........................... 55

Table 3.11. Analysis of variance of environmental data at all four populations for
early( 1), middle (2), and end(3) of the growing season in 2002 .............. 56

Table 3.12. Analysis of variance of environmental data at all four populations for
early(1), middle (2), and end(3) of the growing season in 2003 .............. 57

viii

LIST OF TABLES cont’d

Appendix A. Analysis of variance of environmental data at Hudson and Lott
populations for the a) 2002 and b) 2003 ﬂowering seasons ........ 68

Appendix B: Mean values of light availability, soil pH, soil moisture, and soil
temperature for Hudson (HU) and Lott (LS) at the beginning,
middle, and end of the 2002 and 2003 ﬂowering seasons .......... 70

Appendix C: Correlations among environmental variables in 2002 (above
diagonal) and 2003 (below diagonal) .................................. 71

Appendix D: Mean values of light availability, soil pH, soil moisture, and soil

temperature for all populations at the beginning, middle, and end of
the 2002 and 2003 growing seasons .................................... 72

ix

LIST OF FIGURES

Figure 1: Inﬂorescence height measurement ................................................... 25

Chapter 1. Effects of pollen limitation, resource limitation, and disease on female
reproductive success

Effective sexual reproduction in angiosperrns requires the production of ﬂoral
structures such as pollen, pollinator attractants and/or rewards, and ovules. These
structures impose a resource cost; according to the principle of allocation, individual
plants have a finite amount of resources to use for survival, i.e. grth and maintenance,
as well as reproduction (Kay 1987; Fitter & Setters 1988; Nault & Gagnon 1988; Lovett
Doust 1987; Delph & Meagher 1995; Worley & Harder 1996; Barbour et. a1. 1999).
These resources are allocated based on life history patterns that have evolved in response
to natural selection, which acts to optimize survival and reproduction. Even after plants
evolve a successful strategy of resource allocation for sexual reproduction and produce a
ﬂower or inﬂorescence, they must overcome additional challenges to produce fruit and
seed, i.e. the ultimate goals of sexual reproduction. These limitations on fruit and seed
production, hereafter referred to as female reproductive success, include pollen limitation
and resource limitation.

Numerous studies have focused on systems where pollen limitation and/or
resource limitation is present and identiﬁed the effects of each on female reproductive
success. Female reproductive success is often limited by inadequate pollen deposition
due to variability in pollinator activity and/or efﬁciency (Schemske et al. 1978; Rust
1980; Schemske & Horvitz 1984; Montalvo & Ackerman 1986; Campbell 1987; Young
1988; Zimmerman & Aide 1989; Andersson 1996; Waser et al. 1996; Herrera 2000).
Resource availability has also been shown to limit female reproductive success (Went &
Westergaard 1949; Cloudsley-Thompson & Chadwick 1964; Policansky 1981; Lovett

Doust & Cavers 1982; Willson & Burley 1983; Bierzychudek 1984; Lubbers &

Christensen 1986; Lovett Doust et al. 1986; Ackerman & Montalvo 1990; Niesenbaum
1993).

A few studies have demonstrated that disease, in addition to pollen and resource
limitations, can also affect female reproductive success. Disease can indirectly affect
female reproductive success by altering pollinator behavior or reducing the amount of
resources available for reproduction (Roy 1993 & 1996; Althizer et al. 1998) as well as
directly by attacking reproductive structures of the host plant (Alexander 1987; Burdon
1987; Parker 1987; Clay 1986, 1987, & 1990). In this chapter I will examine pollen
limitation, resource limitation, and disease in detail and discuss the speciﬁc ways that
each can limit female reproductive success.

Pollen limitation

Many plants depend on pollinators for successful pollen transfer/deposition and
ﬂowers have evolved an incredibly wide range of shapes, sizes, colors and scents to
inﬂuence pollinator behavior (Bell 1985; Haynes et al. 1991; Andersson 1996; Conner &
Rush 1996; Menzel & Miiller 1996; Raguso 2001; Pellmyr 2002). Most interactions
between plants and their pollinators are mutualistic and typically include a ﬂoral reward,
e. g. nectar or pollen, in exchange for pollen transfer. However, the plant is under
selection to maximize its attractiveness to highly effective pollinators at the lowest cost.
The two components of pollinator effectiveness are pollinator efﬁciency and quality of
pollinator service. Pollinator efﬁciency simply refers to the number of pollen grains
deposited on stigmas by a particular pollinator during a single visit, while the quality of
pollinator service is determined by observing which insect visitors make contact with

anthers and stigma (Bertin 1989).

Pollen limitation of female reproductive success appears to be common (Burd
1994). The fact that plants commonly invest a large proportion of a ﬂower’s biomass into
pollinator attractants (e. g. scent) (Pleasants & Chaplin 1983; Southwick 1984) suggests
that it is difﬁcult for plants to acquire adequate pollinator service. In his recent review,
Burd (1994) reported that pollen limitation was been documented for 159 of 258 (62%)
plant species. He concluded that: 1) the magnitude of pollen limitation is often large, 2)
percent fruit set and fruit output of hand-pollinated plants are signiﬁcantly higher than
that of unmanipulated controls, and 3) that most species experience variability in the level
of pollen limitation within a season. He also reported that only a few species experienced
a signiﬁcant decline in fecundity as the result of receiving supplemental pollen, which
suggests that negative effects from pollen overabundance are relatively uncommon. He
suggests that pollen limitation may be an adaptive response of resource allocation
strategies to a pollination environment that has large stochastic variation in pollen
delivery and ovule fertilization (Burd 1994).

The majority of Burd’s (1994) data measured pollen limitation by comparing the
percentage fruit set of hand-pollinated plants with that of unmanipulated controls. Pollen
limitation is typically assessed using artiﬁcial pollination. However, it can also be
characterized by low stigmatic loads relative to ovule number (Snow 1986) or
correlations of the number of pollinator visits with seed set (Real & Rathcke 1991).

Spring ephemerals are especially prone to pollen limitation since ﬂowering occurs
between the time when temperatures are warm enough for pollinator activity but prior to
the canopy closure and pollinator abundance often ﬂuctuates seasonally (Schemske et al.

1978). For example, only 33% of unmanipulated ﬂowers of Erythronium albidum set

seed, compared with 78% of ﬂowers outcrossed by hand (Schemske et al. 1978).
Pollinators of E. albidum were active during brief, erratic periods when weather
conditions allowed movement, and peak ﬂowering rarely occurred during peak pollinator
activity (Schemske et al. 1978). Studies conducted in many generalized pollination
systems indicate that temporal variation in pollinator availability and activity reduces
fruit and seed production. Zimmerman & Aide (1989) studied the generalized pollination
system of the orchid Aspasia pricipissa and found that fruit production was limited by
pollinator activity. Speciﬁcally, they observed that fruit set was over six times greater for
hand-pollinated plants compared to unmanipulated controls. Campbell (1987) found that
variation in fruit set among three populations of Veronica cusickii resulted primarily from
differences in pollinator visitation. She observed an overall increase in fruit set for hand-
pollinated plants, however there were differences in the degree of pollinator limitation
among the three populations; in only one population was there was a signiﬁcant
difference in seed set between unmanipulated controls (44%) and hand-pollinated (91%)
plants.

There are numerous examples of pollination systems where not all pollinators are
equally effective (Schemske & Horvitz 1984; Montalvo & Ackerman 1986; Young 1988;
Waser et al. 1996; Herrera 2000). Schemske & Horvitz (1984) found that ﬂoral visitors
to the neotropical orchid Calathea ovandensis differed signiﬁcantly in their ability to
affect fruit set; this variation creates selection pressure on the plant to specialize on the
most effective pollinator. Young (2002) compared the effectiveness of diurnal and
nocturnal pollinators of Silene latifolia which opens its ﬂowers in the evening but closes

them by midmoming of the following day. She found that S. latifolia ﬂowers exposed

only to nocturnal visitors, i.e. moths, produced more seed than ﬂowers exposed only to
diurnal visitors (i.e. bees, ﬂies, and wasps). These results suggest that nocturnal
ﬂowering in S. latifolia is an adaptation for specialization on the more effective nocturnal
pollinators.

According to Bosch & Waser (2001) plant density can indirectly affect female
reproductive success either by changing the quantity and quality of pollination service or
because density and reproduction are in response to the quality of the surrounding
environment. Bosch & Waser (1999) found that seed set declined by one-third in sparse
Delphinium nuttallianum and Aconitum columbianum populations relative to nearby
dense populations. To determine whether reduced seed set was caused by reduced
environmental quality or pollinator quality, Bosch & Waser (2001) manipulated the
density of potted D. nuttallianum and A. columbianum plants and observed interspeciﬁc
differences in how density affected reproductive success. For A. columbianum, visitation
and seed set signiﬁcantly increased with plant density, which is consistent with the
interpretation that pollination quality is lower in sparse populations. In contrast,
pollinator visitation and seed set for D. nuttallianum did not differ between dense and
sparse arrays; this ﬁnding is consistent with the idea that environmental quality is lower
in sparse populations (Bosch & Waser 2001). Rust (1980) found that the density of male
plants inﬂuenced the likelihood of seed set in Arisaema triphyllum (L.) Schott (Araceae);
increasing the density of males signiﬁcantly increased the percentage of females setting

seed but not seed production.

Resource Limitation

Successful pollination alone does not necessarily guarantee female reproductive
success; as shown in the D. nuttallianum example, resource limitation can also inﬂuence
seed production. Resource limitation is most apparent in extreme environments. Desert
annuals must tolerate extremely dry conditions and time germination in response to
sporadic precipitation; the length of time that moisture is available is very short and
reproduction must occur quickly. Most desert annuals require approximately 60 mm of
precipitation within a few hours of germination (Cloudsley-Thompson & Chadwick
1964). Among the fastest plants to complete its life cycle is Boerrhavia repens of the
Sahara Desert. Plants of this species can go from seed to seed in as little as 10 days
(Cloudsley-Thompson & Chadwick 1964).

Resource acquisition in spring ephemerals is restricted by light availability
during the short interval between the ﬁrst warm days of spring and full leaf expansion in
the deciduous forest canopy. Water and nutrients are relatively abundant, but as the
season progresses the amount of available light decreases (Salisbury 1916a&b; Morgan
1971; Taylor & Pearcy 1976; Muller 1978; Schemske et a1. 1978). In Brownﬁeld Woods
near Urbana, Illinois, the amount of available light decreased from 50% of incident
radiation in March to ~30% in mid-April to <10% in early May as leaf expansion
occurred in the deciduous forest canopy (Schemske et al. 1978). Several studies have
demonstrated that photosynthetic rates of spring ephemerals decline in the late spring as
the canopy closes (Salisbury 1916; Anderson & Hubricht 1940; Sparling 1967; Risser &

Cottam 1968; Heinrich 1976; Taylor & Pearcy 1976). Therefore, these plants have a

short period in which to acquire the energy from sunlight and nutrients necessary for
growth and reproduction.

Seed production of the ephemeral T halictrum thalictroides was limited by light
availability late in the season (Lubbers & Christensen 1986). This study found that
artiﬁcially shading plants caused a signiﬁcant reduction in the percentage seed set and
also decreased the probability of ﬂowering again the following year (Lubbers &
Christensen 1986). Niesenbaum (1993) found that fruit set in the understory shrub
Lindera benzoin was limited by available light. Speciﬁcally, he observed that fruit set
was signiﬁcantly greater for branches in the sun versus in the shade, and low light levels
limited ﬂower production in the following growing season (Niesenbaum 1993).

These studies demonstrate that the availability of proximate resources, as
determined by the current environment, can limit female reproductive success. However,
the amount of stored resources can also exert an inﬂuence on female reproductive success
(Policansky 1981; Lovett Doust et al. 1986; Ackerman & Montalvo 1990). Ackerman &
Montalvo (1990) found that although hand-pollination of the pollen-limited orchid
Epidendrum ciliare increased fruit production, the total seed crop mass per fruit
decreased as ﬁ'uit set increased, suggesting a ﬁnite amount of resources that must be
partitioned among all pollinated ovules and developing fruits.

Over the course of its lifetime, the spring ephemeral Arisaema triphyllum
accumulates energy resources in its corm and sexual expression is strongly correlated
with corm size; plants from large corms are female while smaller corms produce male or
vegetative plants (Policansky 1981; Lovett Doust & Cavers 1982; Bierzychudek 1982 &

1984; Lovett Doust et al. 1986). Manipulations of variables such as soil nutrients,

removal of corm tissue, and soil moisture have been shown to alter the sexual expression
of this species (Pickett 1915; Schaffner 1922; Heslop-Harrison 1957). Female
reproductive success in A. triphyllum has been shown to be limited by the amount of
stored resources in the corm (Policansky 1981; Lovett Doust et al. 1986). Studies found
that females with larger corms produced more ﬂowers and had higher rates of successful
fruit and seed production (Policansky 1981; Lovett Doust et al. 1986).

The studies cited above indicate that availability of both proximate and stored
resources can limit female reproductive success. These limitations are most pronounced
in extreme environments but also exert a signiﬁcant inﬂuence on the female reproductive
success of plants found in environments where the amount of available resources declines
over the course of the growing season, e. g. spring ephemerals. Resource limitation can
have a signiﬁcant effect on female reproductive success in spring ephemerals such as T.
thalictroides, L. benzoin and A. triphyllum because they have a short period of time to
acquire the energy from sunlight necessary for growth and seed production.

Disease

Most studies on female reproductive success have concentrated on pollen
limitation and resource availability, however disease can also affect reproduction both
directly and indirectly. Direct effects include the elimination or alteration of reproductive
structures on the plant, while indirect effects can take the form of interference in
pollinator attraction or a reduction in the resources available for seed production.

Pathogens indirectly affect reproduction by altering the attractiveness of the host
plant to potential pollinators. The rust fungus Puccinia monoica produces pseudoﬂowers

on Arabis spp. that resemble true ﬂowers in color, shape, size, nectar production, and

scent (Roy 1993). These pseudoﬂowers attract insects that facilitate the rust’s
reproduction in a way that is analogous to pollination; although insects visited the
pseudoﬂowers equally as often as true ﬂowers, the duration of visits was much longer on
pseudoﬂowers (Roy 1993). Roy (1996) also found that this pathogen also affected
female reproductive success of Anemone patens, which often co-occurs with the Arabis
holboelli host. Pollen loss may occur as insects move between ﬂowers and
pseudoﬂowers, reducing the likelihood of effective pollination and subsequent seed
production in A. patens (Roy 1996). She also observed that the presence of ﬁmgal
spermatia on ﬂower stigmas reduced female reproductive success; there was a signiﬁcant
decrease in seed set when spermatia and pollen were applied at the same time relative to
the hand-pollinated controls (Roy 1996). Althizer et a1. (1998) found that Silene latifolia
plants infected with the anther smut fungus Mycobotrium violaceum often ﬂower earlier
in the season than healthy plants; a high proportion of disease may encourage insects to
visit infected plants and facilitate disease spread early in the season. This study also
discovered that bee pollinators of Silene latifolia showed a strong preference for healthy
over infected ﬂowers but bees with prior exposure to diseased ﬂowers discriminated less
against them than bees exposed only to healthy ﬂowers (Althizer et al. 1998).

Pathogens that infect the roots, stems, and leaves can indirectly affect plant
reproductive success by acting as a drain on energy, altering the allocation of resources
within infected plants and subsequently affecting the quantity and quality of seeds
produced (Burdon 1987). Foliar pathogens have been shown to cause substantial changes
in the carbon allocation patterns among competing energy sinks (Livne and Daly 1966).

This study found that the amount of newly ﬁxed carbon being transported from the ﬁrst

leaves of Phaseolus vulgaris was considerably less if the leaves were infected with the
bean rust Uromyces appendiculatus; diseased leaves accumulated carbon that was ﬁxed
at a different location within the plant, thus acting as a net energy drain (Livne and Daly
1966). Diseased plants can also experience signiﬁcant declines in the rate of
photosynthesis as well as changes in assimilate partitioning and nitrogen metabolism
(Burdon 1987). In some instances, alterations in the assimilate distribution pattern as a
result of infection have taken the form of changes in the dry weight of roots, shoots,
leaves, and reproductive structures.

Pathogens can reduce plant ﬁtness by attacking reproductive structures in the
host plant. For example, M. violaceum sterilizes infected S. latifolia plants by altering
ﬂower structure. In males, the pathogen causes the plant to replace anthers with spore
sacs ﬁlled with teliospores (Baker 1947). Infected females produce male-like ﬂowers
containing spore sacs. Fungal endophytes can also directly alter host reproduction in a
variety of ways. Plants infected by the Balansieae endophytes have substantially lower
reproductive success than their healthy counterparts; infected plants do not produce
viable inﬂorescences and are limited to vegetative reproduction (Clay 1986). However,
other endophytes can actually increase the reproductive success of the host. Clay (1987 &
1990) found that grasses infected by the Acremonium endophytes produced more
inﬂorescences and seeds than uninfected plants. He also observed that seeds from
infected plants germinated more rapidly and the resulting seedlings grew faster than those
produced by healthy plants (Clay 1987).

As is evident from the above studies, the presence of disease can indirectly affect

reproductive success by altering the attractiveness of the host plant to potential

10

pollinators or by reducing the amount of resources available for seed production. Disease
can also directly affect reproductive success of the host by changing or eliminating
reproduction and subsequently altering host ﬁtness. However, in some instances
infection can beneﬁt the host.

The Arisaema triphyllum-Uromyces ari-triphylli system

My study examines pollen limitation and resource limitation in Arisaema
triphyllum (L.) Schott (Araceae). I also examine the inﬂuence of a systemic rust
pathogen Uromyces ari-triphylli (Schw.) Seeler (Basidiomycetes) on female reproductive
success and the growing season of vegetative plants.

Arisaema triphyllum, J ack-in-the-pulpit, is a native spring ephemeral whose
distribution extends from the East Coast westward to the Great Plains states (Gleason &
Cronquist 1991). These herbaceous perennials develop from a corm and possess one or
two tripartite leaves that unfold in late April or early May. Individual plants can be
vegetative, male, or female, but the sex of individual plants is not ﬁxed; at the end of
each growing season the plant makes a developmental "decision" about which sex to be
the following year (Bierzychudek 1984). Larger plants tend to be female, medium sized
plants are usually male, and small plants are vegetative. Heslop-Harrison (1957)
demonstrated the strong causal relationship between corm size and sex; removing
portions of the corm of female plants caused them to ﬂower as males the following
season. Bierzychudek (1984) found that sex determination was based on more than the
current size of a plant; according to her model, current plant size, previous year’s size,
and previous year’s sex were more reliable for predicting the current sex of individuals

than a model containing only current plant size. Pickett (1915) also observed that

11

females switched to males as a result of decreased soil moisture. Schafﬁrer (1922) found
that the addition of soil nutrients, i.e. manure, enabled plants ﬂowering as males to ﬂower
as females the following season. However this effect is not always evident;
Bierzychudek (1984) found that supplementing plants with nutrient fertilizer did not
increase the likelihood that a plant would switch sex.

Males and females produce a single inﬂorescence with numerous ﬂowers.
Inﬂorescences are open for approximately 3-4 weeks; in central Michigan ﬂowering
begins in late April and is completed by the end of June. Arisaema triphyllum has a
kettle trap morphology that uses chemical attractants to deceive its thrip and fungus gnat
pollinators (Rust 1980; Vogel & Martens 2000). In the male spathe, waxy papillae force
insect pollinators down the inﬂorescence, causing them to wade through the pollen
accumulated at the bottom of the inﬂorescence. Insects can escape the inﬂorescence
through a small exit hole formed by a gap in the spathe edges (Rust 1980). The female
spathe appears deceptively similar, however it lacks an exit hole. After depositing pollen
on the stigmas during their efforts to escape, the pollinators become trapped and
eventually die, hence this interaction is antagonistic because the plant provides no reward
and ultimately kills its pollinators.

J ack-in-the-pulpit is self-incompatible and successful cross-pollination is
necessary for females to mature berries, which contain between 1-8 seeds (Lovett Doust
et al. 1986; Jessica Cook unpublished observation). Female reproductive success in A.
triphyllum is often pollen-limited; Bierzychudek (1981) found that hand-pollinated A.
triphyllum produced over an order of magnitude more seeds than naturally pollinated

controls. Rust (1980) found that increasing the density of males did not increase seed

12

production per female but did Signiﬁcantly increase the percentage of females setting
seed.

The quantity of stored resources in the underground corm can also limit female
reproductive success in A. triphyllum. As previously noted, females with larger corms
produce more ﬂowers, fruits, and seed (Policansky 1981; Lovett Doust et al. 1986).
However, resource limitation may not always be evident. Bierzychudek (1982) did not
ﬁnd a signiﬁcant correlation between plant size and seed number or between plant size
and seed weight. Additionally, the amount of resources stored in the corm can affect
female reproductive success in the gender labile A. triphyllum by inﬂuencing the sexual
state of individuals. The resource cost of maturing seed can lead to reductions in corm
size, which increases the likelihood that an individual plant will ﬂower as a male the
following season (Bierzychudek 1984).

The systenric rust, Uromyces ari-triphylli, is common within populations of J ack-
in-the-pulpits (Pady 1939). The rust was resident in nearly all A. triphyllum populations
surveyed in southern Michigan (EE Mason, AM J arosz & K Kampf, unpublished
observation). This pathogen invades the leaves, inﬂorescence, and corm of A. triphyllum;
many populations contain infected plants. Once a plant becomes infected, it remains
diseased for the remainder of its life. The rust is transmitted to asexually produced
cormlets but not to seed or pollen (Parker 1987). Although the mechanism of horizontal
transmission is unknown, adult healthy plants do become infected. The disease
overwinters in the corm and produces bright yellow aeciospores on the leaves and
inﬂorescence (Pady 1939). Infections decrease leaf size and cause leaves to senesce

prematurely, but ﬂowering frequency appears to be equal for healthy and infected

13

individuals (Parker 1987). It is my hypothesis that the disease also causes the
inﬂorescence to senesce prematurely as well, resulting in a shorter ﬂowering period,
which may affect opportunities for pollinator visitation. Parker (1987) found that
infected females produced one-quarter of the number of seeds produced by healthy
females. In addition, the proportion of healthy females setting seed was much greater
than the proportion of diseased females setting seed. Parker (1987) proposed that these
differences could be attributed to either reduced pollinator visitation or resource
limitation in diseased plants.

This study investigated the effects of pollen limitation, resource limitation, and
disease on the female reproductive success and growing season of A. triphyllum. While
other work has investigated these factors singly or in pairs, this is the ﬁrst study to
investigate all three simultaneously. My goal was to investigate the relative importance of
each factor within A. triphyllum populations in Michigan. Speciﬁcally, in this study I
have:

1. Conducted a hand-pollination experiment and compared the seed set frequency of
hand-pollinated versus naturally pollinated females to determine the effects of
pollen limitation;

2. Measured four environmental variables (light availability, soil pH, soil moisture,
& soil temperature) near each plant to determine the effects of resource limitation
on emergence date, growing season length, & ﬂowering period length; and

3. Compared the ﬂowering period length, inﬂorescence size, seed set frequency,
seed number, seed weight, and pollen viability of healthy and infected plants to

determine the effects of disease.

14

Since the disease has the potential to alter many stages of the host life cycle, I also
investigated the effects of disease on the growing season of vegetative plants. This last
study was initiated to compare the relative emergence and growing season length of

healthy and diseased individuals.

15

Literature Cited

Ackerman, J. D. and A. M. Montalvo. 1990. Short- and long-term limitations to fruit
production in a tropical orchid. Ecology 71 :263-272.

Alexander, H. M. 1987. Pollination limitation in a population of Silene alba infected by
the anther-smut fungus, Ustilago violacea. J. Ecol. 75:771-780.

Althizer, S. M., P. H. Thrall, and J. Antonovics. 1998. Vector behavior and the
transmission of anther smut infection in Silene alba. Amer. Mid. Nat. 139: 147-163.

Anderson, E. and L. Hubricht. 1940. A method for describing and comparing blooming-
seasons. Bull. Torrey Bot. Club 67:639-648.

Andersson, S. 1996. Floral display and pollination success in Seneciojacobea
(Asteraceae): interactive effects of head and corymb size. Amer. J. Bot. 83:71-75.

Baker, H. G. 1947. Infection of species of Melandrz‘um by Ustilago violacea (Pers.)
Fuckel and the transmission of the resultant disease. Ann. Bot. 11:333-348.

Barbour, M. G. J. H. Burk, W. D. Pitts, F. S. Gilliam, and M. W. Schwartz. 1999.
Terrestrial plant ecology. Benjamin/Cummings, New York. USA.

Bell, G. 1985. On the function of ﬂowers. Proc. R. Soc. Lond. B. 224:223-265.

Bertin, M. 1989. Pollination biology. Pp. 23—86 in W.G. Abrahamson, ed. Plant-animal
interactions. New York, McGraw-Hill. USA.

Bierzychudek, P. 1981. Pollinator limitation of plant reproductive effort. Am. Nat.117:
838-840.

Bierzychudek, P. 1982. The demography of J ack-in-the-pulpit, a forest perennial that
changes sex. Ecol. Monogr. 52:335-351.

Bierzychudek, P. 1984. Determinants of gender in Jack-in-the-pulpit: the inﬂuence of
plant size and reproductive history. Oecologia 65: 14-18.

Bosch, M. and N. M. Waser. 1999. Effects of local density on pollination and
reproduction in Delphinium nuttallianum and Aconitum columbianum
(Ranunculaceae). Amer. J. Bot. 86:871-879.

Bosch, M. and N. M. Waser. 2001. Experimental manipulation of plant density and its
effect on pollination and reproduction of two confamilial montane herbs. Oecologia
126:76-83.

16

Burd, M. 1994. Bateman’s principle and plant reproduction: the role of pollen limitation
in fruit and seed set. Bot. Rev. 60:83-139.

Burdon, J. J. 1987. Diseases and plant population biology. Cambridge University Press,
Cambridge. United Kingdom.

Campbell, D. R. 1987. Interpopulational variation in fruit production: the role of
pollination limitation in the Olympic mountains. Amer. J. Bot. 74:269-273.

Clay, K. 1986. Induced vivipary in the sedge Cyperus virens and the transmission of the
fungus Balansia cyperi (Clavicipitaceae). Can. J. Bot. 64:2984-2988.

Clay, K. 1987. Effects of fungal endophytes on the seed set and seedling biology of
Lolium perenne and F estuca arundinaceae. Oecologia. 73:358-362.

Clay, K. 1990. Comparative demography of three grarninoids infected by systemic,
clavicipitaceous fungi. Ecology. 71:558-570.

Cloudsley-Thompson, J. L. and M. J. Chadwick. 1964. Life in deserts. Dufour,
Philadelphia. USA.

Conner, J. K. and S. Rush. 1996. Effects of ﬂower size and number on pollinator
visitation to wild radish, Raphanus raphanistrum. Oecologia 105:509-516.

Delph, L. F. and T. R. Meagher. 1995. Sexual dimorphism masks life history trade-offs in
the dioecious plant Silene latifolia. Ecology 76:775-785.

Fitter, A. H. and N. L. Setters. 1988. Vegetative and reproductive allocation of

phosphorus and potassium in relation to biomass in six species of Viola. J. Ecol.
76:61 7-636.

Gleason, H. A. & A. Cronquist. 1991. Manual of vascular plants of northeastern United
States and adjacent Canada. New York Botanical Garden, Bronx. USA.

Haynes, K. F ., A. Latif, and J. Z. Zhao. 1991. Identiﬁcation of ﬂoral compounds from
Abelia grandiﬂora that stimulate upwind ﬂight in cabbage looper moths. J. Chem.
Ecol. 17:637-646.

Heinrich, B. 1976. Flowering phenologies: bog, woodland, and disturbed habitats.
Ecology 57:890-899.

Herrera, C. H. 2000. Flower-to-seedling consequences of different pollination regimes in
an insect-pollinated shrub. Ecology 81 :15-29.

Heslop-Harrison, J. 1957. The experimental modiﬁcation of sex expression in ﬂowering
plants. Biol. Rev 32: 38-90.

17

Kay, Q. 0. 1987. The comparative ecology of ﬂowering. New Phytol. 106:265-281.

Livne, A. and J. M. Daly. 1966. Translocation in healthy and rust-affected beans.
Phytopathology 56:170-175.

Lovett Doust, J. and P. B. Cavers. 1982. Sex and gender dynamics in j ack-in-the-pulpit,
Arisaema triphyllum (Araceae). Ecology 63:797-808.

Lovett Doust, L., J. Lovett Doust, and K. Turi. 1986. Fecundity and size relationships in
jack-in-the-pulpit, Arisaema triphyllum (Araceae). Amer. J. Bot. 73:489-494.

Lovett Doust, J. 1987. Plant reproductive strategies and resource allocation. Trends
Ecol. Evol. 4:230-234.

Lubbers, A. E. and N. L. Christensen. 1986. Intraseasonal variation in seed production
among ﬂowers and plants of Thalictrum thalictroides (Ranunculaceae). Amer. J. Bot.
73: 190-203.

Menzel, R. and U. Mt'iller 1996. Learning and memory in honeybees: from behavior to
neural substrates. Ann. Rev. Neuro. 19:379-404.

Montalvo, A. M. and J. D. Ackerman. 1986. Relative pollinator effectiveness and
evolution of ﬂoral traits in Spathiphyllum friedrichsthalli (Araceae). Amer. J. Bot.
73:1665-1676.

Morgan, M. D. 1971. Life history and energy relationships of Hydrophyllum
appendiculatum. Ecol. Monogr. 41 :329-349.

Muller, R. N. 1978. The phenology, growth, and ecosystem dynamics of Erythronium
americanum in the northern hardwood forest. Ecol. Monogr. 48:1-20.

Nault, A. and D. Gagnon. 1988. Seasonal biomass and nutrient allocation patterns in wild
leek (Allium tricoccum Ait.), a spring geophyte. Bull. Torrey Bot. Club 115:45-54.

Niesenbaum, R. A. 1993. Light or pollen-seasonal limitations on female reproductive
success in the understory shrub Lindera benzoin. J. Ecol. 81:315-323.

Pady, S. M. 1939. Host invasion in systemic infections of Uromyces caladii. Mycologia
3 1 :590-605.

Parker, M. A. 1987. Pathogen impact on sexual vs. asexual reproductive success in
Arisaema triphyllum. Amer. J. Bot. 74: 1758-1763.

Pellmyr, O. 2002. Pollination by animals. Pp. 157-184 in C. H. Herrera, ed. Plant-animal
interactions: an evolutionary approach. Oxford, Malden. United Kingdom.

18

Pickett, F. L. 1915. A contribution to our knowledge ofArisaema triphyllum. Torrey Bot.
Club Memoirs 16:1-55.

Pleasants, J. M. & S. M. Chaplin. 1983. Nectar production rates of Asclepias quadrifolia:
causes and consequences of individual variation. Oecologia 59:232-238.

Policansky, D. 1981. Sex choice and the size advantage model in jack-in the pulpit
(Arisaema triphyllum). Proc. Natl. Acad. Sci. 78:1306-1308.

Raguso, R. A. 2001. Floral scent, olfaction, and scent-driven foraging behavior. Pp. 83-
105 in J. D. Thompson and L. Chittka eds. Cognitive ecology of pollination: animal
behavior and ﬂoral evolution. Cambridge University Press, Cambridge. United
Kingdom.

Real, L. A. and B. J. Rathcke. 1991. Individual variation in nectar production and its
effects on ﬁtness in Kalmia latifolia. Ecology 72:149-155.

Risser, P. and G. Cottam. 1968. Carbohydrate cycles in the bulbs of some spring
ephemerals. Bull. Torrey Bot. Club 95:359-369.

Roy, B. A. 1993. Floral mimicry by a plant pathogen. Nature 362:56-58.

Roy, B. A. 1996. A plant pathogen inﬂuences pollinator behavior and may inﬂuence
reproduction of nonhosts. Ecology 77:2445-2457.

Rust, R. W. 1980. Pollen movement and reproduction in Arisaema triphyllum. Bull.
Torrey Bot. Club 107:539-542.

Salisbury, E. J. 1916a. The oak-hombeam woods of Hertfordshire. Part I.
J. Ecol. 4283-100.

Salisbury, E. J. 1916b. The oak-hombeam woods of Hertfordshire. Part II.
J. Ecol. 4:101-117.

Schaffner, J. H. 1922. Control of the sexual state in Arisaema triphyllum and A.
dracontium. Amer. J. Bot. 9:72-78.

Schemske, D. W., M. F. Willson, M. N. Melampy, L. J. Miller, L. Verner, and K. M.
Schemske. 1978. Flowering ecology of some spring woodland herbs. Ecology 59:
351-366.

Schemske, D. W., and C. C. Horvitz. 1984. Variation among ﬂoral visitors in pollination
ability: a precondition for mutualism specialization. Science 225:519-521.

Snow, A. A. 1986. Pollination dynamics in Epilobium canum (Onagraceae):
consequences for garnetophytic selection. Amer. J. Bot. 73:139-151.

19

Southwick, E. E. 1984. Photosythate allocation to ﬂoral nectar: a neglected energy
investment. Ecology 65:1775-1779.

Sparling, J. H. 1967. Assimilation rates of some woodland herbs in Ontario.
Bot.Gaz. 128:160-168.

Taylor, R. J. and R. W. Pearcy. 1976. Seasonal patterns of the C02 exchange
characteristics of understory plants from a deciduous forest. Can. J. Bot. 54:1094-
1103.

Vogel, S. and J. Martens. 2000. A survey of the function of the lethal kettle traps of
Arisaema (Araceae), with records of pollinating fungus gnats from Nepal. Bot. J.
Linn. Soc. 133:61-100.

Waser, N. M., L. Chittka, M. V. Price, N. M. Williams, and J. Ollerton.l996.
Generalization in pollination systems, and why it matters. Ecology 77:1043-1060.

Went, F. W. and M. Westergaard. 1949. Ecology of Desert Plants HI. Development of
plants in the Death Valley National Monument, CA. Ecology 30:26-38.

Willson, M. F. and N. Burley. 1983. Mate Choice in Plants: Tactics, Mechanisms, &
Consequences. Princeton University Press, Princeton. USA.

Worley, A. C. and L. D. Harder. 1996. Size-dependent resource allocation and the costs
of reproduction in Pinguicula vulgaris (Lentibulariaceae). J. Ecol. 84:195-206.

Young, H. J. 1988. Differential importance of beetle species pollinating Dieﬂenbachia
Iongispatha (Araceae). Ecology 69:832-844.

Young, H. J. 2002. Diurnal and nocturnal pollination of Silene alba (Caryophyllaceae).
Amer. J. Bot. 89:433-440.

Zimmerman, J. K. and T. M. Aide. 1989. Patterns of fruit production in a neotropical
orchid: pollinator vs. resource limitation. Amer. J. Bot. 76:67-73.

20

Chapter 2. Female reproductive success in Arisaema triphyllum: effects of pollen
limitation, resource limitation and disease

Introduction

Life history patterns are shaped by limited energy or nutrient resources that force
trade-offs in the amount of resources that can be devoted to growth, maintenance and
reproduction (Freeman & Herron 2004). The amount of resources a plant devotes to
reproduction should be geared towards optimizing trade-offs under the current
environmental conditions in order to maximize ﬁtness. The production of reproductive
structures, e. g. ovules, is the ﬁrst step toward achieving female reproductive success but
other obstacles must be overcome to successfully produce seed. Pollen limitation,
resource limitation, and disease can reduce female reproductive success.

Variation in pollinator activity and/or efﬁciency can lead to inadequate pollen
deposition on the stigmas of ﬂowers (Schemske et al. 1978; Rust 1980; Bierzychudek
1981; Waser 1983; Schemske & Horvitz 1984; Montalvo & Ackerman 1986; Campbell
1987; Young 1988 & 2002; Zimmerman & Aide 1989; Andersson 1996; Waser et a1.
1996; Herrera 2000). As a consequence, the number of ovules fertilized may be smaller
than the maximum ntunber a plant can mature into seed with the amount of energy and
nutrient resources available. Schemske et al. (1978) observed that ﬂuctuations in
pollinator abundance limited female reproductive success in several spring ephemerals;
hand-pollination increased seed production relative to unmanipulated controls. Two
studies have indicated that female reproductive success in the spring ephemeral Arisaema
triphyllum is pollen-limited (Rust 1980; Bierzychudek 1981). Rust (1980) observed that
seed production of A. triphyllum females signiﬁcantly declined with increasing distance

from potential pollen donors; females within one meter of a male plant had signiﬁcantly

21

higher seed production than females more distant from males. Bierzychudek (1981)
observed a striking difference in female reproductive success between hand—pollinated
and naturally pollinated plants; plants given supplemental pollen produced over an order
of magnitude more seeds than plants relying solely on insect pollinators.

Resource limitation can also inﬂuence female reproductive success, particularly
for plants in extreme environments or those growing in areas where the amount of
available resources, e. g. light or water, declines over the course of the growing season.
Desert annuals have evolved life history strategies to cope with extremely arid conditions
such as germinating in response to infrequent precipitation (Went & Westergaard 1949;
Cloudsley-Thompson & Chadwick 1964). The effect of resource limitation on female
reproductive success is particularly apparent in spring ephemerals whose ﬂowering and
grth are restricted by the changing pattern of light incidence on the forest ﬂoor during
the interval between the ﬁrst warm days of spring and ﬁll] leaf expansion of the
deciduous forest canopy.

At the onset of the growing season, light, water, and nutrients are relatively
abundant but light levels decline dramatically as the season progresses (Salisbury
1916a&b; Morgan 1971; Taylor & Pearcy 1976; Muller 1978; Schemske et a1. 1978).
Schemske et al. (1978) observed that the amount of available light decreased from 50%
of incident radiation in March to ~30% in mid-April to <10% in early May as canopy
expansion occurred in Brownﬁeld Woods near Urbana, Illinois.

Several studies have demonstrated that a decrease in light levels can reduce seed
production in spring ephemerals. Lubbers & Christensen (1986) found that low levels of

light late in the growing season limited seed production of Ihalictrum thalictroides and

22

that artiﬁcial shading resulted in a signiﬁcant reduction in percentage seed set. Fruit set
on individual branches of the understory shrub Lindera benzoin was affected by light
availability; branches in the sun had greater fruit set than those in the shade (N iesenbaum
1993). In addition to proximate resources, the amount of stored resources can affect
female reproductive success; several studies have demonstrated that among female A.
triphyllum, plants with larger corms had higher rates of seed production and matured a
greater number of seeds compared to plants with smaller corms (Policansky 1981; Lovett
Doust et al. 1986).

Although numerous investigations have examined how pollen limitation and/or
resource limitations affects female reproductive success, few have been conducted in
systems where disease is also a potential limitation. Most plant-pathogen interactions are
detrimental to the host plant (Pady 1939; Baker 1947; Alexander 1987; Burdon 1987;
Parker 1987; Clay 1986, 1987, & 1990; Roy 1993; Althizer et.al. 1998) and disease can
limit female reproductive success by reducing or eliminating the reproduction of the host
plant (Pady 1939; Alexander 1987; Burdon 1987; Parker 1987; Clay 1990). Mycobotrium
violaceum, a systemic fungal pathogen of Silene latifolia, sterilizes female ﬂowers by
aborting the pistils and converting it to a male-like ﬂower containing fungal spore sacs
instead of anthers. Infection by Balansieae endophytes reduces female reproductive
success of the host; diseased individuals are unable to produce viable inﬂorescences or
seeds and are limited to vegetative reproduction (Clay 1986). Parker (1987) was the ﬁrst
to investigate the effects of the systemic rust pathogen Uromyces ari-triphylli (Schw.)
Seeler (Basidiomycetes) on female reproductive success of Arisaema triphyllum. The

proportion of diseased females setting seed was signiﬁcantly lower than the proportion of

23

healthy females setting seed and infected females produced signiﬁcantly fewer seeds than
healthy females. However, Parker’s (1987) conclusions were based on the performance
of plants in a single population over a single growing season.

Although the majority of plant-pathogen interactions are negative, there are a few
instances when the presence of a systemic pathogen can actually increase the female
reproductive success of the host (Clay 1987 & 1990). Grasses infected by Acremonium
endophytes produce more seeds than uninfected plants, and have increased rates of seed
germination and seedling growth (Clay 1987).

The spring ephemeral Arisaema triphyllum (L.) Schott (Araceae) is particularly
amenable for investigating limitations on female reproduction, measured here as seed set.
Previous studies have found that seed set is affected by pollen limitation (Rust 1980;
Bierzychudek 1981), resource limitation (Policansky 1981; Lovett Doust et al. 1986) and
disease caused by Uromyces ari-triphylli infections (Parker 1987). However, no study
has investigated these three factors simultaneously. The goal of this study was to
determine the relative importance of pollen limitation, resource limitation and disease on
female reproductive success in A. triphyllum populations in Michigan. Pollen limitation
was assessed by comparing seed set in unmanipulated and hand-pollinated females.
Resource limitation was evaluated by examining correlations between seed set and
environmental factors: light, soil pH, soil, moisture, & soil temperature. The importance
of disease was investigated by comparing seed set of healthy and diseased females.
Materials & Methods

Experiments were conducted in April-September of 2002 & 2003 at two

populations, located in the Lott and Hudson woodlots in central, lower peninsula

24

Michigan. Both sites are part of Michigan State University’s natural areas holdings and
are located within 5 miles campus (http://www.cpp.msu.edu/nat_area/). Incidence (i.e.
percentage of infected plants within a population) of U. ari-triphylli infection was
approximately 50% at each site.

Each year approximately 150 ﬂowering plants were selected randomly in each
population. Flowering phenology of each inﬂorescence was estimated by checking plants
every second day to determine the “start” and “stop” dates for ﬂowering. Flowering
“start” criterion was the complete expansion of the two halves of the spathe with no
visible overlap around the spadix. Flowering was considered to "stop" when the
inﬂorescence itself had senesced or the senescence of the peduncle occurred, causing the
ﬂower to topple over. Inﬂorescence height for each plant was measured as the distance
indicated in Figure 1. Disease status was assessed by checking the spathe and leaves for

the presence of any of the multiple spore types associated with U. ari-triphylli infection.

 

Inﬂorescence
Height (cm)

 

 

 

 

 

Figure 1. Inﬂorescence height measurement. Photo courtesy of Albert G. Richards.

25

The inﬂuence of resources on ﬂowering phenology and seed set was evaluated by
measuring localized environmental variables (i.e., light, soil pH, soil moisture and soil
temperature) at approximately 50 plants in each population. Measurements were made at
the beginning, middle, and end of the ﬂowering season since the environment can
ﬂuctuate over time, e. g., light levels decrease as leaves in the forest canopy expand. In
2002, light was measured as light intensity using an Extech Model 401025 Foot Candle/
Lux Meter placed approximately one inch above the plant leaf. All light measurements
were made on overcast days between 12-4 pm. In 2003, light was measured as percent
canopy cover using a spherical densitometer placed approximately one inch above the
plant leaf. Soil pH and moisture were measured at a depth of three inches using a Kelway
Soil Tester HB-2. Soil temperature was measured at a depth of 2.5 inches using a Tele-
tru thermometer in 2003 only.

Fruits were collected in September of both years and the number of seeds
produced by each inﬂorescence was counted. Seeds were deﬂeshed prior to counting and
weighing. Individual seed weights were recorded in 2002, while only the total weight of
all seed from a single inﬂorescence was recorded in 2003. Individual seed weights were
initially measured because of concern for differences in seed weight distributions
between healthy and diseased plants, however, upon graphical examination of the 2002
data it was determined that the distributions were approximately the same.

Analysis of variance (ANOVA) using Type III tests were carried out using PROC
MD(ED (SAS Institute 1997) to determine the signiﬁcance of environmental factors on
ﬂowering phenology and seed set. Least square means were reported to correct for the

inﬂuence of other main effects in the unbalanced design. Additionally, for inﬂorescence

26

height we also used the Protected Fisher’s LSD (least squares difference) to do pair-wise
mean comparisons.
Hand-pollination experiment

In the observational work described above, any potential inﬂuence of disease and
environmental factors on seed set is confounded with reduced seed set due to limited
pollinator service. A hand-pollination experiment was carried out in both years to
determine whether seed set was pollen-limited. Pollen from ~60 healthy males and ~60
diseased males was used to pollinate 116 healthy females and 97 diseased females using a
2 X 2 factorial design. Control plants (N=166) received no supplemental pollen and
included females designated speciﬁcally for this experiment as well as females marked
for the ﬂowering phenology experiment. In 2002 all females came from Hudson and Lott
populations, but the number of diseased females at Hudson was low (N=14). To ensure
an adequate sample size in 2003, diseased females from a third site, Webber, were added.
These females were used in the analyses for the hand-pollination experiment but were
excluded from the seed production analyses for replication purposes, i.e. so that the seed
data from 2002 (Lott & Hudson females only) and 2003 could be combined for analyses.
Pollen was applied with a camel hair brush when the ﬂowers in the female spathe were
observed to be open. The brush was swabbed around one or more male spathes before
being wiped on the ﬂowers in the female spathe. Plants of a single genet are
incompatible, and it is likely that clumps of plants are the clonal progeny of a single
individual (Bierzychudek 1984). In order to reduce the risk of incompatibility, pollen

donors were selected randomly from a distance of at least 5m away from the recipient

27

female. Females were monitored for seed set and the same seed measurements used in
the ﬂowering period experiment were applied.

Analyses of the data were performed using PROC GENMOD of SAS to do
logistic regressions. Main effects (pOpulation, disease status, year, and sex) were tested
using Likelihood Ratio tests; signiﬁcance of a priori contrasts were assessed by Chi-
Square tests. Contrasts were used to compare 1) seed set frequency of hand-pollinated
vs. unmanipulated control females and healthy vs. diseased females and 2) probability of
setting seed for healthy and diseased females from a healthy vs. diseased pollen donor.
Analysis of variance (ANOVA) using Type 111 tests were carried out using PROC
MIXED (SAS Institute 1997) to determine the signiﬁcance of disease on seed production.
Least square means were reported to correct for the inﬂuence of other main effects in
unbalanced design. PROC GLM (SAS Institute 1997) was used to determine if pollen
loading from hand pollination had a signiﬁcant effect on seed number.

F ungicide trial

It is possible that resource limited plants are more susceptible to infection. If
true, then the effects of resource limitation and infection would be confounded. Curing
plants of disease would allow an independent assessment of the roles of both disease and
resource availability. Just prior to the 2003 ﬁeld season, SOOmL of the fungicide Ti1t®
(propiconazole) at a concentration of 14g/L was applied to the soil surrounding the corms
of 45 females that were infected in 2002. My intent was to eliminate the effects of disease
and pollen limitation (by including treated females in the hand pollination experiment),
thus allowing me to isolate the effects of resource limitation on female reproductive

success. Upon successful treatment, I intended to compare the seed set frequency and

28

seed production of treated females with that of diseased females and females who were
healthy their entire lives. Unfortunately, only a few of the treated plants emerged; none
produced an inﬂorescence and all had malformed leaves.
Results
Healthy plants had a higher probability of being male than diseased plants (X2 =

21.94; p <0.0l), with 66% of healthy ﬂowering plants being male and only 48% of
diseased plants ﬂowering as male.
Inflorescence Height

Inﬂorescence height differed between the sexes and between healthy and diseased
females (Table 2.1). However, signiﬁcant interactions between sex and disease status
and between population and year occurred (Table 2.2). Both healthy and diseased
females were signiﬁcantly larger than males (Table 2.1). However, females were more
affected by disease, since inﬂorescence height for diseased females was signiﬁcantly
smaller (9%) than healthy females. In contrast, inﬂorescence height for healthy and
diseased males did not differ statistically although diseased male inﬂorescences were, on
average, 4% smaller than their healthy counterparts. Inﬂorescences were smaller in 2002
compared to 2003 (Table 2.3). In 2002, mean inﬂorescence height at Hudson was
signiﬁcantly smaller than at Lott (Table 2.3). Inﬂorescence size was larger at both sites

in 2003, and the inﬂorescences at Lott and Hudson did not differ.

29

Table 2.1. Least square mean inﬂorescence height (cm) and standard error values for
healthy & diseased male and female A. triphyllum.

 

 

 

Healthy Diseased Overall
Sex Number Mean Number Mean Number Mean
Male 384 5953:0043 a 101 5.73i0.083 485 5.84i0.046
a

Female 196 7.1410059 b 110 6.50i0.095 306 6.82:t0.057
c
Overall 580 6.552t0.037 211 6.11i0.065

 

' means followed by the same letter do not differ signiﬁcantly

Table 2.2. Analysis of variance for the effects of population, sex, disease status and year
on inﬂorescence height.

 

 

 

 

lMain Effect df Mean Square F Value Pr > F
Population 1 1.50 2.30 0.1295
Sex 1 124.11 190.38 <0.0001
Pop*Sex 1 0.68 1.05 0.3059
Disease status 1 21.98 33.72 < 0.0001
Pop*Disease 1 0.04 0.07 0.7998
Sex*Disease 1 5.66 8.68 0.0033
Year 1 10.16 15.58 <0.0001
Pop*Year 1 2.56 3.93 0.0478
Sex*Year l 1.56 2.39 0.1228
Disease*Year 1 0.35 0.54 0.4626
Error 779 .68

 

 

 

Table 2.3. Least square mean inﬂorescence height (cm) and standard error values for A.
triphyllum at Hudson and Lott populations in 2002 & 2003.

 

 

 

2002 2003
Population Number Size Number Size
Hudson 163 6.08i0.082 a 210 6.47i0.081 c
Lott 234 6.32:l:0.057 b 183 6.4610060 c

 

' means followed by the same letter do not differ signiﬁcantly

30

 

Flowering period length

Flowering period was affected by year, sex, disease status, and population with
signiﬁcant sex by year and population by disease interactions (Table 2.4). The
interaction between year and sex was due to males in 2003 ﬂowering for a shorter period
than ﬂowers from males in 2002 or females (Table 2.5). The ﬂowering period for
females did not differ between years, and was similar to the ﬂowering period for males in

2002.

Table 2.4. Analysis of variance for the effects of population, sex, disease status and year
on ﬂowering period length of healthy & diseased A. triphyllum at Hudson and
Lott populations in 2002 & 2003.

 

 

 

IMain Effect df Mean Square F Value Pr > F
Population 1 407.48 15.96 < 0.0001
Year 1 364.33 14.27 0.0002
Pop*Year 1 53.61 2.10 0.1478
Sex 1 287.99 1 1.28 0.0008
Sex*Year 1 700.07 27.42 < 0.0001
Pop*Sex 1 53.87 2.11 0.1470
Disease status 1 3640.78 142.60 < 0.0001
Disease*Year 1 62.81 2.46 0.1170
Pop*Disease 1 247.40 9.69 0.0019
Sex*Disease 1 0.51 0.02 0.8920
Error 778 25.53

 

 

 

Table 2.5. Least square mean ﬂowering period length (days) and standard error values
for male and female A. triphyllum in 2002 & 2003.

 

2002 2003 Overall

Sex Number Mean Number Mean Number Mean

 

 

Male 201 25.4i0.408 a 182 21.8:t0.377 b 383 23.6:l:0.29l

Female 95 24.9:l:0.471 a 102 25.3:t0.473 a 197 25.1i0.355

Overall 296 25.1:l:0.319 282 23.5:l:O.317

 

' means followed by the same letter do not differ signiﬁcantly

31

 

Healthy plants had signiﬁcantly longer ﬂowering periods than diseased plants,
and the ﬂowering period of healthy plants at each site were similar (Table 2.6). The
disease by pOpulation interaction was due to a differential response of disease at the two
sites. Flowering period of diseased plants at the Hudson site was 26% shorter than
healthy plants. In contrast, diseased plants at Lott were less affected by infection; the
reduction in ﬂowering period was only 15%.

Table 2.6. Least square mean ﬂowering period length (days) and standard error values for

healthy and diseased A. triphyllum at Hudson and Lott populations in 2002 &
2003.

 

Healthy Diseased Overall

Pop. Number Mean Number Mean Number Mean

 

 

Hudson 320 26910324 53 19910720 373 23.410.402

a b a
Lott 260 27310324 157 23.310.411 417 25310261
a c b

Overall 790 27.110.229 210 21 .610.410

 

 

' means followed by the same letter do not differ signiﬁcantly

There was no evidence that proximate resources affected ﬂowering period, since
no single or combination of environmental factors (light availability, soil moisture, soil
pH or soil temperature) had a signiﬁcant effect on ﬂowering period length in either year
and or population (Appendix A). The means for environmental variables in the 2002 and
2003 ﬂowering seasons are in Appendix B.

Seed Set Frequency
Rust infection had a signiﬁcant affect on a female’s ability to set seed (X2: 64.80

and p<0.0001, df =1 ); an infected female was ﬁve times less likely to set seed than a

32

healthy female (Table 2.7). Pollen limitation also affected the probability of a female
setting seed. For healthy females, hand-pollination increased the probability of seed set
by 22% (Table 2.7; X2=12.40 and p=0.0004, df =1). The disease status of the male used
to donate pollen did not inﬂuence the probability of seed set in hand-pollinated healthy
females. When the pollen donor was a healthy male the probability of a healthy female
setting seed was 79% (n=5 8), which was similar to seed set when the pollen donor was a
diseased male (81%, n = 58; X2 = 0.05 and p = 0.816, df =2). Hand-pollination of
diseased females had little effect, increasing seed set probability by a non-signiﬁcant 3%
(X2=0.311 and p>0.25, df =1). Thus, disease undercuts a female’s ability to set seed to the

extent that pollen limitation is no longer a consequence.

In contrast to the signiﬁcant effects of rust infection and pollen limitation, there
were no consistent relationships between seed set and environmental variables. In 2002,
there was no evidence that any environmental factor (light availability, soil moisture, soil
pH or soil temperature) affected seed set. In 2003, seed set increased with higher soil pH

and higher soil moisture (Table 2.8).

Table 2.7. Seed set frequency of unmanipulated controls and hand-pollinated
healthy & diseased A. triphyllum females from Hudson, Lott, and Webber.

 

 

 

Female recipient Naturally pollinated Hand pollinated
Healthy 0.58 (N=102) 0.80 (N=116)
Diseased 0.1 1 (N =64) 0.14 (N =97)

 

 

33

Table 2.8. Logistic regression to test the effects of environmental variables (light, soil
pH, soil moisture, soil temperature) on seed set frequency in 2003.

 

 

 

ain Effect df Chi Square Pr > Chi Square
Population 1 0.48 0.4868
Disease status 1 5.69 0.0171
% Cover 1 0.05 0.8284
Soil pH 1 6.98 0.0082
Soil moisture l 4.49 0.0341
Soil temperature 1 1.04 0.3084

 

 

Seed production

For all females setting seed, regardless of whether or not they were hand-
pollinated or unmanipulated, the number of seed produced was not inﬂuenced by year,
population or disease status (Table 2.9). Pollen loading, however, had a signiﬁcant effect
on seed production; hand-pollination increased the number of seeds produced relative to
their unmanipulated counterparts (Table 2.10). The sample sizes differ between Table 2.7
and Table 2.10 because for 11 females (of the 173 total females setting seed) the seed
could not be collected. Average weight of seed from healthy females was 19% heavier
than seed from diseased females (Table 2.11), indicating that diseased females are
constrained in their ability to provision seed. For healthy females only, seeds produced by
control plants were signiﬁcantly heavier than those produced by hand-pollinated females;
it is possible that the control plants were able to allocate more resources into each seed
because they produced fewer seeds than hand-pollinated plants (Tables 2.10 & 2.11).
Seed weight was also inﬂuenced by a signiﬁcant interaction between population and year
(Table 2.12). Seed weights at Lott did not differ between 2002 and 2003 (Table 2.13). At
Hudson, seed weight in 2003 was greater than seed weights in 2002. Seed weights at

Hudson were generally higher than seed weights at Lott.

34

Table 2.9. Analysis of variance for effects of population, disease status and year on the
number of seeds produced by healthy & diseased A. triphyllum setting seed at
Hudson and Lott populations in 2002 & 2003.

 

 

 

Main Effect df Mean Square F Value Pr > F
Population 1 1084.18 0.79 0.3745
Disease status 1 27.45 0.02 0.8873
Year 1 2538.90 1.85 0.1762
Pop*Disease 1 3677.98 2.68 0.1034
Pop* Year 1 1619.41 1.18 0.2783
Disease*Year 1 45.01 0.34 0.5620
rror 151 1372.38

 

 

 

Table 2.10. Least square mean number of seeds produced by hand-pollinated and
unmanipulated healthy and diseased A. triphyllum females setting seed in
Hudson and Lott populations in 2002 &2003. Numbers represent an average
for those females that set at least one seed on their inﬂorescence.

 

Disease Status Unmanipulated Hand-pollinated

 

 

 

Healthy 33.09 a 47.21 b
N=57 =87

Diseased 15.00 c 37.73 ab
N=7 N=11

 

* means followed by the same letter do not differ signiﬁcantly

Table 2.11. Least square mean seed weight (g) for hand-pollinated & unmanipulated
healthy & diseased A. triphyllum females setting seed in Hudson & Lott
populations.

 

Disease Status Unmanipulated Hand-pollinated

 

 

 

Healthy 0.0497 a 0.0447 b
N=57 =87

Diseased 0.0325 c 0.0336 c
N=7 N=11

 

* means followed by the same letter do not differ signiﬁcantly

35

Table 2.12. Analysis of variance for effects of population, disease status and year on the
average weight of seeds produced by healthy & diseased A. triphyllum at
Hudson and Lott populations in 2002 & 2003.

 

 

 

IMain Effect df Mean Square F Value Pr > F
POpulation 1 0.0002 2.33 0.1288
Disease 1 0.0004 4.21 0.0419
Year 1 0.0003 2.69 0.1031
P0p*Disease 1 0.0000 0.55 0.4602
Pop“ Year 1 0.0006 5.85 0.0168
Disease*Year 1 0.0002 2.09 0.1507
rror 15 1 0.0001

 

 

 

Table 2.13. Least square mean seed weight and standard error for A. triphyllum females
setting seed in Hudson & Lott populations in 2002 & 2003.

 

 

 

2002 2003
Population Number Seed weight (g) Number Seed weight (g)
Hudson 41 0.0397 1 0.004 b 43 0.0504 1 0.004 c
Lott 44 0.0383 1 0.003 a 28 0.0390 1 0.003 ab

 

 

*means followed by the same letter do not differ signiﬁcantly

Discussion
Female reproductive success in Arisaema triphyllum was inﬂuenced

predominantly by disease and pollen limitation. Proximal resources available to the plant
(e. g. light and moisture) had almost no inﬂuence on seed production. Uromyces ari-
triphylli infection had numerous adverse effects on ﬂowering and seed production.
Diseased plants ﬂowered for a shorter period and had smaller inﬂorescences than their
healthy counterparts. Females infected by U. ari-triphylli were less likely to set seed, and
produced lighter seeds. Uromyces ari-triphylli infections are not benign and would be
expected to signiﬁcantly reduce the ﬁtness of diseased plants. Thus, my work suggests

that the Arisaema triphyllum-Uromyces ari-triphylli interaction is typical of many natural

36

plant-pathogen systems where infections decrease host ﬁtness (Burdon 1987; J arosz &
Davelos 1995).

My work corroborates the results of Parker (1987) who found that diseased plants
produced only one-quarter of the number of seeds produced by healthy plants (means: 2.6
and 12.3 seeds, respectively). It was not clear from Parker’s work whether the reduction
in seed number was due to fewer diseased females setting seed or to a smaller number of
seeds being produced by all diseased females. My work indicates that the reduction in
seed number can be attributed to a much lower proportion of diseased females setting any
seed. Diseased females that set seed produced nearly the same number of seeds as their
healthy counterparts, but the average seed weight for diseased plants was signiﬁcantly
lower than for healthy females. This pattern of reduced seed weight suggests that disease
reduces a plant’s ability to provision seed.

Parker (1987) also reported that diseased plants senesced earlier than healthy
plants. As reported in the next chapter, I found a similar trend whereby the growing
season of infected plants was reduced by an average of 2 weeks. My work on ﬂowering
phenology indicates that rust infections also reduce signiﬁcantly the ﬂowering period of
both male and female plants. Parker’s work investigated a single A. triphyllum population
over one growing season, hence it was difﬁcult to interpret the generality of his ﬁndings.
My study incorporated two additional populations over two growing seasons; it indicates
that Parker’s observations are consistent across space and time.

Prior to my study, it was not known whether the lower proportion of diseased
females setting seed was due to reduced pollinator visitation or physiological effects of

disease that reduce a female’s ability to mature seed. Pathogen infection may alter

37

pollinator visitation in a number of ways. Infection could interfere with the production of
olfactory attractants used by A. triphyllum to lure pollinators (Vogel & Martens 2000).
Hence the attractiveness of diseased inﬂorescences may be reduced relative to the
inﬂorescences of healthy plants. Diseased plants ﬂower for a shorter length of time than
healthy plants, which reduces opportunities for pollinator visitation. However, the results
of my hand-pollination experiment suggest that the major effect of U. ari-triphylli
infection is to inhibit the maturation of seed and the reduced ability of infected females to
attract pollinators is of only minor importance to female reproductive success. A scant
eleven percent of diseased females set seed naturally, when supplemental pollen was
provided, the proportion of diseased females setting seed only increased to 14%, a full
44% below the proportion of healthy females setting seed naturally.

The hand-pollination experiment also demonstrated that female reproductive
success is indeed limited by the availability of pollen, which is similar to the ﬁndings of
earlier work by Bierzychudek (1981) and Rust (1980). Bierzychudek (1981) observed
that hand-pollinated females produced over an order of magnitude more seeds than
unmanipulated females (mean = 43.2, 1.0 respectively). My data do not display such
striking trends but I found that pollen loading had a signiﬁcant effect on seed production;
hand-pollinated females produced on average 42.5 seeds, while unmanipulated control
females produced on average 24.0 seeds. Rust (1980) found that A. triphyllum females
within one meter of a male plant produced signiﬁcantly more seeds than females more
distant from males (mean = 33.5, 7.8 respectively). He also observed that pollinator
movement was random within the one meter distance and that increased density of male

plants signiﬁcantly increased the percentage of females setting seed but not the number

38

of seeds produced; he proposed that increased male density may decrease pollen
movement (Rust 1980). I did not take density into consideration in this study nor did I
directly study pollinator movement, but Rust’s (1980) ﬁndings support the idea that
pollen limitation plays a role in limiting the female reproductive success of A. triphyllum.
Given the dynamics of this plant-pollinator interaction, chronic pollen limitation is to be
expected; the reward for insects successfully transferring pollen is death (Rust 1980).
Therefore, pollen transfer is a chance event that occurs only when A. triphyllum manages
to dupe insects into visiting multiple inﬂorescences.

Healthy females in the hand-pollination experiment exhibited signiﬁcant
differences in average seed weight; unmanipulated controls produced heavier seed than
hand-pollinated plants. My ﬁndings suggest a trade-off between the number of seed that
can be produced and the provisioning of individual seed. Control plants, which produced
fewer seeds, are likely able to allocate a larger proportion of resources into each seed,
resulting in a higher average seed weight. Hand-pollinated plants produced more seeds,
but had to allocate available resources among a larger number of progeny, which likely
resulted in a lower average seed weight. Seed weight was also inﬂuenced by population
and year, however, and this suggests that differences in the net amount of resources
available to individual plants resulting from variation across time and space could alter
the proportion of resources that individual plants allocate to seed production. Long-term
studies, possibly with a common garden experiment, are needed to address this issue
further.

I did not ﬁnd that localized environmental conditions had a signiﬁcant effect on

reproduction in A. triphyllum. The only signiﬁcant correlations were between seed set

39

frequency and soil pH and soil moisture. This contrasts with earlier reports that resource
limitations do have strong effects on reproductive success (Lovett Doust & Cavers 1982;
Lovett Doust et a1. 1986). The difference in ﬁndings may be due to the methods by
which the availability of resources was measured. My work concentrated on the effects
of the current physical environment (i.e., light, soil moisture, soil temperature and soil
pH) while earlier studies measured stored resources in the corm. The corm may
effectively ameliorate environmental ﬂuctuations by accumulating nutrients in storage
tissues during environmentally favorable years and releasing nutrients when conditions
are less favorable. Several studies have noted that corms and plants can shrink or
increase in size over time (Lovett Doust & Cavers 1982; Bierzychudek 1984; Lovett
Doust et al. 1986). The ability of the corm to buffer the inﬂuence of the current
environment may explain the limited number of environmental correlations detected in
the current study.

Female reproductive success in A. triphyllum is strongly inﬂuenced by pollen
limitation and disease, and may also be inﬂuenced by the microclimate. Hence it is
apparent that several factors must converge for a female to successfully set seed. This
study reinforces earlier work by Parker (1987) that U. ari-triphylli infections have a
strong adverse affect on female reproductive success in A. triphyllum, and on the
ﬂowering phenology of both male and female plants. My work demonstrates that disease
is just as important as pollen limitation in determining seed set in A. triphyllum. The
inﬂuence of disease has largely been ignored in examinations of potential limitations on
female reproductive success and the general inﬂuence of disease on female reproductive

success remains to be determined. The effect of pathogens is obvious when infections

40

result in sterilization of ﬂowers (Baker 1947; Alexander 1987; Althizer et a1. 1998) or
inhibition of ﬂowering (Clay 1986 & 1987). The potential for foliar pathogens to have
adverse effects on seed set has been demonstrated under greenhouse conditions by J arosz
et al. (1989), however few studies have demonstrated the effects of infection under ﬁeld
conditions (Roy & Bierzychudek 1993), and other work suggests that foliar pathogens
have little inﬂuence on seed set (J arosz & Burdon 1992). This work demonstrates that
the inﬂuence of pathogens is variable across systems, since U. ari-triphylli infections
adversely impact seed set in A. triphyllum under ﬁeld conditions. Disease incidence
within A. triphyllum populations must be taken into account when assessing the
population level consequences of reduced female reproductive success due to disease; in
populations with few diseased plants the adverse effects of U. ari-triphylli infections may
be minimal. As disease incidence increases, population growth may become increasingly

compromised.

41

Literature Cited

Alexander, H. M. 1987. Pollination limitation in a population of Silene alba infected by
the anther-smut fungus, Ustilago violacea. J. Ecol. 75:771-780.

Althizer, S. M., P. H. Thrall, and J. Antonovics. 1998. Vector behavior and the
transmission of anther smut infection in Silene alba. Amer. Mid. Nat. 139:147-163.

Andersson, S. 1996. Floral display and pollination success in Seneciojacobea
(Asteraceae): interactive effects of head and corymb size. Amer. J. Bot. 83:71-75.

Baker, H. G. 1947. Infection of species of Melandrium by Ustilago violacea (Pers.)
Fuckel and the transmission of the resultant disease. Ann. Bot. 11:333-348.

Bierzychudek, P. 1981. Pollinator limitation of plant reproductive effort. Am. Nat.117:
83 8-840.

Bierzychudek, P. 1984. Determinants of gender in J ack-in-the-pulpit: the inﬂuence of
plant size and reproductive history. Oecologia 65:14-18.

Burdon, J. J. 1987. Diseases and plant population biology. Cambridge University Press,
Cambridge. United Kingdom.

Campbell, D. R. 1987. Interpopulational variation in fruit production: the role of
pollination-limitation in the Olympic mountains. Amer. J. Bot. 74:269-273.

Clay, K. 1986. Induced vivipary in the sedge Cyperus virens and the transmission of the
fungus Balansia cyperi (Clavicipitaceae). Can. J. Bot. 64:2984-2988.

Clay, K. 1987. Effects of fungal endophytes on the seed set and seedling biology of
Lolium perenne and F estuca arundinaceae. Oecologia. 73:358-362.

Clay, K. 1990. Fungal endophytes of grasses. Ann. Rev. Ecol. Sys. 21:275-297.

Cloudsley-Thompson, J. L. and M. J. Chadwick. 1964. Life in deserts. Dufour,
Philadelphia. USA.

Freeman, S. & J. C. Herron. 2004. Evolutionary Analysis. Pearson/Prentice Hall, Upper
Saddle River. USA.

Herrera, C. H. 2000. Flower-to-seedling consequences of different pollination regimes in
an insect-pollinated shrub. Ecology 81 :15-29.

J arosz, A. M., J. J. Burdon, and W. J. Muller. 1989. Long term effects of disease
epidemics. J. Appl. Ecol. 26:725-733.

42

J arosz, A. M. & J. J. Burdon. 1992. Host-pathogen interactions in natural populations of
Linum marginale and Melampsora Iini. III. Inﬂuence of pathogen epidemics on host
survivorship and ﬂower production. Oecologia 89:53-61.

J arosz, A. M. & A. L. Davelos. 1995. Effects of disease in wild plant populations and the
evolution of pathogen aggressiveness. New Phytol. 129:371-3 87.

Lovett Doust, J. and P. B. Cavers. 1982. Sex and gender dynamics in jack-in-the-pulpit,
Arisaema triphyllum (Araceae). Ecology 63:797-808.

Lovett Doust, L., J. Lovett Doust, and K. Turi. 1986. F ecundity and size relationships in
jack-in-the-pulpit, Arisaema triphyllum (Araceae). Amer. J. Bot. 73:489-494.

Lubbers, A. E. and N. L. Christensen. 1986. Intraseasonal variation in seed production
among ﬂowers and plants of Thalictrum thalictroides (Ranunculaceae). Amer. J. Bot.
73:190-203.

Michigan State University. (1998). Campus Natural Areas. Lkd. Michigg State
University, at “Division of Campus Park & Planning.”
<http://www.cpp.msu.edu/nat_area> (2004, June 30).

 

Montalvo, A. M. and J. D. Ackerman. 1986. Relative pollinator effectiveness and
evolution of ﬂoral traits in Spathiphyllumfriedrichsthalli (Araceae). Amer. J. Bot.
73: 1665- 1 676.

Morgan, M. D. 1971. Life history and energy relationships of Hydrophyllum
appendiculatum. Ecol. Monogr. 41 :329-349.

Muller, R. N. 1978. The phenology, growth, and ecosystem dynamics of Erythronium
americanum in the northern hardwood forest. Ecol. Monogr. 48: 1-20.

Niesenbaum, R. A. 1993. Light or pollen-seasonal limitations on female reproductive
success in the understory shrub Lindera benzoin. J. Ecol. 81 :315-323.

Pady, S. M. 1939. Host invasion in systemic infections of Uromyces caladii. Mycologia
3 1 :590-605.

Parker, M. A. 1987. Pathogen impact on sexual vs. asexual reproductive success in
Arisaema triphyllum. Amer. J. Bot. 74: 1758-1763.

Policansky, D. 1981. Sex choice and the size advantage model in j ack-in the pulpit
(Arisaema triphyllum). Proc. Natl. Acad. Sci. 78:1306-1308.

Roy, B. A. 1993. Floral mimicry by a plant pathogen. Nature 362:56-58.

43

Roy, B. A. & P. Bierzychudek. 1993. The potential for rust infection to cause natural
selection in apomictic Arabis holboellii (Brassicaceae). Oecologia 95:533-541.

Rust, R. W. 1980. Pollen movement and reproduction in Arisaema triphyllum. Bull.
Torrey Bot. Club 107:539-542.

Salisbury, E. J. 1916a. The oak-hombeam woods of Hertfordshire. Part I.
J. Ecol. 4:83-100.

Salisbury, E. J. 1916b. The oak-hornbeam woods of Hertfordshire. Part II.
J. Ecol. 4:101-117.

SAS Institute Inc., SAS/STA® Software: Changes and Enhancements through
Release 6.12. Cary, NC: SAS Institute Inc., 1997. 1167pp. USA.

Schemske, D. W., M. F. Willson, M. N. Melampy, L .J. Miller, L. Verner, and K. M.
Schemske. 1978. Flowering ecology of some spring woodland herbs. Ecology
59:35 1-3 66.

Schemske, D. W. and C. C. Horvitz. 1984. Variation among ﬂoral visitors in pollination
ability: a precondition for mutualism specialization. Science 225:519-521.

Taylor, R. J. and R. W. Pearcy. 1976. Seasonal patterns of the CO; exchange
characteristics of understory plants from a deciduous forest. Can. J. Bot. 54: 1094-
1 103.

Vogel, S. and J. Martens. 2000. A survey of the function of the lethal kettle traps of
Arisaema (Araceae), with records of pollinating fungus gnats from Nepal. Bot. J.
Linn. Soc. 133:61-100.

Waser, N. M. 1983. Competition for pollination and ﬂoral character differences among
sympatric plant species: a review of evidence. Pp. 227-293 in C. E. Jones and R. J.
Little eds. Handbook of experimental pollination ecology. Van Reinhold, New York.
USA.

Waser, N. M., L. Chittka, M. V. Price, N. M. Williams, and J. Ollerton. 1996.
Generalization in pollination systems, and why it matters. Ecology 77:1043-1060.

Went, F. W. and M. Westergaard. 1949. Ecology of Desert Plants 111. Development of
plants in the Death Valley National Monument, CA. Ecology 30:26-38.

Young, H. J. 1988. Differential importance of beetle species pollinating Dieﬂenbachia
longispatha (Araceae). Ecology 69:832-844.

Young, H. J. 2002. Diurnal and nocturnal pollination of Silene alba (Caryophyllaceae).
Amer. J. Bot. 89:433-440.

44

Zimmerman, J. K. and T. M. Aide. 1989. Patterns of fruit production in a neotropical
orchid: pollinator vs. resource limitation. Amer. J. Bot. 76:67-73.

45

Chapter 3. Effects of disease and resource limitation on the growing season of
vegetative Arisaema triphyllum.
Introduction

The timing of events, e. g. emergence and senescence, over the course of the
growing season can be critical to survival and reproduction (Rathcke & Lacey 1985) and
plants experience trade-offs for early vs. late emergence. Previous studies indicate that
plants emerging early in the growing season have a longer period to attain resources for
growth and reproduction, which leads to larger, more competitive plants with higher
reproductive success than plants emerging late in the season (Baskin & Baskin 1972;
Marks & Prince 1981). The trade-off for early emergence is a higher mortality due to
less favorable environmental conditions (e.g., inadequate light levels, cold temperatures)
early in the growing season (Baskin & Baskin 1972; Marks & Prince 1981). Plants
emerging late in the growing season experience lower mortality rates and more favorable
environmental conditions but have less time to acquire resources (Baskin & Baskin 1972;
Marks & Prince 1981). Early emerging Leavenworthia stylosa had higher mortality than
individuals emerging late in the season because of below average temperatures, but early
plants produced more fruits and seed per plant due to the longer growing season (Baskin
& Baskin 1972). Similarly, Marks & Prince (1981) observed that Lactuca serriola
seedlings emerging in the winter months produced more seeds than individuals that
emerged in the spring or summer because of the duration of the rosette stage, however the
increased reproductive success of the winter seedlings was offset slightly by their higher

mortality rate.

46

The timing of leaf senescence at the end of the growing season is a strictly
controlled physiological process that can be induced by both abiotic (e.g., drought,
nutrient availability) and biotic (e.g., pathogens, herbivory) environmental stresses as
well as endogenous factors such as hormone levels and aging (Munne-Bosch & Alegre
2004). These two types of cues interact in the initiation and progression of leaf
senescence; environmental stresses can speed up senescence by inﬂuencing reproductive
development and hormone levels while endogenous factors can enhance the plant’s
ability to carry out leaf senescence under stress (Munne-Bosch & Alegre 2004). Our
study examined potential environmental stresses rather than endogenous factors. In order
to cope with environmental stresses, plants have evolved mechanisms that allow them to
initiate leaf senescence in order to reallocate nutrients to reproductive organs or eliminate
water consumption by older, less productive leaves; this ability to regulate leaf
senescence is an adaptation that enables plants to complete their life cycle in times of
stress (Gan & Amasino 1997). Although previous ﬁndings indicate that plants in many
ecosystems induce leaf senescence in response to abiotic environmental stresses such as
limited water and nutrients (especially Nitrogen) (de Castri 1981), biotic stresses such as
pathogens could also lead to premature leaf senescence.

Previous research has demonstrated that pathogens do alter host life history
(Alexander 1987; Clay 1986 & 1987; Parker 1987). For example, the anther smut
Mycobotrium violaceum alters the life history of Silene latifolia to maximize pathogen
dispersal by causing infected females to produce anther-like structures containing

teliospores instead of pistillate ﬂowers (Alexander 1987). Clay (1986 & 1987) found that

47

choke pathogens inhibit reproduction by replacing the inﬂorescence of infected plants
with fungal reproductive structures.

In Chapter two, I demonstrated that systemic rust infection by Uromyces ari-
triphylli adversely affected reproduction in Arisaema triphyllum by shortening the
ﬂowering period of infected plants, and reducing both seed set frequency and seed
weight. Parker’s (1987) study suggested that infections should also reduce the growing
period of plants by causing early senescence. My goal in this study was to determine the
extent to which disease reduces the growing season length by causing early senescence
and whether it also had an effect on plant emergence. I compared the emergence and
growing season length of healthy and diseased vegetative A. triphyllum over two growing
seasons. I also attempted to determine whether potential environmental stresses (i.e.
resource limitation) such as light, soil pH, soil moisture, and soil temperature had an
effect on emergence and growing season length.

Materials & Methods

Experiments were conducted in April-September of 2002 & 2003 at four
populations in central, lower peninsula Michigan. Three of the sites (Lott, Hudson, &
Elsesser woodlots) are part of Michigan State University’s natural areas holdings and are
located within 5 miles campus (http://www.cpp.msu.edu/nat_area/). The fourth site,
Webber, is located On private property approximately 20 miles northwest of campus and
permission was obtained from the land owners prior to initiation of the study. Lott &
Hudson populations were highly diseased because the incidence of U. ari-triphylli
infection was approximately 50%. The other two populations, Webber & Elsesser, had a

disease incidence of less than 10%. A series of 1m2 rectangular plots (0.25m X 4m) were

48

set up at various locations in each population to obtain an adequate sampling of
microclimates within each site. Since plants within a clump are likely to be clones, long
rectangular plots were used to ensure that multiple plant genotypes were monitored
within each plot. All plants emerging within each plot were monitored for date of
emergence, date of senescence, and disease status. In 2002 date of emergence was
determined using weekly leaf length measurements and when the measurements were
within 1-2 mm of the previous week’s measurement the leaf was considered fully
expanded. Growth in A. triphyllum is determinant (Bierzychudek 1984), and I expected
leaf expansion to stop within a few weeks of the start of the growing season. However,
leaves kept expanding well into the growing season. Therefore, emergence date criterion
in 2003 was changed to the date when the shoot ﬁrst emerged from the soil. Date of
senescence was determined by weekly monitoring of plants and the senescence criterion
was the total lack of chlorophyll in the leaf. Plants were considered to be diseased if any
of the multiple spore types of U. ari-triphylli were present. Approximately 175 plants
were sampled in each low disease population and 350 plants were sampled in each high
disease population. A larger sample size was used in the high disease populations so
approximately 175 healthy and 175 diseased plants were monitored at each site.

To determine whether environmental conditions inﬂuenced plant emergence and
growing season length, light, soil pH, soil moisture and soil temperature (2003 only)
measurements were taken in the middle and at the ends of each plot. Environmental
measurements were made at the beginning, middle, and end of the growing season
because of potential ﬂuctuations in these factors over the course of the growing season

e. g. overstory canopy closure. In 2002 light intensity was measured using an Extech

49

Model 401025 Foot Candle/ Lux Meter placed approximately one foot above the soil. All
light measurements were made on overcast days between 12-4 pm. In 2003 light was
measured as percent canopy cover using a spherical densiometer (Forestry Suppliers)
placed approximately one foot above the soil. Soil pH and soil moisture were measured
at a depth of three inches using a Kelway Soil Tester. In 2003, soil temperature was
measured at a depth of 2.5 inches using a Tele-tru thermometer.

PROC MIXED in SAS was used to conduct analyses of variance (AN OVAs) and
Type 111 tests to determine the signiﬁcance of the effects of population and disease in the
models for growing season length and emergence date (SAS Institute 1997). A
preliminary analysis to determine whether environmental variables were highly
correlated with each other was carried out using PROC CORR. Environmental variables
early in the season were not highly correlated (i.e., > 0.85) with each other (Appendix 3),
so I decided to exclude interaction terms from the emergence model. Only early
environmental measures were used to evaluate correlations with plant emergence, since
no plant emerged after the mid season measures were taken. With regard to growing
season length, measurements made early, mid and late season were also not highly
correlated; the average correlation between measures was 0.210; the highest correlation
was 0.79 between soil temperature (early) and soil temperature (mid) (Appendix C).
Therefore, I decided to leave all early, mid and late season measurements in the growing
season length model but exclude the interaction terms. PROC GENMOD in SAS was
used to conduct logistic regressions and Type 111 tests were utilized to determine the
effects of environmental variables (light, soil pH, soil moisture, soil temperature) on

emergence and growing season length.

50

Least square means were used because of unbalanced sample sizes. Since
emergence criteria differed between 2002 and 2003, emergence and growing season data
for each year were analyzed separately.

Results
Plant emergence

Plant emergence was inﬂuenced by disease status and population site in both 2002
and 2003 (Table 3.1). Somewhat surprisingly, average emergence date for diseased
plants was signiﬁcantly earlier than healthy plants in both years (Table 3.2). The
inﬂuence of population was more variable across years (Table 3.3). Plants at Elsesser
were among the latest to emerge in 2002 and 2003, while plants at Hudson were in the
earliest group both years. The relative emergence of plants at Webber and Lott changed
between years. Plants at Webber had the second earliest emergence in 2002 and the
earliest emergence in 2003. The Lott population had the most signiﬁcant change across
years. Plants at this site were the latest to emerge in 2002, but grouped with earlier
emergence in 2003.

Table 3.1. Analysis of variance of Julian emergence dates for the effects of population &
disease for vegetative A. triphyllum at all four populations in a)2002 & b)2003.

a) Analysis of variance of Julian emergence dates in 2002.

 

 

 

 

 

 

 

 

 

ain Effect df Mean Square F Value Pr > F
Population 3 3638 37.59 < 0.0001
Disease status 1 5633 58.20 < 0.0001
Pop*Disease 3 134 1.39 0.1291
Error 1 188 96.79
b) Analysis of variance of Julian emergence dates in 2003.

Main Effect df Mean Square F Value Pr > F
opulation 3 756.89 9.69 < 0.0001

Disease status 1 564.73 7.23 0.0073

Pop*Disease 3 101.54 1.30 0.2718
rror 1048 78.1 1

 

 

 

51

Table 3.2. Least square mean Julian emergence dates for healthy and diseased A.

triphyllum in 2002 and 2003.

 

 

 

 

Year Population Julian emergence date*
2002

Healthy 153.6510.36

Diseased 148051064
2003

Healthy 124.9510.33

Diseased 122.9410.67

 

 

 

*means and standard errors. Healthy and diseased values within a year are signiﬁcantly
different.

Table 3.3. Least square mean Julian emergence dates for A. triphyllum in Hudson,

Webber, Elsesser, and Lott populations in 2002 and 2003.

 

 

 

 

Year Population Julian emergence date“
2002
Hudson 146.1710.56 a
Webber 150.2410.95 b
Elsesser 152791080 0
Lott 154.1910.55 c
2003
Webber 121901085 (1
Hudson 122.741063 (1
Lott 123311057 (1
Elsesser 127.841088 e

 

 

 

 

 

*means and standard errors. Values followed by the same letter do not differ signiﬁcantly

Despite differences in environmental measures within and among populations,
none of the environmental factors (light availability, soil moisture, soil pH or soil
temperature) early in the season were found to have a signiﬁcant effect on plant
emergence in 2002 (Table 3.4). However in 2003 percent cover and soil pH early in the
season had signiﬁcant effects on emergence (Table 3.5). The amount of cover early in
the year inﬂuenced plant emergence with individuals in more open microhabitats
emerging later than individuals in closed areas. Soil pH also affected plant emergence;

plants in more alkaline soils tended to emerge earlier than individuals in more acidic

52

soils. Mean values for the environmental variables in 2002 & 2003 can be found in

Appendix D.

Table 3.4. Analysis of variance for the effects of population, disease, and environmental
variables early in the season on plant emergence in 2002.

 

 

 

lMain Effect df Mean Square F Value Pr > F
Population 3 756.89 18.42 < 0.0001
Disease status 1 564.73 57.67 < 0.0001
POp*Disease 3 101.54 1.73 0.1599
Light intensity 1 15.47 0.16 0.6887
Soil pH 1 312.28 3.23 0.0726
Soil moisture 1 125.68 1.30 0.2546
Error 1185 96.68

 

 

 

Table 3.5. Analysis of variance for the effects of population, disease, and environmental
variables early in the season on plant emergence in 2003.

 

 

 

[Main Effect df Mean Square F Value Pr > F
opulation 3 629.76 8.23 < 0.0001
isease status 1 574.66 7.51 0.0063
op*Disease 3 107.13 1.40 0.2402
ight intensity 1 615.98 8.05 0.0046

Soil pH 1 471.36 6.16 0.0133

Soil moisture 1 101.01 1.32 0.2508

Soil temp. 1 30.61 0.40 0.5262

Error 1042 76.52

 

 

 

Growing season length

In 2002, growing season length was inﬂuenced by an interaction between disease
status & population (Table 3.6). Healthy plants always had longer growing seasons than
their diseased counterparts (Table 3.7). On average, the growing season for healthy
plants was more than two weeks longer than the growing season of diseased plants
(mean: healthy = 56.63 days; diseased = 41.28 days). The degree to which disease
affected growing season length varied among populations. At Elsesser, Hudson, and Lott

the growing season for diseased plants was approximately three-fourths that of healthy

53

plants. The effect of disease was more severe at Webber where the growing season of
diseased plants was only 63% of the growing season for healthy plants. The effect of
population was also evident in the healthy plants. At Elsesser, healthy plants had the
longest growing season, which was more than three weeks longer than the shortest
growing season at Lott. For diseased plants, the range across populations for the diseased
plants was 16.75 days. Growing season lengths were similar for diseased plants at the
Hudson and Lott populations. In contrast, the growing season for healthy plants at
Hudson was signiﬁcantly longer than that for healthy plants at Lott.

Table 3.6. Analysis of variance of growing season length for the effects of population and
disease on healthy & diseased vegetative A. triphyllum at all four populations in

 

 

 

2002.
Main Effect df Mean Square F Value Pr > F
Population 3 141173 30.46 < 0.0001
Disease status 1 50958 109.51 < 0.0001
Pop*Disease 3 1316 2.83 0.0375
Error 1188 465.33

 

 

 

Table 3.7. Least square mean growing season length for healthy & diseased vegetative
A. triphyllum at all four populations in 2002.

 

 

 

 

 

 

 

Population Growing season length (days)
Number Healthy1 Number Diseased81 Proportion
Diseased/Healthy
Elsesser 107 71 .4012.08 59 54.081281 0.757
Webber 203 641611.51 31 401913.87 0.626
Hudson 327 528511.19 103 386612.02 0.732
Lott 252 49.201136 114 373312.12 0.759

 

 

 

 

 

 

 

ameans and standard errors. Healthy and diseased values within a population are
signiﬁcantly different.

In 2003, both disease status and population had a signiﬁcant effect on growing
season length but there was no interaction (Table 3.8). The average growing season for

plants, regardless of disease status, differed among the populations (Table 3.9). As in

54

2002, plants at Elsesser had the longest growing season, which was three weeks longer
than the shortest growing season at Webber (Table 3.9). The relationship between
Hudson and Lott populations with respect to growing season length did not remain
constant between years (Tables 3.7 & 3.9). In 2002, plants at Hudson had a longer
growing season than plants at Lott but in 2003 the opposite was true. Diseased plants had
a mean growing season that was signiﬁcantly shorter (15 days) than their healthy
counterparts (Table 3.10).

Table 3.8. Analysis of variance of growing season length for the effects of population and
disease on healthy & diseased vegetative A. triphyllum at all four populations in

 

 

 

2003.
'Main Effect df Mean Square F Value Pr > F
Population 3 896.48 20.99 < 0.0001
Disease status 1 30487 71.28 < 0.0001
Pgﬂ‘Disease 3 85.54 0.20 0.8955
rror 1048 427.71

 

 

 

Table 3.9. Least square mean growing season length and standard errors for vegetative
A. triphyllum at all four populations in 2003.

 

 

 

 

 

 

Population Number Growing season length (days)
Elsesser 135 93.131207 a

Lott 310 79.101134 b

Hudson 401 760011.47 b c

Webber 210 720611.98 c

 

 

 

 

*means followed by the same letter do not differ signiﬁcantly

Table 3.10. Least square mean growing season length and standard errors for healthy &
diseased vegetative A. triphyllum in 2003.

 

 

 

 

Disease Status Number Growing season length (days)
Healthy 853 87 .4410.78
Diseased 203 727111.56

 

 

 

 

Soil pH was the only environmental variable to have a consistent, signiﬁcant

effect on growing period length in both years (Tables 3.11 & 3.12); plants growing in

55

more alkaline soils had a longer growing season than plants growing in more acidic soils.
However, the portion of the growing season in which pH exerted a signiﬁcant effect
varied between years; in 2002 the effect was more pronounced in the beginning and
middle of the growing season (Table 3.11) but in 2003 the effect was predominantly
found in the beginning and end of the growing season (Table 3.12). In 2003, soil
moisture levels in the middle and end of the growing season also had a signiﬁcant effect
on growing period length; plants growing in wetter soils had a longer growing period
length than plants growing in drier soils (Table 3.12).

Table 3.11. Analysis of variance of environmental data at all four populations for
early(1), middle (2), and end(3) of the growing season in 2002.

 

 

 

lMain Effect df Mean Square F Value Pr > F
Population 3 7036.23 15.26 < 0.0001
Disease status 1 50083.6 108.62 < 0.0001
Pop*Disease 3 1152.72 2.50 0.0584
Light 1 1 0.46 0.001 0.9512
Soil pH 1 1 3591.89 7.79 0.0054
Soil moisture 1 1 239.77 0.52 0.4730
Light 2 1 442.65 0.96 0.3282
Soil pH 2 1 1895.08 4.11 0.0427
Soil moisture 2 1 428.81 0.93 0.3340
Light 3 1 1240.33 2.69 0.1014
Soil pH 3 1 1166.56 2.53 0.1119
Soil moisture 3 1 534.86 1.16 0.2807
rror 1168 461.09

 

 

 

56

Table 3.12. Analysis of variance of environmental data at all four populations for
early(1), middle (2), and end(3) of the growing season in 2003.

 

 

 

 

 

 

ain Effect df Mean Square F Value Pr > F
Population 3 2269.89 5.57 0.0009
Disease status 1 31786.56 78.00 < 0.0001
Pop*Disease 3 232.29 0.57 0.6327
% cover 1 1 268.96 0.66 0.4166
Soil pH 1 1 6422.51 15.76 < 0.0001
Soil moisture l 1 12.22 0.03 0.8728
Soil temp. 1 1 497.17 1.22 0.2705
% cover 2 1 419.74 1.03 0.3094
Soil pH 2 l 16.30 0.04 0.8448
Soil moisture 2 1 2436.97 5.98 0.0146
Soil temp. 2 l 73.35 0.18 0.6695
'% cover 3 1 20.38 0.05 0.8266
Soil pH 3 1 2436.97 5.98 0.0147
Soil moisture 3 1 3068.63 7.53 0.0062
Soil temp. 3 1 175.23 0.43 0.5111
Error 1034 407.52
Discussion

Infections by the systemic fungal pathogen, Uromyces ari-triphylli, altered the life
history of vegetative Arisaema triphyllum by shortening the growing season and
promoting early emergence. My ﬁndings support those of Parker (1987) who noted that
ﬂowering A. triphyllum senesced earlier than healthy ﬂowering individuals. Early
emergence does ameliorate the effects of early senescence to some degree, but the
growing season of the average diseased plant remained only three-quarters that of
comparable healthy plants. The combined effects of a shorter growing season and the
nutritional demands of the pathogen probably contribute to the temporal trend where
diseased A. triphyllum decrease in size across years (EE Mason, personal

communication). Overall, resource limitation, as determined by the proximate

57

environment, does not appear to have a strong inﬂuence on emergence and growing
season length of A. triphyllum.

My results present the intriguing possibility that emergence is altered in diseased
A. triphyllum for the beneﬁt of the pathogen. Early emergence by diseased plants may
increase the probability of contagious spread for U. ari-triphylli. Lesions on infected
plants do not erupt and disperse spores until the leaf begins to unfurl. This means that
inoculum for infectious spread of the pathogen is not available until two or more weeks
after the infected individual emerges from the ground. If the pathogen can shift
emergence so that infected plants emerge earlier than healthy individuals, then spread to
succulent new material on healthy plants may be maximized.

Alteration of host life history by pathogens is known for several host-parasite
systems. Life history of Silene latifolia infected by the anther smut, Mycobotrium
violaceum, is altered in a manner that promotes pathogen spread. Infected females
produce anther-like structures containing pathogen teliospores instead of pistillate
ﬂowers. Similarly, the anthers of males produce pathogen teliospores instead of pollen.
Infection also inﬂuences ﬂowering behavior. Healthy males produce more ﬂowers than
healthy females, but female ﬂowers remain on the plants for a longer period of time
(Alexander 1987). Infected plants produce the male number of ﬂowers and ﬂowers are
retained on the plant nearly as long as healthy female (Alexander 1987). Thus, the
pathogen alters the plant to maximize its dispersal. Choke pathogens get their name
because they inhibit reproduction of the plant. The inﬂorescence of infected plants is
replaced by reproductive structures of the pathogen which promote pathogen spread

during the host’s ﬂowering period (Clay 1986 & 1987). Thus, it seems possible that the

58

early emergence exhibited by infected A. triphyllum is an adaptation by the pathogen to
maximize its infectious spread within a population.

My ﬁndings relating to the effect of disease on the growing period corroborate
earlier work by Parker (1987); he observed that diseased A. triphyllum have reduced leaf
longevity, which likely contributed to the shorter growing period length of diseased
plants in my study. A shortened growing season could have pronounced effects on
resource acquisition in A. triphyllum; individuals have only a brief period of time to
utilize available resources for survival and to accumulate photosynthetic reserves in the
corm. Accumulation of these corm resources is essential for vegetative A. triphyllum to
make the transition to ﬂowering as a male or female; previous studies have demonstrated
that sexual expression A. triphyllum is correlated with overall plant size which is directly
related to the amount of corm resources (Heslop-Harrison 1957; Policansky 1981;
Bierzychudek 1982 & 1984; Lovett Doust & Cavers 1982; Lovett Doust et al. 1986;
Parker 1987). The disease probably acts as a drain on a plant’s energy and nutrient
resources, which could further reduce the resources available to produce cormlets or an
inﬂorescence. In consequence, disease could alter population composition by altering the
proportion of vegetative, male, and female plants. Ultimately, disease may decrease the
persistence of A. triphyllum populations in the understory, which could affect the
structure of the forest community.

Environmental variables (light, soil pH, soil moisture, and soil temperature) had
no consistent effect on emergence or growing period length suggesting that proximate
resources are not important inﬂuences on these aspects of the life history. Past studies

have reported that the amount of stored resources in the corm of A. triphyllum can affect

59

the timing of transition from vegetative to ﬂowering as well as female reproductive
success (Lovett Doust & Cavers 1982; Bierzychudek 1984; Lovett Doust et al. 1986).
These stored resources may compensate for reduced nutrient accumulation when
environmental conditions are less favorable. Essentially, the corm may function to
“average” the effects of the proximate environment and buffer a plant’s response to
ﬂuctuations in light incidence, soil pH, soil moisture, and soil temperature from one
growing season to the next. Soil pH was the only environmental variable that consistently
inﬂuenced plant emergence and growing period length. Previous ﬁndings by Giesler et
al. (1998) indicate that there is a strong relationship between soil characteristics (e. g. soil
pH & nitrogen content) and plant productivity. Although my study did not measure
productivity, it is possible that the differences in emergence and growing period length of
plants growing in different soil pH conditions could indirectly inﬂuence productivity. I
observed that plants growing in more alkaline soils emerged earlier and had a longer
growing period than plants growing in acidic soil, which could be associated with an
increase in productivity. This relationship between early emergence/longer growing
period and productivity is purely speculative, however, and ﬁrrther investigation is
needed.

Although light and percent cover did not exert a signiﬁcant inﬂuence on grth
period, I speculate that the signiﬁcant effect of population could be attributed to
differences in light levels among the four sites. I observed that plants at Elsesser
population had the longest growth period in 2003, which could relate to the fact it
consistently had a lower percent canopy cover than any of the other four populations over

the growing season. Similarly, plants at Webber population had the shortest growth

60

period in 2003, which might be due to the fact it consistently had a higher percent canopy
cover than any of the other four populations over the growing season. These statements

are highly speculative, however, and additional research is needed.

61

Literature Cited

Alexander, H. M. 1987. Pollination limitation in a population of Silene alba infected by
the anther-smut fungus, Ustilago violacea. J. Ecol. 75:771-780.

Baskin, J. M. and C. C. Baskin. 1972. Inﬂuence of germination date on survival and seed
production in a natural population of Leavenworthia stylosa. Am. Midl. Nat. 88:318-
332.

Bierzychudek, P. 1982. The demography of Jack-in-the-pulpit, a forest perennial that
changes sex. Ecol. Monogr. 52:335-351.

Bierzychudek, P. 1984. Determinants of gender in J ack-in-the-pulpit: the inﬂuence of
plant size and reproductive history. Oecologia 65:14-18.

Clay, K. 1986. Induced vivipary in the sedge Cyperus virens and the transmission of the
fungus Balansia cyperi (Clavicipitaceae). Can. J. Bot. 64:2984-2988.

Clay, K. 1987. Effects of fungal endophytes on the seed set and seedling biology of
Lolium perenne and F estuca arundinaceae. Oecologia 73:358-362.

di Castri, F. 1981. Mediterranean shrublands of the world. Pp. 1-52 in F. di Castri, D.W.
Goodall, and R. L. Sprecht eds. Mediterranean-type shrublands. Elsevier Scientiﬁc
Publishing, Amsterdam. The Netherlands.

Gan, S. and R. M. Amasino. 1997. Making sense of senescence. Plant Physiol.
113:313-319.

Giesler, R., M. Hogberg, and P. Hogberg. 1998. Soil chemistry and plants in a
Fennoscandian boreal forest as exempliﬁed by a local gradient. Ecology 79:119-137.

Heslop-Harrison, J. 1957 . The experimental modiﬁcation of sex expression in ﬂowering
plants. Biol. Rev. 32:38-90.

Lovett Doust, J. and P. B. Cavers. 1982. Sex and gender dynamics in jack-in-the-pulpit,
Arisaema triphyllum (Araceae). Ecology 63:797-808.

Lovett Doust, L., J. Lovett Doust, and K. Turi. 1986. F ecundity and size relationships in
j ack-in-the-pulpit, Arisaema triphyllum (Araceae). Amer. J. Bot. 73:489-494.

Marks, M. and S. Prince. 1981. Inﬂuence of germination date on survival and fecundity
in wild lettuce Latuca serriola. Oikos 36: 1 85-190.

Munne-Bosch, S. and L. Alegre. 2004. Die and let live: leaf senescence contributes to
plant survival under drought stress. F unct. Plant Biol. 31:203-216.

62

Parker, M. A. 1987. Pathogen impact on sexual vs. asexual reproductive success in
Arisaema triphyllum. Amer. J. Bot. 74: 1758-1763.

Policansky, D. 1981. Sex choice and the size advantage model in jack-in the pulpit
(Arisaema triphyllum). Proc. Natl. Acad. Sci. 78:1306-1308.

Rathcke, B. and E. P. Lacey. 1985. Phenological patterns of terrestrial plants.
Ann. Rev. Ecol. Syst. 16:179-214.

SAS Institute Inc., SAS/STA® Software: Changes and Enhancements through
Release 6.12. Cary, NC: SAS Institute Inc., 1997. 1167pp.

63

Chapter 4. Conclusions

The results of my study indicate that the Arisaema triphyllum-Uromyces ari-
triphylli interaction is typical of many plant-pathogen interactions; infection reduces the
reproductive success of the host (Alexander 1987; Burdon 1987; Parker 1987; J arosz et
al. 1989; J arosz & Burdon 1992; Roy 1993; J arosz & Davelos 1995). Disease sabotages
the efforts of A. triphyllum to maximize female reproduction on two different levels: 1)
limiting the time an inﬂorescence is receptive to pollen, and 2) limiting the ﬂow of
resources to pollinated ovules and developing fruits. My ﬁndings demonstrate the
importance of pathogens in shaping plant communities by reducing individual ﬁtness,
which also has implications at the population level (e. g. reduced competitive ability).

Pollen limitation was also an important limitation on female reproductive success
in A. triphyllum; my results corroborate the ﬁndings of Rust (1980) and Bierzychudek
(1981). For healthy females, hand-pollinated plants were more likely to set seed than
umnanipulated controls and produced more seeds, but the seeds were smaller than those
of the controls. This suggests a trade-off between seed number and the amount of
resources allocated to each seed, and could have implications for seedling survival in A.
triphyllum, but additional studies are needed.

In terms of resource limitation, the proximate environment does not appear to be
as important as disease or the previous environment (as indicated by stored resources in
the corm) in inﬂuencing the female reproductive success or growing season of A.
triphyllum. Past studies have demonstrated that the amount of stored resources affects
female reproductive success but these investigations utilized different methods to assess

resource limitation, e.g. measuring corm size to determine the amount of stored resources

64

(Policansky 1981; Lovett Doust et al. 1986). The inability to correlate proximate
environmental variables with ﬂowering and seed set seems to indicate that the corm
“averages” the environment, and buffers the plant’s response to temporal variability in
factors such as light incidence, soil moisture, soil temperature and soil pH. I would have
liked to test the effects of resource limitation by directly measming the amount of stored
resources as indicated by corm diameter, i.e. replicating the approach of Lovett Doust et
al. (1986), but there were practical and technical constraints. Additionally, if distinct
environmental effects had been observed in the ﬁrst year, I would have initiated more
manipulative measures in the second growing season such as a shading experiment or
artiﬁcial creation of light gaps. My study did suggest that soil pH had small but
signiﬁcant effects on seed set, emergence and growing period. The mechanism by which
soil pH affects these important life history traits of A. triphyllum is unknown, but would
be worth pursuing.

Our ﬁndings with respect to the effects of disease on seed set frequency
corroborate Parker’s (1987) results. This study however has not only expanded the
assessment of the effects of disease by studying multiple populations over two growing
seasons, but also enabled the in-depth examination of the numerous other ways that
disease affects female reproductive success, e. g. ﬂowering phenology. The numerous
effects of disease on the life history of A. triphyllum all point to the reduced ﬁtness of
diseased individuals compared to their healthy counterparts. If a female plant becomes
diseased and is unable to produce seed, reproductive success and genetic contribution to
the next generation might be maintained by ﬂowering as a male in subsequent years; we

observed however that healthy plants had a higher probability of being male than

65

diseased plants. If disease did favor ﬂowering as a male, the structure of A. triphyllum
populations could shift to a more male biased ratio. For infected females able to maintain
seed production, their progeny would be free of disease but have fewer stored resources
both for establishment and sustenance, i.e. the length of time before the seedling would
need to rely on photosynthesis for energy would be shortened.

My work also suggests that infection alters the pattern of emergence in A.
triphyllum. Diseased plants emerge earlier in the season than healthy plants. This pattern
may increase the rate of infectious spread of the pathogen by making spores of the
pathogen available as the healthy plants begin their growing period. Further studies are
needed to conﬁrm this hypothesis. Disease status of individual plants could be
manipulated with inoculation and disease curing experiments that would determine
deﬁnitively if infection alters plant emergence. The clonal nature of A. triphyllum
reproduction could be utilized in this work to examine the change in emergence as clones
are infected.

For further study I would like to examine how the multitude of the above effects
inﬂuence population structure and persistence in A. triphyllum. I do not think that this
disease necessarily poses a threat to the existence of this species for several reasons.
First, there are likely populations within the geographical range of A. triphyllum where
the pathogen is absent. Second, the existence of resistant A. triphyllum genotypes within
infected populations is a distinct (but uninvestigated) possibility. Finally, variation in
disease severity among infected populations suggests that the effects of the pathogen on
the life history of A. triphyllum, e. g. female reproductive success & competitive ability,

pose more of a threat in highly diseased populations than in low disease populations.

66

Literature Cited

Alexander, H. M. 1987. Pollination limitation in a population of Silene alba infected by
the anther-smut fungus, Ustilago violacea. J. Ecol. 75:771-780.

Bierzychudek, P. 1981. Pollinator limitation of plant reproductive effort.
Am. Nat.117: 838-840.

Burdon, J. J. 1987. Diseases and plant population biology. Cambridge University Press,
Cambridge. United Kingdom.

J arosz, A. M., J. J. Burdon, and W. J. Muller. 1989. Long term effects of disease
epidemics. J. Appl. Ecol. 26:725-733.

J arosz, A. M. & J. J. Burdon. 1992. Host-pathogen interactions in natural populations of
Linum marginale and Melampsora lini. III. Inﬂuence of pathogen epidemics on host

survivorship and ﬂower production. Oecologia 89:53-61.

J arosz, A. M. & A. L. Davelos. 1995. Effects of disease in wild plant populations and the
evolution of pathogen aggressiveness. New Phytol. 129:371-387.

Lovett Doust, L., J. Lovett Doust, and K. Turi. 1986. F ecundity and size relationships
in j ack-in-the-pulpit, Arisaema triphyllum (Araceae). Amer. J. Bot. 732489-494.

Parker, M. A. 1987. Pathogen impact on sexual vs. asexual reproductive success in
Arisaema triphyllum. Amer. J. Bot. 74: 1758-1763.

Policansky, D. 1981. Sex choice and the size advantage model in jack-in the pulpit
(Arisaema triphyllum). Proc. Natl. Acad. Sci. 78:1306-1308.

Roy, B. A. 1993. Floral mimicry by a plant pathogen. Nature 362:56-58.

Rust, R. W. 1980. Pollen movement and reproduction in Arisaema triphyllum. Bull.
Torrey Bot. Club 107:539-542.

67

Appendices

Appendix A. Analysis of variance of environmental data at Hudson and Lott

populations for the a) 2002 and b) 2003 ﬂowering seasons.

a) Analysis of variance of environmental data at Hudson and Lott populations for

the 2002 ﬂowering season.

 

 

 

 

Main Effect I (if I Mean Square I F Value I Pr > F
Population 1 87.63 3.16 0.0785
Disease 1 1005.49 36.26 <0.001
Light 1 79.58 2.87 0.0936
Soil pH 1 19.41 0.70 0.4065
Light*pH l 80.42 2.90 0.0918
Soil moisture 1 16.36 0.59 0.4440
Light*moisture 1 71.82 2.59 0.1108
pH*moisture 1 17.75 0.64 0.4261
Light*pH*moisture 1 72.10 2.60 0.1 100
rror ] 99 l 27.73 1 |

 

68

 

b) Analysis of variance of environmental data at Hudson and Lott populations for

the early(1), middle (2), and end(3) of the ﬂowering season in 2003.

 

 

 

  
   
   
 
   

 

Main Effect I df I Mean Square F Value Pr > F
Population 1 61.37 3 .33 0.0722
Sex 1 101.97 5.56 0.0211
Disease 1 163.59 8.92 0.0039
% Cover 1 1 0.18 0.01 0.9078
Soil pH 1 l 3.85 0.21 0.6484
Soil moisture 1 l 19.99 1.09 0.3011
Soil temp. 1 1 2.93 0.16 0.6879
K’/o Cover 1*pH 1 1 1.65 0.09 0.7593
% Cover l*moist.l 1 48.60 2.65 0.1081
% Cover 1*temp.1 1 0.18 0.01 0.9166
pH l*moist.l 1 18.89 1.03 0.3136
pH 1*temp.l 1 1.47 0.08 0.7730
moist.1*temp.1 1 26.96 1.47 0.2298
/o Cover 2 1 2.57 0.14 0.7077
Soil pH 2 1 0.73 0.04 0.8444
Soil moisture 2 1 33.56 1.83 0.1810
Soil temp. 2 1 36.13 1.97 0.1643
/o Cover 2*pH 2 1 0.18 0.01 0.9177
/o Cover 2*moist.2 1 0.18 0.11 0.7431
/o Cover 2*temp 2 1 2.02 2.06 0.1559
H 2*moist.2 1 50.80 2.77 0.1005
H 2*temp 2 1 34.48 1.88 0.1750
oist.2*temp.2 1 41.99 2.29 0.1345
/o Cover 3 1 3.48 0.19 0.6608
Soil pH 3 l 11.00 0.60 0.4411
Soil moisture 3 1 4.03 0.22 0.6398
011 temp. 3 1 19.62 1.07 0.3039
A: Cover 3*pH 3 1 4.95 0.27 0.6082
/o Cover 3*moist.3 l 0.00 0.00 0.9593
/o Cover 3*temp.3 1 16.14 0.88 0.3512
pH 3*moist.3 1 0.00 0.00 0.9564
pH 3*temp.3 1 15.22 0.83 0.3645
moist.3*temp.3 1 1.83 0.10 0.7562
rror | 71 l 18.34

 

 

69

 

Appendix B. Mean values of light availability, soil pH, soil moisture, and soil
temperature for Hudson (HU) and Lott (LS) at the beginning, middle, and
end of the 2002 and 2003 ﬂowering seasons.

 

 

 

 

 

 

 

 

 

 

 

 

Hudson Lott
2002
Early season measures:
Light (Lux) 815.2 309.2
Soil pH 6.6 6.9
Soil Moisture (%) 67.6 64.5
Mid season measures:
Light (Lux) 297.8 429.3
Soil pH 6.0 6.6
Soil Moisture (%) 72.0 78.3
Late season measures:
Light (Lux) 1160.3 1370.5
Soil pH 6.8 7.0
Soil Moisture (%) 61.6 62.0
2003
Early season measures:
% Cover 12.0 12.1
Soil pH 6.8 6.9
Soil Moisture (%) 66.4 65.2
Soil Temperature (°C) 6.9 9.4
Mid season measures:
% Cover 33.6 40.1
Soil pH 6.5 6.9
Soil Moisture (%) 71.1 69.5
Soil Temperature (°C) 9.2 12.0
Late season measures:
% Cover 79.5 74.0
Soil pH 6.5 6.8
Soil Moisture (%) 72.5 74.1
Soil Temperature (°C) 14.0 15.0

 

70

 

Appendix C. Correlations among environmental variables in 2002 (above diagonal) and
2003 (below diagonal).

Light early early early ' mid mid Light late late

0.232 0.376 0.065 0.197 0.376
t

0.427 0.340 '
t i

0.328
t

0.053

0.065 "

0.362
c

0.018

0.387 0.223
I

0.049 0.097 ‘

0.044 0.411
.

 

signiﬁcance for Prob > |r| under HO: Rho=0 at 0.05

71

Appendix D. Mean values of light intensity, soil pH, soil moisture, and soil temperature
(2003 only) for all four populations at the beginning, middle, and end of the
a) 2002 and b) 2003 growing season.

a) mean environmental values in 2002.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Hudson Elsesser Webber Lott
2002
Early season measures:
Light (Lux) 904.7 532.1 446.0 456.2
Soil pH 6.5 6.5 7 7
Soil Moisture (%) 65.0 63.8 46.2 59.5
Mid season measures:
Light (Lux) 272.0 137.1 106.2 342.7
Soil pH 6.0 6.1 6.3 6.5
Soil Moisture (%) 72.3 77.1 73.1 81.0
Late season measures:
Light (Lux) 977.2 844.3 1048.1 1325.8
Soil pH 6.8 6.3 6.2 6.5
Soil Moisture (%) 69.9 68.1 80.5 81.3
b) mean environmental values in 2003.
Hudson Elsesser Webber Lott
2003
Early season measures:
% Cover 12.0 10.4 22.4 16.4
Soil pH 6.7 6.5 6.6 6.7
Soil Moisture (%) 68.6 54.4 62.5 62.5
Soil Temperature (°C) 8.9 6.7 7.8 11.5
Mid season measures:
% Cover 81.6 75.9 80.2 81.8
Soil pH 6.6 6.5 6.6 6.8
Soil Moisture (%) 75.3 67.7 73.3 58.9
Soil Temperature (°C) 13.7 12.3 12.9 14.8
Late season measures:
% Cover 82.4 78.1 88.7 84.3
Soil pH 6.9 6.5 6.7 6.7
Soil Moisture (%) 57.9 70.3 63.5 73.7
Soil Temperature (°C) 21.0 18.8 17.7 16.9

 

 

 

 

 

72

 

 

   

llIllIIIILlllIIIIlIlIlllIlIllell