Citation
Factors affecting dieback in the rare plant Hypericum edisonianum (Edison's St. John's-Wort)

Material Information

Title:
Factors affecting dieback in the rare plant Hypericum edisonianum (Edison's St. John's-Wort)
Creator:
Van de Kerckhove, G. A
Publication Date:
Language:
English
Physical Description:
viii, 124 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Diseases ( jstor )
Fungi ( jstor )
Infections ( jstor )
Mortality ( jstor )
Mycology ( jstor )
Pathogens ( jstor )
Ponds ( jstor )
Seasonal wetlands ( jstor )
Species ( jstor )
Trees ( jstor )
Dissertations, Academic -- Plant Pathology -- UF ( lcsh )
Plant Pathology thesis, Ph.D ( lcsh )
City of Gainesville ( local )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 2002.
Bibliography:
Includes bibliographical references (leaves 107-123).
General Note:
Printout.
General Note:
Vita.
Statement of Responsibility:
by G. A. Van de Kerckhove.

Record Information

Source Institution:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
029237252 ( ALEPH )
51020705 ( OCLC )

Downloads

This item has the following downloads:


Full Text










FACTORS AFFECTING DIEBACK IN THE RARE PLANT Hypericum edisonianum (EDISON'S ST. JOHN'S-WORT)
















BY

G.A. VAN DE KERCKHOVE


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULLFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2002













ACKNOWLEDGMENTS

Dr. H.C. Kistler deserves a great deal of thanks for letting me follow my interests and for the guidance he provided to me (and many other graduate students). We were all very fortunate to have had Corby as an advisor and/or committee member.

Dr. Margaret Smither-Kopperl also deserves my thanks and everlasting gratitude for serving as a very active member of my committee in Gainesville. Margaret was unfailing in her good advice and sense of fair play and I will always count both her professional support and personal friendship as a highlight of my years in Gainesville.

Dr. Eric Menges served as my advisor at Archbold Biological Station and without his interest, input and support it is fair to say this project would have sunk like a stone. I am indebted to him for his patience and allowing me to be a part of his exceptionally productive and hardworking lab.

Dr. Andrew Ogram always provided me with good advice, insights and interesting discussions and I am indebted to him for his contributions as a committee member.

I am also particularly indebted to Dr. Rachel Shireman for her help and insights on how best to deal with institutional quagmires and her offer of the new caddy for a field vehicle. In this same vein, I would also like to thank Dr. Ken Gerhardt for his professional input and guidance to individuals in my department. I am truly appreciative.

I would also like to thank Dr. Mike Thomas, Dr. Wayne Dixon, Dr. Susan Halbert, Dr. Nancy Coile, Robert Leahy, Dr. Tim Schubert, and Penny Campbell for their friendship and support during my employment at the Florida Department of Agriculture and ii








Consumer Services, Division of Plant Industry. My particular thanks go to both Tim and Robert for their time, advice and bench space during the crucial early stages of this project. Dr. John Heppner provided the tentative i.d. on my leafminer, Jeff Loetz photographed both the moth and leaf in Chapter 5 and David Davidson provided media recipes.

Dr. Mark Whitten and Dr. Norris Williams, of the UF Museum of Natural History,

were always willing to help me on short notice. In particular, I would like to thank Mark for his unstinting attention to details and tutoring me in AFLP protocols. I also would like to thank Dr. Edward Hoffmann for making my many years of teaching in the Department of Microbiology a genuinely pleasant and worthwhile endeavor.

Dr. Stephen Mulkey, Dr. Kaoru Kitajima and Dr. Kevin Hogan of the UF Department of Botany provided me with equipment, software, support for the soils work, interesting discussions and kindness borne of old friendships.

I would also like to thank Dr. Hilary Swain, Dr. Mark Deyrup, Roberta Pickert, and Dr. Christine Hawkes for providing me with important feedback, good ideas and a positive point of view during my visits to Archbold Biological Station.

Dr. Richard Braithwaite (CSIRO, Darwin) first introduced me to the potential

problems of pathogens infecting clonal plant species during my visit to his study sites at Kapalga, Northern Territory, Australia and I thank him for his input.

The Florida Department of Forestry supported this research with a grant from the Plant Conservation Program and the Women in Agriculture Club provided support monies by granting me one of their Frances Summerhill awards.


iii








Lastly, I would like to thank Sally, Josh, Mathew, Elaine and Zachary

Dickinson for their support and all those celebrations over the years and, as always, my heartfelt thanks go to Karen Graffius-Ashcraft, Dan Simberloff, C.D. Smith, Diane DeSteven, Miss Beeks, N.N. Head, and the family of I. Bella and K. Kit.


iv















TABLE OF CONTENTS

page

ACKNOWLEDGMENTS................................................................. ii

ABSTRACT................................................................................ vii

CHAPTER

1 INTRODUCTION.................................................................. 1
H o st................................................................................. 12
Environm ent........................................................................ 14
Study Site........................................................................... 14

2 DISEASE AND ITS DISTRIBUTION IN
Hypericum edisonianum POPULATIONS....................................... 16

Introduction......................................................................... 16
Materials and Methods............................................................ 17
R esults.............................................................................. 26
D iscussion........................................................................... 43

3 THE EFFECTS OF HYDROLOGY AND SOIL NUTRIENTS
ON POND POPULATIONS OF Hypericum edisonianum................. 48

Introduction......................................................................... 48
Materials and Methods........................................................... 50
R esults.............................................................................. 52
D iscussion........................................................................... 68

4 GENETIC DIVERSITY IN Hypericum edisonianum.......................... 74

Introduction......................................................................... 74
Materials and Methods........................................................... 78
R esults.............................................................................. 8 1
D iscussion........................................................................... 82


v









CHAPTER page


5 INSECT DAMAGE IN SEVEN POPULATIONS OF
Hypericum edisonianum......................................................... 93

Introduction......................................................................... 93
Materials and Methods........................................................... 94
R esults.............................................................................. 95
Discussion........................................................................... 98

6 SUMMARY AND CONCLUSIONS............................................ 102

APPENDIX FORMULAE FOR FUNGAL CULTURE MEDIA ................ 106

REFERENCES.............................................................................. 107

BIOGRAPHICAL SKETCH.............................................................. 124


vi













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

FACTORS AFFECTING DIEBACK IN THE RARE PLANT
Hypericum edisonianum (EDISON'S ST. JOHN'S-WORT) By

G. A. van de Kerckhove

August 2002

Chairperson: Dr. H.C. Kistler
Major Department: Plant Pathology

Hypericum edisonianum (Small) Adams & Robson (Ascyrum edisonianum Small), a state-endangered plant found in only four counties in central Florida, has been experiencing sporadic diebacks with no known etiology. This study was initiated to determine the causes of diebacks and whether narrow genetic variation combined with unique environmental factors place the plant at risk from disease.

Two newly recorded fungal pathogens attack H edisonianum. Colletotrichum

gloeosporioides causes foliar and stem lesions in H. edisonianum that can ultimately lead to stem death. However, no C gloeosporioides infestations were found in any wild populations. The originally infected stem appears to have been the result of contaminated greenhouse conditions. Sphaeropsis tumefaciens also attacks H edisonianum, causing the formation of woody galls and witches' brooms. Disease incidence ranged from 0 to 83% within seven field populations of H edisonianum however S. tumefaciens infection was only significant in one population from 1998 to 2000.


vii








All study populations of the plant grew in seasonal ponds that were unique in their soil nutrient characters and hydrology. Taller stems of H. edisonianum were found centermost in most ponds and also contained the greatest number of galls. Plant stem heights were correlated with gradients in soil moisture, soil nutrients, organic matter content and soil saturation.

Amplified fragment length polymorphisms (AFLP) indicated considerable genetic diversity in 10 populations of H. edisonianum sampled from 3 counties.

A new, undescribed, leafmining microlepidopteran (Coleotechnites sp.)

(Gelechiidae) inflicts extensive foliar damage on H. edisonianum but does not play a significant role in H. edisonianum stem mortality.

In summary, neither fungal pathogens nor insect infestations are solely responsible for the observed decline in H edisonianum populations. Rather, environmental factors, particularly drought conditions, play a pivotal role in the synergistic effects of both disease and insect damage.


viii















CHAPTER 1
INTRODUCTION

All plant species are subject to pathogens at some time in their life cycle (Burdon 1991) and thousands of studies testify to the ongoing destructive capability of plant pathogens in modem agriculture. Most useful plants do not grow in efficient monocultures however. Rather, at least 60% of plants in global agriculture are cultivated in complex ecosystems that include varietal mixtures, intercropping and traditional subsistence polycultures (Francis, 1986). In West Africa alone at least 80% of cultivated land is in multi-use production while in Latin America staple crops are commonly found growing in polycultures (Francis, 1986). Lennd and Sonoda (1990) reported that more than 1050 million hectares in the tropics and subtropics are used for grazing and much of this land is covered in natural and mixed perennial pastures.

In great contrast to agriculture, the occurrence, dynamics and short and long-term effects of disease in wild plant communities has received far less attention but is by no means any less significant. All important food crops originated in natural plant communities and wild, ancestral stocks are especially important for use in crop improvement as sources for disease resistant genes (Lenn6 et al. 1994, Lenn6 and Wood, 1991). Wild species have played key roles in the improvement of potato, tomato and wheat crops (Lenn6 et al. 1994) and yet these wild host-pathogen systems have received scant attention by plant pathologists (Burdon et al. 1990, Lennd et al. 1994, Lenne and Wood, 1991).


I






2

Endemic pathogens and their associated diseases occur in all types of plants and native plant communities, from salt marshes to forests. In a literature survey of plant pathogens attacking British trees, approximately 15 fungal species were found per host tree taxon (Strong and Levin 1975). In a similar survey for North American plant communities, Strong and Levin (1979) found an average of 15 fungal pathogens per tree species, 7 for shrub species and 5 for herbaceous species.

Most major groups of plant pathogens are well represented in native plant communities. For example, Nienhaus and Castello (1989) compiled the known viruses that attack tree species in native communities and their literature review found that of the 37 tree species assessed, each had at least one viral parasite and some species up to seven. Much earlier MacClement and Richards (1956) conducted a systematic survey of viruses in wild plants and found virus incidences of up to 10%. In single-species surveys of wild populations of Plantago spp., Cooper and MacCallum (1984) found at least 39 different viruses infecting this genus; while earlier, Hammond (1981) reported up to 64% viral infection of Plantago spp. in England. Five wild populations of Primula vulgaris were found to contain from 043% infection with the Arabis mosaic virus (AMV) with the mean infection level of 19% (MacKenzie 1985).

Bacterial pathogens are not well documented in natural plant communities (Jarosz and Davelos 1994) and reports are limited. Nonetheless, bacteria have been reported to affect wild plants in diverse habitats. For example, Foster and Fogleman (1993) found 39 bacterial isolates from Organ pipe cacti, 19 from Saguaro and 16 from Senita cacti and grouped the isolates into 28 conspecific groups based on their fatty acid profiles. They






3


concluded that the patterns of bacterial distributions in cacti were dependent on the chemical compositions of the host species.

The pathogenic bacterium, Ralstonia solanacearum, has been found to affect

eucalyptus trees in Brazil, Australia, China, Venezuela and South Africa (Coutinho et al. 2000) by invading vascular tissue and causing wilt symptoms. This disease was first reported for eucalyptus in 1980 (Coutinho et al. 2000). Bacterial wilts have also been identified as contributing to the decline of an endangered plant, Euphorbia barnardii, a serpentine endemic found only in the Northern Province of South Africa (Knowles and Witkowski 2000). Nevertheless, bacterial diseases are thought to be rather limited in distribution in wild plant populations while fungal pathogens and the diseases they cause are probably most prevalent (Burdon 1994).

Plant diseases have affected the diversity of plant communities (Burdon 1987), the outcome of inter- and intraspecific competition (Brunet and Mundt 2000), the genetic structure of populations (Lennd et al. 1994) and, historically, the distribution of plant species. Perhaps the best-known cases of disease drastically altering the demography of a native plant species on a large scale in North America are the pandemics of the American chestnut (Castenea dentata) blight and Dutch elm disease. The fungus, Cryphonectria parasitica, the causal agent of chestnut blight, was accidentally introduced from Asia into North America in approximately 1904 and arrived roughly 34 years later (in 1938) in Europe (Griffin and Elkins 1986). This fungus kills chestnuts (and some oak species) by causing cankers that ultimately girdle the tree. From its point of origin in New York, the fungus rapidly spread throughout the natural range of C. dentata and by the 1950s close to 100% of all natural forest stands were infected (Brasier 1991). The






4


epidemic spread with equal speed in Europe. However, by the 1950s the European chestnut trees began to spontaneously recover from the disease (Biraghi 1950). This remarkable turnabout in the epidemic was initially attributed to either the ascendancy of a deleterious mutation in the fungal pathogen or resistance in an increasing proportion of trees (Jarosz and Davelos 1995). It was later discovered that the trees were recovering because C. parasitica had been infected with a debilitating, double-stranded RNA hyperparasite (Van Alfen 1975).

The impact of C. parasitica on the community structure of these forests depends

partly on the preinvasion species composition of the forest (Burdon 1994). In some cases the chestnuts were simply replaced by other co-dominant species (Woods and Shanks 1959) while in other forests dominated by chestnut trees, both tree and shrub species diversity increased (Stephenson 1986).

Dutch elm disease, caused by the fungi, Ophiostoma ulmi and 0. novo-ulmi, is

believed to have originated in either the Himalayas or Europe (Brasier 1990) and was first reported in Ohio in the early 1930s. Ophiostoma ulmi had previously appeared in Holland in 1921 and quickly spread across the European continent. This was a relatively mild strain of the disease and a large number of European elms survived infection. A second Ophiostoma pandemic, caused by the much more aggressive 0. novo-ulmi, has rapidly destroyed both European and North American elms and promises to decimate the tree throughout its natural range in the northern hemisphere (Jarosz and Davelos 1994).

Fungal plant pathogens are well documented in wild plant populations. For example, the root rotting fungus, Phellinus weirii, had notable effects on the structure of forest stands in Oregon and British Columbia. As in the cases of the Chestnut blight and Dutch






5


elm disease, the dominant forest species, mountain hemlock (Tsuga mertensiana) and Douglas fir (Pseudotsuga menziesii), are effectively removed from the canopy by this fungal disease. Trees more resistant to the fungus, such as the subdominant species shore pine (Pinus contorta) and Western white pine (P. monticola) replace T. mertensiana. In the aftermath of infection, T. mertensiana constitutes only 5% of the canopy in contrast to 75% prior to infection. The more resistant species subsequently became the dominant conifers in these forests (McCauley and Cook 1980) and formerly rare herbs became common (Holah et al. 1993).

Interestingly, the overall effect P. weirii may have on the restructuring of forest

composition is determined by the hydrological conditions of the particular site infected. In forests where conditions are mesic, mountain hemlock (T mertensiana) and the Pacific silver fir (Abies amabilis) are both particularly susceptible to the fungus however the fir is more resistant to infection. Thus the fir reestablishes at the site sooner but the species diversity remains relatively the same (Cook et al. 1989). In more xeric sites, the abovementioned subdominant, Abies amabilis, doesn't reestablish at the site first but rather a renewed sequence of pine recruitment occurs (Burdon 1991) and the species diversity of these forests is ultimately increased. This is the opposite effect the fungus Phytopthora cinnamomi has had on Australian eucalyptus forests.

In southeastern Australia, the root infecting fungus, Phytophthora cinnamomi, is a

broad host-range pathogen that has caused dieback to occur in whole eucalyptus forests; leaving only grasses and forbs (Weste and Marks 1987) which, in turn, adversely affected the entire assemblage of flora and fauna that are associated with or dependent on these landscapes. As in the case of the Chestnut blight and Dutch elm pandemics, the






6


P. cinnamomi epidemic also appears to be the result of human activity. Before logging, the pathogen was present in these forests but unremarkable in its affects. As logging activities continued however, soil moisture increased as did soil temperature when vegetative cover was removed. Roads established for logging further altered soil drainage patterns and ultimately facilitated the spread of the fungus. Wildfires that frequently occur in these natural eucalypt ecosystems followed logging and may have further enhanced the spread of disease by weakening the remaining trees (Weste and Marks 1987).

Another epidemic caused by fungus, Phaeocryptopus gaeumannii, causal agent of the disease Swiss needle cast (SNC), has had very similar effects at logged sites in the Pacific Northwest, particularly Oregon and Washington. Initially recorded in Switzerland in plantations in 1925 (Gaeumann, 1930) the disease was first observed in the Pacific Northwest in the early 1980s (Stauth 1997). Before the logging of native Sitka spruce and Western hemlock along the Oregon coast, this endemic fungus caused little damage (Meinicke 1939). After logging and particularly after the fires in Tillamook, Oregon, in the 1930s and 1940s, Douglas fir had been cultivated in commercial Christmas tree monocultures using multiple, nonlocal sources (Savonen 2000). Douglas fir is the only known host for this fungal pathogen (Filip 1998). In spite of warnings, commercial growers planted Douglas fir in the fog belt, an area that experiences high humidity from the Pacific coastline to approximately 19 kilometers inland (Gallob 2000). Tillamook is now considered the center or point of origin for this epidemic in the Pacific Northwest (Savonen 2000). In 1996, aerial surveys of the Oregon coastline determined that 130,000 acres were infected with SNC. By 1997, 393,000 acres were infected (Filip






7


1998). Although the fungus does not kill the trees, it retards growth and seriously diminishes their commercial value as Christmas trees or chip products.

Armillaria luteobubalina is another aggressive root-rotting fungus found in Australia's native eucalyptus forests that has drastically affected the vegetative community. The incidence and severity of the disease has also increased at the hand of man with repeated select cutting of older trees in unlogged forest (Edgar et al. 1976). Logging frequency rather than intensity of logging was found to be the critical disease factor because each select-cut exposes remaining trees to inoculum from fresh-cut stumps (Kellas et al. 1987). Another Armillaria (A. ostoyae) has also had serious affects on the various pine species in the U.S.A. in both wild and cultivated systems but primarily on trees in stressed conditions.

Epidemics caused by pathogenic fungi are still a concern for North American forests. Dogwood Anthracnose is currently a serious, spreading fungal epidemic occurring in the Pacific Northwest on Cornus nutalli and in the southeastern Appalachian mountain forests on the native Cornusflorida. First reported in New York in 1980 (Pirone, 1980) and described by Redlin (1991), this fungus (Disculus destructiva) quickly spread from New York and Connecticut to northern Georgia by 1987 and throughout the entire southern Appalachian mountains (Anderson et al. 2001). The symptoms are dieback of lower twigs and branches of mature trees that progresses to the crown, while seedlings and young understory trees are killed outright. The origin of this fungus is unknown but due to the sudden onset and rapid spread of the disease it is assumed that the fungus was introduced (Anderson et al. 2000).






8


In a study of wild dogwoods in the Cactoctin Mountain Park in 1984, Mielke and

Langdon (1986) first reported that only 3% of the native dogwoods remained disease-free while 33% were dead. In a follow-up study in 1988 Schneeberger and Jackson (1989) found that all the native dogwoods were infected and 89% were dead. The same disease was found in Washington on western flowering dogwoods (Cornus nutallii) in 1979 (Byther and Davidson 1979) and by 1983 the disease had spread to Oregon, Idaho and British Columbia (Anderson et al. 2001).

Disease in wild plants is not necessarily as obvious as the foregoing examples may suggest but rather can be pervasive, insidious and host specific. As early as 1933 Sampson observed the colonization of plant tissue by endophytic fungi. Bradshaw (1959) found that Agrostis tenuis and A. stolonifera were widely infected by the pathogen Epichloe typhina, an endophytic fungus that causes parasitic castration. Although field populations of these grasses suffered high levels of infection, there was little indication of this in the vigorous vegetative growth that resulted from early floral abortion. Nevertheless, the fungus is capable of having a profound impact on gene flow in these populations and consequently their population structure. Bradshaw (1959) noted that these endophytic infections remained parasitic under conditions where the host plant reproductive potential was diminished or suppressed. However, in long-lived host species where the environment is stable and seedling establishment is naturally rare, the endophytic relationship becomes more mutualistic than parasitic.

Colonization by Epichloe species, known as e-endophytes, (Schardi and Clay 1995) and their subsequent in situ production of toxic alkaloids, can inhibit both mammalian and insectivorous herbivory, particularly in the grass tall fescue (Festuca arundinacea






9


var. genuina) (Clay 1990). Tall fescue is an introduced, European invasive grass that covers over 15 x106 hectares in the eastern U.S., of which two thirds are infected with the fungal endophyte Neotyphodium coenophialum. Infections are intercellular and produce toxic alkaloids that may contribute to the 'invasiveness" of the plant, its subsequent dominance in native plant communities and the decline of species diversity (Clay and Holah 1999). Secondary metabolites from e-endophytes are also known to facilitate protection of plants from parasitic nematodes by increasing host tolerance to drought conditions and enhancing growth and fecundity.

Thus, endophytic infections span the spectrum from parasitic relationships with their host plants to unconditional mutualisms. Endophytes from the anamorph genus, Acremonium, fit the latter definition. Infection by this fungus has been discovered to increase the relative fitness of its host grass by way of discouraging herbivory while having no effect on the host's sexual reproduction (Clay and Leuchtmann 1989). The fungus exhibits no sexual stages itself and apparently depends on vertical transmission by infected seeds for its own survival. Clay and Leuchtmann (1989) found almost 100% infection by Acremonium species regardless of the presence or absence of herbivores. This is in contrast to the findings of Clay (1990 a, b) where the incidence of endophytic infections are often correlated with herbivore pressure. He found that infections were most often at low levels or absent when herbivory was low and relatively high when herbivore pressure was high.

Pathogenic microflora may also ultimately play an important role in the renowned species diversity and hyper-dispersion of trees in Central American lowland rainforests. In 1970 Janzen, and in 1971, Connell both independently proposed models predicting the






10


impacts of pests in natural systems. The Janzen-Connell model states that when both adult and juveniles are susceptible to the same pests, the pest pressure will be greatest (in a density-dependent manner) closest to the adults and that survival of juveniles will be higher with greater distance from conspecific adults. The net result of these interactions would be a decreased spatial aggregation of individual species. Augspurger (1983) tested the Janzen-Connell model and determined that a soilborne fungus, Pythium spp. can all but eradicate the naturally dense seedling beds beneath maternal trees. She also found an increase in seedling survival with distance from the natal site. In further studies, pathogens were found to play an important role in the density-dependent mortality of 6 of 16 species of rainforest trees (Augspurger and Kelly 1984). However, the adult trees were not the reservoirs of the pathogen and these results did not entirely meet the criteria of the Janzen-Connell hypothesis. Instead, damping-off fungi were the main source of mortality and the microenvironment colonized by dispersed seeds played a key role in the affects the fungi ultimately had on each species. In a study of 23 species of Australian rainforest trees, Connell et al. (1984) found that only one species had high mortality closer to conspecific adults.

Nevertheless, the effects of an unidentified stem canker disease on the tree Ocotea whitei (Lauraceae) in lowland Panamanian rainforest did appear to support the JanzenConnell hypothesis (Gilbert et al. 1994). Both adult and juvenile trees were susceptible to the canker and the disease incidence was host-density dependent. Further, they found that, as the Janzen-Connell model predicts, the coincidence of spatial patterns of canker and host mortality support the role of disease in regulating tree distributions in the forest. The identity of the causal agent for the tree cankers had not been confirmed, however an






11


isolate from the fungal genus Phialophora was found to induce cankers in Phoebe cinnamomifolia (Lauraceae).

In contrast to the species-rich forests of Central America, the widely occurring

monotypic stands of Spartina are also subject to pathogen attack with notably different consequences. The ergot fungus, Claviceps purpurea, was first recorded as infecting S. alterniflora marshes in North America in 1895 (Eleuterius 1970) and much later Boyle (1976) reported the fungus in Irish populations of S. townsendii in 1960 (Gray et al. 1990). From the early 1960s, when the fungus was first detected on S. anglica swards along the English coast (Hubbard 1970), the infection has spread rapidly and reached epidemic levels (Gray et al. 1990). Spartina anglica contains very little genetic variation (Raybould 1989, Raybould et al. 1990) and this character, along with its demonstrated susceptibility to the fungal pathogen, would seem to consign this plant species to inevitable population declines throughout the English coast. This is not necessarily the case however if host populations manage to escape the pathogen during years of unusual climatic extremes where the fungus is limited (Mantle, 1980) or where phenological or morphological changes in the host preclude infection (Parker 1988). Gray et al. (1990) further argued that S. anglica populations may not be in jeopardy of massive declines because of the plant's low reliance on seedling recruitment for population sustainability and the absence of other, competitive plant species that might otherwise invade. They report that a hyperparasite, Fusarium heterosporum, has been discovered to colonize C. purpurea sclerotia and that this interaction may serve as a natural biological control of this epidemic.






12


What role do plant pathogens play in endangered plant species with extremely

limited geographic distributions? If disease strikes discrete populations of a rare plant species, particularly those with populations with (probable) limited genetic variability (Menges 1999) is local population decline and extinction their irreversible fate? To address these questions, an endangered plant, Hypericum edisonianum (Small) P. Adams & Robson, commonly known as Edison's St. John's-Wort, was selected to investigate small population epidemics. The overall objectives of this study were to determine whether the sporadic diebacks observed to occur in this narrowly endemic plant are solely attributable to the action of plant pathogens; and to determine the environmental factors that may contribute to this plants decline.

A 3-year demographic study was initiated in seven small, seasonal- pond populations of H. edisonianum. Within these seven study sites permanent transects were established and annual measures taken on disease incidence, plant mortality, plant fitness characters, environmental stresses and herbivore damage. Preliminary assays on the genetic variability of H. edisonianum were also conducted using amplified fragment length polymorphisms (AFLP).

Host

This study investigates the diebacks observed in the endangered plant Hypericum edisonianum (Small) P. Adams & Robson (Clusiaceae). Edison's St. John's-Wort is a shrub with opposite, entire, glabrous leaves that are 15 to 25 mm long. The yellow flowers have four petals (12 to 18 mm) and 4 sepals (10 to 12 mm with cordate bases) and occur year- round, in sparse distribution but sometimes in great profusion at the Archbold Biological Station, Lake Placid, Florida. The flowers have numerous stamens






13

and the fruits consist of small capsules with 3 to 4 locules (Ward 1979). The thin bark is smooth, reddish brown to gray and often shows prominent leaf scars on stems with circumferences that can grow to 2.5 cm. The vegetative habit of the plant often generates hundreds of stems that are unbranched until the upper third of the stem, where they become multiply branched and spreading. It is common to find very dense clusters of stems in patchy distributions within seasonal ponds.

As a member of the Order Theales and Family Clusiaceae, the Hypericum family

consists of at least 68 species (USDA Plant Database 2002), with 21 species occurring in Florida. Ward (1979) suggested that the closest relative to H edisonianum is H. stans, a much more widespread species found in the southeastern coastal plain. He argued that H. edisonianum is derived from H. stans as a result of Pleistocene flooding that isolated the southernmost distribution of H stans in the Lake Wales Ridge region of Florida when the rest of the peninsula was still below sea level. The basis of this view was the persistence of both dry soil and fresh water ecotypes of H. edisonianum that evolved in Pleistocene refugia.

Hypericum edisonianum is reported as occurring in only 3 counties (Glades, De Soto and Highlands) in the state of Florida (Wunderlin 1998) however a single population does persist in Polk County on the Avon Park Bombing Range. This particular population has been subjected to both fire and conversion to bedded pine plantation and continues to persist under a pine canopy.

Hypericum edisonianum populations frequently grow in seasonal ponds (Abrahamson et al. 1984) that are found by the tens of thousands throughout the plant's 4-county geographic distribution. These ponds characteristically fill with rainwater during the






14


rainy season in Florida (June to September) and then slowly percolate to dry, shallow basins when the rains subside. Although H. edisonianum also can be found growing in low, moist swales, this study was conducted only in seasonal ponds where plant populations are locally very abundant.

Environment

Hypericum edisonianum is a narrow endemic that occurs only in the state of Florida, particularly in the Lake Wales Ridge (LWR) region of the state. The ridge is 160 km long (Menges 1999) and characterized by the presence of ancient sand dunes and terraces formed when Pleistocene sea levels were much higher than today. As sea levels subsided to present day elevations, these paleo dunes and crests remained and presently follow a north-south configuration through central and south central Florida (Brooks 1981). Deep, nutrient-poor sands overlay marine sands, marls, clays and sandy limestone that constitute the Miocene Hawthorne formations (Puri and Vernon 1964).

Fire, caused by lightning strikes, has had long-term effects on the Lake Wales Ridge vegetative community. Both spatial and temporal distributions of fires may have been instrumental in shaping the present-day mosaic of scrub vegetation in Florida (Menges 1999). Drought conditions often arise in these scrub communities and are most important during the winter and early spring months (Menges and Gallo 1991) however frequent dense fogs may ameliorate drought stress during these months (Menges 1994).

Study Site

This study was conducted at the Archbold Biological Station (ABS), Lake Placid,

Florida at 27 11' N and 81'21' W (Menges and Kohfeldt 1995). This research station is located in a major paleo dune system in the southern Lake Wales Ridge (LWR) (White






15

1958) and is dedicated to long-term ecological studies. The ABS privately manages 2081 hectares (5,141acres) of undisturbed land that resides on coarse to fine beach and dune sands of Plio-Pleistocene age (Abrahamson et al. 1984). The vegetative communities at ABS are part of the Florida scrub ecosystem and include Southern Ridge Sandhill, Sand Pine Scrub, Scrubby flatwoods, bayheads and seasonal ponds (Abrahamson et al. 1984). These plant communities have remained intact and persisted for an estimated 50,000 years (Watts and Hansen 1994).

Although the species diversity of Florida scrub systems is relatively low they

contain a significant number of endemic flora and fauna (Menges 1999). The number of endemic and endangered plants found in the LWR region of Florida rivals that of any other ecosystem in the United States.














CHAPTER 2
DISEASE AND ITS DISTRIBUTION IN Hypericum edisonianum POPULATIONS Introduction

Hypericum edisonianum (Edison's St. John's-Wort) is a globally rare plant that is only locally abundant in the protected land holdings of the Archbold Biological Station (ABS) in Highlands County, Florida. The plant's occurrence outside of the ABS properties is much more sporadic due, in part, to the rapid and continuing destruction of its habitat in south-central Florida.

This shrubby plant forms conspicuous, sometimes massive stands in shallow, seasonal ponds found on the ABS property. Over time, a number of ABS scientists have observed diebacks of unknown etiology in these Hypericum ponds, as they are referred to on station vegetation maps (Abrahamson et al. 1984) yet there appear to be no temporal or spatial patterns of note. Hypericum edisonianum is one of the many rare and endemic plants found on the Lake Wales Ridge in south central Florida (Christman and Judd 1990). Until this study, H. edisonianum has received little research attention.

Edison's St. John's-Wort is related to the species H. perforatum, (St. John's-Wort), the plant currently in favor and commercially grown for homeopathic products. Hypericum perforatum, while cultivated primarily for its antidepressant compounds (Barnes et al. 2001, Goodnick et al. 2001, Kim et al. 1999, Shelton et al. 2001) also has received substantial research interest for its antimicrobial activity (Pistelli et al. 2000), antibacterial (Schempp-Christoph et al. 1999), anti-fungal (Warr et al. 1992)) and anti-viral properties (Lavie et al. 1995).


16






17


Hypericum perforatum, while the subject of ongoing research for its medicinal

properties, is at the same time the target of extensive eradication efforts in regions where the plant is considered a serious pest (Campbell and Nicol 2000, Morrison et al. 1998). Also known as Klamath weed in North America, H perforatum invasion into pastures has been controlled by the release of the beetles Chrysolina quadrigemina (Huffaker 1967, Huffaker and Kennett 1959), Chrysolina hyperici and the host-specific pathogenic fungus, Colletotrichum gloeosporioides (Morrison et al. 1998).

Hypericum edisonianum displays no such invasive characters. Moreover, H

edisonianum potential for therapeutic applications is unknown. The urgent need for developing a better understanding of this rare plant and the potential disease processes underway in its many small populations was the catalyst for this study. In 1997, E. Menges, of the Archbold Biological Station, first suggested the need for further information on seasonal-pond diebacks of H edisonianum and shortly thereafter the search for disease in several pond populations was initiated.

Materials and Methods
Disease Screening

The first step in assessing the potential for disease presence in the pond populations of H edisonianum entailed an ABS-wide (5,000 acre) search for disease symptoms and possible disease foci. A heavy-duty, 4-wheel drive vehicle, provided by ABS, was used to check Hypericum ponds in all tracts of the main station property. In each of 24 ponds visited a walking transect was started at the pond margin and all H. edisonianum stems encountered, in progress towards the pond center, were examined for disease and/or insect damage. At the pond center, a new transect was begun, heading back toward the pond margin and ending approximately 5 meters from the previous start point at the pond






18


margin. This search pattern essentially partitioned the pond into wedges wherein any Hypericum stem with observable symptoms or damage was collected. Several different symptoms were used to determine whether plant tissue was collected (Table 2-1, after Fox 1993). In ponds where diebacks had previously been observed or appeared to be presently underway, symptomatic stems were also uprooted and root tissue collected.

Table 2-1. Disease and insect damage indicators in field surveys

Chlorosis Lesions Galls
Discoloration Pustules Abnormal growth
Speckling Pitting Constrictions
Spotting Holes Leaf curling, wrinkling
Chewing Mines Exit holes


All samples of plant tissue collected during these pond inspections (leaves, branches, stems and roots) were placed in new, appropriately labeled plastic bags and were stored temporarily in a cooler until return to the on-site plant ecology laboratory of E. Menges. These surveys were performed at irregular intervals from May 1997 to May 1999 with tissue sampling concentrated from May to August of each year.

There are a multitude of diseases one might search for and various methods available to diagnose them in this plant species. As fungi are the most frequent causal agents of plant diseases (Burdon 1994), a decision was made to screen only for fungal pathogens with commonly used culture media: (water agar, potato dextrose agar (PDA), acid potato dextrose agar (APDA) and corn meal agar (CMA)). Selective media specific for culturing the fungal root pathogens Pythium spp. and Phytopthora spp. (PARP plates) (courtesy of D. Mitchell, University of Florida and D. Davidson, Division of Plant Industry, Florida Department of Agriculture and Consumer Services) were also used to screen root samples. The appendix contains recipes for all culture media.






19


Symptomatic and asymptomatic tissue samples (stems and leaves) were surface

sterilized with a 10% CloroxTM solution (0.5% sodium hypochlorite) for approximately 1 to 3 minutes (depending on how woody they were), rinsed with sterile distilled water; and blotted on autoclaved paper towels and then small pieces of tissue aseptically placed on culture plates. These plates were then sealed with ParafilmTM and left at room temperature in the laboratory under ambient light conditions. Tissue samples were also placed in sterile glass petri dishes with autoclaved, moistened paper towels. All culture plates and moist chambers were transported to the laboratory of Dr. Tim Schubert (Florida Department of Agriculture and Consumer Services, Disease Diagnostic Lab, Gainesville) for further incubation and identification. Selection of Ponds for Study

Each seasonal pond selected for further study was chosen on the basis of its size and physical location on the ABS property, using digitally corrected aerial photographs (Figures 2-1 and 3-15). An attempt was made to select ponds that were roughly similar in size. Ponds are individually numbered at ABS and those used in this study were: 7-50, 7-56, 7-64, 30-35, and 31-45. The first number is the tract number on the station and the second number is the numeric count of that pond within the tract. Two unnumbered ponds in the northeast (known as the Red Hill tract) of the property were also included as study ponds and were designated Pond 8 and Pond 88. These are the two northernmost ponds in this study and they are located within 200 meters of each other beneath a partial pine canopy. Ponds 7-50, 7-56 and 7-64 are found in the north central portion of the ABS property within 500 meters of one another (Figure 2-1) and, except for 7-50, receive full sunlight. (Pond 7-50 is partially shaded by pines.) The distance between this cluster and Ponds 8 and 88 to the north are at least 1,900 meters. Pond 30-35 is located at the fence






20


VC %
'Al
~i4%









get 4t7.
Irvs 0I
14

~V
old-1

Fiue21 as-oo eilpoorp f rhodBooia tto hwn he
study ~~ pod 75,76 n -0 ice rmlett ihrsetvl) hs od
were~~ seetdfrtercoepoxmt ooeaohradthsgetrpoaiiyo disas spea an eefo frltrsuis.Alte td od eea es n





fige f-theFalse-co lorg ae phooapih souAhoster Bioiarttion howiryng thre





approximately 5,700 meters from Ponds 7-50, 7-56 and 7-64. It receives full sunlight. The southernmost pond in this study, 31-45, is separated from Pond 30-35 by 1,300 meters and also receives full sunlight. The distance between the northernmost ponds (8 and 88)


and southernmost pond (31-45) is 8,456 meters.






21


Study Transects

In the seven selected seasonal ponds, meter-wide belt transects were randomly placed in each pond and then all individual H. edisonianum stems were tagged within these transects (Figure 2-2).

To place the transects, a number was blindly selected from a random number table and used as a compass point to position an aluminum angle stake at the palmetto edge of the pond margin. From this stake, a meter tape was reeled out in a straight line to a PVC (polyvinyl chloride) pipe located in the deepest center of the pond. This pipe had been previously established by Kevin Main, land manager of ABS, for long-term hydrological studies. Two ponds in this study have no PVC pipe (Ponds 8 and 88) and their centers were estimated by measuring the pond diameter and then placing the transect line from the palmetto edge to the midpoint of that diameter. No long-term hydrological data are available for these two ponds.

A wire flag was placed at every meter interval of these transects and then at one-meter widths to form a meter-wide belt transect (Figure 2-2). The length and number of transects varied by pond given the differences in individual pond sizes. All H. edisonianum stems within these transects were permanently tagged using 17-gauge galvanized wire to secure individually number-stamped aluminum disks. During the 1998-2000 censuses (between June and July of each year) the categories of data recorded were stem height, crown size, number of flowers, disease symptoms, insects present, mortality and new shoots (Table 2-2).

Field Transplants

The number of plants that could be removed for pathogenicity tests was limited as this






22


is a State endangered species (Coile 2000). A State collecting permit was obtained to allow removal of a limited number of plants from the field for research purposes.

Small (20 to 30 cm) apparently healthy, intact, vegetative shoots were collected from various ponds and transplanted to 15 cm diameter pots (Figure 2-3). A small tiling spade was used to punch four cuts into the root mass around the selected stem. Upon the fourth punch the whole root mass and stem were levered up and transferred by hand to sterile plastic pots that contained a shallow bed of autoclaved pine chips. The hollows around the root mass were filled in with additional field soil, firmly tamped down and













5L.I


S-A





Figure 2-2. Meter-wide transect (defined by yellow flags) in Pond 30-35 during dry period in 1998. The foreground is bare sand with sparse vegetation. The light green vegetation is healthy, flowering H edisonianum (with some grasses and Lacnanthes caroliniana). The gray area in the center is dead H edisonianum. Photo was taken by author standing at the palmetto edge of pond. then the whole pot soaked with pond water (or tap water when pond water was not available). Potted transplants were placed in shade during actual digging and then






23


transferred back to University of Florida greenhouses after one to two days of recovery on the front veranda of the ABS Plant Laboratory. In the University of Florida greenhouses, the transplants were watered every other day for approximately one month or until the small shoots stabilized and began growing. Thereafter they were watered to saturation every week. After six months, the transplants (Figure 2-4) were moved to the Florida State University Greenhouse facility and watered every week using tap water until placement in a shallow, artificial pond located beneath a pine canopy on the FSU property. The transplants were not treated with nutrient supplements or pesticides.

Table 2-2. Annual Pond Population Census Measures, 1998-2000*

Stem height (from ground level to highest leaf in cm)
Crown size (longest and shortest axis measured in cm (these values
were multiplied and the square root used as the area of the crown)
Flowers (old flowers, new flowers and buds)
Disease symptoms (Table 2-1) Insects present/insect damage
Survival or mortality of stem for year
New shoots (recruits of the year)

* All measures were taken between June and July of each year. Colletotrichum Isolation and Pathogenicity Tests

Leaves with lesions were collected from only one transplanted stem that displayed these symptoms, surface sterilized and plated onto CMA and APDA as previously described. After incubation for four days at room temperature, the culture was scraped from the surface of the plate and placed in 50 ml of sterile water. Quadrant streaks were made on fresh APDA plates using these water slurries. Single spores were subsequently collected from these plates after 15 to 18 hours incubation at room temperature. Spore collection was done under magnification by cutting individual, germinating spores from the medium using a sterile needle and transferring them to fresh APDA plates. After incubation for four days






24


these cultures were scraped from the medium and suspended in sterile water. Conidial spore suspensions (106 spores/ml) from these cultures were then used to spray-inoculate entire, greenhouse-grown, healthy plants to run off with a hand sprayer. Control plants were sprayed in the same manner with sterile water only. Both control and experimental plants were enclosed in plastic bags 24 hours before and after spraying to facilitate infection. Three separate trials were performed using these procedures and the same initial fungal isolate. This fungal isolate was kept in continuous culture on APDA plates that were stored at room temperature under ambient light conditions. Sphaeropsis Isolation and Pathogenicity Tests

Field-collected stem galls were also taken from H. edisonianum, surface sterilized, blotted and thin-sectioned using a sterile razor. Small fragments (2 to 3 mm long) of these sections were placed on APDA plates, sealed with ParafilmTM and incubated at room temperature. The uniformly dark mycelial cultures from these plates were identified as Sphaeropsis tumefaciens. These initial isolates were used to inoculate artificial wounds made on healthy appearing H edisonianum stems (previously transplanted from the field). Wounds were created by making shallow, downward cuts into the stem (approximately 5 mm in length) using a sterile razor. Small squares of the growing fungus were then sliced out of the culture plate and embedded into the angled stem cut. The entire inoculated wound was wrapped in a small band of ParafilmTM to exclude contaminants and prevent desiccation. Control plants were similarly wounded and wrapped but received no fungal inoculum. Control and test plants were placed in greenhouses at the Florida State University, Tallahassee, and were maintained only with weekly watering for approximately six months before the pots were transferred to the edge of an artificial outdoor pond for another six months.






25


Figure 2-3 Four shoots of Hypericum edisonianum taken from wild populations at ABS property, four days after transplantation.




























Figure 2-4. Greenhouse-grown stems of Hypericum edisonianum. These stems were collected from the field when they were approximately 30 cm tall (see above photo) and are now adult size (note meter stick in foreground) after six months of greenhouse residence.






26.


Results
Disease Screening

A number of saprophytic fungi were repeatedly cultured from leaf, stem and root samples of field-collected specimens. These fungi belonged to the genera listed in Table 2-3. Approximately 500 plates with six separate tissue samples each were used to isolate these fungi over an approximate 2-year period. These genera consisted of saprophytic species and were isolated repeatedly but sporadically (data not shown). There was no visible correlation between isolation of these fungi and presence of any dieback symptoms. Therefore, these fungi were not considered further as causal agents of the H. edisonianum dieback.

In-field visual inspection of dozens roots gave no indication of root disease. When the roots from dead stems were inspected, the tissues were still intact, however root hairs were absent. The stems were cross-sectioned in the field with a sharp jackknife and only dry, buff colored woody tissue was observed. PARP, APDA and CMA culture plates yielded no isolates of Pythium or Phytophthora from root samples.

Table 2-3. Genera of saprophytic fungi isolated from H. edisonianum tissues


Cladosporium sp. Nigrospora sp.
Curvularia sp. Pestalotiopsis sp.
Phomopsis sp. Epicoccum sp.
Trichoderma sp.


Koch's Postulates for Colletotrichum gloeosporioides

Within 3 weeks of being placed on greenhouse benches, one of the new field

transplants (selected in the field based upon its healthy appearance) displayed sporulating lesions on the upper surface of two leaves. These lesions first appeared as small brown specks that quickly expanded to approximately IxI cm in 1 week (Figure 2-5).






27


The fungus, Colletotrichum gloeosporioides, was isolated from these lesions in pure culture on APDA plates and single spore cultures were used to experimentally sprayinoculate healthy appearing Hypericum edisonianum stems. This resulted in new lesions (the same in appearance as those lesions first observed on the original plant stem), appearing on experimental plants within 1 week. These lesions also first appeared as small brown specks that quickly expanded. Few new lesions were observed after the first week, indicating that the high humidity found within the bagged plants was essential for infection. Infected leaves remained on the plant for approximately 3 weeks before browning and falling. Control plants showed no signs of infection. The fungus recovered from the experimental plants was identified again as Colletotrichum gloeosporioides (Penz.) Penz. & Sacc. in Penz. and this identification was confirmed by Dr. Tim Schubert and Mr. Robert Leahy (Florida Department of Agriculture and Consumer Services).

In preliminary trials, four plants were used for inoculations. Two plants were placed in water filled-basins while another two were left on the greenhouse bench with only occasional watering after receiving the spray inoculation. Two control stems were used for each treatment and were sprayed with distilled water. The experimental plants in saturated soils died within 4 weeks of inoculation with the fungal spore suspensions (106 spores/ml) while the two plants in drier soils survived for approximately 9 weeks before succumbing. The control plants showed no signs of infection or decline.

The second round of trial inoculations (using spore suspensions prepared in the same manner as in preliminary inoculations and used at 106 spores/ml, resulted in 9 of the 10 experimental plants becoming infected. None of the 10 control plants exhibited lesions. In the third round of trial inoculations eight of the 10 experimental plants showed lesion development while none of the control plants were symptomatic. In both the second and






28


third inoculation trials, all plants had to be moved out of the greenhouses and placed on benches outdoors in November temperatures. Lesions that were enlarging while in the greenhouse ceased to do so after placement on the outdoor benches and experimental plants did not die.






















Figure 2-5. Colletotrichum gloeosporioides stem and leaf lesions on H. edisonianum plant from Koch's Postulates trials.

The fungal pathogen Colletotrichum gloeosporioides was reisolated on APDA plates from inoculated plants and confirmed by microscopic inspection to consist of 1-celled, hyaline conidia (Sutton 1980) considered to be the causal agent of these lesions in H edisonianum plants in all experimental trials. Repeated field searches for symptomatic plants throughout the study (1998 to 2000) were unsuccessful and C. gloeosporioides was never recovered from any wild population. The single transplant shoot initially found with C. gloeosporioides lesions (Figure 2-5) appears to have been infected within the first 3 weeks of residence in the University of Florida greenhouse.






29


Koch's Postulates for Sphaeropsis tumefaciens

Woody galls, ranging in size from slight swellings to large (5 cm in length), fissured

growths were frequently observed on stems in Pond 30-35 (Figure 2-6) from 1998 to 2000. The conducting tissue above and below large field-collected galls (3+ cm in length) appeared to become stained, perhaps by the colonization of the dark hyphae of the fungus. This dark staining of tissue progressed several centimeters in each direction from the original point of infection (Figure 2-7) and was found in both naturally occurring and artificially induced infections. Experimental wound inoculations of healthy H. edisonianum plants (in Materials and Methods section) resulted in the formation of small galls (2 cm or less in length) after approximately 1 year of incubation (Figure 2-8).

The pycnidia of S. tumefaciens were dark brown and gave rise to mostly unicellular conidia that were 20 to 34 mm long. Plating fragments of surface-sterilized stems galls consistently yielded pure cultures of the fungus. Sphaeropsis tumefaciens was not isolated from control stems. The causal fungus was identified as Sphaerospsis tumefaciens Hedges (Hedges 1911, Hedges and Tenny 1912, Sutton 1980) and this identification was confirmed by Dr. Tim Schubert and Mr. Robert Leahy.

Three separate inoculation trials were conducted. In the first, eight experimental and eight control stems were utilized. Seven of the eight experimental stems contained galls after a 1-year incubation. No control stems had galls. The second trial consisted of six experimental and six control stems. All experimental stems contained galls while controls had none after 1-year incubation. The last trial inoculations used the remaining seven mature, greenhouse-grown stems (five experimental, two-control). After 8 months of incubation, four of the five experimental stems had galls and the control stems none. All experimentally induced galls were collected, surface sterilized and small fragments were






30


plated on APDA media which were later used for microscopic identification of the fungus; thus it is unknown whether these galls would have eventually caused stem death. Field Infections of S. tumefaciens

Pond 30-35 had the only population of H edisonianum with a high proportion of stems infected by S. tumefaciens (Table 2-4). In 1998, 64% of the stems censused in the transect carried galls. In the following year (1999) approximately half of the stems in this same transect were infected. This decrease in percentage of stems infected was a reflection of the increase in newly produced stems (stem recruitment). In 2000, the population of H. edisonianum in Pond 30-35 crashed (Figure 2-9) and this was reflected in the transect sample data where 83% stem mortality was recorded (Table 2-4). Mortality in 2000 was higher than in 1999 for most populations of H. edisonianum without S. tumefaciens infections or any other detectable disease (Table 2-4), such as in Pond 8 (47%) and Pond 3145 (44%) (Figure 2-9).

Spatial Distribution of S. tumefaciens Infections

On an ABS-wide scale, Pond 30-35 appears to be the infection focus for S. tumefaciens infection in H edisonianum on the ABS property. During the three years of this study a continuing ABS-wide vigilance for infection was maintained and yet no other pond population of H edisonianum was observed to contain more stem and branch galls.

The greatest percentage of stems infected in Pond 30-35 during 1998 was near the pond margin and then in 1999 and 2000 shifted towards the center (20 to 25 meters from the pond margin) (Figure 2-11). When the status of stems was examined in a cross-section of the pond (in a transect that spanned across the entire pond from edge to edge) mortality was






31


Figure 2-6. Multiple galls caused by Sphaeropsis tumefaciens on wild Hypericum edisonianum plants in Pond 30-35 at Archbold Biological Station.









1 2 3 4
Figure 2-7. Staining of tissue that occurred above and below field collected and experimentally induced stem galls. The first stem section at left (1) is from below gall and second stem section (2) was excised from above the same gall. The third section from the left (3) was taken from below the gall and the fourth section (4) from left was taken from above the same gall.






32


Figure 2-8. Experimentally induced gall on stem (left) and lack of gall formation on control stem (right) after one year of incubation. Arrow indicates where control stem received wound but no inoculum.



Table 2-4. Percentage of stems infected and mortality in H. edisonianum

Stems Stems Stems Stem Stem
Pond Infected Infected Infected Mortality Mortality
I.D. 1998 (%) n 1999 (%) n 2000 (%) n 98 -99 (%) 99-00 (%)

8 0 19 0 19 0 17 21.0 47.0
88 0 63 0 47 0 80 25.5 32.5
7-50 0 89 0 89 0 89 12.0 38.0
7-56 0.01 86 0.007 137 0.008 128 18.0 26.0
7-64 0 106 0 107 0.008 116 19.0 19.0
30-35 64.0 150 48 179 24.0 188 4.0 83.0
31-45 0 51 0 9 0 89 27.0 44.0

Note: the number of stems and galls varied from year to year due to stem mortality, missing
stems and recruitment of new stems into each pond population each year.






33

concentrated in the center in 1999 (20 to 25 meters from the pond margin) and then was throughout the pond in 2000 (Figure 2-12).

Stem mortality was examined in all study ponds (Figures 2-13 through 2-18) and the

only other ponds to experience the most stem mortality in the center was Pond 7-64 (Figure 2-17) in 1998 to 1999 and Pond 88. These ponds had no S. tumefaciens galls.

Stem mortality in Pond 8 increased at the pond margin from 1998-2000 as it did in Ponds 7-50 and 7-56 (Figures 2-13, 2-15 and 2-16). Stem mortality increased in all but the very center of Pond 31-45 (Figure 2-18).




\4*





Figure 2-9. Uprooted stems from a single pond showing galls at same relative heights.

The survivorship of stems in Pond 30-35 was not significantly affected by the presence of S. tumefaciens infections. The presence of stem galls, caused by S. tumefaciens, in 1998 wasn't correlated with stem mortality in 1998 to 1999 (Pearson ) =.357, df =1, p > .10). There was no significant relationship between stem galls recorded in 1999 and stem mortality in 1999 to 2000 (Pearson X=2.245, df= 1, p > .10).

The means by which S. tumefaciens is dispersed within Pond 30-35 is unknown,

however, it appears that water may play a key role. Figure 2-9 shows multiple, uprooted stems of H edisonianum that were taken from a single pond. These stems have galls at the same relative height that suggest spore dispersal may be facilitated by floating on the water surface.







34


Stem survival, 1998-1999
140

120

100.
E 2)
80.
0 ()
. 60,
Dead
z
40
4- a liv e 20 ' missng

0 new recruit
8 88 7-50 7-56 7-64 30-35 31-45

Pond by Archbold I.D. number






Stem survival, 1999-2000
120


100


E 80
0)


(D Mdead
E
z 40 Ve

E]nissing 20


0 pre hously dead
8 88 7-50 7-56 7-64 30-35 31-45

Pond by Archbold I.D. number




Figure 2-10. From 1998 to 1999 Pond 30-35 had the lowest proportion of dead stems in the permanent transect of all study ponds (4%). By 2000 Pond 30-35 suffered the greatest proportionate mortality of all ponds (83%).


























U


11131110 L L Meters from pond margin


30%0) 0%



= 0%.


0 10 0

Meters from pond margin


25%



1
0

(i













0 '0

Met


I


2C 30

ers from pond margin


Figure 2-11. Distribution of S. tumefaciens infections (galls) from 1998 to 2000 in Pond 30-35 (percentages represent stems infected for each meter of transect). Occurrence of galls increased with distance from the pond margin (origin). .


35


Co S) 15%'






Cn
CRr


0







36


Pond 30-35, 1998-1999



(D 14,

12 10

E 8 06

41

2 da
0)ni
C liii I 11 malv


1 7 9 11 13 15 17 19 21 23 25 27 Distance from pond margin (m)



Pond 30-35, 1999-2000
12


-10

C
8
(D

E6


C,



2L alive

0 P dead
1 7 9 11 13 15 17 19 21 23 25 27 Distance from pond margin (m)




Figure 2-12. Distribution of dead and living stems in Pond 30-35 for 1999 and 2000. Percentages are based upon total number of stems found in each category (alive, dead). Meter I starts at margin of pond and Meter 27 ends at other margin of pond.








37


10 20 30

Meters from pond margin









Dead, 2000
















10 20 30

Meters from pond margin


Alive 1999


I


50%40%*






20%


10%


-I











0


30


Figure 2-13. Distribution of dead stems in Pond 8 from 1999 to 2000 was greatest at or near the pond margin (pond center is located at transect meter 11).


Dead 1999


2M'


S,
CO 16


U,



0


I I


0 10 20

Meters from pond margin


0 20 30

Meters from pond margin









Alive, 2000


















Dead 1999


















20 30


Meters from pond margin


Alive 1999


















10 20 30

Meters from pond margin


0


Dead, 2000


20%



E 1%






0


20 30


0 10


20 30


Meters from pond margin


Meters from pond margin


Figure 2-14. The distribution of dead stems in Pond 88 increased with distance from the pond margin from 1998 to 1999 and then became more widely distributed throughout the transect from 1999 to 2000.


38


30%-


CO
20%
0


10%.


0


10


Alive, 2000


I


I


2


10


















Dead 1999


10 20 30

Meters from pond margin


10 20 30

Meters from pond margin


100%. 75%. 50 25%


Figure 2-15. Distribution of dead stems in Pond 7-50 was greatest near the pond margin from 1998 to 2000.


39


0


Alive 1999

















o 30

Meters from pond margin









Alive, 2000

















0 10 20 30

Meters from pond margin


75%.

2
C')
= 50%0
0-0


Dead, 2000


0



















Al


















Akl"


2M%15%.




"6


Meters from pond margin










Dead, 2000


10 20 30

Meters from pond margin


Meters from pond margin










Alive, 2000



















0 10 20 30

Meters from pond margin


Figure 2-16. Stem mortality in Pond 7-56 occurred throughout the transect from 1998 to 1999 and then became more prevalent at the margin and midtransect from 1999 to 2000.


40


Dead 1999


ive 1999


0


20


25%



2M' C0
E
o 15%C/)

10%.



5%-


0


30 0 10 20 3


10

















50%40%-


E C')

0


Dead 1999


















0 10 30

Meters from pond margin









Dead, 2000


















0 10 20 30

Meters from pond margin


I


10


20 30


Meters from pond margin


Figure 2-17. Dead stems in Pond 7-64 were most numerous near the pond center from 1998 to 1999 and then became more prevalent throughout the transect from 1999 to 2000 (the pond center is located at meter 15).


Alive 1999


















0 10 20 30

Meters from pond margin









I Alive, 2000


41


p


20%10%-


15%" E,
E C)10%0%


0

















Dead 1999














I.
10 2 0


Meters from pond margin






Dead, 2000


75%

E


0

25%



0 10 20

Meters from pond margin


Alive 1999


II


III


0 10 2 0


Meters from pond margin






Alive, 2000


6 12 A0


Meters from pond margin


Figure 2-18. Dead stems were most concentrated at the margins of Pond 31-45 from 1998 to 1999 and then where found throughout the transect from 1999 to 2000.


42


25%"


20%'
E
U)
15%
0
10%.


5%-


0






43


Discussion

In this study two fungal pathogens, C. gloeosporioides and S. tumefaciens, were

discovered to cause diseases in the rare plant H. edisonianum. This is the first record for both of these pathogens on this host.

Although the C. gloeosporioides infection appeared to be contained on greenhousegrown transplanted stems in this study, this pathogen may pose a serious threat to the small, extant populations of H. edisonianum in the state of Florida. Trial inoculations of healthy plants using spore suspensions of this fungus resulted in consistently high infections (100%, 90%, and 80% respectively for the three different trials). This fungus is already well known for its virulence in H. perforatum and has been under development as a natural biological control agent for Hypericum spp. in the U.S., Canada and Australia (Hildebrand and Jensen 1991, Shepherd 1995, Morrison et al. 1998).

Colletotrichum gloeosporioides was first isolated from citrus plants in 1886 in Florida and causes the disease Post bloom Fruit Drop (Liyanage et al. 1992, McMillan, Jr. and Timmer 1989, Agostini and Timmer 1994). This is a wide host-range pathogen that causes disease in numerous plant species in Florida (Alfieri et al. 1994, Timmer et al. 1994) and worldwide (Bernstein et al. 1995, Freeman et al. 1998). In citrus, the disease causes necrotic petal lesions and premature fruit drop while in H. edisonianum the leaves and stems develop quickly expanding, necrotic lesions.

The Archbold Biological Station property is located in the midst of an extremely

dense concentration of commercial citrus operations (Figure 4-2). Over time, the upland native plant communities, especially sandhill and Florida scrub, have been converted to citrus. Furthermore, it has recently been discovered that C. gloeosporioides causes premature fruit drop in saw palmetto (Serenoa repens) (Carrington et al. 200 1). a






44


ubiquitous species found throughout Florida scrub plant communities and commonly found at the margins of seasonal ponds. Thus, the potential for a continuous source of C. gloeosporioides inoculum is ever present for the remaining H. edisonianum populations.

This fungal pathogen was only recovered from one transplanted stem of H. edisonianum during a six-month residence in the University of Florida teaching greenhouse during 1998. Isolates of this single infected stem were used in all three Koch's Postulates trials. However, it is quite possible that undetected infestations may be present in populations of H. edisonianum not sampled in the ABS property or in distant ponds outside of the ABS property.

The fungal pathogen Sphaeropsis tumefaciens was also discovered to infect H.

edisonianum. Sphaeropsis tumefaciens is also a citrus pathogen (Rodiguez and Melendez 1984), having been introduced into the U.S. sometime in the 1930's (Holliday and Punithalingam, 1970). In citrus and H. edisonianum alike, the fungus causes hypertrophic tissue or galls and witches' brooms to form that were observed to persist for years on both the primary stems and upper branches of H. edisonianum (Figure 2-6). Over time these galls slowly enlarge and become fissured. Sinclair et al. (1987) report that galls shed conidia (mitospores) from embedded pycnidial structures, which serve as a chronic source of inoculum.

Sphaeropsis tumefaciens has a broad host range and has been recorded as attacking numerous plant species in Florida (Alfieri et al. 1994, Marlatt and Ridings 1974, Marlatt and Ridings 1976) and is of particular interest as a possible biological control agent for the Brazilian pepper (Schinus terebinthifolius) (Marlatt and Ridings 1979).






45


In a station-wide search at ABS this disease was found in several seasonal pond

populations of the plant however not more than 10 stems were affected in any one pond except for Pond 30-35.

Of the seven pond populations of H. edisonianum censused, only those in Pond 30-35 contained numerous stems with S. tumefaciens galls. Infection of stems was greatest near the pond center from 1999 to 2000 where presumably plants would experience the greatest flooding stress.

How S. tumefaciens infections were distributed within Pond 30-35 is of interest because of the dynamic character of water levels in all seasonal ponds. Stems of H. edisonianum at the pond center experience much longer hydroperiods than those stems at the pond margin. Could this disparity of soil saturation times affect the distribution of S. tumefaciens infections? Survivorship of stems was least at the center of the pond from 1999 to 2000, further suggesting that water stress and S. tumefaciens infections may play a role in H. edisonianum decline in this pond.

As a rule, flooding stress incapacitates or kills terrestrial plants more quickly than

drought or soil moisture depletion (Larcher 1995). Those genera tolerant of anoxic soils, such as Taxodium, Nyssa, and Salix have evolved both functional and morphological adaptations to cope with the multitude of abiotic factors that are inevitably coupled with low or absent soil oxygen due to flooding. These factors, such as increased soil acidity with subsequent deficiencies in nitrogen and increases in metal oxides in solution, may work in concert to debilitate plant communities found in seasonal ponds. Such debilitation or weakening of plants is commonly viewed as predisposing them to pathogenic invasion. However, populations of H. edisonianum in seasonal ponds are unusual in the sense that their clonal, vegetative form may result in each pond population






46


representing a single individual (genet). Acting as a single individual, stems of a clonal plant have the capacity to maintain a multitude of physiological connections between distant ramets, such as translocation of resources, control of intra-clonal competition by regulation of ramet production and buffering capability in heterogeneous and/or stressful microsites (J6nsd6ttir and Watson 1997). The advantages of physiological integration between ramets, particularly in resource-poor environments such as seasonal ponds are: 1) a means of conserving scarce resources by sequestering and reallocating to various clonal fragments; 2) the evolution of developmental divisions of labor among ramets (J6nsd6ttir et al. 1996) whereby old ramets function as nutrient storage units and new ramets in acquisition of carbon and nutrients; and 3) "sampling of new environments" by ramets in environments of patchy resource distribution (J6nsd6ttir and Watson 1997). The degree of physiological connections and interactions in H. edisonianum clones is unknown.

The presence of S. tumefaciens does not have a significant an impact on the H.

edisonianum stems growing in Pond 30-35. This pond contained the only population of H. edisonianum with approximately 50% infection in 1999 and, during the extreme drought conditions of 2000; this infection coincided with high stem mortality. However, Ponds 7-64 and 88 also suffered the greatest percentage of stem mortality towards the pond center without any apparent infection by the fungus.

Given the clonal nature of H edisonianum, it may be erroneous to assume S. tumefaciens infection ultimately leads to death of the long-lived genet. Perhaps galls act to block assimilates from the affected aerial shoot and this signals the rhizome to redirect resources to dormant meristems. Dying H edisonianum stems affected with galls frequently 'die-back' by a gradual browning and loss of leaves. Whether this is






47


attributable to a straightforward blockage of conducting tissue by the galls that is exacerbated by drought conditions or a more complex and integrated physiological response to infection by the clone is unknown at this time. The observed dark staining of the conducting tissues may be indicative of the movement of the fumgus in xylem tissue. Such movement may be responsible, in part, for the occurrence of gall formation high up in the branches of H. edisonianum, however this was not investigated further.

Given the physiological adaptations known to occur in clonal plants, it became apparent that H. edisonianum response to seasonal flooding, drought, and attack by parasites may not easily fit the a general paradigm of pathogen-related mortality. Therefore, the hydrology and nutrient status of the seasonal ponds containing H. edisonianum were further investigated to better understand the plants response to stress and how fitness characters varied with environmental extremes.















CHAPTER 3
HYDROLOGY AND SOIL NUTRIENT CHARACTERS IN SEASONAL
POND POPULATIONS OF Hypericum edisonianum Introduction

The seasonal ponds found in south Florida, by definition, fill with water in response to rainfall in the summer rainy season and then gradually percolate down in the dry winter months to reach drought conditions by spring (Myers and Ewel, 1992). The depth and duration of flooding of Florida marshes and seasonal ponds are highly variable (Abrahamson et al. 1984, Duever et al. 1975, Pesnall and Brown, 1977) and within individual ponds the duration of standing water changes markedly from the palmettoedged margin to the center. Approximately 11% of the mapped land area of the Archbold Biological Station (ABS) property is represented by seasonal ponds (Abrahamson et al. 1984). These ponds occur in the presence of impermeable confining layers or where the water table emerges through the sandy substrates (Myers and Ewel, 1992).

Common vegetative features of seasonal ponds are saw palmetto margins (Serenoa repens) that enclose four major types of plant associations. These associations are maidencane (Panicum hemitomon), Hypericum (Hypericum edisonianum), cutthroat Grass (Panicum abscissum) and broomsedge (Andropogon brachystachys) (Abrahamson et al. 1984). Seasonal ponds, especially those near bayheads, are often invaded by tree species, particularly in the upper elevational zones (Landman and Menges 1999).


48






49


Soils of the ABS are deep sand, low in clay and silt content, with nutrients and drainage characteristics typical of dunes and sea terraces formed in the Pleistocene. Sanibel, Sellers and Placid soils are the only soil types that have a deep muck layer overlying mineral soils (Abrahamson et al. 1984). Hypericum edisonianum grows most frequently in Pompano depressional, Immokalee depressional and Placid soil types (Abrahamson et al. 1984).

Most Florida marshes have highly buffered waters as a result of the underlying limestone or calcareous substrates, with pH levels generally occurring in the neutral range (pH 7) except in flatwoods ponds where acidic groundwaters may be present (Myers and Ewel 1992). Nutrients are low in most acidic ponds where rainfall rather than upland runoff is the primary nutrient source (Myers and Ewel 1992). Peat soils in particular are low in the minor elements (copper, manganese, zinc and boron), high in organic nitrogen and low in phosphorous (Forsee 1940, Bryan 1958).

Fire has an enduring and profound effect on the landscape and vegetative communities of south central Florida and the ABS. Of the estimated 2100 to 2600 lightning strikes that occurred within the ABS property boundaries in a 14-year period, 30 strikes caused fires (Abrahamson et al. 1984). The frequency of fires (or fire return intervals) in these native scrub communities exert a strong selective pressure on the evolution of plant life history and reproductive strategies (Keely 1981, Ostertag and Menges 1994), species diversity (Johnson and Abrahamson 1990, Menges et al. 1993), landscape-level patterns of plant species abundance and their interactions (Menges and Hawkes 1998), as well as limit the invasion of woody vegetation (Myers and Ewel 1992).

In this study the hydrological and soil nutrient characters of seven seasonal ponds were examined to investigate the pattern of mortality in the Hypericum edisonianum






50


population of Pond 30-35. The greatest amount of infection by the fungal pathogen, Sphaeropsis tumefaciens occurred near the center of pond 30-35 in 1999 to 2000 (Chapter 2) and this distribution of disease suggests that environmental stress may predispose H. edisonianum plants to parasitic attack. Therefore, elements of the pond microenvironment likely to contribute to plant stress were investigated and these results used to evaluate their effects on plant fitness characters. The assumption was that fitness characters (stem heights, crown areas, flowering, new stem shoots) would reflect increasing stress effects of standing water from the pond margin to center.

Wild plant populations are frequently beset with multiple environmental stressors. For example, oxygen depletion in soils during flooding often leads to anaerobic microorganisms creating strongly reducing conditions, which in turn, can lead to toxic concentrations of Fe2+, Mn2, and H2S (Larcher 1995). The hypothesis was that stems near the center of Pond 30-35, known to contain the highest number of pathogenic fungal infections, also undergo the greatest extremes in soil chemical parameters (of all study ponds) due to flood conditions. Therefore, stem fitness characters would be most diminished. Such stress conditions may further lead to opportunistic insect herbivory which may then further contribute to pathogenic infections and stem mortality in Pond 30-35.

Materials and Methods

Meter-wide belt transects were randomly placed in seven study ponds (Ponds 8, 88, 750, 7-56, 7-64, 30-35 and 31-45) and wireflags were placed at one-meter intervals (Figure 2-2). Growth, reproduction and mortality measurements were made for each stem in these transects (Table 2-2).






51


On 20 February 2000, soil samples were collected at 2-meter intervals along each

permanent transect from the pond margin to the pond center. Each sample was labeled with the appropriate pond identification number and meter location, and left to air dry on paper towels in the ABS laboratory of E. Menges. Dried soil samples were then transported to Gainesville, Florida, and tested at the University of Florida Analytic Research Laboratory (Table 3-1).


Table 3-1. Soil analyses of samples taken from seven study ponds


Extractable Elements:
Macro nutrients: Phosphorus, Potassium, Calcium and Magnesium
Micro nutrients: Copper, Iron, Manganese, Zinc
Aluminum and Sodium
Water Extractable ions: pH, electrical conductivity, Cl-, NH4-N, N03-N
Organic matter content



The method used for extractable elements was as follows: 5.0 grams of mineral soils (or 1.25 grams organic soil) were mixed with 20 ml Mehlich-1 extractant (0.05 N HCl in

0.025 N H2SO4) shaken for 5 minutes and filtered. All elements were analyzed in this filtrate by inductively coupled argon plasma (ICAP) spectroscopy (Page et al. 1982).

Organic matter was determined using two different methods. Soils with 6% or greater organic matter were analyzed by loss on ignition (samples were heated in ovens at 450* for six hours.) Soils containing less than 4% organic matter were determined by the Walkley-Black dichromate method (Page et al. 1982)

Soil water content was measured along transects in each of the seven study ponds over the course of one late afternoon on 11 July 2000. Using a Cambridge Delta-T Devices Thetameter (Type HH11), in situ soil moisture was measured by inserting a probe






52

approximately 10 cm into the soil in the center of alternating m2 of the belt transect. Soil moisture (a soil/water volumetric ratio) was instantaneously recorded for both organic and mineral soils.

Long-term records of bi-weekly water depths in seasonal ponds were provided by ABS. These data consist of water depths collected at a PVC (polyvinyl chloride) pipe permanently located in the deepest portion of each pond. There were no long-term data for water depths from the pond margins to pond centers to use in conjunction with permanent belt transects established for this study in 1998. Therefore, soil moisture gradients were used to estimate how study ponds retained water along these transects.

Data analyses were performed using SPSS version 10 (SPSS, Inc.). Spearman's 2tailed correlation tests were used on soils data. Two-tailed tests were utilized because they are the most conservative for this nonparametric test and because sample sizes were unavoidably different for each pond. Simple linear regressions were used to examine plant fitness characters for the years 1998 to 2000.

A Trimble Global Positioning System (GPS) was used to measure the areas of five of the seven study ponds. The ponds located in the Red Hill tract of ABS (Ponds 8 and 88) were not measured using GPS because the pine canopy interfered with satellite signal capture. ArcView software was used for modifying (with permission) GIS (Global Information System) maps created by Roberta Pickert of the ABS/GIS Laboratory and for measuring linear distances between ponds.

Results
Hydrology

Mean water depths were compiled for 25 seasonal ponds from 1989 to 1999 and, as

expected, found to be quite variable among ponds (Figure 3-1). Only two of the 25 ponds






53

had mean water depths of greater than 40 cm. Eight ponds had mean depths of between 20 to 30 cm, ten ponds between 10 to 20 cm and five ponds with less than 10 cm. Seasonal ponds, however, often undergo substantial hydrological fluctuations relevant to vegetative communities, from year to year and month to month. An example of the yearly variation of mean water depth from 1991 to 1999 is presented (Figure 3-2) for five of the seven seasonal ponds (Ponds 7-50, 7-56, 7-64, 30-35 and 31-45) used in this study. The two ponds not included (Ponds 8 and 88) are remote, unrecorded and unmapped ponds on the ABS property. In 1997, the mean water depth for the five ponds was below

5 cm while during the following El Nina year, 1998, the same ponds all had mean depths of at least 20 cm. Pond 30-35, containing the highest frequency of fungus-infected H. edisonianum stems, was exceptional in that its mean depth was over 50 cm in 1998.

The seasonal ponds at ABS fill with water in response to rainfall. Monthly mean

water depths are generally at their lowest ebb in May or June and then gradually increase in response to summer storms (Fig.3-3). These five study ponds are located in the intraridge valley portion of ABS (Abrahamson et al. 1984) and are likely to be filled by ascent of a perched water table.

Soil moisture (or water content) measurements, taken from the palmetto-edged margin of the ponds to their center, were used to obtain an indirect estimate of the pattern of water retention or the basin configuration, of each of the seven study ponds in this study. All measurements were taken while the ponds were dry, during the drought of 2000. For example, in Figure 3-4 the soil transect in Pond 7-50 exhibited a short, steep gradient of increasing subsurface water content from the pond margin to meter 12 and then roughly constant water content from meter 12 to meter 33 (pond center). This profile matches field observations of this pond having a bowl-like basin that abruptly changes to a grassy,






54


dry, marginal shelf of <50 cm height. Hypericum edisonianum stems growing at the margin experienced very different hydrological conditions than those stems in the 'basin' of this pond.


50



40 -


C)
c


30



20-


10 -il


0


H


H


F1I


H


H


nih


Archbold Ponds


Figure 3-1. Mean water levels (in cm) of 25 ABS ponds, 1989 to 1999. Data provided by the Archbold Biological Station.

The study ponds varied in the distribution of soil moisture from edge to center

(Figure 3-4). Ponds 88, 7-56, 7-64 and 30-35 contain a relatively shallow subsurface soil moisture gradient from the margin and then a sudden increase of water content, as though there was a sandy mantle surrounding a large hole in the center of the pond. Field observations match these soil moisture profiles for at least three ponds. Pond 30-35









55


50



40 30



20 10



0
91


98 99


Year



Figure 3-2. Mean water depths for five study ponds at ABS from 1991 to 1999. Mean water depths in each pond vary from year to year. (Data were not available for some ponds in some months in 1989 to 1990. Data were provided by the Archbold Biological Station.





40


mar apr may


jun jul aug sep


oct nov dec


MONTH


Figure 3-3. Mean water depths (cm) in five study ponds from 1991 to 1999 by month. Pond water depths reflect seasonal rainfall patterns. Data provided by the Archbold Biological Station.


C WU W)


92 93 94 95 96 97


CL (D


30'





20 10


% . - - --


0 1


jan feb


7-50 7-56 7-64 30-35 31-45


7-50

7-64 7-56 30-35 31-45






56

particularly matches this soil moisture profile in that it has a gently sloping sandy margin and then water depths (during the rainy season) that increase markedly near the pond center (Figure 2-2). Pond 7-56 soil moisture readings do not entirely agree with field observations, however. In the field, the pond appears to have a gentle gradient wherein water levels increase very gradually from margin to center without there being a deep hole at the center. These field observations were confirmed after fires burned off all the vegetation in this pond during 12-13 February 2001 (Figures 3-16 and 3-17) to reveal only a shallow, sandy basin.

Ponds 8 and 31-45 have very shallow soil moisture gradients throughout the entire

transect (Figure 3-4). Pond 8 was not observed to contain standing water for the duration of this study (1997 to 2000), while Pond 31-45 is often rather dry and grassy during the rainy season except for a small, deep, water-filled hole at the very center of the pond.

As already noted, these soil moisture readings were taken in July 2000, when there

was no standing water in any pond. Both 1999 and 2000 were exceptionally dry years at ABS.

Seasonal Pond Soils

Analyses of soil samples taken in the permanent transects in each study pond showed that organic matter accumulated and was most concentrated in the center of Pond 30-35 (Spearman's rho P<.01, r2 = 0.52, Figure 3-5). Only two other study ponds showed this organic matter gradient, Pond 88 (Spearman's rho P<.01, r2 = 0.4) and Pond 7-64 (Spearman's rho P<.05, r2 = 0.8). These three ponds all were described above as having a shallow sandy matle with a 'hole' in the center. The other four study ponds do not accumulate organic matter in this pattern.






57


.600


.500 3 0a


.400


.300 x

+
.200 9 + + +

+ + + + + +
+ 4+;4+ .100 + +
..


.000 _


10


20


30


0


ci
0

x


+


+

+


4 +


.P8

* P88 + P750 x P756

+ P764

- P3035

-+ P3145


40


Meters from pond Margin


Figure 3-4. Soil moisture gradients in all seven study ponds, 11 July 2000. Pond 7-50 (P750) has a very steep soil moisture gradient and then levels off. Ponds 88, 7-56, 7-64 and 30-35 have very shallow soil moisture gradients that sharply increase towards the pond center. Ponds 8 and 31-45 display very small soil moisture gradients from margin to center. Note: pond margin at origin and final value is at pond center.


O)

E




U)

.0
(n


a













35.0
-0-- Pond 8 30.0 -- Pond 88
-A-- Pond 7-50 .250 - Pond 7-56
E )- Pond 7-64 20.0 --- Pond 30-35

X ---- Pond 31-45


0.

5.0

0 0


2 4 6 8 10 12 14 16 18 20 22 24 26 Meters from pond Margin


28 30 32 34 36


Figure 3-5. Variation in concentration of organic matter in pond soils. Meter 1 is at the pond margin. Increasing meter value indicates increasing proximity to the pond center.


25.0

-A- organic matter
20.0 -- potassium


15.0


10.0


5.0 .0


/

0

7.


1 2 3 4 5 6 7 8 9 10 11

Sample Collection Number

Figure 3-6. Pond 30-35 organic matter and potassium concentrations sampled at every other meter in 21-meter long permanent transect. Sample #1 was taken at meter 1 (located at pond margin) and Sample #2 was taken at meter 3, etc.


58








59


6.0


5.0


4.0


E3.0 U)U
2.0


NH4
--- N03

.0
2 4 6 8 10 12 14 16 18 20 22
Meters from pond Margin


Figure 3-7. Pond 30-35 soil concentrations of N114-N and N03-N by meter location in permanent transect in pond. Meter 1 is at pond margin.






250.0 -,


200.0


150.0


S100.0
0)

50.0 + Calcium

--M-Aluminum .0---
2 4 6 8 10 12 14 16 18 20 22
Meters from pond Margin


Figure 3-8. Pond 30-35 soil concentrations of aluminum and calcium by meter location in permanent transect in pond. Meter 1 is at pond margin.






60

The concentration of potassium in pond soil is closely correlated with organic matter in Pond 30-35 (Spearman's rho P< .01, r2 = 0.13) and follows the same pattern of increase towards the pond center (Figure 3-6). This relationship also occurs in all other ponds except those described as having very shallow soil moisture gradients (Pond 8 Spearman's rho P>0.35, r2 = 0.0008; Pond 31-45 Spearman's rho P>0.70, r2 =0.02).

Concentrations of NI4-N and N03-N are coupled together in all study ponds, and Pond 30-35 has a significant concentration gradient for both (Spearman's rho P< .01, r2

0.05) from the margin of the pond towards the center (Figure 3-7); however, only NIL-N is correlated with organic matter (Spearman's rho P< .01, r2 = 0.75).

The soil pH of Pond 30-35 decreases with distance from the pond margin to the center (from 5 to 3.8) and under such increasingly acidic conditions soil aluminum concentrations also increased. Surprisingly, the soils in Pond 30-35 also showed an increasing concentration of calcium towards the center that is contrary to the norm of acidic soils, which generally have very low calcium content (Figure 3-8). This calcium gradient did not occur in any other study pond. Stem Heights, Flowering and Recruitment Stem heights

Hypericum edisonianum stems in Pond 30-35 grew tallest in the center of the pond in 1999 and 2000 (Figure 3-9) where water depths (and the assumed hydric stress) were greatest and presumably of longest duration. Mortality associated with S. tumefaciens infections in the center of the pond was low (4%) in 1999 and then soared to 83% in 2000 (Table 2-6).






61


Regression of stem heights along all permanent sampling transects showed that stems increased with water depth (p<0.05) in all ponds except for Pond 8, the small dry pond in the northeastern portion of ABS (Figure 3-10).


Table 3-2. Mean stem heights for H.

Mean stem
Pond height (cm) 1998

8 29.2 (5)
88 53.0 (77)
7-50 70.2 (23)
7-56 61.5 (58)
7-64 74.3 (106)
30-35 82.9 (60)
31-45 44.4 (34)

Sample size in parentheses


edisonianum in seven study ponds

Mean stem Mean stem
height (cm) 1999 height (cm) 2000

78.0 (1)* 84.0 (1)*
62.1 (31) 71.1(60)
81.9 (22) 85.4 (19)
56.2 (88) 70.24 (80)
82.0 (81) 73.4 (113)
85.3(61) 83.4(41)
46.1 (37) 62.5 (17)


The two driest ponds (8 and 31-45, see Figure 3-4) also had the shortest mean stem heights in 1998 and the pond with the deepest mean water levels (Pond 30-35) had the tallest mean stem heights (Table 3-2 and Figures 3-9 and 3-10).

The population of H. edisonianum stems in Pond 8 appears to be in a fatal decline. Stems were small and scattered in this dry pond and four of the five stems found in the study transect in 1998 have since died. However, over the past two years new shoots have been appearing at the outside edge of this pond and are growing towards a lower, wetter depression in the adjacent abandoned two-track road. Crown area

The crown sizes of H. edisonianum stems in Pond 30-35 were correlated with stem height (Spearman's rho P< .05, r2= 0.08) and increased in size as stems grew taller toward the pond center (Figure 3-11). Although stems in Pond 30-35 were, on average, the tallest stems in this study, their crowns were not the largest. Crown areas of stems










62


Pond 30-35


a U









a I




a


a
m a


160


140 120 100


80 60


40 20


0


10


20


* 1998

Rsq = 0.0043

* 1999

Rsq = 0.6975

* 2000

Rsq = 0.6666


30


Distance from pond margin (m)


Pond 8


8

80

E
a
On


I A* *998
Rsq - 02752

* * . 1999
- - * Rsq 0.1265

02000
__ Rs- -0 1112
10 15 20 2.5 30 35 40 4.5

Distance frcm pond margin (in)


120 100 80 600


40~ 20


Pond 88








* mm


2 4 6 6 10 12 14


R*q= 0 164
0 1948 Rq= 01001

*2000
Rtq= 0_20


1


Distance from pond margin (m)


Figure 3-9. Regressions of stem height by transect meter for Pond 30-35 Hypericum edisonianum (top) and Ponds 8 and 88. Stem heights in Pond 30-35 were greatest at the pond center (meter 28) in 1999 to 2000.


Do A
flu


II)

E


0


7060,

50 401 30,
E

20, 10


a
















Pond 7-56
1201


j I



1998
Rsq=00445 1999
Rsq-0.1492 2000
Rsq = 0.1683
7 9 9 10 11 12


100



80 00 E 60


40


20.


o a




a a g aa


10 20 30


Distance from pond margin (m)







Pond 7-64
120

110


100
E







a a E.
a Rq - 0.05
. " . " -2000 s0 * .q-0.0216
0 2 4 6 8 10 12 14 16

Distance from pond margin (m)


Distance from pond margin (m)


140 120 1001 80. 01


401


20.


Pond 31-45


2 4 6 8 10 12 14 16 18


899- 0.4395 a1999 Raq - 0 5526 2000
RSq -0.4994


Distance from pond margin (m)


Figure 3-10. Linear regressions of stem height by meter in ponds 7-50, 7-56, 7-64 and 31-45. All populations of H. edisonianum showed increased stem heights towards the center of the ponds except for Pond 7-50.







Pond 30-35


1999 Rq= 0 .830


2000
" ggggReq = 00759
10 20 30

Distance from pond margin (m)


Figure 3-11. The relationship between crown size by meter location in Pond 30-35.


Pond 7-50
1401 ,120


00


63


100 90.



40. 20,

0 j


Rsq - 0.2121
* 1999 Rsq -0.1906 a 2000 Rsq - 02681


60 50 140 ~30


0 10


0
0


*


" .






64


Table 3-3. Crown areas of H. edisonianum stems from 1998 to 2000 Mean crown Mean crown Mean crown
Pond area (cm) 1998 area (cm) 1999 area (cm) 2000

8 9.5 (5) 29.6(1) 34.8(1)
88 10.0 (77) 18.7 (28) 22.1 (53)
7-50 17.7 (23) 23.7 (22) 28.9 (17)
7-56 15.4 (58) 13.9 (88) 20.1 (58)
7-64 20.4(106) 25.7(81) 21.9(91)
30-35 0 16.3(61) 22.5(10)
31-45 7.1 (34) 12.1 (19) 25.8 (2)

Sample size in parentheses


were compared across pond populations of H. edisonianum from 1998 to 2000. Those stems in Ponds 8 and 31-45 (the driest ponds) had smaller crowns than those in the other five study populations in 1998 (Table 3-3). The increase in size toward the pond center in Pond 8 is based upon only one surviving stem (Figure 3-12). The size increase of crowns at the pond center also occurred in Ponds 88, 7-50 and 31-45. Pond 31-45 is the pond with the soil moisture readings indicating that it is relatively dry with a deep, waterfilled hole in the center and Pond 88 is a pond with a sandy mantle with a hole in the center.

Flowering

Hypericum edisonianum flowers throughout the year and the flowering phenology of individual pond populations are quite variable. In this study populations were sampled only during June, July and August of each year and therefore it is probable that the flowering periods of some study populations were missed entirely. Figure 3-13 shows the mean number of stems in flower in the permanent transects for only four of the seven study ponds (Ponds 88, 7-50, 7-64 and 31-45). In all cases, flowering increased as a whole from 1998 to 1999 and then decreased from 1999 to 2000 except for Pond 88.






65


During the summer of 1998 there was high water levels in seasonal ponds and a subsequent drought in 1999 and 2000 when no standing water observed in any study pond.

Stem Recruitment

New shoots that appeared in Pond 30-35 were first clustered at the midway point of the permanent transect in 1999 and the following year were concentrated at the center of the pond (Figure 3-14). This pattern may reflect the general drying trend in the pond during 1999 to 2000, where as the pond basin lost soil moisture content at the margins the new growth followed the moisture gradient towards the center. Fire effects

In February 2001 a passing train set off sparks that led to an intensely hot wildfire that rapidly progressed through the drought-stricken landscape of ABS. Two study ponds, Pond 7-56 and 7-64 were in the path of this fire (Figure 3-15) and were entirely burned in less than a day. Figures 3-16 and 3-17 show Pond 7-56 as it occurred during the time of this study from 1998 to 2000 and then shortly after the fires of February 12-13, 2001.

On 16 June 2002 the post fire responses of H edisonianum populations in Ponds 7-56 and 7-64 were evaluated in a nonsystematic manner. Although all boundary flags of the permanent transects had been incinerated, the aluminum identification tags previously placed on each stem within the study transects still remained on the ground and were used as a guide to the former study population. In both ponds there were very abundant and robust stems of H. edisonianum. Most notable were the similar heights of the resprouts and the lack of any observable parasitic damage.















Pond 88 601


50


Rq = 0. 1417 o1999
Rsq = 00783


Rsq = 0.0110
5 1.0 1.5 2.0 2.5 3.0 35 4.0 45


40 30


20 10


0


RSq = 0.0353

* 0 1999
RSq=0.0084

2000
. I* * Rsq= 0.0207


i


2 4 6 8 10 12 14


Distance from pond margin (m)


Pond 7-50 60,


7


6(


5(


4C 13C


S20


10


5'
E






2(


Rsq = 0.0327 01999 Rsq= 0.135E O 2000


I Rsq=0.270E
6 7 8 9 10 11 12


Distance from pond margn (m)




Pond 7-56




aI



0



S1998
Raq=0.002' ' 1999

0 I Raq = 0.071i

g~ ~~~ i 2000
H , ,Rsq =0.003,


10 20 30


Distance from pond margin (m) Pond 7-64


Distance from pond magn (m) Pond 31-45


50.

0401




e i i o "1998 20 , s =


101 a * o Rsq = 0 1250
S 20M0
0 Rsq = 0.0591
0 2 4 6 8 10 12 14 16 Distance from pond margin (m)


60







0


o




00


6 8 10 12 14 16 18 20 22

Distance from pond margin (m)


Figure 3-12. Relationship between crown area and distance from pond margin (origin) to pond center. Ponds were sampled annually for three years. Data shown as linear regressions.


Pond 8 301 .


66


20


0


O 1996 Rsq= 0.1186 01969
Rsq = 0.2193


Req = 0.1643


a


g

a








-
a a g a


0


1(


0






































F ,r w7V


88 7-50


706050.


400 30a


20.


7-64 31-45


Pond (by I.D. number)



Figure 3-13. Flowering patterns in Ponds 88, 7-50, 7-64 and 31-45. The overall flowering of pond populations of H. edisonianum increased from 1998 to 1999 and then decreased in 2000. (New flowers included buds and old flowers included ovules only.)


2D. 154


10 M
D~SWU ft-i PoId mas in


I
21

3
* 4*
C


0-


1 x 31
Distanc fromn pond rmarin (mn)


Figure 3-14. Pond 30-35 recruitment of new stems 1998 to 1999 (left) and 1999 to 2000 (right). New stems were clustered in the middle of the transect in 1998 to 1999 and then shifted to the center of the pond in 1999 to 2000 (meter 27).


67


[1106d and new flowers,
1998

[Z W and new flowers,
1999

~d and new flOwers, U 2000


U)



0


E
<3


10


0


z






68


Discussion

The hydroperiods and water depths of each of the seven study ponds were

demonstrably variable from year to year, month to month and from the pond margin to the pond center. The H. edisonianum populations in Pond 30-35 experienced unusually wet conditions in 1998 where the mean water depths in the center were 50 cm. In the following years, this same plant population suffered record-breaking drought conditions. In this same time frame, the smallest population of H. edisonianum in the smallest pond (Pond 8) endured increasingly extreme hydric conditions. Even in the wettest year (1998) no standing water was observed in this pond. As drought conditions worsened in 1999 and 2000 the population of H. edisonianum in Pond 8 went into decline without any evidence of pathogenic interactions (Figure 2-4).

The soils of all the study ponds play a significant role in the duration of standing

water and size gradients in plant growth. Ponds 8 and 88 are located in the northeast Red Hill tract of Archbold Biological Station, where the sands of the region are deepest, and it is probable that the water table underlying Ponds 88 is perched. The sandy soils of Pond

8 rapidly percolated rainfall, whereas Pond 88, a larger pond, retained rainwater longer, perhaps due to the organic layer overlaying the sandy substrates below. Pond 30-35 also contained a bed of organic matter and retained water at greater depths and for longer periods than other ponds in this study.

The accumulation of organic matter in the seasonal ponds where H. edisonianum

grows appears to play a pivotal role in the microsite conditions that dictate plant growth. In the annual dry down of seasonal ponds there comes a point where hydric conditions become most favorable for soil microorganisms to decompose organic matter and thus make available essential plant nutrients (Comanor and Staffeldt 1978). In this study a







69





Pond 88





Pond 7-56 Pond 7-50

. Pon 7-64








414










V.J



it i






Pon 30O.p~hontb 9'~ LaU t

Pond 31-45. = a




Figure 3-15. Aerial photograph of Archbold Biological station showing station boundaries (wide white lines), study pond locations (blue symbols), and extent of Feb. 12, 2000 fire (yellow lines). Star is ignition point. Note extensive citrus plantings to the east (linear patterns created by rows of trees).






70


L


Figure 3-16. Pond 7-56 in July 1998. Permanent transects begin at the pond margin and end at the center of the pond (where the white PVC pipe is located).


>


Figure 3-17. Pond 7-56 three weeks after fires that occurred on February 12-13, 2001. The PVC pipe at the center of the pond was incinerated as were all transect flags. The center of the pond has the most new vegetation. Note the saw palmetto at the pond margins has also resprouted.






71


LM.








Figure 3-18. Multiple stems arising from rhizome (arrow) in Pond 31-45 (left photo). Hypericum edisonianum resprouting in Pond 7-56 three weeks after February 12-13 fires. Resprouting typically occurred within approximately 20 cm from charred adult stem (right photo).

significant organic matter gradient was found to exist from the pond margin to the center in Ponds 30-35, 88 and 7-64. These ponds were described as having a shallow soil moisture gradient from the pond margin and then a rather abrupt increase in water content as though there were a hole in the pond basin. The duration of standing water was greatest at the pond centers. Incomplete degradation of organic matter added to the retention of water (additional organic matter accumulation) as well as contributed to the gradient of decreasing sodl pH.

The soil pH levels in the seven study ponds never rose above pH 5.0 and were

typically approximately pH 4.0. Perhaps the seasonal water fluctuations contribute to the already acid conditions by means of truncating organic breakdown in each pond over many years.






72


Acidic soils will often limit the nutrients available to plants. The deep sands of the LWR and ABS are in themselves extremely acid and nutrient-deficient (Abrahamson et al. 1984) and the acidic conditions found in seasonal ponds may exacerbate the lack of available nutrients to H. edisonianum populations. In soils with pH levels of 4.0 or less, clay minerals are broken down and metal hydrous oxides are brought into solution (Larcher 1994). As a result, increasing levels of free ions of aluminum and heavy metals are released (Brunet 1994). This phenomenon appeared to be occurring in Pond 8 where aluminum concentrations of soils reached more than 4,000 mg/kg. Possibly the litter fall and leaching from the pine canopy contributed to these conditions. As acid conditions continue in these seasonal ponds, the breakdown of organic matter is further inhibited and nitrification also decreases. In a study of the effects of increasing Al, Mn and Fe at high acidity (pH 3.8) on a wide-ranging plant, Succisa pratensis Moench, Pegtel (1986) found that both xeric and mesic populations showed the same response curve. The author argued that these results may be indicative of the plant genetically differentiating into edaphic ecotypes that are able tolerate phytotoxic concentrations of these microelements.

Nitrogen uptake in plants is essential for the formation of amino acids that are used in the synthesis of nucleic acids and proteins. The low pH conditions of the seven ponds in this study appear to have limited nitrifying bacteria, resulting in consistently lower amounts of NO3 than NI-4 in the pond soils. The ratio of NH4-N and N03-N remained relatively constant in the seven seasonal ponds while the relative amounts of both nutrients increased in a concentration gradient towards the pond center. In a study using sand-culture grown H. perforatum, Briskin et al. (2000) found that the production of hypericins (hypericin and pseudohypericin) in the leaves of plants was correlated with decreased nitrogen levels. When supplemental nitrogen was provided to these same






73


stems the production of hypericins decreased. If stems of H edisonianum at the margin of Pond 30-35 are producing more hypericins than those stems in the center then it may account for the concomitant decrease in S. tumefaciens infections.

According to Liebig's "Law of the Minimum" plant growth is affected by the nutrient that is most limited in its availability. Plants that are limited by the lack of essential nutrients and minerals will often develop a dwarf growth form as a deficiency-stress strategy when cell elongation is limited (Grime 1979). This may explain, in part, why smaller stems of H edisonianum always occurred at the margin of all study ponds while those stems found growing at the pond center were consistently larger.

The crown sizes of stems were not predictable by their position in the pond in general. This may be due in part to stem density and/or genetic parameters, however crown size increased with stem height towards the pond center in Pond 30-35, negating the density argument.

The post fire response of Ponds 7-56 and 7-64 in June 2002 was that of abundant and robust regrowth of H edisonianum with no evidence of disease symptoms of any description.

The findings of this study do not support the argument that H edisonianum stems

found growing in the center of Pond 30-35 were predisposed to pathogenic attack (by S. tumefaciens) due to stress of inundation. Rather, stems at the center of this pond were taller, had larger crowns and regenerated new shoots significantly more than stems at the pond margin and this was consistently correlated with more abundant nutrient and mineral resources.













CHAPTER 4
GENETIC DIVERSITY IN Hypericum edisonianum Introduction

Disease is one of the many agents that affect the vital rates of a plant population

(Caswell, 1989). As a group, plant pathogens are capable of inflicting a wide range of damage in wild plant populations. Some fungal pathogens, such as the damping-off genera (e.g., Pythium or Rhizoctonia) are relatively nonselective and cause mortality mainly at germination and seedling stages. However most pathogens show some sort of genetic specificity that restricts them to a small range of hosts (Burdon, 1994). Although pathogens affect the individual plant genotype, it is the change that takes place in the various plant fitness characters as a whole that utlimately affect the size of the population.

There are numerous and often very specfic ways in which plant diseases affect hostplant fitness but, given that they are often unobserved in wild plant communities, their effects are probably greatly underestimated (Burdon, 1994). For example, in cases such as Ustilago infecting the flowers of Silene alba (Alexander, 1989) there is no question that the host plant fecundity is diminished by anther infection and, thus, genotypic variability in the next generation is also decreased.

Population-level genetic processes also affect a plant population's vital rates by

change in the presence or organization of alleles; manifested by the frequency of alleles and levels of heterozygosity (Schmeske et al. 1994). For example, in a study of the


74






75


critically endangered sentry milk-vetch (Astragalus cremnophylax var. cremnophylax) Travis et al. (1996) found that the smallest population of this plant suffered from an extreme lack of genetic diversity. The authors attributed this to a severe lack of suitable habitat (on the South Rim of the Grand Canyon) and a pronounced founder effect. Larger populations of milk-vetch on the South Rim also suffered from lower than expected genetic diversity; which was thought due, in part, to the extremely stressful site conditions and the periodic crashes this population experiences. Conversely, the allelerichness of a population can theoretically contribute to the population's increase by enhancing its evolutionary potential to survive unpredictable environmental change. Further, populations with the greatest amount of genetic variation are thought to suffer least from inbreeding depression and/or the effects of genetic drift (Burdon and Shattock 1980, Dinoor and Eshed 1984, Menges and Dolan 1998, Segal et al. 1980).

In a survey of literature involving research on rare and endemic plant species,

Schemeske et al. (1994) found that of 78 papers published between 1987 and 1992, only six papers addressed genetic variation in quantitative characters. These studies, for the most part, infer evolutionary potential of plant populations by using isozymes and polymorphic DNA to estimate allele diversity and heterozygosity levels. These observations on genetic variation are then used to formulate conservation and management plans for individual plant species (Van Treuren et al 1993, Watson et al. 1994, Menges and Dolan 1998).

In spite of the theoretical relationships believed to exist between genetic diversity and a species' persistence in nature, the relationship between the molecular markers used in these studies and fitness characters is often obscure and there have been no empirical






76


studies that directly link the genetic composition of wild plant populations with their growth rate or survival (Schemske et al. 1994).

Stochastic environmental processes that occur outside of a plant population have profound effects on its long-term survival. Theoretically, in small populations, chance events effectively remove alleles more often than in larger populations (Nei et al. 1975). However, population size is not necessarily the key factor that determines outcome during stochastic events in plant populations. Using matrix projection models on metapopulations of Furbish's lousewort (Pedicularisfurbishiae) Menges (1990) found that the persistence of this plant species could not be assured by simply protecting individual populations. Rather, to insure the long-term survival of P. furbishiae it was recommended that enough original habitat be protected to allow for continual recolonization of numerous, small, patches. This balance between extant plant populations and their local extinction and recolonization of habitat patches is referred to as metapopulation dynamics. Metapopulations can be described as a collection of individual populations that shift in their presence or absence in a landscape mosaic of available and appropriate habitat. If too many of the original habitat patches are lost, then the plant species in question is unable to periodically recolonize these patches and declines to the point where extinction is inescapable. Kareiva and Wennergren (1995) likened this process to the collapse of an epidemic following the threshold loss of susceptible host plants.

Thus, the long-term survival of rare and endemic species (such as H. edisonianum)

may be more dependent upon foresight in land management and conservation then in the plant's genetic character (Lande 1998, Menges 1991, Schmeske et al. 1994). Given the






77

rate of development in south central Florida this question is now becoming urgent for this and many other narrowly endemic plant species. Hypericum edisonianum populations

Hypericum edisonianum is one of the many endemic plant species found only in the

Lake Wales Ridge (LWR) region of Florida (Christman and Judd 1990). Its populations are widely scattered throughout patches of scrub habitat and are often rather small and isolated. This patchy distribution of scrub habitat in the LWR is attributable, in part, to the historical mosaic of fire occurrences throughout the region (Menges 1990) and human disturbance. Landscape-level dynamics affect how individual patches of scrub vegetation function and the overall abundance patterns of scrub species (Menges 1999, Menges and Hawkes 1998).

If disease effectively removes only a small proportion of genes from future

generations of H. edisonianum will the survival of this species be in jeopardy? Given that H edisonianum is a long-lived, clonal plant that is capable of capturing substrates in seasonal ponds with fast-growing ramets, it is possible that some individual ponds may contain only a few, unique genetic individuals (clones) of the plant. The pond populations of H. edisonianum, known only to occur in 4 counties in Florida, may constitute extant populations that have been greatly reduced from a formerly larger distribution. If this is so, then these pond populations are expected to suffer losses of fitness from the fixation of old mutations (carried by the founders) that are equivalent to the effect of many generations of new mutations (Lande et al. 1994). This loss of fitness may possibly have occurred in light of the ongoing destruction of H edisonianum populations and their habitat. In the last 50 years, the state of Florida has lost






78


approximately eight million acres of forest and wetlands from the expansion of human populations and subsequent clearing of land (Cox et al. 1994). The loss of scrub habitat, that contains the seasonal ponds that support H edisonianum in Florida, has been massive. Davis (1967) estimated that scrub habitat formerly covered approximately 417,000 ha (1.03 million acres) in the state. Later, in 1993, Kautz et al. found the land area in scrub habitat to be only 170,850 ha (422,000 acres).

Founder effects may constitute an important risk of extinction for populations that

have been suddenly reduced to very small effective size or that are already near the point of genetic inviability (Lande 1994). If populations are more gradually reduced in size however, or that initially have a substantial maximum growth rates, then the fixation of new mutations poses a more serious risk of eventual extinction (Lande et al. 1994).

Overall however, it may be the effects of short-term impacts (such as disease) on H edisonianum that affect its chances of reproductive success and long-term survival (Lande 1998, Menges 1991, Schemske et al. 1994).

In this study amplified fragment length polymorphisms (AFLP) (Vos et al. 1995)

were used to screen ten populations of H edisonianum. It was predicted that individual pond populations of this plant would have limited genotypic variation due to environmental barriers to gene flow in isolated pond populations and/or the clonal habit of H edisonianum.

Materials and Methods

Newly emerged leaves were collected from individually tagged H edisonianum stems in the permanent transects placed in each study pond. Each sample was placed in appropriately labeled, new plastic bags and stored briefly in a cooler until return to the






79


ABS labs. These leaves were cut in half and only the tips were used for extraction of whole genomic DNA. Tips of new leaves were used with the assumption that endophytic infection, if present, would be greatly minimized as well as would possible stores of secondary metabolites.

DNA extraction from H. edisonianum was very difficult. At least three different

procedures using C-TAB extractions were used without success. Grinding leaf or flower tissue with either laboratory-grade quartz sand or in liquid nitrogen does not result in good DNA yields. Upon the recommendation of Mark Whitten (UF Museum of Natural History) samples containing only two leaf tips were ground in mortars using only 1.2 ml of warm C-TAB buffer and 8 pl of mercaptoethanol. The tissue homogenate was then transferred to new 1.5 ml tubes and incubated at 650 C in a heat block for a minimum of 6 hours with occasional shaking. Longer incubation periods resulted in better DNA yields however M. Whitten has been successful using incubations of less than an hour. Moreover, attempting to extract large quantities of DNA from correspondingly large volumes of leaf tissue (in Oakridge tubes) consistently failed.

After incubation the tubes were briefly vortexed and then 500 jil of 24:1

chloroform/isoamyl alcohol added to each tube and vortexed again. Samples were then centrifuged at 14,000 rpm for 5 minutes to separate the phases.

The aqueous phase (containing the DNA) was pipetted into new 1.5 ml tubes using 100 pl lots to measure total volumes. Then 3 M sodium acetate was added to each sample using the following formula: sodium acetate volume in jsl = volume aqueous phase in gl X 0.04. The new volume was then used to calculate the amount of 100% isopropanol to add using this formula: isopropanol volume in l = volume aqueous






80


phase X 0.65. The samples were gently rolled to mix all the constituents together and then placed in a 40 freezer for approximately 1 to 2 weeks to allow for DNA precipitation.

After precipitation the samples were centrifuged at 14,000 rpm for 20 minutes and the alcohol poured off. The remaining pellet was washed with 1 ml of 70% ethanol and then left to dry on the laboratory bench.

Each sample received 25-50 pl of IX TE storage buffer and then placed in a 4' C

freezer for storage. This 'mini-prep' protocol usually resulted in yields from 5-35 pg/ml of DNA.

Amplified fragment length polymorphisms (AFLP)

Four different primer sets were used to generate DNA fingerprints for ten different populations of H. edisonianum. The first step in performing AFLP was to generate random restriction (DNA) fragments using two endonucleases (EcoRI and MseI). Then 3-base "adaptor" sequences were ligated to the ends of this DNA using T4 ligase in the second step. The DNA fragments with attached adaptor sequences served as templates for the first "pre-selective" amplification using polymerase chain reaction (PCR). Only DNA fragments with adaptors on each end will amplify exponentially (disposing of all other fragments).

After the preselective amplification, there were still far too many fragments present to allow for clear visualization of bands on an acrylamide gel. Therefore, another "selective" amplification was carried out using two more, different EcoRI and MseI primers. At this step fluorescent dye labels were attached to the EcoRI primer but not the MseI primer. The DNA fragments with EcoRI at both ends do not amplify well and fragments with only MseI are not visualized on acrylamide gels for lack of fluorescence.






81


Only DNA fragments with EcoRI and MseI at either end amplify well and appear on acrylamide gels due to the fluorescent dye marker on the EcoRI primer. After the second PCR amplification the number of bands appearing on acrylamide gels were greatly reduced from the previous restriction/amplification step and band scoring made feasible. Acrylamide gels were laser-scanned using an ABI 373 autosequencer to produce precise basepair measures of each DNA fragment peak (band).

For this study the most successful selective primer sets were:

EcoRl 3A: GACTGCGTACCAATTCACA EcoRI 4A: GACTGCGTACCAATTCACG MseI 3C: GATGAGTCCTGAGTAACAA MseI 3C: GATGAGTCCTGAGTAACAA EcoRl 4A: GACTGCGTACCAATTCACG EcoRI 3A: GACTGCGTACCAATTCACA MseI 4C: GATGAGTCCTGAGTAACTA MseI 2C: GTATATACAAATTATATAA Results

The AFLP screening gels revealed H edisonianum populations to be much more genetically diverse than anticipated. Unique haplotypes were scored for almost each individual in each population found in the Archbold Biological Station (ABS) property (Figure 4-1) while populations outside of ABS in Glades and DeSoto counties appeared to be slightly more homogeneous (Table 4-1). This was found to be the case using each of the four primer sets used to screen these samples.

Using the gel results that represented primers "3a3c" as an example (Figure 4-1), the ten most frequent polymorphic loci were tallied across all pond populations and a table constructed of their frequency of occurrence (Table 4-1). The Glades and DeSoto county populations consistently shared more bands then those populations at ABS. Pond 31-45 was the only ABS pond that appeared to be more similar to Glades and DeSoto populations then to ABS ponds in bandsharing. Interestingly, this pond is at the southernmost extreme of the ABS property nearest these county boundaries. Ponds 7-50






82


and 7-56 had virtually identical band patterns at each locus in every sample. These two ponds are in very close proximity to one another on the ABS property and these patterns suggest gene flow between the two ponds (Figure 4-2).

Pond 30-35, the population of H. edisonianum heavily infested with S. tumefaciens,

appeared to be the least similar to any other ponds studied in the ABS property. The total number of fragments appearing in lanes 13-15 and also those automatically called in electrophaerograms was quite limited. This may be evidence of genuinely lower genetic diversity in this pond or DNA samples that contained a greater amount of impurities than all other samples.

When the multilocus haplotypes for these same ten populations were broken down by individual plant samples per pond (Table 4-2) it can be seen that stems from pond populations (excluding Glades and DeSoto county populations) differed from one another by location in the pond. Stems from the edge (or pond margin) exhibited a higher frequency of band occurrences that those from the pond center. This pattern became even more marked when multilocus haplotypes from nine individual stems taken from a single transect (from pond margin to pond center) in Pond 7-64 were examined (Figure 43, Table 4-3). Those stems from meters 9, 13 and 15 (at the pond center) shared very few bands with stems in meters 6, 7, 8 and 10.

Discussion

The AFLP findings of this study are still very preliminary. Initially several leaf

samples were collected from each study pond along permanent transects and the DNA extracted. Unfortunately unforeseen events led to the curtailment of this study and the










6 7 8 9 10 11 12 13 14 15


16 17 18 19 20 21 22 23


Lane Plant # Location Plant location in population


Glades Pop. #1 Glades Pop. #1 Glades Pop. #1 DeSoto Pop. #1 DeSoto Pop. #1 DeSoto Pop. #1 Glades Pop. #2 Glades Pop. #2 Glades Pop. #2 Pond 88 Pond 88 Pond 88 Pond 30-35 Pond 30-35 Pond 30-35 Pond 8 Pond 8 Pond 7-56 Pond 31-45 Pond 31-45 Pond 31-45 Pond 7-64 Pond 7-64


High elevation Mid elevation Low elevation High elevation Mid elevation Low elevation High elevation Mid elevation Low elevation Pond margin Mid transect Center of pond Pond margin Mid transect Center of pond Center of pond Mid transect Mid transect Pond margin Mid transect Center of pond Pond margin Mid transect


Figure 4-1. Example of AFLP screening gel used to detect genetic differences among populations of H. edisonianum (blue bands). Red bands are size markers used for automated band calling (laser scans). Primer set 3a3c was utilized.


83


1
2
3
4
5
6
7
8
9 10
11
12 13
14 15
16 17 18 19
20 21 22 23


1
2
3
1
2
3
1
2
3
55
101
"c" 119 131 190 13 27 31
504 526 527 501
540










Table 4-1. Number of times alleles (fragments) occur in 10 populations of Hypericum edisonianum. Each population contained three samples and value in column indicates number of occurrences in these three samples.


Fragment

F71 F239 F100 F223 F87 F191 F198 F300 F321 F89 Pond


I I I . I

2 2 2 1

. . 3

. . 3


1 1


1 I I


1 1

2 2


3

3


3

3


I1


1 2 2


31-45 3 3 2 2 3 2 2 2 3 1
----------------------------------------------Glades 1 3 2 3 2 3 1 3 2 3 2


Glades 2 DeSoto


3 2 2 2 2 1 1 1 3 1

3 3 2 1 2 2 2 3 3 1


* 17/30 16/30 15/30 15/30 14/30 12/30 12/30 12/30 12/30 12/30
** 57% 53% 50% 50% 47% 40% 40% 40% 40% 40%
Primers 3a3c
* first value is total number of occurrences of fragment (band) in all samples and second value is total number of samples.
** the percentage occurrence of fragment (band) in the total number of samples.


8


88


7-50 7-56

7-64


30-35


I


I






85


1 1) 2 A 1 r 7 2 0 In i1 1)


ILane P1ant #


1
2
3
4
5
6
7
8
9 10 11
12


28 33
34 28 33
34 38 100
200 38 100
200


Sample JAcationl in po ihiin


Pond 7-50 Pond 7-50 Pond 7-50 Pond 7-50 Pond 7-50 Pond 7-50 Pond 7-56 Pond 7-56 Pond 7-56 Pond 7-56 Pond 7-56 Pond 7-56


Mid transect Center of pond Center of pond Mid transect Center of pond Center of pond Margin of pond Mid transect Center of pond Margin of pond Mid transect Center of pond


Figure 4-2. Comparison of banding patterns between Pond 7-50 and Pond 7-56. Lanes 1-6 are from Pond 7-50. Primer set 3a3c was used in the samples in the first 3 rows and primer set 4a4c was used in samples in lanes 4-6. Pond 7-56 samples are in lanes 7-12. Primer set 3a3c was used in the samples in lanes 7-9 and primer set 4a4c used in samples in lanes 10-12. These pond populations of H. edisonianum were sampled at the 'Margin', the middle of the transect and at the pond center. Ponds 7-50 and 7-56 are within 500 meters of each other and share more bands than any other ponds in this study.


T~ ~ ~ pi,-Pnt1 R h ]ctn npplto


Sam le location









Table 4-2. Multilocus haplotypes of 30 different leaf samples from 10 different Hypericum edisonianum populations.

Pond Plant # Loc. 71 87 89 100 191 198 223 239 300 321

8 13 C 0 0 0 0 0 0 0 0 0 0
27 E I 1 0 1 0 1 0 1 1 0
88 55 E 0 0 0 0 0 0 0 0 0 0
101 M 1 1 0 1 0 1 0 1 1 0
200 C 1 0 0 1 0 1 1 1 1 0
7-50 28 M 0 0 1 0 1 0 1 0 0 0
33 C 0 0 1 0 1 0 1 0 0 0
34 C 0 0 1 0 1 0 1 0 0 0
7-56 31 E 0 0 1 0 1 0 1 0 0 0
38 E 0 0 1 0 1 0 1 0 0 0
100 M 0 0 1 0 1 0 1 0 0 0
200 C 0 0 1 0 1 0 1 0 0 0
7-64 1 E 1 1 0 1 0 1 1 1 1 0
40 M 0 0 1 0 0 0 0 0 0 0
200 C 0 0 0 0 0 0 0 0 0 0
30-35 19 E 0 0 0 1 0 0 0 1 0 0
31 M I 1 0 1 0 0 0 1 0 0
90 C 0 0 0 0 0 0 0 0 0 0
31-45 4 E I 1 1 0 1 0 1 1 0 1
26 M 1 1 0 1 1 1 1 1 1 1
27 C I 1 0 1 0 1 0 1 1 1

Gladesl 1 1 1 1 1 0 1 1 0 0 1
2 1 1 0 1 0 1 1 1 1 1
3 1 1 1 1 1 1 0 1 1 1
Glades2 1 1 1 0 1 1 0 0 1 1 1
2 1 0 1 0 0 1 1 0 0 1
3 1 1 0 1 0 0 1 1 0 1
DeSoto 1 1 I 1 1 I 1 0 1 1 1
2 1 0 0 1 0 0 1 1 1 1
3 1 1 0 0 1 1 0 1 1 1


E=pond margin leaf sample, M= midtransect leaf sample, C=pond center leaf sample. The Glades 1&2 and DeSoto County populations were not in ponds but along roadways. Primers 3a3c.
























C)









0
Ia


-t

04


1.0












0

0


64- 14 1 jL0 - ) - A
* -A

*IJ -hpa






pfb~


o- o V) U









S~. 5' 44
00 8 9 . . 0a0


t'J ~~.1


a,


00










Table 4-3. Multilocus haplotypes of nine different H. edisonianum stems (leaves) from Pond 7-64 permanent transect. Meter 1 is at the pond margin and Meter 15 is at the center of the pond.

Fragment (in basepair units)

70 78 126 132 152 164 174 178 312 332 340 Unique alleles
Meter


0 0 0 0 0 0 0 0 0 0 0

1 1 1 1 1 1 1 1 1 1 1

0 1 1 1 0 1 1 1 1 1 1

1 1 1 1 1 1 1 1 1 1 1

1 1 1 1 1 1 1 1 1 1 1


2

3

6

5

2


1 0 0 0 0 0 0 0 0 0 0 23


1 1 1 1 1 1 1 1 1 1 1

0 0 0 0 0 0 0 0 0 0 0


5

6


15 0 0 0 0 1 0 0 0 0 0 0 18
F= DNA fragment size in basepairs Unique alleles are additional (not shown) fragments found only for that particular sample Primers 3a3c.


00
00


1


6

7

8

9

9 10


13






89


use of these additional samples. In most studies using dominant marker fingerprints sample sizes are larger per population allowing Hardy-Weinberg frequencies to be inferred. The sample sizes in this study were, at most, three stems per pond and sometimes it appeared that DNA impurities might have caused fewer or false polymorphisms to be generated in some samples. Nevertheless, the preliminary screening gels in this study do represent randomly restricted and amplified fragments of DNA that, at least in general, serve to illustrate the diverse genetic character of H. edisonianum populations.

Perhaps the most parsimonious explanation for this diversity is that pollinators

facilitate gene flow among ponds and that at least some viable seeds are able to survive. Although H. edisonianum is primarily clonal in its growth habit it does flower abundantly in some pond populations throughout most of the year. Many species of bees were observed to visit H edisonianum flowers throughout the course of this study however they regularly moved from flower to flower within the ponds more often than trap-line ponds.

In an unsystematic field survey done in 1998 it was not unusual to find seedlings free of rhizomatous connections to other stems in the area. In one circumstance several seedlings of H edisonianum were discovered growing outside the boundaries of ABS in experimental plots in old cow pastures. Low, wet areas had been converted to pasture approximately thirteen years before and H. edisonianum seedlings appeared only when the turf had been entirely removed. It is not known whether this constitutes evidence of an enduring seed bank or incidental seed dispersal.

In studies of rare and endangered plants found in vernal pools, Jain (1994) found that






90


species of meadowfoam (Limnanthes alba and L. douglasii) that had wide-ranging metapopulations were more variable and heterozygous than the narrowly distributed and inbreeding L. floccosa and L. bakeri. In the latter two species the seed bank played a critical role in the persistence of L. floccosa, particularly for very small populations.

Another possible explanation of the observed genetic diversity in the study

populations of H. edisonianum may have to do with somatic mutation occurring over long periods of time. The Florida scrub communities in which the seasonal pond populations of H. edisonianum grow, are believed to have persisted for at least 50,000 years (Watts and Hansen 1994). Although the age of individual populations of H. edisonianum are not known at this time it is reasonable to argue that some may be of great age. Some clonal plants have been recorded as living for extraordinarily long periods of time. For example, Kemperman and Barnes (1976) reported that a single clone of trembling aspen (Populus tremuloides) may have been more than 10,000 years old and covered an area of 81 hectares. In another case Steinger et al. (1996) documented clones of the sedge, Carex curvula, as being 2,000 years old. Klekowski (1997) argues that long-lived clonal organisms may increasingly accumulate somatic mutations and that it is this genetic load that may ultimately lead to the decline in sexuality (as offspring accumulate defective or lethal genotypes) and/or the extinction of the clone itself. The pond populations of H. edisonianum display vastly different flowering phenologies throughout the year. Whether this is simply a reflection of microenvironment conditions or genetic load is debatable. Nevertheless the results of the screening gels in this study do point out that each pond population of H edisonianum does contain numbers of unique haplotypes that may reflect a long history of accumulated somatic mutations.






91


Ultimately the cause of so many unique haplotypes in each population of H.

edisonianum cannot be explained by this study. However, the effects of this diversity may be very significant with regard to damage by pathogens. The success of mixed plantings (mixed genotypes) has been studied extensively in agriculture with the majority of cases showing that disease severity is attenuated. However, as Burdon (1987) points out, very few agricultural studies examine the effect of disease reduction coupled with the change in the plant's reproductive performance over the long term. A recent exception to this is a study by Brunet and Mundt (2000). Using wheat genotypes susceptible to different races of the pathogen Puccinia strfiformis (wheat rust), they investigated the effects of disease and competition on the overall fitness of the host genotype. They found that there were few significant interactions between host fitness and disease or competition.

Moreover, very few varietal mix studies have used the proper disease-free controls necessary to evaluate true disease reduction. Thus, if the resource requirements for the different varieties do not sufficiently overlap, then the survival of the mixtures will almost certainly always be greater than the component varieties (Burdon 1987). A notable exception to this is the recent work by Zhu et al. (2000). Using a comprehensive and controlled design that encompassed all rice fields in five townships in 1998 and ten townships in 1999 in the Yunnan Province of China, the authors tested the effectiveness of genetic diversity in planted rice fields. Those fields planted with a genetically diversified mix of resistant and disease susceptible varieties of rice had a 94% decrease in the blast disease caused by Magnaporthe grisea when compared to monotypic fields.

The infection of H. edisonianum stems by S. tumefaciens in Pond 30-35 presents an






92

interesting case. While this population suffered very high infection and mortality during the year 2000, there was very vigorous new shoot recruitment in the areas of the pond where most mortality took place. Perhaps the new shoots represent genotypes more resistant to the pathogen. Equally plausible is the argument that the clone is simply "cutting its losses" by dispensing with diseased stems and translocating nutrients to new shoots in the same nutrient microsite. Assessing these two opposing suppositions constitutes the future experimental work planned for this rare plant.




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EK31MBPKF_IS7K9K INGEST_TIME 2015-04-13T20:38:46Z PACKAGE AA00029982_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

FACTORS AFFECTING DIEBACK IN THE RARE PLANT Hypericum edisonianum (EDISON'S ST. JOHNS-WORT) BY G.A. VAN DE KERCKHOVE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULLFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2002

PAGE 2

ACKNOWLEDGMENTS Dr. H.C. Kistler deserves a great deal of thanks for letting me follow my interests and for the guidance he provided to me (and many other graduate students). We were all very fortunate to have had Corby as an advisor and/or committee member. Dr. Margaret Smither-Kopperl also deserves my thanks and everlasting gratitude for serving as a very active member of my committee in Gainesville. Margaret was unfailing in her good advice and sense of fair play and I will always count both her professional support and personal friendship as a highlight of my years in Gainesville. Dr. Eric Menges served as my advisor at Archbold Biological Station and without his interest, input and support it is fair to say this project would have sunk like a stone. I am indebted to him for his patience and allowing me to be a part of his exceptionally productive and hardworking lab. Dr. Andrew Ogram always provided me with good advice, insights and interesting discussions and I am indebted to him for his contributions as a committee member. I am also particularly indebted to Dr. Rachel Shireman for her help and insights on how best to deal with institutional quagmires and her offer of the new caddy for a field vehicle. In this same vein, I would also like to thank Dr. Ken Gerhardt for his professional input and guidance to individuals in my department. I am truly appreciative. I would also like to thank Dr. Mike Thomas, Dr. Wayne Dixon, Dr. Susan Halbert, Dr. Nancy Coile, Robert Leahy, Dr. Tim Schubert, and Penny Campbell for their friendship and support during my employment at the Florida Department of Agriculture and ii

PAGE 3

Consumer Services, Division of Plant Industry. My particular thanks go to both Tim and Robert for their time, advice and bench space during the crucial early stages of this project. Dr. John Heppner provided the tentative i.d. on my leafminer, Jeff Loetz photographed both the moth and leaf in Chapter 5 and David Davidson provided media recipes. Dr. Mark Whitten and Dr. Norris Williams, of the UF Museum of Natural History, were always willing to help me on short notice. In particular, I would like to thank Mark for his unstinting attention to details and tutoring me in AFLP protocols. I also would like to thank Dr. Edward Hoffmann for making my many years of teaching in the Department of Microbiology a genuinely pleasant and worthwhile endeavor. Dr. Stephen Mulkey, Dr. Kaoru Kitajima and Dr. Kevin Hogan of the UF Department of Botany provided me with equipment, software, support for the soils work, interesting discussions and kindness borne of old friendships. I would also like to thank Dr. Hilary Swain, Dr. Mark Deyrup, Roberta Pickert, and Dr. Christine Hawkes for providing me with important feedback, good ideas and a positive point of view during my visits to Archbold Biological Station. Dr. Richard Braithwaite (CSIRO, Darwin) first introduced me to the potential problems of pathogens infecting clonal plant species during my visit to his study sites at Kapalga, Northern Territory, Australia and I thank him for his input. The Florida Department of Forestry supported this research with a grant from the Plant Conservation Program and the Women in Agriculture Club provided support monies by granting me one of their Frances Summerhill awards. iii

PAGE 4

Lastly, I would like to thank Sally, Josh, Mathew, Elaine and Zachary Dickinson for their support and all those celebrations over the years and, as always, my heartfelt thanks go to Karen Graffius-Ashcraft, Dan Simberloff, CD. Smith, Diane DeSteven, Miss Beeks, N.N. Head, and the family of I. Bella and K. Kit. iv

PAGE 5

TABLE OF CONTENTS Eige ACKNOWLEDGMENTS ii ABSTRACT vii CHAPTER 1 INTRODUCTION 1 Host 12 Environment 14 Study Site 14 2 DISEASE AND ITS DISTRIBUTION IN Hypericum edisonianum POPULATIONS 16 Introduction 16 Materials and Methods 17 Results 26 Discussion 43 3 THE EFFECTS OF HYDROLOGY AND SOIL NUTRIENTS ON POND POPULATIONS OF Hypericum edisonianum 48 Introduction 48 Materials and Methods 50 Results 52 Discussion 68 4 GENETIC DIVERSITY IN Hypericum edisonianum 74 Introduction 74 Materials and Methods 78 Results 81 Discussion 82 v

PAGE 6

CHAPTER page 5 INSECT DAMAGE IN SEVEN POPULATIONS OF Hypericum edisonianum 93 Introduction 93 Materials and Methods 94 Results 95 Discussion 98 6 SUMMARY AND CONCLUSIONS 1 02 APPENDIX FORMULAE FOR FUNGAL CULTURE MEDIA 106 REFERENCES 107 BIOGRAPHICAL SKETCH 124 vi

PAGE 7

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy FACTORS AFFECTING DIEBACK IN THE RARE PLANT Hypericum edisonianum (EDISONS ST. JOHN'S-WORT) By G. A. van de Kerckhove August 2002 Chairperson: Dr. H.C. Kistler Major Department: Plant Pathology Hypericum edisonianum (Small) Adams & Robson (Ascyrum edisonianum Small), a state-endangered plant found in only four counties in central Florida, has been experiencing sporadic diebacks with no known etiology. This study was initiated to determine the causes of diebacks and whether narrow genetic variation combined with unique environmental factors place the plant at risk from disease. Two newly recorded fungal pathogens attack H edisonianum. Colletotrichum gloeosporioides causes foliar and stem lesions in H. edisonianum that can ultimately lead to stem death. However, no C. gloeosporioides infestations were found in any wild populations. The originally infected stem appears to have been the result of contaminated greenhouse conditions. Sphaeropsis tumefaciens also attacks H. edisonianum, causing the formation of woody galls and witches' brooms. Disease incidence ranged from 0 to 83% within seven field populations of H. edisonianum however S. tumefaciens infection was only significant in one population from 1998 to 2000. vii

PAGE 8

All study populations of the plant grew in seasonal ponds that were unique in their soil nutrient characters and hydrology. Taller stems of H. edisonianum were found centermost in most ponds and also contained the greatest number of galls. Plant stem heights were correlated with gradients in soil moisture, soil nutrients, organic matter content and soil saturation. Amplified fragment length polymorphisms (AFLP) indicated considerable genetic diversity in 10 populations of H. edisonianum sampled from 3 counties. A new, undescribed, leafmining microlepidopteran (Coleotechnites sp.) (Gelechiidae) inflicts extensive foliar damage on H. edisonianum but does not play a significant role in H. edisonianum stem mortality. In summary, neither fungal pathogens nor insect infestations are solely responsible for the observed decline in H. edisonianum populations. Rather, environmental factors, particularly drought conditions, play a pivotal role in the synergistic effects of both disease and insect damage. viii

PAGE 9

CHAPTER 1 INTRODUCTION All plant species are subject to pathogens at some time in their life cycle (Burdon 1991) and thousands of studies testify to the ongoing destructive capability of plant pathogens in modern agriculture. Most useful plants do not grow in efficient monocultures however. Rather, at least 60% of plants in global agriculture are cultivated in complex ecosystems that include varietal mixtures, intercropping and traditional subsistence polycultures (Francis, 1986). In West Africa alone at least 80% of cultivated land is in multi-use production while in Latin America staple crops are commonly found growing in polycultures (Francis, 1986). Lenne and Sonoda (1990) reported that more than 1050 million hectares in the tropics and subtropics are used for grazing and much of this land is covered in natural and mixed perennial pastures. In great contrast to agriculture, the occurrence, dynamics and short and long-term effects of disease in wild plant communities has received far less attention but is by no means any less significant. All important food crops originated in natural plant communities and wild, ancestral stocks are especially important for use in crop improvement as sources for disease resistant genes (Lenne et al. 1994, Lenne and Wood, 1991). Wild species have played key roles in the improvement of potato, tomato and wheat crops (Lenne et al. 1994) and yet these wild host-pathogen systems have received scant attention by plant pathologists (Burdon et al. 1990, Lenne et al. 1994, Lenne and Wood, 1991). 1

PAGE 10

Endemic pathogens and their associated diseases occur in all types of plants and native plant communities, from salt marshes to forests. In a literature survey of plant pathogens attacking British trees, approximately 1 5 fungal species were found per host tree taxon (Strong and Levin 1975). In a similar survey for North American plant communities, Strong and Levin (1979) found an average of 15 fungal pathogens per tree species, 7 for shrub species and 5 for herbaceous species. Most major groups of plant pathogens are well represented in native plant communities. For example, Nienhaus and Castello (1989) compiled the known viruses that attack tree species in native communities and their literature review found that of the 37 tree species assessed, each had at least one viral parasite and some species up to seven. Much earlier MacClement and Richards (1956) conducted a systematic survey of viruses in wild plants and found virus incidences of up to 10%. In single-species surveys of wild populations of Plantago spp., Cooper and MacCallum (1984) found at least 39 different viruses infecting this genus; while earlier, Hammond (1981) reported up to 64% viral infection of Plantago spp. in England. Five wild populations of Primula vulgaris were found to contain from 043% infection with the Arabis mosaic virus (AMV) with the mean infection level of 19% (MacKenzie 1985). Bacterial pathogens are not well documented in natural plant communities (Jarosz and Davelos 1994) and reports are limited. Nonetheless, bacteria have been reported to affect wild plants in diverse habitats. For example, Foster and Fogleman (1993) found 39 bacterial isolates from Organ pipe cacti, 19 from Saguaro and 16 from Senita cacti and grouped the isolates into 28 conspecific groups based on their fatty acid profiles. They

PAGE 11

3 concluded that the patterns of bacterial distributions in cacti were dependent on the chemical compositions of the host species. The pathogenic bacterium, Ralstonia solanacearum, has been found to affect eucalyptus trees in Brazil, Australia, China, Venezuela and South Africa (Coutinho et al. 2000) by invading vascular tissue and causing wilt symptoms. This disease was first reported for eucalyptus in 1980 (Coutinho et al. 2000). Bacterial wilts have also been identified as contributing to the decline of an endangered plant, Euphorbia barnardii, a serpentine endemic found only in the Northern Province of South Africa (Knowles and Witkowski 2000). Nevertheless, bacterial diseases are thought to be rather limited in distribution in wild plant populations while fungal pathogens and the diseases they cause are probably most prevalent (Burdon 1994). Plant diseases have affected the diversity of plant communities (Burdon 1 987), the outcome of interand intraspecific competition (Brunet and Mundt 2000), the genetic structure of populations (Lenne et al. 1 994) and, historically, the distribution of plant species. Perhaps the best-known cases of disease drastically altering the demography of a native plant species on a large scale in North America are the pandemics of the American chestnut (Castenea dentata) blight and Dutch elm disease. The fungus, Cryphonectria parasitica, the causal agent of chestnut blight, was accidentally introduced from Asia into North America in approximately 1904 and arrived roughly 34 years later (in 1938) in Europe (Griffin and Elkins 1986). This fungus kills chestnuts (and some oak species) by causing cankers that ultimately girdle the tree. From its point of origin in New York, the fungus rapidly spread throughout the natural range of C. dentata and by the 1950s close to 100% of all natural forest stands were infected (Brasier 1991). The

PAGE 12

4 epidemic spread with equal speed in Europe. However, by the 1950s the European chestnut trees began to spontaneously recover from the disease (Biraghi 1950). This remarkable turnabout in the epidemic was initially attributed to either the ascendancy of a deleterious mutation in the fungal pathogen or resistance in an increasing proportion of trees (Jarosz and Davelos 1995). It was later discovered that the trees were recovering because C. parasitica had been infected with a debilitating, double-stranded RNA hyperparasite (Van Alfen 1975). The impact of C. parasitica on the community structure of these forests depends partly on the preinvasion species composition of the forest (Burdon 1994). In some cases the chestnuts were simply replaced by other co-dominant species (Woods and Shanks 1959) while in other forests dominated by chestnut trees, both tree and shrub species diversity increased (Stephenson 1986). Dutch elm disease, caused by the fungi, Ophiostoma ulmi and O. novo-ulmi, is believed to have originated in either the Himalayas or Europe (Brasier 1990) and was first reported in Ohio in the early 1930s. Ophiostoma ulmi had previously appeared in Holland in 1921 and quickly spread across the European continent. This was a relatively mild strain of the disease and a large number of European elms survived infection. A second Ophiostoma pandemic, caused by the much more aggressive O. novo-ulmi, has rapidly destroyed both European and North American elms and promises to decimate the tree throughout its natural range in the northern hemisphere (Jarosz and Davelos 1994). Fungal plant pathogens are well documented in wild plant populations. For example, the root rotting fungus, Phellinus weirii, had notable effects on the structure of forest stands in Oregon and British Columbia. As in the cases of the Chestnut blight and Dutch

PAGE 13

5 elm disease, the dominant forest species, mountain hemlock (Tsuga mertensiana) and Douglas fir (Pseudotsuga menziesii), are effectively removed from the canopy by this fungal disease. Trees more resistant to the fungus, such as the subdominant species shore pine (Pinus contorta) and Western white pine (P. monticola) replace T. mertensiana. In the aftermath of infection, T. mertensiana constitutes only 5% of the canopy in contrast to 75% prior to infection. The more resistant species subsequently became the dominant conifers in these forests (McCauley and Cook 1980) and formerly rare herbs became common (Holah et al. 1993). Interestingly, the overall effect P. weirii may have on the restructuring of forest composition is determined by the hydrological conditions of the particular site infected. In forests where conditions are mesic, mountain hemlock (J. mertensiana) and the Pacific silver fir {Abies amabilis) are both particularly susceptible to the fungus however the fir is more resistant to infection. Thus the fir reestablishes at the site sooner but the species diversity remains relatively the same (Cook et al. 1989). In more xeric sites, the abovementioned subdominant, Abies amabilis, doesn't reestablish at the site first but rather a renewed sequence of pine recruitment occurs (Burdon 1991) and the species diversity of these forests is ultimately increased. This is the opposite effect the fungus Phytopthora cinnamomi has had on Australian eucalyptus forests. In southeastern Australia, the root infecting fungus, Phytophthora cinnamomi, is a broad host-range pathogen that has caused dieback to occur in whole eucalyptus forests; leaving only grasses and forbs (Weste and Marks 1987) which, in turn, adversely affected the entire assemblage of flora and fauna that are associated with or dependent on these landscapes. As in the case of the Chestnut blight and Dutch elm pandemics, the

PAGE 14

6 P. cinnamomi epidemic also appears to be the result of human activity. Before logging, the pathogen was present in these forests but unremarkable in its affects. As logging activities continued however, soil moisture increased as did soil temperature when vegetative cover was removed. Roads established for logging further altered soil drainage patterns and ultimately facilitated the spread of the fungus. Wildfires that frequently occur in these natural eucalypt ecosystems followed logging and may have further enhanced the spread of disease by weakening the remaining trees (Weste and Marks 1987). Another epidemic caused by fungus, Phaeocryptopus gaeumannii, causal agent of the disease Swiss needle cast (SNC), has had very similar effects at logged sites in the Pacific Northwest, particularly Oregon and Washington. Initially recorded in Switzerland in plantations in 1925 (Gaeumann, 1930) the disease was first observed in the Pacific Northwest in the early 1980s (Stauth 1997). Before the logging of native Sitka spruce and Western hemlock along the Oregon coast, this endemic fungus caused little damage (Meinicke 1939). After logging and particularly after the fires in Tillamook, Oregon, in the 1930s and 1940s, Douglas fir had been cultivated in commercial Christmas tree monocultures using multiple, nonlocal sources (Savonen 2000). Douglas fir is the only known host for this fungal pathogen (Filip 1998). In spite of warnings, commercial growers planted Douglas fir in the fog belt, an area that experiences high humidity from the Pacific coastline to approximately 19 kilometers inland (Gallob 2000). Tillamook is now considered the center or point of origin for this epidemic in the Pacific Northwest (Savonen 2000). In 1 996, aerial surveys of the Oregon coastline determined that 130,000 acres were infected with SNC. By 1997, 393,000 acres were infected (Filip

PAGE 15

7 1998). Although the fungus does not kill the trees, it retards growth and seriously diminishes their commercial value as Christmas trees or chip products. Armillaria luteobubalina is another aggressive root-rotting fungus found in Australia's native eucalyptus forests that has drastically affected the vegetative community. The incidence and severity of the disease has also increased at the hand of man with repeated select cutting of older trees in unlogged forest (Edgar et al. 1976). Logging frequency rather than intensity of logging was found to be the critical disease factor because each select-cut exposes remaining trees to inoculum from fresh-cut stumps (Kellas et al. 1987). Another Armillaria {A. ostoyae) has also had serious affects on the various pine species in the U.S.A. in both wild and cultivated systems but primarily on trees in stressed conditions. Epidemics caused by pathogenic fungi are still a concern for North American forests. Dogwood Anthracnose is currently a serious, spreading fungal epidemic occurring in the Pacific Northwest on Cornus nutalli and in the southeastern Appalachian mountain forests on the native Cornus florida. First reported in New York in 1980 (Pirone, 1980) and described by Redlin (1991), this fungus {Disculus destructiva) quickly spread from New York and Connecticut to northern Georgia by 1987 and throughout the entire southern Appalachian mountains (Anderson et al. 2001). The symptoms are dieback of lower twigs and branches of mature trees that progresses to the crown, while seedlings and young understory trees are killed outright. The origin of this fungus is unknown but due to the sudden onset and rapid spread of the disease it is assumed that the fungus was introduced (Anderson et al. 2000).

PAGE 16

8 In a study of wild dogwoods in the Cactoctin Mountain Park in 1984, Mielke and Langdon (1986) first reported that only 3% of the native dogwoods remained disease-free while 33% were dead. In a follow-up study in 1988 Schneeberger and Jackson (1989) found that all the native dogwoods were infected and 89% were dead. The same disease was found in Washington on western flowering dogwoods (Cornus nutallii) in 1979 (Byther and Davidson 1979) and by 1983 the disease had spread to Oregon, Idaho and British Columbia (Anderson et al. 2001). Disease in wild plants is not necessarily as obvious as the foregoing examples may suggest but rather can be pervasive, insidious and host specific. As early as 1933 Sampson observed the colonization of plant tissue by endophytic fungi. Bradshaw (1959) found that Agrostis tenuis and A. stolonifera were widely infected by the pathogen Epichloe typhina, an endophytic fungus that causes parasitic castration. Although field populations of these grasses suffered high levels of infection, there was little indication of this in the vigorous vegetative growth that resulted from early floral abortion. Nevertheless, the fungus is capable of having a profound impact on gene flow in these populations and consequently their population structure. Bradshaw (1959) noted that these endophytic infections remained parasitic under conditions where the host plant reproductive potential was diminished or suppressed. However, in long-lived host species where the environment is stable and seedling establishment is naturally rare, the endophytic relationship becomes more mutualistic than parasitic. Colonization by Epichloe species, known as e-endophytes, (Schardl and Clay 1995) and their subsequent in situ production of toxic alkaloids, can inhibit both mammalian and insectivorous herbivory, particularly in the grass tall fescue {Festuca arundinacea

PAGE 17

9 var. genuind) (Clay 1990). Tall fescue is an introduced, European invasive grass that covers over 15 xlO 6 hectares in the eastern U.S., of which two thirds are infected with the fungal endophyte Neotyphodium coenophialum. Infections are intercellular and produce toxic alkaloids that may contribute to the 'invasiveness" of the plant, its subsequent dominance in native plant communities and the decline of species diversity (Clay and Holah 1999). Secondary metabolites from e-endophytes are also known to facilitate protection of plants from parasitic nematodes by increasing host tolerance to drought conditions and enhancing growth and fecundity. Thus, endophytic infections span the spectrum from parasitic relationships with their host plants to unconditional mutualisms. Endophytes from the anamorph genus, Acremonium, fit the latter definition. Infection by this fungus has been discovered to increase the relative fitness of its host grass by way of discouraging herbivory while having no effect on the host's sexual reproduction (Clay and Leuchtmann 1989). The fungus exhibits no sexual stages itself and apparently depends on vertical transmission by infected seeds for its own survival. Clay and Leuchtmann (1989) found almost 100% infection by Acremonium species regardless of the presence or absence of herbivores. This is in contrast to the findings of Clay (1990 a, b) where the incidence of endophytic infections are often correlated with herbivore pressure. He found that infections were most often at low levels or absent when herbivory was low and relatively high when herbivore pressure was high. Pathogenic microflora may also ultimately play an important role in the renowned species diversity and hyper-dispersion of trees in Central American lowland rainforests. In 1970 Janzen, and in 1971, Connell both independently proposed models predicting the

PAGE 18

impacts of pests in natural systems. The Janzen-Connell model states that when both adult and juveniles are susceptible to the same pests, the pest pressure will be greatest (in a density-dependent manner) closest to the adults and that survival of juveniles will be higher with greater distance from conspecific adults. The net result of these interactions would be a decreased spatial aggregation of individual species. Augspurger (1983) tested the Janzen-Connell model and determined that a soilborne fungus, Pythium spp. can all but eradicate the naturally dense seedling beds beneath maternal trees. She also found an increase in seedling survival with distance from the natal site. In further studies, pathogens were found to play an important role in the density-dependent mortality of 6 of 16 species of rainforest trees (Augspurger and Kelly 1984). However, the adult trees were not the reservoirs of the pathogen and these results did not entirely meet the criteria of the Janzen-Connell hypothesis. Instead, damping-off fungi were the main source of mortality and the microenvironment colonized by dispersed seeds played a key role in the affects the fungi ultimately had on each species. In a study of 23 species of Australian rainforest trees, Connell et al. (1984) found that only one species had high mortality closer to conspecific adults. Nevertheless, the effects of an unidentified stem canker disease on the tree Ocotea whitei (Lauraceae) in lowland Panamanian rainforest did appear to support the JanzenConnell hypothesis (Gilbert et al. 1994). Both adult and juvenile trees were susceptible to the canker and the disease incidence was host-density dependent. Further, they found that, as the Janzen-Connell model predicts, the coincidence of spatial patterns of canker and host mortality support the role of disease in regulating tree distributions in the forest. The identity of the causal agent for the tree cankers had not been confirmed, however an

PAGE 19

isolate from the fungal genus Phialophora was found to induce cankers in Phoebe cinnamomifolia (Lauraceae). In contrast to the species-rich forests of Central America, the widely occurring monotypic stands of Spartina are also subject to pathogen attack with notably different consequences. The ergot fungus, Claviceps purpurea, was first recorded as infecting S. alterniflora marshes in North America in 1895 (Eleuterius 1970) and much later Boyle (1976) reported the fungus in Irish populations of & townsendii in 1960 (Gray et al. 1990). From the early 1960s, when the fungus was first detected on S. anglica swards along the English coast (Hubbard 1970), the infection has spread rapidly and reached epidemic levels (Gray et al. 1990). Spartina anglica contains very little genetic variation (Raybould 1989, Raybould et al. 1990) and this character, along with its demonstrated susceptibility to the fungal pathogen, would seem to consign this plant species to inevitable population declines throughout the English coast. This is not necessarily the case however if host populations manage to escape the pathogen during years of unusual climatic extremes where the fungus is limited (Mantle, 1980) or where phenological or morphological changes in the host preclude infection (Parker 1988). Gray et al. (1990) further argued that S. anglica populations may not be in jeopardy of massive declines because of the plant's low reliance on seedling recruitment for population sustainability and the absence of other, competitive plant species that might otherwise invade. They report that a hyperparasite, Fusarium heterosporum, has been discovered to colonize C. purpurea sclerotia and that this interaction may serve as a natural biological control of this epidemic.

PAGE 20

12 What role do plant pathogens play in endangered plant species with extremely limited geographic distributions? If disease strikes discrete populations of a rare plant species, particularly those with populations with (probable) limited genetic variability (Menges 1 999) is local population decline and extinction their irreversible fate? To address these questions, an endangered plant, Hypericum edisonianum (Small) P. Adams & Robson, commonly known as Edison's St. John'sWort, was selected to investigate small population epidemics. The overall objectives of this study were to determine whether the sporadic diebacks observed to occur in this narrowly endemic plant are solely attributable to the action of plant pathogens; and to determine the environmental factors that may contribute to this plants decline. A 3 -year demographic study was initiated in seven small, seasonalpond populations of H. edisonianum. Within these seven study sites permanent transects were established and annual measures taken on disease incidence, plant mortality, plant fitness characters, environmental stresses and herbivore damage. Preliminary assays on the genetic variability of H edisonianum were also conducted using amplified fragment length polymorphisms (AFLP). Host This study investigates the diebacks observed in the endangered plant Hypericum edisonianum (Small) P. Adams & Robson (Clusiaceae). Edison's St. John'sWort is a shrub with opposite, entire, glabrous leaves that are 1 5 to 25 mm long. The yellow flowers have four petals (12 to 18 mm) and 4 sepals (10 to 12 mm with cordate bases) and occur yearround, in sparse distribution but sometimes in great profusion at the Archbold Biological Station, Lake Placid, Florida. The flowers have numerous stamens

PAGE 21

13 and the fruits consist of small capsules with 3 to 4 locules (Ward 1 979). The thin bark is smooth, reddish brown to gray and often shows prominent leaf scars on stems with circumferences that can grow to 2.5 cm. The vegetative habit of the plant often generates hundreds of stems that are unbranched until the upper third of the stem, where they become multiply branched and spreading. It is common to find very dense clusters of stems in patchy distributions within seasonal ponds. As a member of the Order Theales and Family Clusiaceae, the Hypericum family consists of at least 68 species (USDA Plant Database 2002), with 21 species occurring in Florida. Ward (1979) suggested that the closest relative to H. edisonianum is H. stans, a much more widespread species found in the southeastern coastal plain. He argued that H. edisonianum is derived from H. stans as a result of Pleistocene flooding that isolated the southernmost distribution of H. stans in the Lake Wales Ridge region of Florida when the rest of the peninsula was still below sea level. The basis of this view was the persistence of both dry soil and fresh water ecotypes of H. edisonianum that evolved in Pleistocene refugia. Hypericum edisonianum is reported as occurring in only 3 counties (Glades, De Soto and Highlands) in the state of Florida (Wunderlin 1998) however a single population does persist in Polk County on the Avon Park Bombing Range. This particular population has been subjected to both fire and conversion to bedded pine plantation and continues to persist under a pine canopy. Hypericum edisonianum populations frequently grow in seasonal ponds (Abrahamson et al. 1984) that are found by the tens of thousands throughout the plant's 4-county geographic distribution. These ponds characteristically fill with rainwater during the

PAGE 22

14 rainy season in Florida (June to September) and then slowly percolate to dry, shallow basins when the rains subside. Although H. edisonianum also can be found growing in low, moist swales, this study was conducted only in seasonal ponds where plant populations are locally very abundant. Environment Hypericum edisonianum is a narrow endemic that occurs only in the state of Florida, particularly in the Lake Wales Ridge (LWR) region of the state. The ridge is 160 km long (Menges 1999) and characterized by the presence of ancient sand dunes and terraces formed when Pleistocene sea levels were much higher than today. As sea levels subsided to present day elevations, these paleo dunes and crests remained and presently follow a north-south configuration through central and south central Florida (Brooks 1981). Deep, nutrient-poor sands overlay marine sands, marls, clays and sandy limestone that constitute the Miocene Hawthorne formations (Puri and Vernon 1964). Fire, caused by lightning strikes, has had long-term effects on the Lake Wales Ridge vegetative community. Both spatial and temporal distributions of fires may have been instrumental in shaping the present-day mosaic of scrub vegetation in Florida (Menges 1999). Drought conditions often arise in these scrub communities and are most important during the winter and early spring months (Menges and Gallo 1991) however frequent dense fogs may ameliorate drought stress during these months (Menges 1994). Study Site This study was conducted at the Archbold Biological Station (ABS), Lake Placid, Florida at 27° 1 1' N and 81 °21' W (Menges and Kohfeldt 1995). This research station is located in a major paleo dune system in the southern Lake Wales Ridge (LWR) (White

PAGE 23

15 1958) and is dedicated to long-term ecological studies. The ABS privately manages 2081 hectares (5,141 acres) of undisturbed land that resides on coarse to fine beach and dune sands of Plio-Pleistocene age (Abrahamson et al. 1984). The vegetative communities at ABS are part of the Florida scrub ecosystem and include Southern Ridge Sandhill, Sand Pine Scrub, Scrubby flatwoods, bayheads and seasonal ponds (Abrahamson et al. 1984). These plant communities have remained intact and persisted for an estimated 50,000 years (Watts and Hansen 1994). Although the species diversity of Florida scrub systems is relatively low they contain a significant number of endemic flora and fauna (Menges 1999). The number of endemic and endangered plants found in the LWR region of Florida rivals that of any other ecosystem in the United States.

PAGE 24

CHAPTER 2 DISEASE AND ITS DISTRIBUTION IN Hypericum edisonianum POPULATIONS Introduction Hypericum edisonianum (Edison's St. John'sWort) is a globally rare plant that is only locally abundant in the protected land holdings of the Archbold Biological Station (ABS) in Highlands County, Florida. The plant's occurrence outside of the ABS properties is much more sporadic due, in part, to the rapid and continuing destruction of its habitat in south-central Florida. This shrubby plant forms conspicuous, sometimes massive stands in shallow, seasonal ponds found on the ABS property. Over time, a number of ABS scientists have observed diebacks of unknown etiology in these Hypericum ponds, as they are referred to on station vegetation maps (Abrahamson et al. 1984) yet there appear to be no temporal or spatial patterns of note. Hypericum edisonianum is one of the many rare and endemic plants found on the Lake Wales Ridge in south central Florida (Christman and Judd 1990). Until this study, H. edisonianum has received little research attention. Edison's St. John'sWort is related to the species H. perforatum, (St. John'sWort), the plant currently in favor and commercially grown for homeopathic products. Hypericum perforatum, while cultivated primarily for its antidepressant compounds (Barnes et al. 2001, Goodnick et al. 2001, Kim et al. 1999, Shelton et al. 2001) also has received substantial research interest for its antimicrobial activity (Pistelli et al. 2000), antibacterial (Schempp-Christoph et al. 1999), anti-fungal (Warr et al. 1992)) and anti-viral properties (Lavieetal. 1995). 16

PAGE 25

17 Hypericum perforatum, while the subject of ongoing research for its medicinal properties, is at the same time the target of extensive eradication efforts in regions where the plant is considered a serious pest (Campbell and Nicol 2000, Morrison et al. 1998). Also known as Klamath weed in North America, H. perforatum invasion into pastures has been controlled by the release of the beetles Chrysolina quadrigemina (Huffaker 1967, Huffaker and Kennett 1959), Chrysolina hyperici and the host-specific pathogenic fungus, Colletotrichum gloeosporioides (Morrison et al. 1998). Hypericum edisonianum displays no such invasive characters. Moreover, H. edisonianum potential for therapeutic applications is unknown. The urgent need for developing a better understanding of this rare plant and the potential disease processes underway in its many small populations was the catalyst for this study. In 1997, E. Menges, of the Archbold Biological Station, first suggested the need for further information on seasonal-pond diebacks of H. edisonianum and shortly thereafter the search for disease in several pond populations was initiated. Materials and Methods Disease Screening The first step in assessing the potential for disease presence in the pond populations of H. edisonianum entailed an ABS-wide (5,000 acre) search for disease symptoms and possible disease foci. A heavy-duty, 4-wheel drive vehicle, provided by ABS, was used to check Hypericum ponds in all tracts of the main station property. In each of 24 ponds visited a walking transect was started at the pond margin and all H edisonianum stems encountered, in progress towards the pond center, were examined for disease and/or insect damage. At the pond center, a new transect was begun, heading back toward the pond margin and ending approximately 5 meters from the previous start point at the pond

PAGE 26

18 margin. This search pattern essentially partitioned the pond into wedges wherein any Hypericum stem with observable symptoms or damage was collected. Several different symptoms were used to determine whether plant tissue was collected (Table 2-1, after Fox 1993). In ponds where diebacks had previously been observed or appeared to be presently underway, symptomatic stems were also uprooted and root tissue collected. Table 2-1. Disease and insect damage indicators in field surveys Chlorosis Lesions Galls Discoloration Pustules Abnormal growth Speckling Pitting Constrictions Spotting Holes Leaf curling, wrinkling Chewing Mines Exit holes All samples of plant tissue collected during these pond inspections (leaves, branches, stems and roots) were placed in new, appropriately labeled plastic bags and were stored temporarily in a cooler until return to the on-site plant ecology laboratory of E. Menges. These surveys were performed at irregular intervals from May 1997 to May 1999 with tissue sampling concentrated from May to August of each year. There are a multitude of diseases one might search for and various methods available to diagnose them in this plant species. As fungi are the most frequent causal agents of plant diseases (Burdon 1994), a decision was made to screen only for fungal pathogens with commonly used culture media: (water agar, potato dextrose agar (PDA), acid potato dextrose agar (APDA) and corn meal agar (CMA)). Selective media specific for culturing the fungal root pathogens Pythium spp. and Phytopthora spp. (PARP plates) (courtesy of D. Mitchell, University of Florida and D. Davidson, Division of Plant Industry, Florida Department of Agriculture and Consumer Services) were also used to screen root samples. The appendix contains recipes for all culture media.

PAGE 27

19 Symptomatic and asymptomatic tissue samples (stems and leaves) were surface sterilized with a 10% Clorox™ solution (0.5% sodium hypochlorite) for approximately 1 to 3 minutes (depending on how woody they were), rinsed with sterile distilled water; and blotted on autoclaved paper towels and then small pieces of tissue aseptically placed on culture plates. These plates were then sealed with Parafllm™ and left at room temperature in the laboratory under ambient light conditions. Tissue samples were also placed in sterile glass petri dishes with autoclaved, moistened paper towels. All culture plates and moist chambers were transported to the laboratory of Dr. Tim Schubert (Florida Department of Agriculture and Consumer Services, Disease Diagnostic Lab, Gainesville) for further incubation and identification. Selection of Ponds for Study Each seasonal pond selected for further study was chosen on the basis of its size and physical location on the ABS property, using digitally corrected aerial photographs (Figures 2-1 and 3-15). An attempt was made to select ponds that were roughly similar in size. Ponds are individually numbered at ABS and those used in this study were: 7-50, 7-56, 7-64, 30-35, and 31-45. The first number is the tract number on the station and the second number is the numeric count of that pond within the tract. Two unnumbered ponds in the northeast (known as the Red Hill tract) of the property were also included as study ponds and were designated Pond 8 and Pond 88. These are the two northernmost ponds in this study and they are located within 200 meters of each other beneath a partial pine canopy. Ponds 7-50, 7-56 and 7-64 are found in the north central portion of the ABS property within 500 meters of one another (Figure 2-1) and, except for 7-50, receive full sunlight. (Pond 7-50 is partially shaded by pines.) The distance between this cluster and Ponds 8 and 88 to the north are at least 1,900 meters. Pond 30-35 is located at the fence

PAGE 28

20 Figure 2-1 . False-color aerial photograph of Archbold Biological Station showing three study ponds (7-56, 7-64 and 7-50, circled from left to right, respectively). These ponds were selected for their close proximity to one another and thus greater probability of disease spread and gene flow (for later studies). All other study ponds were at least one mile from this cluster. Arrow indicates north. line of the property along State Road 8 in the southeastern portion of the property and is approximately 5,700 meters from Ponds 7-50, 7-56 and 7-64. It receives full sunlight. The southernmost pond in this study, 31-45, is separated from Pond 30-35 by 1,300 meters and also receives full sunlight. The distance between the northernmost ponds (8 and 88) and southernmost pond (31-45) is 8,456 meters.

PAGE 29

21 Study Transects In the seven selected seasonal ponds, meter-wide belt transects were randomly placed in each pond and then all individual H. edisonianum stems were tagged within these transects (Figure 2-2). To place the transects, a number was blindly selected from a random number table and used as a compass point to position an aluminum angle stake at the palmetto edge of the pond margin. From this stake, a meter tape was reeled out in a straight line to a PVC (polyvinyl chloride) pipe located in the deepest center of the pond. This pipe had been previously established by Kevin Main, land manager of ABS, for long-term hydrological studies. Two ponds in this study have no PVC pipe (Ponds 8 and 88) and their centers were estimated by measuring the pond diameter and then placing the transect line from the palmetto edge to the midpoint of that diameter. No long-term hydrological data are available for these two ponds. A wire flag was placed at every meter interval of these transects and then at one-meter widths to form a meter-wide belt transect (Figure 2-2). The length and number of transects varied by pond given the differences in individual pond sizes. All H. edisonianum stems within these transects were permanently tagged using 17-gauge galvanized wire to secure individually number-stamped aluminum disks. During the 1998-2000 censuses (between June and July of each year) the categories of data recorded were stem height, crown size, number of flowers, disease symptoms, insects present, mortality and new shoots (Table 2-2). Field Transplants The number of plants that could be removed for pathogenicity tests was limited as this

PAGE 30

22 is a State endangered species (Coile 2000). A State collecting permit was obtained to allow removal of a limited number of plants from the field for research purposes. Small (20 to 30 cm) apparently healthy, intact, vegetative shoots were collected from various ponds and transplanted to 1 5 cm diameter pots (Figure 2-3). A small tiling spade was used to punch four cuts into the root mass around the selected stem. Upon the fourth punch the whole root mass and stem were levered up and transferred by hand to sterile plastic pots that contained a shallow bed of autoclaved pine chips. The hollows around the root mass were filled in with additional field soil, firmly tamped down and Figure 2-2. Meter-wide transect (defined by yellow flags) in Pond 30-35 during dry period in 1998. The foreground is bare sand with sparse vegetation. The light green vegetation is healthy, flowering H. edisonianum (with some grasses and Lacnanthes caroliniana). The gray area in the center is dead H. edisonianum. Photo was taken by author standing at the palmetto edge of pond. then the whole pot soaked with pond water (or tap water when pond water was not available). Potted transplants were placed in shade during actual digging and then

PAGE 31

23 transferred back to University of Florida greenhouses after one to two days of recovery on the front veranda of the ABS Plant Laboratory. In the University of Florida greenhouses, the transplants were watered every other day for approximately one month or until the small shoots stabilized and began growing. Thereafter they were watered to saturation every week. After six months, the transplants (Figure 2-4) were moved to the Florida State University Greenhouse facility and watered every week using tap water until placement in a shallow, artificial pond located beneath a pine canopy on the FSU property. The transplants were not treated with nutrient supplements or pesticides. Table 2-2. Annual Pond Population Census Measures, 1998-2000* Stem height (from ground level to highest leaf in cm) Crown size (longest and shortest axis measured in cm (these values were multiplied and the square root used as the area of the crown) Flowers (old flowers, new flowers and buds) Disease symptoms (Table 2-1) Insects present/insect damage Survival or mortality of stem for year New shoots (recruits of the year) * All measures were taken between June and July of each year. Colletotrichum Isolation and Pathogenicity Tests Leaves with lesions were collected from only one transplanted stem that displayed these symptoms, surface sterilized and plated onto CMA and APDA as previously described. After incubation for four days at room temperature, the culture was scraped from the surface of the plate and placed in 50 ml of sterile water. Quadrant streaks were made on fresh APDA plates using these water slurries. Single spores were subsequently collected from these plates after 15 to 18 hours incubation at room temperature. Spore collection was done under magnification by cutting individual germinating spores from the medium using a sterile needle and transferring them to fresh APDA plates. After incubation for four days

PAGE 32

24 these cultures were scraped from the medium and suspended in sterile water. Conidial spore suspensions (10 6 spores/ml) from these cultures were then used to spray-inoculate entire, greenhouse-grown, healthy plants to run off with a hand sprayer. Control plants were sprayed in the same manner with sterile water only. Both control and experimental plants were enclosed in plastic bags 24 hours before and after spraying to facilitate infection. Three separate trials were performed using these procedures and the same initial fungal isolate. This fungal isolate was kept in continuous culture on APDA plates that were stored at room temperature under ambient light conditions. Sphaeropsis Isolation and Pathogenicity Tests Field-collected stem galls were also taken from H. edisonianum, surface sterilized, blotted and thin-sectioned using a sterile razor. Small fragments (2 to 3 mm long) of these sections were placed on APDA plates, sealed with Parafilm™ and incubated at room temperature. The uniformly dark mycelial cultures from these plates were identified as Sphaeropsis tumefaciens. These initial isolates were used to inoculate artificial wounds made on healthy appearing H. edisonianum stems (previously transplanted from the field). Wounds were created by making shallow, downward cuts into the stem (approximately 5 mm in length) using a sterile razor. Small squares of the growing fungus were then sliced out of the culture plate and embedded into the angled stem cut. The entire inoculated wound was wrapped in a small band of Parafilm™ to exclude contaminants and prevent desiccation. Control plants were similarly wounded and wrapped but received no fungal inoculum. Control and test plants were placed in greenhouses at the Florida State University, Tallahassee, and were maintained only with weekly watering for approximately six months before the pots were transferred to the edge of an artificial outdoor pond for another six months.

PAGE 33

25 Figure 2-3 Four shoots of Hypericum edisonianum taken from wild populations at ABS property, four days after transplantation. Figure 2-4. Greenhouse-grown stems of Hypericum edisonianum. These stems were collected from the field when they were approximately 30 cm tall (see above photo) and are now adult size (note meter stick in foreground) after six months of greenhouse residence.

PAGE 34

26Results Disease Screening A number of saprophytic fungi were repeatedly cultured from leaf, stem and root samples of field-collected specimens. These fungi belonged to the genera listed in Table 2-3. Approximately 500 plates with six separate tissue samples each were used to isolate these fungi over an approximate 2-year period. These genera consisted of saprophytic species and were isolated repeatedly but sporadically (data not shown). There was no visible correlation between isolation of these fungi and presence of any dieback symptoms. Therefore, these fungi were not considered further as causal agents of the H. edisonianum dieback. In-field visual inspection of dozens roots gave no indication of root disease. When the roots from dead stems were inspected, the tissues were still intact, however root hairs were absent. The stems were cross-sectioned in the field with a sharp jackknife and only dry, buff colored woody tissue was observed. PARP, APDA and CMA culture plates yielded no isolates of Pythium or Phytophthora from root samples. Table 2-3. Genera of saprophytic fungi isolated from H. edisonianum tissues Cladosporium sp. Curvularia sp. Phomopsis sp. Trichoderma sp. Nigrospora sp. Pestalotiopsis sp. Epicoccum sp. Koch's Postulates for Colletotrichum gloeosporioides Within 3 weeks of being placed on greenhouse benches, one of the new field transplants (selected in the field based upon its healthy appearance) displayed sporulating lesions on the upper surface of two leaves. These lesions first appeared as small brown specks that quickly expanded to approximately lxl cm in 1 week (Figure 2-5).

PAGE 35

27 The fungus, Colletotrichum gloeosporioides, was isolated from these lesions in pure culture on APDA plates and single spore cultures were used to experimentally sprayinoculate healthy appearing Hypericum edisonianum stems. This resulted in new lesions (the same in appearance as those lesions first observed on the original plant stem), appearing on experimental plants within 1 week. These lesions also first appeared as small brown specks that quickly expanded. Few new lesions were observed after the first week, indicating that the high humidity found within the bagged plants was essential for infection. Infected leaves remained on the plant for approximately 3 weeks before browning and falling. Control plants showed no signs of infection. The fungus recovered from the experimental plants was identified again as Colletotrichum gloeosporioides (Penz.) Penz. & Sacc. in Penz. and this identification was confirmed by Dr. Tim Schubert and Mr. Robert Leahy (Florida Department of Agriculture and Consumer Services). In preliminary trials, four plants were used for inoculations. Two plants were placed in water filled-basins while another two were left on the greenhouse bench with only occasional watering after receiving the spray inoculation. Two control stems were used for each treatment and were sprayed with distilled water. The experimental plants in saturated soils died within 4 weeks of inoculation with the fungal spore suspensions (1 0 6 spores/ml) while the two plants in drier soils survived for approximately 9 weeks before succumbing. The control plants showed no signs of infection or decline. The second round of trial inoculations (using spore suspensions prepared in the same manner as in preliminary inoculations and used at 10 6 spores/ml, resulted in 9 of the 10 experimental plants becoming infected. None of the 10 control plants exhibited lesions. In the third round of trial inoculations eight of the 10 experimental plants showed lesion development while none of the control plants were symptomatic. In both the second and

PAGE 36

third inoculation trials, all plants had to be moved out of the greenhouses and placed on benches outdoors in November temperatures. Lesions that were enlarging while in the greenhouse ceased to do so after placement on the outdoor benches and experimental plants did not die. Figure 2-5. Colletotrichum gloeosporioides stem and leaf lesions on H. edisonianum plant from Koch's Postulates trials. The fungal pathogen Colletotrichum gloeosporioides was reisolated on APDA plates from inoculated plants and confirmed by microscopic inspection to consist of 1 -celled, hyaline conidia (Sutton 1980) considered to be the causal agent of these lesions in H. edisonianum plants in all experimental trials. Repeated field searches for symptomatic plants throughout the study (1998 to 2000) were unsuccessful and C gloeosporioides was never recovered from any wild population. The single transplant shoot initially found with C. gloeosporioides lesions (Figure 2-5) appears to have been infected within the first 3 weeks of residence in the University of Florida greenhouse.

PAGE 37

29 Koch's Postulates for Sphaeropsis tumefaciens Woody galls, ranging in size from slight swellings to large (5 cm in length), fissured growths were frequently observed on stems in Pond 30-35 (Figure 2-6) from 1998 to 2000. The conducting tissue above and below large field-collected galls (3+ cm in length) appeared to become stained, perhaps by the colonization of the dark hyphae of the fungus. This dark staining of tissue progressed several centimeters in each direction from the original point of infection (Figure 2-7) and was found in both naturally occurring and artificially induced infections. Experimental wound inoculations of healthy H. edisonianum plants (in Materials and Methods section) resulted in the formation of small galls (2 cm or less in length) after approximately 1 year of incubation (Figure 2-8). The pycnidia of S. tumefaciens were dark brown and gave rise to mostly unicellular conidia that were 20 to 34 mm long. Plating fragments of surface-sterilized stems galls consistently yielded pure cultures of the fungus. Sphaeropsis tumefaciens was not isolated from control stems. The causal fungus was identified as Sphaerospsis tumefaciens Hedges (Hedges 1911, Hedges and Tenny 1912, Sutton 1980) and this identification was confirmed by Dr. Tim Schubert and Mr. Robert Leahy. Three separate inoculation trials were conducted. In the first, eight experimental and eight control stems were utilized. Seven of the eight experimental stems contained galls after a 1-year incubation. No control stems had galls. The second trial consisted of six experimental and six control stems. All experimental stems contained galls while controls had none after 1-year incubation. The last trial inoculations used the remaining seven mature, greenhouse-grown stems (five experimental, two-control). After 8 months of incubation, four of the five experimental stems had galls and the control stems none. All experimentally induced galls were collected, surface sterilized and small fragments were

PAGE 38

30 plated on APDA media which were later used for microscopic identification of the fungus; thus it is unknown whether these galls would have eventually caused stem death. Field Infections of S. tumefaciens Pond 30-35 had the only population of H. edisonianum with a high proportion of stems infected by S. tumefaciens (Table 2-4). In 1998, 64% of the stems censused in the transect carried galls. In the following year (1999) approximately half of the stems in this same transect were infected. This decrease in percentage of stems infected was a reflection of the increase in newly produced stems (stem recruitment). In 2000, the population of H. edisonianum in Pond 30-35 crashed (Figure 2-9) and this was reflected in the transect sample data where 83% stem mortality was recorded (Table 2-4). Mortality in 2000 was higher than in 1999 for most populations of H. edisonianum without S. tumefaciens infections or any other detectable disease (Table 2-4), such as in Pond 8 (47%) and Pond 3145 (44%) (Figure 2-9). Spatial Distribution of & tumefaciens Infections On an ABS-wide scale, Pond 30-35 appears to be the infection focus for S. tumefaciens infection in H. edisonianum on the ABS property. During the three years of this study a continuing ABS-wide vigilance for infection was maintained and yet no other pond population of//, edisonianum was observed to contain more stem and branch galls. The greatest percentage of stems infected in Pond 30-35 during 1998 was near the pond margin and then in 1999 and 2000 shifted towards the center (20 to 25 meters from the pond margin) (Figure 2-11). When the status of stems was examined in a cross-section of the pond (in a transect that spanned across the entire pond from edge to edge) mortality was

PAGE 39

31 Figure 2-6. Multiple galls caused by Sphaeropsis tumefaciens on wild Hypericum edisonianum plants in Pond 30-35 at Archbold Biological Station. I II 12 3 4 Figure 2-7. Staining of tissue that occurred above and below field collected and experimentally induced stem galls. The first stem section at left (1) is from below gall and second stem section (2) was excised from above the same gall. The third section from the left (3) was taken from below the gall and the fourth section (4) from left was taken from above the same gall.

PAGE 40

Figure 2-8. Experimentally induced gall on stem (left) and lack of gall formation on control stem (right) after one year of incubation. Arrow indicates where control stem received wound but no inoculum. Table 2-4. Percentage of stems infected and mortality in H. edisonianum Stems Stems Stems Stem Stem Pond Infected Infected Infected Mortality Mortality ID. 1998 (%) n 1999 (%) n 2000 (%) n 98 -99 (%) 99-00 (%) 8 0 19 0 19 0 17 21.0 47.0 88 0 63 0 47 0 80 25.5 32.5 7-50 0 89 0 89 0 89 12.0 38.0 7-56 0.01 86 0.007 137 0.008 128 18.0 26.0 7-64 0 106 0 107 0.008 116 19.0 19.0 30-35 64.0 150 48 179 24.0 188 4.0 83.0 31-45 0 51 0 9 0 89 27.0 44.0 Note: the number of stems and galls varied from year to year due to stem mortality, missing stems and recruitment of new stems into each pond population each year.

PAGE 41

33 concentrated in the center in 1999 (20 to 25 meters from the pond margin) and then was throughout the pond in 2000 (Figure 2-12). Stem mortality was examined in all study ponds (Figures 2-13 through 2-18) and the only other ponds to experience the most stem mortality in the center was Pond 7-64 (Figure 2-17) in 1998 to 1999 and Pond 88. These ponds had no S. tumefaciens galls. Stem mortality in Pond 8 increased at the pond margin from 1998-2000 as it did in Ponds 7-50 and 7-56 (Figures 2-13, 2-15 and 2-16). Stem mortality increased in all but the very center of Pond 31-45 (Figure 2-18). Figure 2-9. Uprooted stems from a single pond showing galls at same relative heights. The survivorship of stems in Pond 30-35 was not significantly affected by the presence ofS. tumefaciens infections. The presence of stem galls, caused by S. tumefaciens, in 1998 wasn't correlated with stem mortality in 1998 to 1999 (Pearson =.357, df =1, p > .10). There was no significant relationship between stem galls recorded in 1999 and stem mortality in 1999 to 2000 (Pearson £ =2.245, df= 1, p > .10). The means by which 5. tumefaciens is dispersed within Pond 30-35 is unknown, however, it appears that water may play a key role. Figure 2-9 shows multiple, uprooted stems of//, edisonianum that were taken from a single pond. These stems have galls at the same relative height that suggest spore dispersal may be facilitated by floating on the water surface.

PAGE 42

34 Stem survival, 1998-1999 140120' 100' n Pond by Archbold I D. number Stem survival, 1999-2000 120' 100' 8 88 7-50 7-56 7-64 30-35 31-45 Pond by Archbold I D. number Figure 2-10. From 1998 to 1999 Pond 30-35 had the lowest proportion of dead stems in the permanent transect of all study ponds (4%). By 2000 Pond 30-35 suffered the greatest proportionate mortality of all ponds (83%).

PAGE 43

35 Figure 2-11. Distribution of S. tumefaciens infections (galls) from 1998 to 2000 in Pond 30-35 (percentages represent stems infected for each meter of transect). Occurrence of galls increased with distance from the pond margin (origin). .

PAGE 44

36 Pond 30-35, 1998-1999 16 > 14 TO pu 12 ro 13 co 10 dj •o w E 8 S H— o 6 o> cn 2 4 c 0) o l_ 0) 2 CL flU i a i n HI dead I I alive 1 7 9 11 13 15 17 19 21 23 25 27 Distance from pond margin (m) Pond 30-35, 1999-2000 1 7 9 11 13 15 17 19 21 23 25 27 Distance from pond margin (m) Figure 2-12. Distribution of dead and living stems in Pond 30-35 for 1999 and 2000. Percentages are based upon total number of stems found in each category (alive, dead). Meter 1 starts at margin of pond and Meter 27 ends at other margin of pond.

PAGE 45

37 B 55 To "S 20%Dead 1999 1 10 20 30 Meters from pond margin Alive 1999 Meters from pond margin en E aj 3o%co "5 o _ Dead, 2000 20 Meters from pond margin J 10 Alive, 2000 30 Meters from pond margin Figure 2-13. Distribution of dead stems in Pond 8 from 1999 to 2000 was greatest at or near the pond margin (pond center is located at transect meter 1 1).

PAGE 46

38 Meters from pond margin Meters from pond margin Figure 2-14. The distribution of dead stems in Pond 88 increased with distance from the pond margin from 1998 to 1999 and then became more widely distributed throughout the transect from 1999 to 2000.

PAGE 47

39 in E Dead 1999 Meters from pond margin Meters from pond margin U5 E £ 55 ^ Dead, 2000 L 10 20 30 Meters from pond margin Meters from pond margin Figure 2-15. Distribution of dead stems in Pond 7-50 was greatest near the pond margin from 1998 to 2000.

PAGE 48

Meters from pond margin 10 20 30 Meters from pond margin Figure 2-16. Stem mortality in Pond 7-56 occurred throughout the transect from 1998 to 1999 and then became more prevalent at the margin and midtransect from 1999 to 2000.

PAGE 49

41 Alive 1999 M . 0 10 20 30 0 10 20 30 Meters from pond margin Meters from pond margin Meters from pond margin Meters from pond margin Figure 2-17. Dead stems in Pond 7-64 were most numerous near the pond center from 1998 to 1999 and then became more prevalent throughout the transect from 1999 to 2000 (the pond center is located at meter 15).

PAGE 50

42 Meters from pond margin Meters from pond margin Meters from pond margin Meters from pond margin Figure 2-18. Dead stems were most concentrated at the margins of Pond 31-45 from 1998 to 1999 and then where found throughout the transect from 1999 to 2000.

PAGE 51

43 Discussion In this study two fungal pathogens, C. gloeosporioides and S. tumefaciens, were discovered to cause diseases in the rare plant H. edisonianum. This is the first record for both of these pathogens on this host. Although the C. gloeosporioides infection appeared to be contained on greenhousegrown transplanted stems in this study, this pathogen may pose a serious threat to the small, extant populations of H. edisonianum in the state of Florida. Trial inoculations of healthy plants using spore suspensions of this fungus resulted in consistently high infections (100%, 90%, and 80% respectively for the three different trials). This fungus is already well known for its virulence in H. perforatum and has been under development as a natural biological control agent for Hypericum spp. in the U.S., Canada and Australia (Hildebrand and Jensen 1991, Shepherd 1995, Morrison et al. 1998). Colletotrichum gloeosporioides was first isolated from citrus plants in 1886 in Florida and causes the disease Post bloom Fruit Drop (Liyanage et al. 1992, McMillan, Jr. and Timmer 1989, Agostini and Timmer 1994). This is a wide host-range pathogen that causes disease in numerous plant species in Florida (Alfieri et al. 1994, Timmer et al. 1994) and worldwide (Bernstein et al. 1995, Freeman et al. 1998). In citrus, the disease causes necrotic petal lesions and premature fruit drop while in H. edisonianum the leaves and stems develop quickly expanding, necrotic lesions. The Archbold Biological Station property is located in the midst of an extremely dense concentration of commercial citrus operations (Figure 4-2). Over time, the upland native plant communities, especially sandhill and Florida scrub, have been converted to citrus. Furthermore, it has recently been discovered that C. gloeosporioides causes premature fruit drop in saw palmetto (Serenoa repens) (Carrington et al. 2001), a

PAGE 52

44 ubiquitous species found throughout Florida scrub plant communities and commonly found at the margins of seasonal ponds. Thus, the potential for a continuous source of C. gloeosporioides inoculum is ever present for the remaining H. edisonianum populations. This fungal pathogen was only recovered from one transplanted stem of H. edisonianum during a six-month residence in the University of Florida teaching greenhouse during 1998. Isolates of this single infected stem were used in all three Koch's Postulates trials. However, it is quite possible that undetected infestations may be present in populations of H. edisonianum not sampled in the ABS property or in distant ponds outside of the ABS property. The fungal pathogen Sphaeropsis tumefaciens was also discovered to infect H. edisonianum. Sphaeropsis tumefaciens is also a citrus pathogen (Rodiguez and Melendez 1984), having been introduced into the U.S. sometime in the 1930's (Holliday and Punithalingam, 1970). In citrus and H. edisonianum alike, the fungus causes hypertrophic tissue or galls and witches' brooms to form that were observed to persist for years on both the primary stems and upper branches of H. edisonianum (Figure 2-6). Over time these galls slowly enlarge and become fissured. Sinclair et al. (1987) report that galls shed conidia (mitospores) from embedded pycnidial structures, which serve as a chronic source of inoculum. Sphaeropsis tumefaciens has a broad host range and has been recorded as attacking numerous plant species in Florida (Alfieri et al. 1994, Marlatt and Ridings 1974, Marlatt and Ridings 1976) and is of particular interest as a possible biological control agent for the Brazilian pepper (Schinus terebinthifolius) (Marlatt and Ridings 1979).

PAGE 53

45 In a station-wide search at ABS this disease was found in several seasonal pond populations of the plant however not more than 10 stems were affected in any one pond except for Pond 30-35. Of the seven pond populations of H. edisonianum censused, only those in Pond 30-35 contained numerous stems with S. tumefaciens galls. Infection of stems was greatest near the pond center from 1999 to 2000 where presumably plants would experience the greatest flooding stress. How S. tumefaciens infections were distributed within Pond 30-35 is of interest because of the dynamic character of water levels in all seasonal ponds. Stems of H. edisonianum at the pond center experience much longer hydroperiods than those stems at the pond margin. Could this disparity of soil saturation times affect the distribution of 5. tumefaciens infections? Survivorship of stems was least at the center of the pond from 1999 to 2000, further suggesting that water stress and S. tumefaciens infections may play a role in H. edisonianum decline in this pond. As a rule, flooding stress incapacitates or kills terrestrial plants more quickly than drought or soil moisture depletion (Larcher 1995). Those genera tolerant of anoxic soils, such as Taxodium, Nyssa, and Salix have evolved both functional and morphological adaptations to cope with the multitude of abiotic factors that are inevitably coupled with low or absent soil oxygen due to flooding. These factors, such as increased soil acidity with subsequent deficiencies in nitrogen and increases in metal oxides in solution, may work in concert to debilitate plant communities found in seasonal ponds. Such debilitation or weakening of plants is commonly viewed as predisposing them to pathogenic invasion. However, populations of H. edisonianum in seasonal ponds are unusual in the sense that their clonal, vegetative form may result in each pond population

PAGE 54

46 representing a single individual (genet). Acting as a single individual, stems of a clonal plant have the capacity to maintain a multitude of physiological connections between distant ramets, such as translocation of resources, control of intra-clonal competition by regulation of ramet production and buffering capability in heterogeneous and/or stressful microsites (Jonsdottir and Watson 1997). The advantages of physiological integration between ramets, particularly in resource-poor environments such as seasonal ponds are: 1) a means of conserving scarce resources by sequestering and reallocating to various clonal fragments; 2) the evolution of developmental divisions of labor among ramets (Jonsdottir et al. 1996) whereby old ramets function as nutrient storage units and new ramets in acquisition of carbon and nutrients; and 3) "sampling of new environments" by ramets in environments of patchy resource distribution (Jonsdottir and Watson 1997). The degree of physiological connections and interactions in H. edisonianum clones is unknown. The presence of S. tumefaciens does not have a significant an impact on the H. edisonianum stems growing in Pond 30-35. This pond contained the only population of H. edisonianum with approximately 50% infection in 1999 and, during the extreme drought conditions of 2000; this infection coincided with high stem mortality. However, Ponds 7-64 and 88 also suffered the greatest percentage of stem mortality towards the pond center without any apparent infection by the fungus. Given the clonal nature of H. edisonianum, it may be erroneous to assume S. tumefaciens infection ultimately leads to death of the long-lived genet. Perhaps galls act to block assimilates from the affected aerial shoot and this signals the rhizome to redirect resources to dormant meristems. Dying H. edisonianum stems affected with galls frequently 'die-back' by a gradual browning and loss of leaves. Whether this is

PAGE 55

47 attributable to a straightforward blockage of conducting tissue by the galls that is exacerbated by drought conditions or a more complex and integrated physiological response to infection by the clone is unknown at this time. The observed dark staining of the conducting tissues may be indicative of the movement of the fungus in xylem tissue. Such movement may be responsible, in part, for the occurrence of gall formation high up in the branches ofH. edisonianum, however this was not investigated further. Given the physiological adaptations known to occur in clonal plants, it became apparent that H. edisonianum response to seasonal flooding, drought, and attack by parasites may not easily fit the a general paradigm of pathogen-related mortality. Therefore, the hydrology and nutrient status of the seasonal ponds containing H. edisonianum were further investigated to better understand the plants response to stress and how fitness characters varied with environmental extremes.

PAGE 56

CHAPTER 3 HYDROLOGY AND SOIL NUTRIENT CHARACTERS IN SEASONAL POND POPULATIONS OF Hypericum edisonianum Introduction The seasonal ponds found in south Florida, by definition, fill with water in response to rainfall in the summer rainy season and then gradually percolate down in the dry winter months to reach drought conditions by spring (Myers and Ewel, 1992). The depth and duration of flooding of Florida marshes and seasonal ponds are highly variable (Abrahamson et al. 1984, Duever et aL 1975, Pesnall and Brown, 1977) and within individual ponds the duration of standing water changes markedly from the palmettoedged margin to the center. Approximately 1 1% of the mapped land area of the Archbold Biological Station (ABS) property is represented by seasonal ponds (Abrahamson et al. 1984). These ponds occur in the presence of impermeable confining layers or where the water table emerges through the sandy substrates (Myers and Ewel, 1992). Common vegetative features of seasonal ponds are saw palmetto margins (Serenoa repens) that enclose four major types of plant associations. These associations are maidencane (Panicum hemitomori), Hypericum {Hypericum edisonianum), cutthroat Grass {Panicum abscissum) and broomsedge (Andropogon brachystachys) (Abrahamson et al. 1984). Seasonal ponds, especially those near bayheads, are often invaded by tree species, particularly in the upper elevational zones (Landman and Menges 1999). 48

PAGE 57

Soils of the ABS are deep sand, low in clay and silt content, with nutrients and drainage characteristics typical of dunes and sea terraces formed in the Pleistocene. Sanibei, Sellers and Placid soils are the only soil types that have a deep muck layer overlying mineral soils (Abrahamson et al. 1984). Hypericum edisonianum grows most frequently in Pompano depressional, Immokalee depressional and Placid soil types (Abrahamson et al. 1984). Most Florida marshes have highly buffered waters as a result of the underlying limestone or calcareous substrates, with pH levels generally occurring in the neutral range (pH 7) except in flatwoods ponds where acidic groundwaters may be present (Myers and Ewel 1992). Nutrients are low in most acidic ponds where rainfall rather than upland runoff is the primary nutrient source (Myers and Ewel 1 992). Peat soils in particular are low in the minor elements (copper, manganese, zinc and boron), high in organic nitrogen and low in phosphorous (Forsee 1940, Bryan 1958). Fire has an enduring and profound effect on the landscape and vegetative communities of south central Florida and the ABS. Of the estimated 2100 to 2600 lightning strikes that occurred within the ABS property boundaries in a 14-year period, 30 strikes caused fires (Abrahamson et al. 1984). The frequency of fires (or fire return intervals) in these native scrub communities exert a strong selective pressure on the evolution of plant life history and reproductive strategies (Keely 1981, Ostertag and Menges 1994), species diversity (Johnson and Abrahamson 1990, Menges et al. 1993), landscape-level patterns of plant species abundance and their interactions (Menges and Hawkes 1998), as well as limit the invasion of woody vegetation (Myers and Ewel 1992). In this study the hydrological and soil nutrient characters of seven seasonal ponds were examined to investigate the pattern of mortality in the Hypericum edisonianum

PAGE 58

50 population of Pond 30-35. The greatest amount of infection by the fungal pathogen, Sphaeropsis tumefaciens occurred near the center of pond 30-35 in 1999 to 2000 (Chapter 2) and this distribution of disease suggests that environmental stress may predispose H. edisonianum plants to parasitic attack. Therefore, elements of the pond microenvironment likely to contribute to plant stress were investigated and these results used to evaluate their effects on plant fitness characters. The assumption was that fitness characters (stem heights, crown areas, flowering, new stem shoots) would reflect increasing stress effects of standing water from the pond margin to center. Wild plant populations are frequently beset with multiple environmental stressors. For example, oxygen depletion in soils during flooding often leads to anaerobic microorganisms creating strongly reducing conditions, which in turn, can lead to toxic concentrations of Fe 2 +, Mn 2 , and H 2 S (Larcher 1995). The hypothesis was that stems near the center of Pond 30-35, known to contain the highest number of pathogenic fungal infections, also undergo the greatest extremes in soil chemical parameters (of all study ponds) due to flood conditions. Therefore, stem fitness characters would be most diminished. Such stress conditions may further lead to opportunistic insect herbivory which may then further contribute to pathogenic infections and stem mortality in Pond 30-35. Materials and Methods Meter-wide belt transects were randomly placed in seven study ponds (Ponds 8, 88, 750, 7-56, 7-64, 30-35 and 31-45) and wireflags were placed at one-meter intervals (Figure 2-2). Growth, reproduction and mortality measurements were made for each stem in these transects (Table 2-2).

PAGE 59

51 On 20 February 2000, soil samples were collected at 2-meter intervals along each permanent transect from the pond margin to the pond center. Each sample was labeled with the appropriate pond identification number and meter location, and left to air dry on paper towels in the ABS laboratory of E. Menges. Dried soil samples were then transported to Gainesville, Florida, and tested at the University of Florida Analytic Research Laboratory (Table 3-1). Table 3-1 . Soil analyses of samples taken from seven study ponds Extractable Elements: Macro nutrients: Phosphorus, Potassium, Calcium and Magnesium Micro nutrients: Copper, Iron, Manganese, Zinc Aluminum and Sodium Water Extractable ions: pH, electrical conductivity, C1-, NH4-N, N0 3 -N Organic matter content The method used for extractable elements was as follows: 5.0 grams of mineral soils (or 1.25 grams organic soil) were mixed with 20 ml Mehlich-1 extractant (0.05 N HC1 in 0.025 N H2SO4) shaken for 5 minutes and filtered. All elements were analyzed in this filtrate by inductively coupled argon plasma (ICAP) spectroscopy (Page et al. 1982). Organic matter was determined using two different methods. Soils with 6% or greater organic matter were analyzed by loss on ignition (samples were heated in ovens at 450° for six hours.) Soils containing less than 4% organic matter were determined by the Walkley-Black dichromate method (Page et al. 1982) Soil water content was measured along transects in each of the seven study ponds over the course of one late afternoon on 1 1 July 2000. Using a Cambridge Delta-T Devices Thetameter (Type HH1), in situ soil moisture was measured by inserting a probe

PAGE 60

52 approximately 10 cm into the soil in the center of alternating m 2 of the belt transect. Soil moisture (a soil/water volumetric ratio) was instantaneously recorded for both organic and mineral soils. Long-term records of bi-weekly water depths in seasonal ponds were provided by ABS. These data consist of water depths collected at a PVC (polyvinyl chloride) pipe permanently located in the deepest portion of each pond. There were no long-term data for water depths from the pond margins to pond centers to use in conjunction with permanent belt transects established for this study in 1998. Therefore, soil moisture gradients were used to estimate how study ponds retained water along these transects. Data analyses were performed using SPSS version 10 (SPSS, Inc.). Spearman's 2tailed correlation tests were used on soils data. Two-tailed tests were utilized because they are the most conservative for this nonparametric test and because sample sizes were unavoidably different for each pond. Simple linear regressions were used to examine plant fitness characters for the years 1998 to 2000. A Trimble Global Positioning System (GPS) was used to measure the areas of five of the seven study ponds. The ponds located in the Red Hill tract of ABS (Ponds 8 and 88) were not measured using GPS because the pine canopy interfered with satellite signal capture. Arc View software was used for modifying (with permission) GIS (Global Information System) maps created by Roberta Pickert of the ABS/GIS Laboratory and for measuring linear distances between ponds. Results Hydrology Mean water depths were compiled for 25 seasonal ponds from 1989 to 1999 and, as expected, found to be quite variable among ponds (Figure 3-1). Only two of the 25 ponds

PAGE 61

53 had mean water depths of greater than 40 cm. Eight ponds had mean depths of between 20 to 30 cm, ten ponds between 10 to 20 cm and five ponds with less than 10 cm. Seasonal ponds, however, often undergo substantial hydrological fluctuations relevant to vegetative communities, from year to year and month to month. An example of the yearly variation of mean water depth from 1991 to 1999 is presented (Figure 3-2) for five of the seven seasonal ponds (Ponds 7-50, 7-56, 7-64, 30-35 and 31-45) used in this study. The two ponds not included (Ponds 8 and 88) are remote, unrecorded and unmapped ponds on the ABS property. In 1997, the mean water depth for the five ponds was below 5 cm while during the following El Nino year, 1998, the same ponds all had mean depths of at least 20 cm. Pond 30-35, containing the highest frequency of fungus-infected H. edisonianum stems, was exceptional in that its mean depth was over 50 cm in 1998. The seasonal ponds at ABS fill with water in response to rainfall. Monthly mean water depths are generally at their lowest ebb in May or June and then gradually increase in response to summer storms (Fig.3-3). These five study ponds are located in the intraridge valley portion of ABS (Abrahamson et al. 1984) and are likely to be filled by ascent of a perched water table. Soil moisture (or water content) measurements, taken from the palmetto-edged margin of the ponds to their center, were used to obtain an indirect estimate of the pattern of water retention or the basin configuration, of each of the seven study ponds in this study. All measurements were taken while the ponds were dry, during the drought of 2000. For example, in Figure 3-4 the soil transect in Pond 7-50 exhibited a short, steep gradient of increasing subsurface water content from the pond margin to meter 12 and then roughly constant water content from meter 12 to meter 33 (pond center). This profile matches field observations of this pond having a bowl-like basin that abruptly changes to a grassy,

PAGE 62

54 dry, marginal shelf of <50 cm height. Hypericum edisonianum stems growing at the margin experienced very different hydrological conditions than those stems in the "basin' of this pond. 50 40 E o S 30 CL 0) 0) -*— ' c 03 20 10 n n ^/VVWVVWVWV\>^'\©*o'5p y & 7 e te'h'h** ****** ** Archbold Ponds Figure 3-1. Mean water levels (in cm) of 25 ABS ponds, 1989 to 1999. Data provided by the Archbold Biological Station. The study ponds varied in the distribution of soil moisture from edge to center (Figure 3-4). Ponds 88, 7-56, 7-64 and 30-35 contain a relatively shallow subsurface soil moisture gradient from the margin and then a sudden increase of water content, as though there was a sandy mantle surrounding a large hole in the center of the pond. Field observations match these soil moisture profiles for at least three ponds. Pond 30-35

PAGE 63

55 60 Year Figure 3-2. Mean water depths for five study ponds at ABS from 1991 to 1999. Mean water depths in each pond vary from year to year. (Data were not available for some ponds in some months in 1989 to 1990. Data were provided by the Archbold Biological Station. 40 jan feb mar apr may jun jul aug sep oct nov dec MONTH Figure 3-3. Mean water depths (cm) in five study ponds from 1991 to 1999 by month. Pond water depths reflect seasonal rainfall patterns. Data provided by the Archbold Biological Station.

PAGE 64

56 particularly matches this soil moisture profile in that it has a gently sloping sandy margin and then water depths (during the rainy season) that increase markedly near the pond center (Figure 2-2). Pond 7-56 soil moisture readings do not entirely agree with field observations, however. In the field, the pond appears to have a gentle gradient wherein water levels increase very gradually from margin to center without there being a deep hole at the center. These field observations were confirmed after fires burned off all the vegetation in this pond during 12-13 February 2001 (Figures 3-16 and 3-17) to reveal only a shallow, sandy basin. Ponds 8 and 31-45 have very shallow soil moisture gradients throughout the entire transect (Figure 3-4). Pond 8 was not observed to contain standing water for the duration of this study (1997 to 2000), while Pond 31-45 is often rather dry and grassy during the rainy season except for a small, deep, water-filled hole at the very center of the pond. As already noted, these soil moisture readings were taken in July 2000, when there was no standing water in any pond. Both 1999 and 2000 were exceptionally dry years at ABS. Seasonal Pond Soils Analyses of soil samples taken in the permanent transects in each study pond showed that organic matter accumulated and was most concentrated in the center of Pond 30-35 (Spearman's rho P< .01, r 2 = 0.52, Figure 3-5). Only two other study ponds showed this organic matter gradient, Pond 88 (Spearman's rho P<.01, r 2 = 0.4) and Pond 7-64 (Spearman's rho P<.05, r 2 = 0.8). These three ponds all were described above as having a shallow sandy matle with a "hole' in the center. The other four study ponds do not accumulate organic matter in this pattern.

PAGE 65

57 .600 .500 o CO 1 .400 B S I .300 2 in 5 .200 E o .100 .000 — — P8 P88 o P750 X P756 P764 P3035 P3145 10 20 30 Meters from pond Margin 40 Figure 3-4. Soil moisture gradients in all seven study ponds, 1 1 July 2000. Pond 7-50 (P750) has a very steep soil moisture gradient and then levels off. Ponds 88, 7-56, 7-64 and 30-35 have very shallow soil moisture gradients that sharply increase towards the pond center. Ponds 8 and 31-45 display very small soil moisture gradients from margin to center. Note: pond margin at origin and final value is at pond center.

PAGE 66

58 -D— Pond 8 Pond 88 -A— Pond 7-50 Pond 7-56 Pond 7-64 • Pond 30-35 Pond 31-45 10 12 14 16 18 20 22 24 26 Meters from pond Margin 28 30 32 34 36 Figure 3-5. Variation in concentration of organic matter in pond soils. Meter 1 is at the pond margin. Increasing meter value indicates increasing proximity to the pond center. 25.0 3456789 10 11 Sample Collection Number Figure 3-6. Pond 30-35 organic matter and potassium concentrations sampled at every other meter in 21 -meter long permanent transect. Sample #1 was taken at meter 1 (located at pond margin) and Sample #2 was taken at meter 3, etc.

PAGE 67

59 Figure 3-7. Pond 30-35 soil concentrations of NH4-N and NO3-N by meter location in permanent transect in pond. Meter 1 is at pond margin. 250.0 200.0 I 150.0 0 100.0 50.0 8 10 12 14 16 Meters from pond Margin 18 20 22 Figure 3-8. Pond 30-35 soil concentrations of aluminum and calcium by meter location in permanent transect in pond. Meter 1 is at pond margin.

PAGE 68

60 The concentration of potassium in pond soil is closely correlated with organic matter in Pond 30-35 (Spearman's rho P< .01, r 2 = 0.13) and follows the same pattern of increase towards the pond center (Figure 3-6). This relationship also occurs in all other ponds except those described as having very shallow soil moisture gradients (Pond 8 Spearman's rho P>0.35, r 2 = 0.0008; Pond 31-45 Spearman's rho P>0J0, r 2 =0.02). Concentrations of NH4-N and NO3-N are coupled together in all study ponds, and Pond 30-35 has a significant concentration gradient for both (Spearman's rho P< .01, r 2 = 0.05) from the margin of the pond towards the center (Figure 3-7); however, only NH4-N is correlated with organic matter (Spearman's rho P< .01, r 2 = 0.75). The soil pH of Pond 30-35 decreases with distance from the pond margin to the center (from 5 to 3.8) and under such increasingly acidic conditions soil aluminum concentrations also increased. Surprisingly, the soils in Pond 30-35 also showed an increasing concentration of calcium towards the center that is contrary to the norm of acidic soils, which generally have very low calcium content (Figure 3-8). This calcium gradient did not occur in any other study pond. Stem Heights, Flowering and Recruitment Stem heights Hypericum edisonianum stems in Pond 30-35 grew tallest in the center of the pond in 1999 and 2000 (Figure 3-9) where water depths (and the assumed hydric stress) were greatest and presumably of longest duration. Mortality associated with S. tumefaciens infections in the center of the pond was low (4%) in 1999 and then soared to 83% in 2000 (Table 2-6).

PAGE 69

61 Regression of stem heights along all permanent sampling transects showed that stems increased with water depth (p<0.05) in all ponds except for Pond 8, the small dry pond in the northeastern portion of ABS (Figure 3-10). Table 3-2. Mean stem heights for H. edisonianum in seven study ponds Mean stem Mean stem Mean stem Pond height (cmU 998 height (cm) 1999 height (cm) 2000 8 29.2 (5) 78.0 (1)* 84.0(1)* 88 53.0 (77) 62.1 (31) 71.1(60) 7-50 70.2 (23) 81.9 (22) 85.4(19) 7-56 61.5 (58) 56.2 (88) 70.24(80) 7-64 74.3(106) 82.0 (81) 73.4(113) 3035 82.9(60) 85.3 (61) 83.4(41) 3145 44.4(34) 46.1 (37) 62.5(17) Sample size in parentheses The two driest ponds (8 and 3 1-45, see Figure 3-4) also had the shortest mean stem heights in 1998 and the pond with the deepest mean water levels (Pond 30-35) had the tallest mean stem heights (Table 3-2 and Figures 3-9 and 3-10). The population of H. edisonianum stems in Pond 8 appears to be in a fatal decline. Stems were small and scattered in this dry pond and four of the five stems found in the study transect in 1998 have since died. However, over the past two years new shoots have been appearing at the outside edge of this pond and are growing towards a lower, wetter depression in the adjacent abandoned two-track road. Crown area The crown sizes of//, edisonianum stems in Pond 30-35 were correlated with stem height (Spearman's rho P< .05, r 2 = 0.08) and increased in size as stems grew taller toward the pond center (Figure 3-11). Although stems in Pond 30-35 were, on average, the tallest stems in this study, their crowns were not the largest. Crown areas of stems

PAGE 70

62 Figure 3-9. Regressions of stem height by transect meter for Pond 30-35 Hypericum edisonianum (top) and Ponds 8 and 88. Stem heights in Pond 30-35 were greatest at the pond center (meter 28) in 1999 to 2000.

PAGE 71

63 Pond 7-50 7 6 9 10 1 Distance from pond margin (m) • 1999 Rsq0 1492 • 2000 Rsg = 0.1683 Pond 7-56 10 20 Distance from pond margin (m) Pond 7-64 Pond 31-45 2 4 6 8 10 12 14 16 Distance from pond margin (m) 1998 Rsq = 0 4395 1999 R»q * 0 5526 " 2000 Rsq = 0 4994 6 8 10 12 14 16 18 Distance from pond margin (m) Figure 3-10. Linear regressions of stem height by meter in ponds 7-50, 7-56, 7-64 and 31-45. All populations of H. edisonianum showed increased stem heights towards the center of the ponds except for Pond 7-50. Figure 3-11. The relationship between crown size by meter location in Pond 30-35.

PAGE 72

64 Table 3-3. Crown areas of H. edisonianum stems from 1998 to 2000 Pond Mean crown area (cm) 1998 Mean crown Mean crown area (cm) 1999 area (cm) 2000 8 9.5 (5) 10.0 (77) 17.7(23) 15.4 (58) 20.4(106) 0 29.6(1) 18.7 (28) 23.7 (22) 13.9 (88) 25.7 (81) 16.3 (61) 12.1 (19) 34.8(1) 22.1 (53) 28.9(17) 20.1 (58) 21.9 (91) 22.5 (10) 25.8 (2) 88 7-50 7-56 7-64 3035 3145 7.1 (34) Sample size in parentheses were compared across pond populations of H. edisonianum from 1998 to 2000. Those stems in Ponds 8 and 3 1-45 (the driest ponds) had smaller crowns than those in the other five study populations in 1998 (Table 3-3). The increase in size toward the pond center in Pond 8 is based upon only one surviving stem (Figure 3-12). The size increase of crowns at the pond center also occurred in Ponds 88, 7-50 and 3 1-45. Pond 3 1-45 is the pond with the soil moisture readings indicating that it is relatively dry with a deep, waterfilled hole in the center and Pond 88 is a pond with a sandy mantle with a hole in the center. Flowering Hypericum edisonianum flowers throughout the year and the flowering phenology of individual pond populations are quite variable. In this study populations were sampled only during June, July and August of each year and therefore it is probable that the flowering periods of some study populations were missed entirely. Figure 3-13 shows the mean number of stems in flower in the permanent transects for only four of the seven study ponds (Ponds 88, 7-50, 7-64 and 3 1-45). In all cases, flowering increased as a whole from 1998 to 1999 and then decreased from 1999 to 2000 except for Pond 88.

PAGE 73

65 During the summer of 1998 there was high water levels in seasonal ponds and a subsequent drought in 1999 and 2000 when no standing water observed in any study pond. Stem Recruitment New shoots that appeared in Pond 30-35 were first clustered at the midway point of the permanent transect in 1999 and the following year were concentrated at the center of the pond (Figure 3-14). This pattern may reflect the general drying trend in the pond during 1999 to 2000, where as the pond basin lost soil moisture content at the margins the new growth followed the moisture gradient towards the center. Fire effects In February 2001 a passing train set off sparks that led to an intensely hot wildfire that rapidly progressed through the drought-stricken landscape of ABS. Two study ponds, Pond 7-56 and 7-64 were in the path of this fire (Figure 3-15) and were entirely burned in less than a day. Figures 3-16 and 3-17 show Pond 7-56 as it occurred during the time of this study from 1998 to 2000 and then shortly after the fires of February 12-13, 2001. On 16 June 2002 the post fire responses of H. edisonianum populations in Ponds 7-56 and 7-64 were evaluated in a nonsystematic manner. Although all boundary flags of the permanent transects had been incinerated, the aluminum identification tags previously placed on each stem within the study transects still remained on the ground and were used as a guide to the former study population. In both ponds there were very abundant and robust stems of H. edisonianum. Most notable were the similar heights of the resprouts and the lack of any observable parasitic damage.

PAGE 74

66 Pond 7-50 Pond 7-56 8 9 10 11 12 0 10 20 Distance from pood margin (m) Distance from pond margin (m) Figure 3-12. Relationship between crown area and distance from pond margin (origin) to pond center. Ponds were sampled annually for three years. Data shown as linear regressions.

PAGE 75

67 j E C 70 i 60 g 50 "5 40 30 c J 20 10 n I l otd and new flowers, 1998 I bid and new flcwers. 1999 Id and new flowers, 2000 88 7-50 7-64 31-45 Pond (by I.D. number) Figure 3-13. Flowering patterns in Ponds 88, 7-50, 7-64 and 31-45. The overall flowering of pond populations of H. edisonianum increased from 1998 to 1999 and then decreased in 2000. (New flowers included buds and old flowers included ovules only.) Dtettnce from pond rrmtfn(m) m 20 30 Distance from pond margin (m) Figure 3-14. Pond 30-35 recruitment of new stems 1998 to 1999 (left) and 1999 to 2000 (right). New stems were clustered in the middle of the transect in 1998 to 1999 and then shifted to the center of the pond in 1999 to 2000 (meter 27).

PAGE 76

68 Discussion The hydroperiods and water depths of each of the seven study ponds were demonstrably variable from year to year, month to month and from the pond margin to the pond center. The H. edisonianum populations in Pond 30-35 experienced unusually wet conditions in 1998 where the mean water depths in the center were 50 cm. In the following years, this same plant population suffered record-breaking drought conditions. In this same time frame, the smallest population of H. edisonianum in the smallest pond (Pond 8) endured increasingly extreme hydric conditions. Even in the wettest year (1998) no standing water was observed in this pond. As drought conditions worsened in 1999 and 2000 the population of H. edisonianum in Pond 8 went into decline without any evidence of pathogenic interactions (Figure 2-4). The soils of all the study ponds play a significant role in the duration of standing water and size gradients in plant growth. Ponds 8 and 88 are located in the northeast Red Hill tract of Archbold Biological Station, where the sands of the region are deepest, and it is probable that the water table underlying Ponds 88 is perched. The sandy soils of Pond 8 rapidly percolated rainfall, whereas Pond 88, a larger pond, retained rainwater longer, perhaps due to the organic layer overlaying the sandy substrates below. Pond 30-35 also contained a bed of organic matter and retained water at greater depths and for longer periods than other ponds in this study. The accumulation of organic matter in the seasonal ponds where H. edisonianum grows appears to play a pivotal role in the microsite conditions that dictate plant growth. In the annual dry down of seasonal ponds there comes a point where hydric conditions become most favorable for soil microorganisms to decompose organic matter and thus make available essential plant nutrients (Comanor and Staffeldt 1978). In this study a

PAGE 77

69 Figure 3-15. Aerial photograph of Archbold Biological station showing station boundaries (wide white lines), study pond locations (blue symbols), and extent of Feb. 12, 2000 fire (yellow lines). Star is ignition point. Note extensive citrus plantings to the east (linear patterns created by rows of trees).

PAGE 78

Figure 3-16. Pond 7-56 in July 1998. Permanent transects begin at the pond margin and end at the center of the pond (where the white PVC pipe is located). Figure 3-17. Pond 7-56 three weeks after fires that occurred on February 12-13, 2001. The PVC pipe at the center of the pond was incinerated as were all transect flags. The center of the pond has the most new vegetation. Note the saw palmetto at the pond margins has also resprouted.

PAGE 79

71 Figure 3-18. Multiple stems arising from rhizome (arrow) in Pond 3 1-45 (left photo). Hypericum edisonianum resprouting in Pond 7-56 three weeks after February 12-13 fires. Resprouting typically occurred within approximately 20 cm from charred adult stem (right photo). significant organic matter gradient was found to exist from the pond margin to the center in Ponds 30-35, 88 and 7-64. These ponds were described as having a shallow soil moisture gradient from the pond margin and then a rather abrupt increase in water content as though there were a hole in the pond basin. The duration of standing water was greatest at the pond centers. Incomplete degradation of organic matter added to the retention of water (additional organic matter accumulation) as well as contributed to the gradient of decreasing soil pH. The soil pH levels in the seven study ponds never rose above pH 5.0 and were typically approximately pH 4.0. Perhaps the seasonal water fluctuations contribute to the already acid conditions by means of truncating organic breakdown in each pond over many years.

PAGE 80

72 Acidic soils will often limit the nutrients available to plants. The deep sands of the LWR and ABS are in themselves extremely acid and nutrient-deficient (Abrahamson et al. 1984) and the acidic conditions found in seasonal ponds may exacerbate the lack of available nutrients to H. edisonianum populations. In soils with pH levels of 4.0 or less, clay minerals are broken down and metal hydrous oxides are brought into solution (Larcher 1994). As a result, increasing levels of free ions of aluminum and heavy metals are released (Brunet 1994). This phenomenon appeared to be occurring in Pond 8 where aluminum concentrations of soils reached more than 4,000 mg/kg. Possibly the litter fall and leaching from the pine canopy contributed to these conditions. As acid conditions continue in these seasonal ponds, the breakdown of organic matter is further inhibited and nitrification also decreases. In a study of the effects of increasing Al, Mn and Fe at high acidity (pH 3.8) on a wide-ranging plant, Succisa pratensis Moench, Pegtel (1986) found that both xeric and mesic populations showed the same response curve. The author argued that these results may be indicative of the plant genetically differentiating into edaphic ecotypes that are able tolerate phytotoxic concentrations of these microelements. Nitrogen uptake in plants is essential for the formation of amino acids that are used in the synthesis of nucleic acids and proteins. The low pH conditions of the seven ponds in this study appear to have limited nitrifying bacteria, resulting in consistently lower amounts of N0 3 than NFL, in the pond soils. The ratio of NH4-N and NO3-N remained relatively constant in the seven seasonal ponds while the relative amounts of both nutrients increased in a concentration gradient towards the pond center. In a study using sand-culture grown H. perforatum, Briskin et al. (2000) found that the production of hypericins (hypericin and pseudohypericin) in the leaves of plants was correlated with decreased nitrogen levels. When supplemental nitrogen was provided to these same

PAGE 81

73 stems the production of hypericins decreased. If stems of H. edisonianum at the margin of Pond 30-35 are producing more hypericins than those stems in the center then it may account for the concomitant decrease in S. tumefaciens infections. According to Liebig's "Law of the Minimum" plant growth is affected by the nutrient that is most limited in its availability. Plants that are limited by the lack of essential nutrients and minerals will often develop a dwarf growth form as a deficiency-stress strategy when cell elongation is limited (Grime 1979). This may explain, in part, why smaller stems of//, edisonianum always occurred at the margin of all study ponds while those stems found growing at the pond center were consistently larger. The crown sizes of stems were not predictable by their position in the pond in general. This may be due in part to stem density and/or genetic parameters, however crown size increased with stem height towards the pond center in Pond 30-35, negating the density argument. The post fire response of Ponds 7-56 and 7-64 in June 2002 was that of abundant and robust regrowth of//, edisonianum with no evidence of disease symptoms of any description. The findings of this study do not support the argument that H. edisonianum stems found growing in the center of Pond 30-35 were predisposed to pathogenic attack (by S. tumefaciens) due to stress of inundation. Rather, stems at the center of this pond were taller, had larger crowns and regenerated new shoots significantly more than stems at the pond margin and this was consistently correlated with more abundant nutrient and mineral resources.

PAGE 82

CHAPTER 4 GENETIC DIVERSITY IN Hypericum edisonianum Introduction Disease is one of the many agents that affect the vital rates of a plant population (Caswell, 1 989). As a group, plant pathogens are capable of inflicting a wide range of damage in wild plant populations. Some fungal pathogens, such as the damping-off genera (e.g., Pythium or Rhizoctonia) are relatively nonselective and cause mortality mainly at germination and seedling stages. However most pathogens show some sort of genetic specificity that restricts them to a small range of hosts (Burdon, 1994). Although pathogens affect the individual plant genotype, it is the change that takes place in the various plant fitness characters as a whole that utlimately affect the size of the population. There are numerous and often very specfic ways in which plant diseases affect hostplant fitness but, given that they are often unobserved in wild plant communities, their effects are probably greatly underestimated (Burdon, 1994). For example, in cases such as Ustilago infecting the flowers of Silene alba (Alexander, 1989) there is no question that the host plant fecundity is diminished by anther infection and, thus, genotypic variability in the next generation is also decreased. Population-level genetic processes also affect a plant population's vital rates by change in the presence or organization of alleles; manifested by the frequency of alleles and levels of heterozygosity (Schmeske et al. 1 994). For example, in a study of the 74

PAGE 83

critically endangered sentry milk-vetch (Astragalus cremnophylax var. cremnophylax) Travis et al. (1996) found that the smallest population of this plant suffered from an extreme lack of genetic diversity. The authors attributed this to a severe lack of suitable habitat (on the South Rim of the Grand Canyon) and a pronounced founder effect. Larger populations of milkvetch on the South Rim also suffered from lower than expected genetic diversity; which was thought due, in part, to the extremely stressful site conditions and the periodic crashes this population experiences. Conversely, the allelerichness of a population can theoretically contribute to the population's increase by enhancing its evolutionary potential to survive unpredictable environmental change. Further, populations with the greatest amount of genetic variation are thought to suffer least from inbreeding depression and/or the effects of genetic drift (Burdon and Shattock 1980, Dinoor and Eshed 1984, Menges and Dolan 1998, Segal et al. 1980). In a survey of literature involving research on rare and endemic plant species, Schemeske et al. (1994) found that of 78 papers published between 1987 and 1992, only six papers addressed genetic variation in quantitative characters. These studies, for the most part, infer evolutionary potential of plant populations by using isozymes and polymorphic DNA to estimate allele diversity and heterozygosity levels. These observations on genetic variation are then used to formulate conservation and management plans for individual plant species (Van Treuren et al 1993, Watson et al. 1994, Menges and Dolan 1998). In spite of the theoretical relationships believed to exist between genetic diversity and a species' persistence in nature, the relationship between the molecular markers used in these studies and fitness characters is often obscure and there have been no empirical

PAGE 84

76 studies that directly link the genetic composition of wild plant populations with their growth rate or survival (Schemske et al. 1994). Stochastic environmental processes that occur outside of a plant population have profound effects on its long-term survival. Theoretically, in small populations, chance events effectively remove alleles more often than in larger populations (Nei et al. 1975). However, population size is not necessarily the key factor that determines outcome during stochastic events in plant populations. Using matrix projection models on metapopulations of Furbish's lousewort {Pedicularis furbishiae) Menges (1990) found that the persistence of this plant species could not be assured by simply protecting individual populations. Rather, to insure the long-term survival of P. furbishiae it was recommended that enough original habitat be protected to allow for continual recolonization of numerous, small, patches. This balance between extant plant populations and their local extinction and recolonization of habitat patches is referred to as metapopulation dynamics. Metapopulations can be described as a collection of individual populations that shift in their presence or absence in a landscape mosaic of available and appropriate habitat. If too many of the original habitat patches are lost, then the plant species in question is unable to periodically recolonize these patches and declines to the point where extinction is inescapable. Kareiva and Wennergren (1995) likened this process to the collapse of an epidemic following the threshold loss of susceptible host plants. Thus, the long-term survival of rare and endemic species (such as H. edisonianum) may be more dependent upon foresight in land management and conservation then in the plant's genetic character (Lande 1998, Menges 1991, Schmeske et al. 1994). Given the

PAGE 85

77 rate of development in south central Florida this question is now becoming urgent for this and many other narrowly endemic plant species. Hypericum edisonianum populations Hypericum edisonianum is one of the many endemic plant species found only in the Lake Wales Ridge (LWR) region of Florida (Christman and Judd 1990). Its populations are widely scattered throughout patches of scrub habitat and are often rather small and isolated. This patchy distribution of scrub habitat in the LWR is attributable, in part, to the historical mosaic of fire occurrences throughout the region (Menges 1990) and human disturbance. Landscape-level dynamics affect how individual patches of scrub vegetation function and the overall abundance patterns of scrub species (Menges 1999, Menges and Hawkes 1998). If disease effectively removes only a small proportion of genes from future generations of H edisonianum will the survival of this species be in jeopardy? Given that H. edisonianum is a long-lived, clonal plant that is capable of capturing substrates in seasonal ponds with fast-growing ramets, it is possible that some individual ponds may contain only a few, unique genetic individuals (clones) of the plant. The pond populations of H. edisonianum, known only to occur in 4 counties in Florida, may constitute extant populations that have been greatly reduced from a formerly larger distribution. If this is so, then these pond populations are expected to suffer losses of fitness from the fixation of old mutations (carried by the founders) that are equivalent to the effect of many generations of new mutations (Lande et al. 1994). This loss of fitness may possibly have occurred in light of the ongoing destruction of H. edisonianum populations and their habitat. In the last 50 years, the state of Florida has lost

PAGE 86

78 approximately eight million acres of forest and wetlands from the expansion of human populations and subsequent clearing of land (Cox et al. 1994). The loss of scrub habitat, that contains the seasonal ponds that support H. edisonianum in Florida, has been massive. Davis (1967) estimated that scrub habitat formerly covered approximately 41 7,000 ha (1 .03 million acres) in the state. Later, in 1 993, Kautz et al. found the land area in scrub habitat to be only 170,850 ha (422,000 acres). Founder effects may constitute an important risk of extinction for populations that have been suddenly reduced to very small effective size or that are already near the point of genetic inviability (Lande 1994). If populations are more gradually reduced in size however, or that initially have a substantial maximum growth rates, then the fixation of new mutations poses a more serious risk of eventual extinction (Lande et al. 1994). Overall however, it may be the effects of short-term impacts (such as disease) on H. edisonianum that affect its chances of reproductive success and long-term survival (Lande 1998, Menges 1991, Schemske et al. 1994). In this study amplified fragment length polymorphisms (AFLP) (Vos et al. 1995) were used to screen ten populations of H. edisonianum. It was predicted that individual pond populations of this plant would have limited genotypic variation due to environmental barriers to gene flow in isolated pond populations and/or the clonal habit of H. edisonianum. Materials and Methods Newly emerged leaves were collected from individually tagged H. edisonianum stems in the permanent transects placed in each study pond. Each sample was placed in appropriately labeled, new plastic bags and stored briefly in a cooler until return to the

PAGE 87

79 ABS labs. These leaves were cut in half and only the tips were used for extraction of whole genomic DNA. Tips of new leaves were used with the assumption that endophytic infection, if present, would be greatly minimized as well as would possible stores of secondary metabolites. DNA extraction from H. edisonianum was very difficult. At least three different procedures using C-TAB extractions were used without success. Grinding leaf or flower tissue with either laboratory-grade quartz sand or in liquid nitrogen does not result in good DNA yields. Upon the recommendation of Mark Whitten (UF Museum of Natural History) samples containing only two leaf tips were ground in mortars using only 1.2 ml of warm C-TAB buffer and 8 ul of mercaptoethanol. The tissue homogenate was then transferred to new 1 .5 ml tubes and incubated at 65° C in a heat block for a minimum of 6 hours with occasional shaking. Longer incubation periods resulted in better DNA yields however M. Whitten has been successful using incubations of less than an hour. Moreover, attempting to extract large quantities of DNA from correspondingly large volumes of leaf tissue (in Oakridge tubes) consistently failed. After incubation the tubes were briefly vortexed and then 500 ul of 24:1 chloroform/isoamyl alcohol added to each tube and vortexed again. Samples were then centrifuged at 14,000 rpm for 5 minutes to separate the phases. The aqueous phase (containing the DNA) was pipetted into new 1.5 ml tubes using 100 ul lots to measure total volumes. Then 3 M sodium acetate was added to each sample using the following formula: sodium acetate volume in ul = volume aqueous phase in jil X 0.04. The new volume was then used to calculate the amount of 100% isopropanol to add using this formula: isopropanol volume in ul = volume aqueous

PAGE 88

80 phase X 0.65. The samples were gently rolled to mix all the constituents together and then placed in a 4° freezer for approximately 1 to 2 weeks to allow for DNA precipitation. After precipitation the samples were centrifuged at 14,000 rpm for 20 minutes and the alcohol poured off. The remaining pellet was washed with 1 ml of 70% ethanol and then left to dry on the laboratory bench. Each sample received 25-50 ul of IX TE storage buffer and then placed in a 4° C freezer for storage. This 'mini-prep' protocol usually resulted in yields from 5-35 ug/ml of DNA. Amplified fragment length polymorphisms (AFLP) Four different primer sets were used to generate DNA fingerprints for ten different populations ofH. edisonianum. The first step in performing AFLP was to generate random restriction (DNA) fragments using two endonucleases (EcoN and Msel). Then 3-base "adaptor" sequences were ligated to the ends of this DNA using T4 ligase in the second step. The DNA fragments with attached adaptor sequences served as templates for the first "pre-selective" amplification using polymerase chain reaction (PCR). Only DNA fragments with adaptors on each end will amplify exponentially (disposing of all other fragments). After the preselective amplification, there were still far too many fragments present to allow for clear visualization of bands on an acrylamide gel. Therefore, another "selective" amplification was carried out using two more, different EcoW and Msel primers. At this step fluorescent dye labels were attached to the £coRI primer but not the Msel primer. The DNA fragments with EcoSl at both ends do not amplify well and fragments with only Msel are not visualized on acrylamide gels for lack of fluorescence.

PAGE 89

81 Only DNA fragments with EcoRl and Msel at either end amplify well and appear on acrylamide gels due to the fluorescent dye marker on the EcoKl primer. After the second PCR amplification the number of bands appearing on acrylamide gels were greatly reduced from the previous restriction/amplification step and band scoring made feasible. Acrylamide gels were laser-scanned using an ABI 373 autosequencer to produce precise basepair measures of each DNA fragment peak (band). For this study the most successful selective primer sets were: EcoRl 3A: GACTGCGTACCAATTCACA EcoRl 4A: GACTGCGTACCAATTCACG Msd3C: GATGAGTCCTGAGTAACAA MseI3C: GATGAGTCCTGAGTAACAA EcoRl 4A: GACTGCGTACCAATTCACG EcoRl 3 A: GACTGCGTACCAATTCACA Msel AC: GATGAGTCCTGAGTAACTA Msel 2C: GTATATACAAATTATATAA Results The AFLP screening gels revealed H. edisonianum populations to be much more genetically diverse than anticipated. Unique haplotypes were scored for almost each individual in each population found in the Archbold Biological Station (ABS) property (Figure 4-1) while populations outside of ABS in Glades and DeSoto counties appeared to be slightly more homogeneous (Table 4-1). This was found to be the case using each of the four primer sets used to screen these samples. Using the gel results that represented primers "3a3c" as an example (Figure 4-1), the ten most frequent polymorphic loci were tallied across all pond populations and a table constructed of their frequency of occurrence (Table 4-1). The Glades and DeSoto county populations consistently shared more bands then those populations at ABS. Pond 3 1-45 was the only ABS pond that appeared to be more similar to Glades and DeSoto populations then to ABS ponds in bandsharing. Interestingly, this pond is at the southernmost extreme of the ABS property nearest these county boundaries. Ponds 7-50

PAGE 90

82 and 7-56 had virtually identical band patterns at each locus in every sample. These two ponds are in very close proximity to one another on the ABS property and these patterns suggest gene flow between the two ponds (Figure 4-2). Pond 30-35, the population of H. edisonianum heavily infested with S. tumefaciem, appeared to be the least similar to any other ponds studied in the ABS property. The total number of fragments appearing in lanes 13-15 and also those automatically called in electrophaerograms was quite limited. This may be evidence of genuinely lower genetic diversity in this pond or DNA samples that contained a greater amount of impurities than all other samples. When the multilocus haplotypes for these same ten populations were broken down by individual plant samples per pond (Table 4-2) it can be seen that stems from pond populations (excluding Glades and DeSoto county populations) differed from one another by location in the pond. Stems from the edge (or pond margin) exhibited a higher frequency of band occurrences that those from the pond center. This pattern became even more marked when multilocus haplotypes from nine individual stems taken from a single transect (from pond margin to pond center) in Pond 7-64 were examined (Figure 43, Table 4-3). Those stems from meters 9, 13 and 15 (at the pond center) shared very few bands with stems in meters 6, 7, 8 and 10. Discussion The AFLP findings of this study are still very preliminary. Initially several leaf samples were collected from each study pond along permanent transects and the DNA extracted. Unfortunately unforeseen events led to the curtailment of this study and the

PAGE 91

83 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Lane Plant # Location Plant location in population 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1 2 3 1 2 3 1 2 3 55 101 v 119 131 190 13 27 31 504 526 527 501 540 Glades Pop. #1 Glades Pop. #1 Glades Pop. #1 DeSoto Pop. #1 DeSotoPop. #1 DeSoto Pop. #1 Glades Pop. #2 Glades Pop. #2 Glades Pop. #2 Pond 88 Pond 88 Pond 88 Pond 30-35 Pond 30-35 Pond 30-35 Pond 8 Pond 8 Pond 7-56 Pond 3 1-45 Pond 31-45 Pond 31-45 Pond 7-64 Pond 7-64 High elevation Mid elevation Low elevation High elevation Mid elevation Low elevation High elevation Mid elevation Low elevation Pond margin Mid transect Center of pond Pond margin Mid transect Center of pond Center of pond Mid transect Mid transect Pond margin Mid transect Center of pond Pond margin Mid transect Figure 4-1 . Example of AFLP screening gel used to detect genetic differences among populations of//, edisonianum (blue bands). Red bands are size markers used for automated band calling (laser scans). Primer set 3a3c was utilized.

PAGE 92

84 a u t oo — — ci a p oo oo o IT) «/"> .J ° r~m
PAGE 93

85 1 2 3 4 5 6 7 8 9 10 11 12 Lane Plants SamDle location Sample location in population l 28 Pond 7-50 Mid transect 2 33 Pond 7-50 Center of pond 3 34 Pond 7-50 Center of pond 4 28 Pond 7-50 Mid transect 5 33 Pond 7-50 Center of pond 6 34 Pond 7-50 Center of pond 7 38 Pond 7-56 Margin of pond 8 100 Pond 7-56 Mid transect 9 200 Pond 7-56 Center of pond 10 38 Pond 7-56 Margin of pond 1! 100 Pond 7-56 Mid transect 12 200 Pond 7-56 Center of pond Figure 4-2. Comparison of banding patterns between Pond 7-50 and Pond 7-56. Lanes 1-6 are from Pond 7-50. Primer set 3a3c was used in the samples in the first 3 rows and primer set 4a4c was used in samples in lanes 4-6. Pond 7-56 samples are in lanes 7-12. Primer set 3a3c was used in the samples in lanes 7-9 and primer set 4a4c used in samples in lanes 10-12. These pond populations of H. edisonianum were sampled at the 'Margin', the middle of the transect and at the pond center. Ponds 7-50 and 7-56 are within 500 meters of each other and share more bands than any other ponds in this study.

PAGE 94

86 CM OOOOOOOOOOOOOOOOOO — — I c I Eg '-5 o I * •Si — Q. E C3 c B o 1*1 8 * B J* 3 11 31 x> .3 o o CI en (N ri oo o o oo oo 8 -J O — O — — OOOOOOO — OOOOOO — — ©,— ___©o_„._ o — o — — ooooooo— -oo — — o — — — o o o o OOOOO — — O !— — oo — — © — o O — O — — 'OOOOOOO — OOOOOO — I o o o o o OOOOOO — — o — — — o — o — o — oo — — oo — o — o — o — — ooooooo — oo — — oo — — ! — — — — o — — — o ooooo — — — — — — — O — oooo — oo o — o — oooooooo — ooo — o — — — o — o — — ooooooo — ooo — o — — — or-xi — ooor^^— oooo — ooo\ — o_*^ot-—
PAGE 95

87 123 456789 II Lane Plant # Sample location l 23 Margin of pond 2 35 Margin of pond 3 41 Mid transect 4 47 Mid transect 5 56 Mid transect 6 74 Center of pond 7 104 Center of pond 8 1 Margin of pond 9 15 Margin of pond Figure 4-3. Pond 7-64 transect samples taken from the pond margin, mid transect and pond center. See Table 4-3 also.

PAGE 96

88 £3 /— v O n g JB s 3 E 5 o CO • — < U oa is p 1 1 8 o •a o, o 5 U cO JB M 3 O o r=3 — U S a C U co m a Q. 09 a c 1 3) CO "cO u a C o CN CN CO 00 CN CN CN 00 o CN cn CN O —i o — O r-H — , ~H O ~ o — < O 1-H O —< O — i — I O — i ^ o ~ «-C ^ O O «-« f-M o ~h O ~H ON On O oo a. t/3 J O o 3 C g I [0 CO C CJ C/3 | SB 0 1 o CO o .S u .a 03 3 00 < 53 Q & — — CO CD CO 03 O m CO CO 09 U

PAGE 97

89 use of these additional samples. In most studies using dominant marker fingerprints sample sizes are larger per population allowing HardyWeinberg frequencies to be inferred. The sample sizes in this study were, at most, three stems per pond and sometimes it appeared that DNA impurities might have caused fewer or false polymorphisms to be generated in some samples. Nevertheless, the preliminary screening gels in this study do represent randomly restricted and amplified fragments of DNA that, at least in general, serve to illustrate the diverse genetic character ofH. edisonianum populations. Perhaps the most parsimonious explanation for this diversity is that pollinators facilitate gene flow among ponds and that at least some viable seeds are able to survive. Although H. edisonianum is primarily clonal in its growth habit it does flower abundantly in some pond populations throughout most of the year. Many species of bees were observed to visit H. edisonianum flowers throughout the course of this study however they regularly moved from flower to flower within the ponds more often than trap-line ponds. In an unsystematic field survey done in 1998 it was not unusual to find seedlings free of rhizomatous connections to other stems in the area. In one circumstance several seedlings ofH. edisonianum were discovered growing outside the boundaries of ABS in experimental plots in old cow pastures. Low, wet areas had been converted to pasture approximately thirteen years before and H. edisonianum seedlings appeared only when the turf had been entirely removed. It is not known whether this constitutes evidence of an enduring seed bank or incidental seed dispersal. In studies of rare and endangered plants found in vernal pools, Jain (1994) found that

PAGE 98

90 species of meadowfoam (Limnanthes alba and L. douglasii) that had wide-ranging metapopulations were more variable and heterozygous than the narrowly distributed and inbreeding L. floccosa and L. bakeri. In the latter two species the seed bank played a critical role in the persistence of L. floccosa, particularly for very small populations. Another possible explanation of the observed genetic diversity in the study populations of H. edisonianum may have to do with somatic mutation occurring over long periods of time. The Florida scrub communities in which the seasonal pond populations of//, edisonianum grow, are believed to have persisted for at least 50,000 years (Watts and Hansen 1994). Although the age of individual populations of H. edisonianum are not known at this time it is reasonable to argue that some may be of great age. Some clonal plants have been recorded as living for extraordinarily long periods of time. For example, Kemperman and Barnes (1976) reported that a single clone of trembling aspen (Populus tremuloides) may have been more than 10,000 years old and covered an area of 81 hectares. In another case Steinger et al. (1996) documented clones of the sedge, Carex curvula, as being 2,000 years old. Klekowski (1997) argues that long-lived clonal organisms may increasingly accumulate somatic mutations and that it is this genetic load that may ultimately lead to the decline in sexuality (as offspring accumulate defective or lethal genotypes) and/or the extinction of the clone itself. The pond populations of//, edisonianum display vastly different flowering phenologies throughout the year. Whether this is simply a reflection of microenvironment conditions or genetic load is debatable. Nevertheless the results of the screening gels in this study do point out that each pond population of H. edisonianum does contain numbers of unique haplotypes that may reflect a long history of accumulated somatic mutations.

PAGE 99

91 Ultimately the cause of so many unique haplotypes in each population of H. edisonianum cannot be explained by this study. However, the effects of this diversity may be very significant with regard to damage by pathogens. The success of mixed plantings (mixed genotypes) has been studied extensively in agriculture with the majority of cases showing that disease severity is attenuated. However, as Burdon (1987) points out, very few agricultural studies examine the effect of disease reduction coupled with the change in the plant's reproductive performance over the long term. A recent exception to this is a study by Brunet and Mundt (2000). Using wheat genotypes susceptible to different races of the pathogen Puccinia striiformis (wheat rust), they investigated the effects of disease and competition on the overall fitness of the host genotype. They found that there were few significant interactions between host fitness and disease or competition. Moreover, very few varietal mix studies have used the proper disease-free controls necessary to evaluate true disease reduction. Thus, if the resource requirements for the different varieties do not sufficiently overlap, then the survival of the mixtures will almost certainly always be greater than the component varieties (Burdon 1 987). A notable exception to this is the recent work by Zhu et al. (2000). Using a comprehensive and controlled design that encompassed all rice fields in five townships in 1998 and ten townships in 1999 in the Yunnan Province of China, the authors tested the effectiveness of genetic diversity in planted rice fields. Those fields planted with a genetically diversified mix of resistant and disease susceptible varieties of rice had a 94% decrease in the blast disease caused by Magnaporthe grisea when compared to monotypic fields. The infection of//, edisonianum stems by S. tumefaciens in Pond 30-35 presents an

PAGE 100

92 interesting case. While this population suffered very high infection and mortality during the year 2000, there was very vigorous new shoot recruitment in the areas of the pond where most mortality took place. Perhaps the new shoots represent genotypes more resistant to the pathogen. Equally plausible is the argument that the clone is simply "cutting its losses" by dispensing with diseased stems and translocating nutrients to new shoots in the same nutrient microsite. Assessing these two opposing suppositions constitutes the future experimental work planned for this rare plant.

PAGE 101

CHAPTER 5 INSECT DAMAGE IN SEVEN POPULATIONS OF Hypericum edisonianum Introduction In general, plant-herbivore interactions are assumed to be detrimental to the host plant, leading to the evolution of a multitude of resistance mechanisms thought to protect the plant from further attack. For example, plant characters affected by insect depredations include inductive production of secondary metabolites (Rhoades 1979, 1983, Tallamy and Krischik 1989), constitutive production of secondary compounds (Bazzaz et al. 1987, Ehrlich and Raven 1964, Fraenkel 1959, Levin 1976, Southwood et al. 1986), tissue-nutritive values (Feeny 1976, Neuvonen and Haukioja 1984, Southwood 1972), seed size and fruit morphology (Bradford and Smith 1977, Janzen 1969, 1971, 1975) and leaf flushing and growth rates (Aide 1988). Environmental factors are thought to affect the presence of herbivores and the varied intraspecific damage they incur in targeted host plants. Within-site environmental factors found to be correlated with damage are light (Collinge and Louda 1998, Harrison 1987, Huffaker 1970), soil moisture content (Bernays and Lewis 1986, Lewis 1984, OluomiSadeghi et al. 1988), soil nutrient content (Onuf et al. 1977, Rhodes 1983), and the density of plant conspecifics (Duggan 1985, Paulissen 1987, Stanton 1983). In this study the occurrence of, and damage caused by, herbivorous arthropods and parasites on H. edisonianum were annually assessed from 1998 to 2000 in seven study pond populations. The intent was to investigate the potential causal relationship between 93

PAGE 102

94 herbivore damage and plant stress that may further correlate with disease incidence in H. edisonianum. Methods Surveys of Parasite Damage Soil samples from study ponds were randomly collected beneath H. edisonianum stems, placed in plastic bags and transported to the Florida Department of Agriculture and Consumer Services (FDACS) in Gainesville for identification of parasitic nematodes. This screening was performed once in 1999. Insects, and their damage to Hypericum edisonianum, were assessed annually from 1998 to 2000 by examining all stems found within the permanent transects that were established in each study pond. These belt transects began at the pond margin and ended at the pond center and every individual stem of H. edisonianum within these transects was tagged with unique identification numbers (see Chapter Two methods). Damage thought attributable to insects was almost always in the form of leafminer mines. These mines were recorded in the field by inspecting the leaves of each stem in the study transect and by scoring each mine one time. Some stems contained multiple mines. Collection of Insects for Identification Small branches with suspected insect damage (mines, holes, evidence of chewing) were collected for rearing-out of any persisting larvae or eggs. These branches were collected randomly from H. edisonianum in permanent transects and then placed in large plastic bags, inflated with air, sealed, and suspended from overhead fixtures in the plant laboratory at Archbold Biological Station. Plastic bags containing branches with numerous leaf mines were monitored daily. The time elapsed from collection of the

PAGE 103

95 mines to the appearance of the lepidoptera varied from one day to seven days, reflecting the different ages of larvae at the time of collection. Moths were allowed to remain in the plastic bag with the plant material until found dead. An inspection of the bottom of the bag every morning would often yield two to three moths and as well as unidentified parasitic wasps. All specimens from plastic bags were collected for later identification. Results Hypericum edisonianum is utilized by a number of parasites (Table 5-1). Occasionally heavy outbreaks of white fly Aleuroplatus plumosis (Aleyrodidae) (Russell 1944) were observed in large seasonal ponds experiencing prolonged flooding on the ABS property, but they were infrequently encountered on the stems in the permanent transects of this study. The cottony-cushion scale Icerya purchasi (Margarodidae) (Hamon 1992) was also found in high numbers in very large, southern ABS ponds where standing water remained for most of 1998. The Hypericum scale (Aonidomytilus hyperici) was found only occasionally on H edisonianum stems and was not associated with any particular environmental condition. All other parasites other than the wasps and leaf miners listed in Table 3-1 were infrequently observed on H. edisonianum. Only one genus of nematode {Tylenchorhyncus) was found in pond soils and in numbers too low to cause damage (personal communication, R. Inserra, Florida Department of Agriculture and Consumer Services). Parasitic wasps in the family Bethylidae were frequently associated with leaf miners from field-collected stems of H. edisonianum. These small wasps are known parasites of lepidopteran larvae (Borror et al. 1989) however no interactions with leaf miners or

PAGE 104

moths were observed or recorded. Their taxonomic identity awaits determination, as do other micro-hymenoptera in the Bracsonidae (Table 5-1). Distribution of Leafminer Damage Outbreaks of a leafininer, a larval form of a microlepidopteran tentatively named Coleotechnites nigra Busck (Gelechiidae) (personal communication, John Heppner, Florida Department of Agriculture and Consumer Services), occurred frequently in pond populations of H. edisonianum from 1998 to 2000. The presence of leafrniners declined from 1998 to 2000 in Pond 30-35, as they did in Ponds 7-50, 7-64 and 31-45, (Figure 5-1). The decline of leafrniners in Pond 30-35 coincided with the dieback of the H. edisonianum population in 2000 when less than 25% of the stems remained standing in the permanent transect (Figure 5-1). As stems fell the leafrniners became more concentrated on the remaining standing stems. The numbers of miners increased in Ponds 88 and 7-56 from 1998 to 2000. To investigate whether stem mortality was associated with the presence of leafrniners, excluding infection by the fungal pathogen, S. tumefaciens, Pearson's Chi-square test was performed using one degree of freedom There was no significant relationship between the presence of leafrniners in 1998 and stem mortality 1998 in Pond 30-35 (x*= .036, df =1, p> .9). Nor were leafrniners a significant factor in stem mortality in 1999 = 3.86, df = 1, p > .05). In 2000, there was an apparent relationship between leafminer presence and stem mortality (x 2 = .688, df = 1 , p < .05).

PAGE 105

"2 « o .2 I) 00 e/2 1 < f o o a* 3 tr a* u a 1 a*>s S Jg S £ S S
PAGE 106

98 6* 8 88 7-50 7-56 7-64 30-35 31-45 Archbold pond number Figure 5-1. Leaf miner populations by pond from 1998 to 2000. All data were collected from permanent transects that were randomly placed in each of seven study ponds. Discussion Hypericum edisonianum is host to a number of parasites that are, in general, infrequent visitors. However it should be noted that insect collections came from individually picked branches of H. edisonianum and not the result of vigorous sweep netting. Sweep netting was not undertaken because of the damage this method would cause to the rather brittle branches of H. edisonianum and the desire to sample only parasites. Thus the findings of this study probably under represent the insect communities that may exist.

PAGE 107

99 Figure 5-2. Female moth (top). Female and male moths are silvery in appearance (not shown in photo) and are approximately 7-8 mm in length. Damage done to H. edisonianum leaf by leafminer C. nigra (bottom). This leaf miner typically chews into the leaf from the leaf margin and excavates all tissues between the dermal layers. The exit holes are visible at the midrib.

PAGE 108

100 The most commonly encountered herbivorous insect was the leaf miner Coleotechnites nigra. Adult moths have black antennae with indistinct silvery annulations. Busck (1903) first described this moth (formerly Recurvaria nigra) as having a black face, head and thorax with hind wings nearly black. However the specimens found at ABS appear to be a lighter form and/or a new species (personal communication J. Heppner). The occurrence of this leafminer, if indeed it is confirmed as C. nigra, is the first record for the state of Florida and a range extension from the District of Columbia, where the species was first recorded growing on Hypericum fruticosa (Busck 1903). This newly discovered insect-plant association between C. nigra and H edisonianum may be a highly specialized interaction where the moth exclusively feeds on H edisonianum. The moth was not observed to feed on any other Hypericum species however this is only from casual inspection of plants in the field and not a systematic survey. Adult moths were kept alive in plastic bags for approximately two weeks by providing fresh H edisonianum branches. The male moths have a very distinctive 'hair pencil' or coremata, a scent organ that they wave above them in a rotary fashion. Coremata have been observed in Creatonotos moths to attract females. The male moths only produce suitable chemical compounds to attract females when the males ingest specific plant alkaloids. When the male moths are fed diets deficient in these alkaloids (nonpreferred plants) the males fail to produce pheromones or develop fully functional coremata (Schneider et al. 1982).

PAGE 109

101 Parasitic microhymenopterans were routinely collected from the same incubation chambers (plastic bags) as the leaf miner larvae and moths. Whether these wasps are specialized on C. nigra was not determined. Although the C. nigra caused gross tissue damage to the leaves it colonized its overall effect on H. edisonianum populations was negligible. Although there was a significant Chi-square value for the presence of leafrniners and stem mortality in Pond 30-35 in 2000 it would probably be erroneous to assume a causal relationship. In the year 2000, there were extreme drought conditions and stems in Pond 30-35 suffered very high mortality (83%) despite the leafminer's population being at its lowest ebb in 3 years in that pond. Thus, the findings of this study do not support the argument that H. edisonianum stems found growing in the center of Pond 30-35 were predisposed to pathogenic attack by S. tumefaciens due to the stress of herbivory.

PAGE 110

CHAPTER 6 SUMMARY AND CONCLUSIONS Two fungal pathogens were found to cause disease in H. edisonianum; Sphaeropsis tumefaciens Hedges and Colletotrichum gloeosporioides (Penz.) Penz. & Sacc. Sphaeropsis tumefaciens was occasionally found throughout the property holdings of the Archbold Biological Station, Lake Placid, Florida, where this study was conducted but only one seasonal pond population of H. edisonianum contained a severe infestation of this fungus. Colletotrichum gloeosporioides also causes disease in H. edisonianum, but this pathogen was only isolated from one greenhouse-grown stem of the plant and was not observed in or isolated from wild populations. Nevertheless, this pathogen may present a serious threat to H. edisonianum given that saw palmetto is widespread in the landscape, shown to be susceptible to the pathogen, and could serve as a source of inoculum. Demographic data were collected over a three-year period and provided insights on how each pond population of this plant endured climatological extremes at the southern end of the Lake Wales Ridge as well as within seasonal pond microenvironments. Meterwide belt transects were randomly placed in each of seven study ponds and all individual stems of the plant within these transects were censused annually. The prediction that flooding stress in the middle of Pond 30-35 (containing the most infected population ofH. edisonianum) predisposed stems to disease was not supported by the 102

PAGE 111

103 data. Rather, the most robust stems were found in the center of the pond and the least vigorous stems found at the pond margin, where drought effects were most severe. Analyses of long-term hydrology data obtained from the Archbold Biological Station showed that the seasonal ponds in this study reflected their dynamic response to climate. All but one pond (Pond 8) filled with water as a result of seasonal rainfall. Soil moisture transects taken in each of the seven study ponds accurately reflected the soil water-retaining character of the basin except for Pond 7-56. This pond had a shallow moisture gradient throughout rather than the water-holding pocket in the center. Ponds with shallow moisture gradients that converged into a deep pocket at the center concentrated organic matter in the center. This layer of matter correlated with decreased pH and increased concentrations of nitrogen, potassium and aluminum. Soil nutrient parameters were investigated in each of the seven study ponds and concentration gradients were found to exist from the pond margin to the pond center for pH, nitrogen, potassium, organic matter, aluminum and calcium In Pond 30-35 these gradients were correlated with the hydrology of the pond basin. Hypericum edisonianum is a clonal species that sends up dozens to thousands of vegetative shoots (ramets) in a clumped pattern in numerous seasonal ponds in the Lake Wales Ridge region. It is generally assumed that shoots or ramets of clonal plants that arise from a single genet are genetically identical. This character, along with the plants frequent isolation in discrete pond populations within a white-sand landscape, led to the prediction that H. edisoniaum pond populations would have low genotypic variation. To investigate the predicted lack of genetic diversity in H. edisonianum, amplified fragment length polymorphisms (AFLP) methods were employed to examine ten study

PAGE 112

104 populations. The resulting acrylamide electrophoretic gels and automated band-calling electrophaerograms showed that H. edisonianum populations contained considerable genotypic diversity. How then, can one explain the periodic diebacks that have occurred in this plant? The answer is a complex one. There can be no doubt that extremes in both pond hydrology and climatic conditions (drought in particular) have deleterious effects in H. edisonianum populations. In 1998, the seasonal ponds at ABS experienced extended wet periods that were followed by extreme drought in 1999 through 2000. Ponds in this study that did not contain diseased H. edisonianum stems suffered up to 47% mortality. The only pond in this study with a high incidence of disease from infection by S. tumefaciens also had the greatest mortality (83%) in 2000. Plant pathogens that attack H. edisonianum may be occurring in a metapopulation pattern where spatial or temporal conditions in seasonal ponds are fleetingly conducive to infestation in this harsh landscape but cannot endure due to episodic fires, drought and/or flood conditions. Alternatively, the great genetic diversity discovered in this plant species may serve to diminish the effects of pathogens and parasites and thus only the occasional epidemic, as seen in Pond 30-35, arises. A new, undescribed leafrnining microlepidopteran (Coleotechnites sp.) was found to cause foliar damage to several pond populations of H. edisonianum. Although the damage caused by this leafrniner can be extensive there was no significant relationship between leafrniner presence and stem mortality. The landscape of the Lake Wales Ridge and that of Archbold Biological Station is one shaped by wildfire over the millennia. In the aftermath of the station's most intensely hot fires in 2001, two study pond populations (Pond 7-56 and 7-64) of//, edisonianum were

PAGE 113

105 seemingly destroyed. Yet, within three weeks there was abundant resprouting of this plant. By June 2002 both of these ponds contained robust, even-aged populations of H. edisonianum. Perhaps as long as wildfires continue in this arid landscape, the incidence of disease in the many rare and endemic plant species will be kept at a minimum.

PAGE 114

APPENDIX FORMULAE FOR FUNGAL CULTURE MEDIA Dave Mitchel's Selective Medium for Phytopltthora and Pythium (PARP 1 liter of corn meal agar amended with: 5 mg pimaricin 250 mg ampicillin 10 mg rifampicin (dissolved in 1 mL ethanol) 100 mg pentachloronitrobenzene DPI Selective Medium for Phytopthora and Pythium (CMA/PPA) per 500 mL Difco corn meal agar add: 1 mL pimaricin 1 mL rifampicin 0.5 cc ampicillin 5 ml pentachloronitrobenzene or 'Terraclor' DPI Acid PDA To 500 mL of Difco Potato Dextrose Agar add: 0.85 ml 50% lactic acid 106

PAGE 115

REFERENCES Abrahamson, W.G. 1980. Demography and vegetative reproduction. In: Demography and Evolution in Plant Populations. O.T. Solbrig (ed), pp. 89-106. Blackwell Scientific Publications, Oxford. Abrahamson, W.G., Johnson, A.F., Layne, J.N. and Peroni, P. A. 1984. Vegetation of the Archbold Biological Station, Florida: An example of the southern Lake Wales Ridge. Florida Scientist 47(4):209-250. Abrahamson, W.G. and McCrea, K.D. 1986. Nutrient and biomass allocation in Solidago. Effects of two stem gall makers and ramet isolation. Oecologia 68: 174-1 80. Agostini, J.P. and Timmer, L.W. 1994. Population dynamics and survival of strains of Colletotrichum gloeosporioides on citrus in Florida. Phytopathology 84:420-425. Aide, T.M. 1988. Herbivory as a selective agent on the timing of leaf production in a tropical understory community. Nature 336:574-575. Alexander, H.M. 1989. An experimental field study of anther-smut disease of Silene alba caused by Ustilago violacea: genotypic variation and disease incidence. Evolution 43:835-847. Alfieri, Jr., S.A., Landon, K.R., Kimbrough, J.W., El-Gholl, N.E. and Wehlburg, C., 1994. Diseases and Disorders of Plants in Florida. Bulletin No. 14, Div. Plant Industry, Fla. Dept. Agric. Consumer Serv., Gainesville, Fla. Anderson, R.L., Knighten, J.L., Windham, M. Langdon, K, Hedrix, F. and Roncadori, R. 2001. Dogwood Anthracnose and its Spread in the South. USDA website (2001): http://fhpr8.srs.fs.fed.us/pubs/dogwood. Arnholdt-Schmitt, B. 2000. RAPD analysis: a method to investigate aspects of the reproductive biology of Hypericum perforatum L. Theo. Appl. Gen. 100(6):906-91 1. Augspurger, C.K. 1983. Seed dispersal of the tropical tree Platypodium elegans, and the escape of its seedlings from fungal pathogens. J. Ecol. 71 :7 '59-771. Augspurger, C.K. 1984. Seedling survival of tropical tree species: interactions of dispersal distance, light-gaps, and pathogens. Ecology 65(6): 1 7051712 107

PAGE 116

108 Augspurger, C.K. and Kelly, C.K. 1984. Pathogen mortality of tropical tree seedlings: experimental studies of the effects of dispersal distance, seedling density and light conditions. Oecologia 61:211-217. Barnes, J., Anderson, L. A. and Phillipson, J.D. 2001. St. Johns' wort (Hypericum perforatum L.): a review of its chemistry, pharmacology and clinical properties. J. Pharm Pharmacol. 53(5):583-600. Bernays, E.A. and Lewis, A.C. 1986. The effect of wilting on palatability of plants to Schistocerca gregaria, the desert locust. Oecologia 70:132-135. Bernstein, B., Zehr, E.I., Dean, R.A. and Shabi, E. 1995. Characteristics of Colletotrichum from peach, apple, pecan and other hosts. Plant Dis. 79:478-482. Biraghi, A. 1950. Caratteri di resistenze in Castanea sativa nei confronti di Endothia parasitica. Bollettino della Stazione di patologia vegetale di Roma 7:161-171. Blatchley, W.S. and Leng, C.W., 1916. Rhynchophora or Weevils of North Eastern America, The Nature Publishing Co., Indianapolis. Borror, D.J., Triplehorn, C.A. and Johnson, N.F. 1989. An Introduction to the Study of Insects, 6th Ed., Harcourt Brace College Publications, New York. Boyce, M.S. 1992. Population viability analysis. Ann. Rev. Ecol. and Syst. 23:481-506. Boyle, P.J. 1976. Ergot epiphytotic on Spartina spp. in Ireland. Irish J. Agric. Res. 15:419-424. Bradford, D.F. and Smith, C.C. 1977. Seed predation and seed number in Scheelea palm fruits. Ecology 58:667-673. Bradshaw, A.D. 1959. Population differentiation in Agrostis tenuis Sibth. II. The incidence and significance of infection by Epichloe typhina. New Phytol. 58:3 10-315. Brasier, CM. 1987. Rapid changes in genetic structure of epidemic populations of Ophistoma ulmi. Nature 232:538-541. Brasier, CM. 1990. China and the origins of Dutch elm disease: an appraisal Plant Path. 39:5-16. Brasier, CM. 1991. Ophiostoma novo-ulmi sp. nov. causative agent of current Dutch elm disease pandemics. Mycopathologia 115:151-161. Brasier, CM. 1996 Episodic selection as a force in fungal microevolution, with special reference to clonal speciation and hybrid introgression. Can. J. Bot. 73:1212-1221.

PAGE 117

109 Briskin, D.P., Leroy, A., and Gawienowski, M. 2000. Influence of nitrogen on the production of hypericins by St. John'sWort. Plant Physiol, and Biochem. 38(5):413-420. Brooks, H.K., 1981. Physiographic divisions of Florida. Florida Coop. Extension Service, Inst. Food and Agric. Sci., Univ. Florida, Gainesville. Brunet, J. 1994. Interacting effects of pH, aluminum and base cations on growth and mineral composition of the woodland grasses Bromus benekenii and Hordelymus europaeus. Plant and Soil 161 (2): 157-1 66. Brunet, J. and Mundt, C. 2000. Combined effects of disease and competition on plant fitness. Can. J. Bot. 78:646-654. Bryan, O.C. 1958. The soils of Florida and their crop adaptation. Fla. Dept. Agric. Bull. No. 42. Burdon, J.J. 1980. Variation in disease-resistance within a population of Trifolium repens. J. Ecol. 68:737-744. Burdon, J.J. 1987. Diseases and Plant Population Biology. Cambridge University Press, Cambridge. Burdon, J.J. 1991 . Fungal pathogens as selective forces in plant populations and communities. Aust. J. Ecol. 16:423-432. Burdon, J.J. 1994. The role of parasites in plant populations and communities. In: Biodiversity and Ecosystem Function. Mooney, H. (ed), SpringerVerlag, Berlin. Burdon, J. J., Brown, A.H. and Jarosz, A.M. 1990. The spatial scale of genetic interactions in host-pathogen coevolved systems. In: Pests, Pathogens and Plant Communities, Burdon, J.J. and Leather, S.R. (eds), Blackwell Scientific Publications Oxford. Burdon, J.J. and Chilvers, G. A. 1982. Host density as a factor in plant disease ecology Ann. Rev. Phytopath. 20:143-166. Burdon, J.J. and Jarosz, A.M. 1992. Temporal variation in the racial structure of flax rust (Melampsora lini) populations growing on natural stands of wild flax (Linum marginale): local versus metapopulation dynamics. Plant Path. 41 : 165-79. Burdon, J.J. and Shattock, R.C. 1980. Disease in plant communities. Appl. Bio. 5:145Burgman, M.A. and Lamont, B.B. 1 992. A stochastic model for the viability of Banksia cuneata populations: environmental, demographic and genetic effects. J. Annl Ecol

PAGE 118

110 Busck, A., 1903. A revision of the American moths of the family Gelechidae with descriptions of new species, The Proceedings of the U.S. Natl. Museum, Vol. XXV, pp. 767-938, No. 1304, Washington Gov. Printing Office. Byther, R.S., and Davidson, R.M. 1979. Dogwood anthracnose. Ornamentals Northwest Newsl. 3(2):20-21. Campbell, M.H. and Nicol, H.I. 2000. Effect of split applications of herbicides on the control of Hypericum perforatum L. (St. John's wort) and regeneration of native grasses and annual clovers on nonarable land. Plant Prot. Quart. 15(3):1 19-122. Carrington, M.E., Roberts, P.D., Urs, N.V.R.R., McGovern, R.J., Seijo, T.E. and Mullahey, J.J. 2001 . Premature fruit drop in saw palmettos caused by Colletotrichum gloeosporioides. Plant Dis. 85:122-125. Caswell, H. 1989. Matrix Population Models. Sinauer, Sunderland, Mass. Christman, S.P. and Judd, W.S. 1990. Notes on plants endemic to Florida scrub. Florida Sci. 53:52-73. Clay, K. 1988. Clavicipitaceous fungal endophytes of grasses: coevolution and the change from parasitism to mutualism In: Coevolution of fungi with plants and animals, Pirozynski, K.A. and Hawksworth, D.L. (eds). London: Academic Press. Clay, K. 1988. Fungal endophytes of grasses: a defensive mutualism between plants and fungi. Ecology 69(1): 10-16. Clay, K. 1990. Fungal endophytes of grasses. Ann. Rev. Ecol. Syst. 21:275-295. Clay, K. 1990a. Insects, endophyte fungi and plants. In: Pests, Pathogens and Plant Communities, Burdon, J.J. and Leathers, S.R. (eds). Blackwell Scientific Publications, Oxford. Clay, K. 1990b. The impact of parasitic and mutualistic fungi on competitive interactions among plants. In: Perspectives on Plant Competition, Grace, J.B. and Tilman, D. (eds), Academic Press, San Diego. Clay, K. 1990c. Fungal endophytes of grasses. Ann. Rev. Ecol. and Syst. 21:275-297. Clay, K. and Leuchtmann, A. 1989. Infection of woodland grasses by fungal endophytes Mycologia 81:805-811. Coile, N. 2000. Notes on Florida's Endangered and Threatened Plants. Fla. Dept. Agric. Cons. Serv., Div. Plant Ind., Gainesville, Florida.

PAGE 119

Ill Collinge, S.K. and Louda, S.M. 1988. Herbivory by leafininers in response to experimental shading of a native crucifer. Oecologia 75:559-566. Comanor, P.L. and Staffeldt, E.E., 1978. Decomposition of plant litter in two western north american deserts. In: Nitrogen in Desert Ecosystem, US/IBP Synthesis Series 9, West, N.E. and Skujins, J.J. (eds), Dowden, Hutchinson and Ross, Inc., Stroudsburg, Pennsylvania. Commonwealth Mycological Institute. 1964-present. Descriptions of Pathogenic Fungi and Bacteria No.l. Commonwealth Agriculural Bureau, Kew, Surrey, England. Commonwealth Mycological Institute. 1970. Descriptions of Pathogenic Fungi and Bacteria No. 278. Sphaeropsis tumefaciens. Commonwealth Mycological Institute, The Eastern Press Ltd., London and Reading. Connell, J.H. 1 971 . On the role of natural enemies in preventing competitive exclusion in some marine animals and in rain forest trees. In: Dynamics of numbers in populations (Proceedings of the Advanced Study Institute), Boer, P.J. and van der, Gradwell G.R. (eds). Centre for Agricultural Publication and Documentation, Wageningen. Connell, J.H., Tracey, J.H. and Webb, L.J. 1984. Compensatory recruitment, growth and mortality as factors maintaining rainforest tree diversity. Ecol. Monogr. 54:141-164. Cooper, J.I. and MacCallum, F.O. 1984. Viruses and the Environment. Chapman and Hall, London. Coutinho, T.A., Roux, J. RiedeL K.H. Terblanch, J. and Wingfield, M.J. 2000. First report of bacterial wilt caused by Ralstonia solanacearum on eucalypts in South Africa Forest Path. 20(4):205-210. Cox, J., Kautz, R, MacLaughlin, M. and Gilbert, T., 1994. Closing the Gaps in Florida's Wildlife Habitat Conservation System. Recommendations to meet minimum conservation goals for declining wildlife species and rare plant and animal communities, Office of Environmental Services, FL Game and Fresh Water Fish Commission. Davelos, A.L. 1993. Ecological genetic interactions of pathogenic rust fungi and their host plants. Masters thesis. Univ. Kansas, Lawrence, Kansas. Davis, J.H. 1967. General map of natural vegetation of Florida. Circular S178. Agricultural Experiment Station, IF AS, University of Florida, Gainesville. Dinoor, A. and Eshed, N. 1984. The role and importance of pathogens in natural plant communities. Ann. Rev. Phytopath. 22: 443-466. Douglas, D.A. 1989. Clonal growth of Salix setchelliana on glacial river gravel bars in Alaska. J. Ecol. 77:112-126.

PAGE 120

112 Duever, M.J., Carlson, J.E. and Riopelle, L.A.. 1975. Ecosystem analysis at Corkscrew Swamp. In: Cypress Wetlands for Water Management, Recycling and Conservation. 2nd Annual Report to Nat. Sci. Found, and Rockefells Found.,Center for Wetlands, Univ. Florida, Gainesville. Duggan, A.E. 1 985. Pre-dispersal seed predation by Anthocharis cardamines (Pieridae) in the population dynamics of the perennial Cardamine pratensis (Brassicaceae). Oikos 44:99-106. Edgar, J.G., Kile, G.A. and Almond, C.A. 1976. Tree decline and mortality in selectively logged eucalypt forests in central Victoria. Aust. For. 39:288-303. Ehrlich, P.R. and Raven, P.H. 1964. Butterflies and plants: A study in coevolution. Evolution 18:586-608. Eleuterius, L.N. 1970. Observations on Claviceps purpurea on Spartina alterniflora. Gulf Research Report 3(1): 105-109. Eleuterius, L.N. and Meyers, S.P. 1974. Claviceps purpurea on Spartina in coastal marshes. Mycologia 66:978-986. Ellstrand, N.C. and Elam, D.R. 1993. Population genetic consequences of small population size: implications for plant conservation. Ann. Rev. Ecol. Syst. 24:217-242. Erikson, P. 1994. Pine needle cast infections discolor lodgepole pine stands in the Cariboo forest regions. FID Pest report Canadian Forest Service, Pacific Forestry Centre, Victoria. Feeny, P.P. 1976. Plant apparency and chemical defense. In: J. W. Wallace and R.L. Mansell, eds, Biochemical interaction between plants and insects, Plenum, New York. Fiedler, P.I. 1987. Life history and population dynamics of rare and common mariposa lilies (Calochorus Pursh: Liliaceae). J. Ecol. 75:977-995. Fields, P.G., Arnason, J.T., Philogene, B.J.R., Aucoin, R.R. Morand, P. and Soucybreau, C. 1991. Phototoxins as insecticides and natural plant defenses. Memoirs of the Entomological Society of Canada. 1 59:29-38. Forbes, W.T.M. 1923. The Lepidoptera of New York and Neighboring States; Primitive Forms, Microlepidoptera, Pyraloids, Bombyces. Cornell Univ. Agric. Exp. Sta Memoir 68, Ithaca, New York. Forsee, Jr., W.T. 1940. Recent plant responses to some of the microelements on Everglades peat. Proc. Soil Sci. Soc. Fla. 2:53-58.

PAGE 121

113 Foster, J.L.M. and Fogleman, J.C. 1993. Identification and ecology of bacterial communities associated with necroses of three cactus species. Appl. Environ. Microbiol. 59:1-6. Fox, R.T.V., 1993. Principles of Diagnostic Techniques in Plant Pathology, CAB International, Wallingford, United Kingdom. Francis, C.A. 1986. Introduction: distributions and importance of multiple cropping. In: Multiple Cropping Systems, Francis, F. A. (ed), Macmillan, New York. FrankeL J.F. Shugart, H.H. and Harmon, M.E. 1987. Tree death as an ecological process. Bioscience. 37:550-556. Franklin, I.R. 1980. Evolutionary change in small populations. In: Conservation Biology: An Evolutionary-Ecologiical Perspective, (eds) M.E. Soule and B.A. Wilcox, Sinauer Assoc., Sunderland, MA. Freeman, S., Katan, T. and Shabi, E. 1998. Characterization of Colletotrichum species responsible for anthracnose diseases of various fruits. Plant Dis. 82:596-605. Gallob, J. 2000. Swiss needle cast cuts through Douglas Fir. Newport News-Times, May 24, Newport, Oregon, http://www.newportnewstimes.com . (2001) Gaumann, E. 1930. Uber eine neue Krankheit der Douglasien. Schweiz. Ztschr. F.Forstw., Jahrg. 63. Gilbert, G.S., HubbelL S.P. and Foster, R.B. 1994. Density and distance-to-adult effects of a canker disease of trees in a moist tropical forest. Oecologia 98: 100-108. Godfrey, R.K. 1988. Trees, Shrubs, and Woody Vines of Northern Florida and Adjacent Georgia and Alabama. The University of Georgia Press, Athens and London. Goodnick, P., Keitner, S., Stahl, M. and Halbreich, U. 2001. Effectiveness of St. John's wort in major depression: a randomized controlled trial. JAMA 285(15): 1978-1986. Gray, A. J., Drury, M. and Raybould, A.F. 1990. Spartina and the ergot fungus Claviceps purpurea, a singular contest? In: Pests, pathogens and plant communities, Burdon, J.J. and Leather, S.R. (eds). Oxford: Blackwell Scientific Publications. Griffin, G.J. and J.R. Elkins. 1986. Chestnut blight. In: Chestnut blight, other endothia disease, and the genus Endothia, Roane M.K., Griffin, G.J. and Elkins, J.R. (eds.) St. Paul's Am. Phytopath. Soc. 1-26. Grime, J.P., 1979. Plant Strategies and Vegetation Processes. Wiley, Chichester. Haack, R.A. and Byler, J.W. 1993. Insects and pathogens: Regulators of forest ecosystems. J. For. 91:32-37.

PAGE 122

114 Hamon, A., 1992. The cottonycushion scale, Icerya purchasi, in Florida (Homoptera: Margarodidae). Entom. Circular #352, Fla. Dept. Agric. Cons. Serv., Gainesville, Florida. Hammond, J. 1981. Viruses occurring in Plantago species in England. Plant Path. 30:237-243. Harrison, E. 1987. Treefall gaps versus forest understory as environments for a defoliating moth on a tropical forest shrub. Oecologia 72:65-68. Hedges, F. 191 1. Sphaeropsis tumefaciens, no v. sp., the cause of the lime and orange knot. Phytopathology 1 :63-65. Hedges, F. and Tenny, L. 1912. A knot of citrus trees caused by Sphaeropsis tumefaciens. U.S. Dept. Agric. Bull., pp 247-274. Hildebrand, P.D. and Jensen, K.I.N. 1991. Potential for the biological control of St. John'sWort (Hypericum perforatum) with an endemic strain of Colletotrichum gloeosporioides). Can. J. Plant Path. (Revue Canadienne de Phytopathologie) 13(1):6070. Holah, J.C. Wilson, M.V. and Hansen, F.M. 1993. Effects of a native forest pathogen, Phellinus weirii, on Douglas fir forest composition in western Oregon. Can. J. For. Res. 23:2473-2480. Holliday, P. and Punithalingam, E. 1970. Sphaerospsis tumefaciens. Descriptions of Pathogenic Fungi and Bacteria. No. 278. Commonwealth Mycol. Institute, Kew, Surrey, England. Houston, D.R., Parker, E.J. and Lonsdale, D. 1979. Beech bark disease: Patterns of spread and development of the initiating agent Cryptococcus fagisuga. Can. J. For. Res. 9:336-344. Hubbard, J.C.E. 1970. Effects of cutting and seed production in Spartina anglica. J. Ecol. 58:329-334. Huffaker, C.B. 1967. A comparison of the status of biological control of St. Johnswort in California and Australia. Mushi 39 (Suppl.): 51-73. Huffaker, C.B. and Kennett, C.E. 1959. A ten-year study of vegetational changes associated with biological control of Klamath weed. J. Range Management 12:69-82. Jacobi, V. and Castello, J.D. 1992. Infection of red spruce, black spruce and balsam fir seedlings with tomato mosaic virus. Can. J. For. Res. 22:919-924.

PAGE 123

115 Jacobi, F.J., Castello, J.D., and Flachmann, M. 1992. Isolation of tomato mosaic virus from red spruce. Plant Dis. 76:518-522. Jain, S.K. 1994. Genetics and demography of rare plants and patchily distributed colonizing species. In: Conservation Genetics, Loeschcke, V.J., Tomiuk, J., and Jain, S.K (eds.), Birkhauser Verlag, Basel, Boston. Janzen, D.H. 1969. Seed-eaters versus seed size, number, toxicity and dispersal. Evolution 23:1-27. Janzen, D.H. 1970. Herbivores and the number of tree species in tropical forests Am Nat. 104:501-528. Janzen, D. H. 1971. Seed predation by animals. Ann. Rev. Ecol. Syst. 2:465-492. Janzen, D.H. 1975. Behavior of Hymenaea courbaril when its predispersal seed predator is absent. Science 189:145-147. Jarosz, A.M. and Davelos, A.L. 1995. Tansley Review No. 81. Effects of disease in wild plant populations and the evolution of pathogen aggressiveness. New Phytologist 129:371-387. Johnson, A.F. and Abrahamson, W.C. 1990. A note on the fire responses of species in rosemary scrubs on the souther Lake Wales Ridge. Florida Scientist 53(2): 138-143. Jonsdottir, I.S., Callaghan, T.V. and Headley, A.D. 1996. Resource dynamics within arctic clonal plants. Ecol. Bulletins 45:53-64. Jonsdottir, LS. and Watson, M.A. 1997. Extensive physiological integration: an adaptive trait in resource-poor environments. In: The Ecology and Evolution of Clonal Plants, de Kroon, H. and van Groenendael, J. (eds), Backhuys Publishers, Leiden, The Netherlands. Kareiva, P. and Wennergren, U., 1995. Connecting landscape patterns to ecosystem and population processes. Nature 373:299-302. Kautz, R.S., Gilbert, D.T. and Mauldin, G.M. 1993. Vegetative cover in Florida based on 1985-1989 Landsat Thematic Mapper imagery. Florida Scientist 56:135-154. Kellas, J.D., G. A. Kile, R.G. Jarrett and Morgan, B.J. 1987. The occurrence and effects of Armillaria luteobubalina following partial cutting in mixed eucalypt stands in the Wombat Forest, Victoria. Aust. For. Res. 17:263-276. Kemperman, J. A. and Barnes, B.V. 1976. Clone size in American aspens. Can J Bot 54:2603-2607. Kephart, S.R. and Paladino, C. 1997. Demographic change and microhabitat variability in a grassland endemic, Silene douglasii var. oraria (Caryophyllaceae) Am. J Bot 84:179-189. '

PAGE 124

116 Kile, G.A. 1981. Armillaria luteobubalina: a primary cause of decline and death of trees in mixed species eucalypt forests in central Victoria. Aust. For. Res. 1 1 :63-77. Kim, H., Streltzer, J. and Goebert, D. 1999. St. John's wort for depression: A metaanalysis of well-defined clinical trials. J. Nervous Mental Dis. 187(9):532-539. Klekowski, E.J. 1997. Somatic mutation theory of clonality. In: The Ecology and Evolution of Clonal Plants, de Kroon, H. and van Groenendael, J. (eds.), Backhuys Publishers, Leiden, The Netherlands. Knowles, L. and Witkowski, E.T.F. 2000. Conservation biology of the succulent shrub, Euphorbia barnardii, a serpentine endemic of the Northern Province, South Africa. Aust. Ecol. 25(3):241-252. Lande, R. 1988. Genetics and demography in biological conservation. Science 241 : 1455-1460. Lande, R. 1 994. Risk of population extinction from fixation of new deleterious mutations. Evolution 48(5): 1460-1469. Landman, G. and Menges, E. 1999. Dynamics of woody bayhead invasion into seasonal ponds in south central Florida. Castanea 64(2): 1 301 37. Larcher, W. 1995. Physiological Plant Ecology. Third edition, Springer, Austria. Lavie, G., Mazur, Y., Lavie, D., Prince, A.M., Pascual, D., Liebes, L., Levin, B., and Meruelo, D. 1 995. Hypericin as an inactivator of infectious viruses in blood components. Transfusion 35(5):392-400. Lenne, J.M. and Sonoda, R.M. 1990. Tropical pasture pathology: a pioneering and challenging endeavor. Plant Dis. 74:945-951. Lenne, J.M., Teverson, D.M. and Jeger, M.J. 1994. Evaluation of plant pathogens in complex ecosystems. In: Ecology of Plant Pathogens, Blakeman, J.P. and Williamson, B. (eds), British Society for Plant Pathology, CAB International, Wallingford, UK. Lenne, J.M. and Wood, D.G. 1991. Plant diseases and the use of wild germplasm. Ann. Rev. Phytopathology 29:35-63. Levin, D.A. 1976. The chemical defenses of plants to pathogens and herbivores Ann Rev. Ecol. Syst. 7:121-159. Liyanage, H.D., McMillan, Jr., R.T. and Kistler, H.C. 1992. Two genetically distinct populations of Colletotrichum gloeosporioides from citrus. Phytopathology 82(1 1):13711376.

PAGE 125

117 MacClement, W.D. and Richards, M.G. 1956. Virus in wild plants. Can. J. Bot. 34:793799. Mackenzie, S. 1985. Reciprocal transplantation to study local specialisation and the measurements of components of fitness. Ph.D. Thesis, University of Wales. Mai, W.F. and Lyon, H.H. 1975. Pictorial Key to Genera of Plant Parasitic Nematodes, 4th Ed., Cornell Univ. Press, New York. Mantle, P.G. 1980. Effects of weed grasses on the ecology and distribution of Claviceps purpurea. In: Pests, Pathogens and Vegetation, Thresh, J.M. (ed), Pitman, New York. Marlatt, R.B. and Ridings, W.H. 1974. Sphaeropsis gall of Bottle brush tree, Callistemon vimminalis, a new host. Phytopathology 64: 1 00 1 1 003 . Marlatt, R.B., and Ridings, W.H. 1976. Sphaeropsis disease of Carissa grandiflora. Plant Dis. Rep. 60:842-843, Marlatt, R.B. and Ridings, W.H. 1979. Sphaeropsis gall of Schinus Xerebinthifolius, a new host. Plant Dis. Rep. 63:786-787. Matson, P.A. and Waring, R.H. 1984. Effects of nutrient and light limitation on mountain hemlock: susceptibility to laminated root rot. Ecology. 65:1517-1524. McCauley, K.J. and Cook, S.A. 1980. Phellinus weirii infestations of two mountain hemlock forests in the Oregon Cascades. For. Sci. 26:23-29. McMillan, Jr., R.T. and Timmer, L.W. 1989. Outbreak of citrus postbloom fruit drop caused by Colletotrichum gloeosporioides in Florida. Plant Dis. 73:81. Mehrhoff, L.A. 1989. The dynamics of declining populations of an endangered orchid, Isotria medeoloides. Ecology 70:783-786. Meinecke, E.P. 1939. The Adelopus needle cast of Douglas fir on the Pacific Coast. A report for the Dept. Nat. Res., Div. For.,Sacramento, California. Menges, E.S. 1990. Population viability analysis for an endangered plant. Cons. Biol 4:52-62. Menges, E.S. 1994. Fog temporarily increases water potential in Florida scrub oaks. Florida Scientist 57:65-74. Menges, E.S., 1999. Ecology and Convservation of Florida Scrub. In: Savannas, Barrens, and Rock Outcrops Plant Communities. Anderson, J., Fralish, S. and Baskin, J.M. (eds). Cambridge University Press.

PAGE 126

118 Menges, E.S., Abrahamson, W.G., Givens, K.T., Gallo, N.P. and Layne, J.N. 1993. Twenty years of vegetation change in five long-unburned Florida plant communities. J. Veg. Sci. 4:375-386. Menges, E.S. and Dolan, R.W. 1998. Demographic viability of populations oiSilene regia in midwestern prairies: relationships with fire management, genetic variation, geographic location, population size and isolation. J. Ecol. 86:63-78. Menges, E.S. and Gallo, N.P. 1991. Water relations of scrub oaks on the Lake Wales Ridge, Florida. Florida Scientist 54:69-79. Menges, E.S. and Hawkes, C.V. 1998. Interactive effects of fire and microhabitat on plants of Florida scrub. Ecol. Applications 8(4):935-946. Menges, E.S. and Kohfeldt, N. 1995. Life history strategies of Florida scrub plants in relation to fire. Bull. Torrey Bot. Club 122(4):282-297 Mielke, M.E. and Langdon, K. 1986. Dogwood Anthracnose fungus threatens Catoctin Mountain Park. USDI National Park Service, Park Service, Winter:6-8. Morrison, K.D., Reekie, E.G. and Jensen-Klaus, I.N. 1998. Biocontrol of common St. Johnswort {Hypericum perforatum) with Chrysolina hyperici and a host-specific Colletotrichum gloeosporioides. Weed Technology 12(3):426-435. Myers, R.L. and Ewel, J. J., 1992. Ecosystems of Florida, University of Central Florida Press, Orlando. Nault, A. and Gagnon, D. 1993. Ramet demography of Allium tricoccum, a spring ephemeral, perennial forest herb. J. Ecol. 81:101-119. Nei, M. 1975. Molecular Population Genetics and Evolution. North-Holland Research Monograph: Frontiers in Biology, Vol. 40. North-Holland Publishing Co., Amsterdam. Neuvonen, S., and Haukioja, E. 1984. Low nutritive quality as defence against herbivores: Induced responses in birch. Oecologia 63:71-74. Nienhaus, F. and Castello, J.D. 1989. Viruses in forest trees. Ann. Rev. Phytopath 27:165-186. Nunney, L. and Campbell, K.A. 1993. Assessing minimum viable population size: demography meets population genetics. Trends Ecol. Evol. 8:234-239. Oluomi-Sadeghi, H., Helm, C.G., Kogan, M. and Schoeneweiss, F.F. 1988. Effect of water stress on abundance of twospotted spider mite on soybeans under greenhouse conditions. Entomologia Experimentalis et Applicata 48:85-90.

PAGE 127

119 Onuf, CP., Teal, J.M. and Valiela, I. 1977. Interactions of nutrients, plant growth and herbivory in a mangrove system. Ecology 58:514-526. Ostertag, R. and Menges, E.S. 1994. Patterns of reproductive effeort with time since last fire in Florida scrub plants. J. Veg. Sci. 5:303-310. Page, A.L., Miller, R.H. and Keenay, D.R. 1982. Methods of Soil Analysis Part 2. Chemical and Microbiological Properties, 2nd Edition. American Society of Agronomy, Inc., Soil Science Society of America, Inc., Madison, Wisconsin. Parker, M.A. 1988. Genetic uniformity and disease resistance in a clonal plant. Am. Nat. 132(4):538-549. Paulissen, M.A. 1 987. Exploitation by and effects of caterpillar grazers on the annual, Rudbeckia hirta (Compositae). Am. Midi. Nat. 1 17:439-441. Pegtel, D.M. 1986. Responses of plants to aluminum, manganese and iron with particular reference to Succisa pratensis. Plant and Soil 93(l):43-56. Pesnall, G.L. and Brown Dtt, R.T. 1977. The major plant communities of Lake Okeechobee, Florida and their associated inundation characteristics as determined by gradient analysis. S. Fla. Water Manage. Dist. Tech. Publ. No. 77-1. Pirone, P.P., 1980. Parasitic fungus affects region's dogwood. The New York Times, Feb. 24:34-37. Pistelli, L., Bertoli, A., Zucconelli, S. Morelli, I., Panizzi, L. and Menichini, F. 2000. Antimicrobial activity of crude extracts and pure compounds of Hypericum hircinum. Fitoterapia 71 (Supplement 1):138-140. Puri, H.S. and Vernon, R.O. 1964. Summary of the geology of Florida and a guidebook to the classic exposures. Florida Geol. Surv. Spec. Publ. 5. Raybould, A.F., 1989. The population genetics of Spartina anglica (C.E. Hubbard) Ph.D Thesis, University of Birmingham. Raybould, A.F., Gray, A. J., Lawrence, J.J. and Marshall, D.F. 1991 . The evolution of Spartina anglica: Origin and genetic variability. Bio. J. Linn. Soc. 43:111-126. Redlin, S.C., 1991. Discula destructiva sp. nov., cause of Dogwood Anthracnose Mycologia 83(5):633-642, Rhoades, D.F. 1979. Evolution of plant chemical defenses against herbivory. In: Herbivores: Their interaction with secondary plant metabolites, G.A. Rosenthal and D.H. Janzen (eds.), Academic Press, New York.

PAGE 128

120 Rhoades, D.F. 1983. Herbivore population dynamics and plant chemistry. In: Variable Plants and Herbivores in Natural and Managed Systems, R.F. Denno and M.S. McClure (eds.), Academic Press, New York. Rodriguez, R. and Melendez, P.L. 1984. Occurrence of Sphaeropsis knot on citron {Citrus medica L.) in Puerto Rico. J. Agric. Univ. P.R. 68:179-183. Russell, L.M. 1944. Description of nine species of Aleuroplatus from eastern North America (Homoptera: Aleyrodiadae). J. Wash. Acad. Sci. 34(10):333-341. Sampson, K. 1933. The systemic infections of grasses by Epichloe typinina (Pers.) Tul. Trans. Brit. Mycol. Soc. 18:30-47. Schardl, C.L. and Clay, K. 1995. Evolution of mutualistic endophytes from plant pathogens. In: The Mycota, Vol. VI: Plant Relationships, Carroll, G.C. and Tudzynski, P. (eds.) SpringerVerlag, Berlin. Schempp-Christoph, M., Pelz, K., Wittmer, A., Schoept, E. and Simon, J.C. 1999. Antibacterial activity of hyperforin from St. John's wort, against multiresistant Staphylococcus aureus and gram-positive bacteria. Lancet 353(91 70):2129. Schemske, D.W., Husband, B.C., Ruckelshaus, M.H., Goodwillie, C, Parker, I.M. and Bishop, J.G. 1994. Evaluating approaches to the conservation of rare and endangered plants. Ecology 75(3):584-606. Schneeberger, N.F. and Jackson, W., 1989. Impact of Dogwood Anthracnose on flowering dogwood at Catoctin Mountain Park. Plant Diagn. Quart. 10:30-43. Schneider, D., Boppre, M., Zqeig, J., Horsley, S.B., BelL T.W., Meinwald, J., Hand, K. and Diehl, E.W. 1982. Scent organ development in Creatonotos moths: regulation by pyrrolizidine alkaloids. Science 215:1 2641 265. Shepherd, R.C.H. 1995. A Canadian isolate of Colletotrichum gloeosporoides as a potential biological control agent for St. John's wort {Hypericum perforatum) in Australia. Plant Prot. Quart. 10(4):1 48-151. Silander, J. A. 1979. Microevolution and clone structure in Spartina patens. Science 203:658-660. Sinclair, W.A., Lyon, H.H, and Johnson, W.T. 1987. Diseases of Trees and Shrubs. Cornell University Press, Ithaca and London. Son, S.H., Moon, H.K. and Hall, R.B. 1993. Somaclonal variation in plants regenerated from callus-culture of hypbrid aspen {Populus alba lxp grandidentata Michx) Plant Science 90(l):89-94.

PAGE 129

121 Soule, M.E. 1987. Viable Populations for Conservation. Cambridge University Press. Southwood, T.R.E. 1972. The insect/plant relationship: An evolutionary perspective. In: Insect/Plant Relationships, H.F. van Emden, (ed.), Blackwell Scientifc, Oxford. SPSS Base 10.0 Applications Guide. 2001. SPSS Inc., Chicago. Stanton, M.L. 1983. Spatial patterns in the plant community and their effects upon insect search. In: Host-seeking Behaviour and Mechanisms, S. Ahmad (ed.), Academic Press, New York. Stauth, D. 1997. Cooperative to tackle forest disease threat. News release, Oregon State University News and Communications Services. Steinger, T., Korner, C. and Schmid, B. 1996. Long-term persistence in a changing climate: DNA analysis suggests very old ages of clones of alpine Carex curvula. Oecologia 105:94-99. Stephenson, S.L. 1986. Changes in a former chestnut-dominated forest after a half century of succession. Am. Midi. Nat. 1 16:173-179. Strong, D.R. and Levin, D.A. 1975. Species richness of the parasitic fungi of British trees. Proc. Natl. Acad. Sci. (USA):72:21 16-21 19. Strong, D.R. and Levin, D.A. 1979. Species richness of plant parasites and growth form of their hosts. Am Midi. Nat. 1 14(l):l-22 Sutton, B.C. 1980. The Coelomycetes: Fungi Imperfecti with Pycnidia, Acervuli and Stromata. Commonwealth Mycological Institute, Kew, Surrey, England. Tallamy, D.W. and Krischik, V.A. 1989. Variation and function of cucurbitacins in Cucurbita: An examination of current hypotheses. Am Nat. 133:766-786. Timmer, L.W., Agostini, J.P., Zitko, S.E. and Zulfiqar, M. 1994. Postbloom fruit drop, an increasingly prevalent disease of citrus in the Americas. Plant Dis. 78:329-334. Travis, S.E., Maschinski, I. and Keim, P. 1996. An analysis of genetic variation in Astragalus cremnophylax var. cremnophylax, a critically endangered plant, using AFLP markers. Molecular Ecol. 5:735-745. Twery, M. J. and Patterson, W. A. 1984. Variations in beech bark disease and its effects on species composition and structure of northern hardwood stands in central New England. Can. J. For. Res. 14:565-574. Van Alfen, N.K., Jaynes, R.A., Anagnostakis, L.L. and Day, P.R. 1975. Chestnut blight: biological control by transmissible hypo-virulence in culture. Science 189:890-891.

PAGE 130

122 Van Treuren, R., Biilsma, R., Ouborg, J.J. and Van Delden, W. 1993. The significance of genetic erosion in the process of extinction. IV. Inbreeding depression and heterosis effects caused by selfing and outcrossing Scabiosa columbaria. Vos, P, Hogers, R., Bleeker, M, Reijans, M., van de Lee, T., Homes, M., Frijters, A., Pot, J., Peleman, J., Kuiper, M., and Zabeau, M. 1995. AFLP: a new technique for DNA fingerprinting. Nucleic Acids Res. 23:4407-4414. Wallace, A, Romney, E.M., Kleinkopf, G.E. and Soufi, S.M. 1978. Uptake of mineral forms of nitrogen by desert plants. In: Nitrogen in Desert Ecosystems, US/IBP Synthesis Series 9, West, N.E. and Skujins, J.J. (eds), Dowden, Hutchinson and Ross, Inc., Stroudsburg, Pennsylvania. Ward, D. 1979. Rare and endangered biota of Florida. Volume 5. Plants. University Presses of Florida, Gainesville, Florida. Warr, S.J., Thompson, K., Kent, M. 1992. Antifungal activity in seed coat extracts of woodland plants. Oecologia 92(2):296-298. Watson, L.E., Uno, G.E., McCarty, N. A. and Kornkven, A.B. 1994. Conservation biology of a rare plant species Eriocaulon kornickianum (Eriocaulaceae). Am J. Bot. 81:980-986. Watts and Hansen, 1994. Pre-Holocene and Holocene pollen records of vegetation history from the Florida peninsula and their climatic implications. Palaeogeography, Palaeoclimatology, and Palaeoecology 109:163-176. Weste, G. 1981. Changes in the vegetation of sclerophyll shrubby woodland associated with invasion by Phytophthora cinnamomi. Aust. J. Bot. 29:261-276. Weste, G. 1986. Vegetation changes associated with invasion by Phytophthora cz>j/7fl/wo7w/ of defined plots in the Brisbane Ranges, Victoria, 1975-1985. Aust J Bot 34:633-648. Weste, G. and Marks, G.C. 1987. The biology of Phytophthora cinnamomi in Australasian forests. Ann. Rev. Phytopathol. 25:207-229. White, W.A. 1958. Some geomorphic features of central peninsular Florida. Florida Geol. Surv. Geol. Bull. 41:1-92. Whiteway, J.A., Alphey, T.J.W., Mathias, P.L., Southney, J.F. 1982. Computer mapping of records of beet cyst nematode (Heterodera schachtii). Plant Path. 31:157-162. Willis, A.J., Ash, J.E. and Groves, R.H. 1993. Combined effects of two arthropod herbivores and water-stress on growth of Hypericum species. Oecologia 96(4):5 17-525.

PAGE 131

123 Willis, A.J., Groves, R.H., and Ash, J.E. 1998. Interactions between plant competition and herbivory on the growth of Hypericum species: a comparison of glasshouse and field results. Aust. J. Bot. 46(5-6):707-721. Woods, F.W. and Shanks, R.E. 1959. Natural replacement of chestnut by other species in the Great Smoky Mountains National Park. Ecology 40(3):349-361 . Wunderlin, R.P. 1998. Guide to the Vascular Plants of Florida. Univ. Fla. Press, Gainesville, Florida. Zhu, Y., Chen, H., Fan, J., Wang, Y., Li, Y., Chen, J., Fan, J., Yang, S., Hu, L., Leung, H., Mew, T., Teng, P., Wang, Z. and Mundt, C. 2000. Genetic diversity and disease control in rice. Nature 406:718-722.

PAGE 132

BIOGRAPHICAL SKETCH G. A. van de Kerckhove hails from the Motor City and received a Bachelor of Science degree from the University of Michigan and a Master of Science degree from Florida State University before arriving in Gainesville. 124

PAGE 133

] certify that 1 have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ,/ jj H. C //> < r rc^ H. C. Kistler, Chair Professor of Plant Pathology I certify that 1 have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate in scope and quality, as a dissertation for the degree of Doctor of Philosophy. _ » Chair, Dept. of Plant Pathology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate in scope and quality, as a dissertation for the degree of Doctor of Philosophy. /n Andrew V. Ogram Associate Professor of Soil and Water Science ] certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Eric S. Menges Director, Plant Laboratory, Archbold Biological Station I certify that 1 have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate in scope and quality, as a dissertation for the degree of Doctor of Philosophy. l^i in. S<^yO£x~-/^~"/yy< Margaret Smither-Kopperl ft Entomos Biological Systems, Gainesville, FL This dissertation was submitted to the Graduate Faculty of the College of Agricultural and Life Sciences and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August 2002 j\ Dean, College of Agricultural Dean, Graduate School