Citation
Association of Culicoides mississippiensis Hoffman with Ilex vomitoria Aiton, the yaupon holly

Material Information

Title:
Association of Culicoides mississippiensis Hoffman with Ilex vomitoria Aiton, the yaupon holly
Creator:
Stewart, Robert Gordon, 1949-
Publication Date:
Language:
English
Physical Description:
xiii, 174 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Chemicals ( jstor )
Essential oils ( jstor )
Female animals ( jstor )
Flowering ( jstor )
Flowers ( jstor )
Hammocks ( jstor )
Insects ( jstor )
Nectar ( jstor )
Phenyls ( jstor )
Species ( jstor )
Ceratopogonidae ( lcsh )
Dissertations, Academic -- Entomology and Nematology -- UF
Entomology and Nematology thesis, Ph. D
Hollies -- Diseases and pests ( lcsh )
City of Palmetto ( local )
Genre:
bibliography ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1996.
Bibliography:
Includes bibliographical references (leaves 162-172).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Robert Gordon Stewart.

Record Information

Source Institution:
University of Florida
Holding Location:
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:
023836975 ( ALEPH )
35809167 ( OCLC )

Downloads

This item has the following downloads:


Full Text











ASSOCIATION OF CULICOMDS MISSISSIPPIENSIS HOFFMAN
WITH ILEX VOMITORIA AITON, THE YAUPON HOLLY


















By

ROBERT GORDON STEWART


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

UNIVERSITY OF FLORIDA 1996


UNIVERSITY OF FLORIDA LIBRARIES
































accc.:~c ~ j-' :A2.0~zy

Margaret K. Stewart, whose warmth, encouragement, simplicity and unwavering love are the cornerstones of my life.














ACKNOWLEDGMENTS


Appreciation is extended to my major professor, Dr. Daniel L. Kline, for providing an exceptional opportunity to learn and apply diverse methods used in vector control, for his guidance and financial support. Many thanks also to my committee members, Dr. E. Greiner, Dr. D. Hall, Dr. H. McAuslane, and Dr. W. Wirth (posthumously) for direction and helpful advice. Thanks to the United States Department of Agriculture-ARS CMAVE and especially Dr. D. Barnard for the use of facilities and for providing funds for the duration of my program. Heartfelt thanks to H. T. McKeithen for his -main man' technical advice on mud, pesticides and computers, and to E. Rountree for administrative assistance, humorous barbs and weather reports.

To J. Harrison thanks for many hours of statistical consultation. Thanks also to the many who helped in my work, including for gas chromatography D. Milne, Dr. M. Whitten, R.

Heath and B. Deuben; for scanning electron microscopy D. Duzak; for arthropod identification D. Amalin, Dr. W. Grogan,


iii








Dr. V. Gupta, Dr. S. Halbert, Dr. A. Hamon, C. Tipping, and

Dr. C. Welborn. For excellent technical advice and much appreciated support thanks to Dr. S. Allan and Dr. G. Hu.

To all my friends in Gainesville, especially James and Grace Okine, Dini Miller, Likui Yang, Mbulaheni Nthangeni, Hilary George, Alicia Daniel, and Christine Masson, words cannot express enough the appreciation for bringing normalcy to my life in hectic and often stressful times.

My mother and father were a constant source of moral support. To them and to my sisters, Diana and Carol, my warm

appreciation is extended for their encouragement throughout my academic and professional career.
























iv


I















TABLE OF CONTENTS


. . . . . . . . . . . . . . . . . iii


ACKNOWLEDGMENTS . . . LIST OF TABLES . . . LIST OF FIGURES . . .


viii


ABSTRACT . . . . . . . . . . . . . . . . . . . . . . .

CHAPTERS

1 INTRODUCTION . . . . . . . . . . . . . . . . . .

2 ILEX VOMITORIA AITON . . . . . . . . . . . .
Introduction . . . . . . . . . . . . . . . . . . .
Materials and Methods . . . . . . . . . . . . . .
Distribution Mapping . . . . . . . . . . . .
Spatial and Temporal Patterns of Flowering
Sample Group Selection and Evaluation . . . .
Results . . . . . . . . . . . . . . . . . . . . .
Distribution Mapping . . . . . . . . . . . .
Spatial and Temporal Patterns of Flowering . . .
Sample Group Selection and Evaluation . . . . .
Discussion . . . . . . . . . . . . . . . . . . .
Distribution Mapping . . . . . . . . . . . .
Spatial and Temporal Patterns of Flowering
Sample Group Selection and Evaluation.....


. . . . . . . . . . . . . . . . . . x


xii


1


.7
7
16 16 19 19
21 21 21
34
39 39
42 42


3 ILEX VOMITORIA FEEDING GUILD . . . . . .
Introduction . . . . . . . . . . . . . .
Materials and Methods . . . . . . . . . .
Pre-flowering Samples . . . . . . . .
Flowering Season Samples . . . . . .
Visitation Rhythms . . . . . . . . .
Anthrone Tests . . . . . . . . . . .


. . . . 44
. . . . 44
. . . . 49
. . . . 49
. . . . 50
. . . . 52
. . . . 53


v










Results . . . . . . . . . .
Pre-flowering Samples . .
Flowering Season Samples


Uncommon visitors . . . . . .
Common visitors . . . . . . .
very common visitors . . . .
Visitation Rhythms . . . . . . . .
Culicoides mississipniensis
Dasyhelea mutabilis . . . . .
Thysanoptera . . . . . . . .
Anthrone Tests . . . . . . . . . .
Biting Rates of Q. mississiDienli Discussion . . . . . . . . . . . . . .
Pre-flowering Samples . . . . . . .
Flowering Season Samples . . . . .
Uncommon visitors . . . . . .
Common visitors . . . . . . .
Very common visitors . . . .
Visitation Rhythms . . . . . . . .
Culicoides mississioniensis
Dasyhelea mutabilis
Thysanoptera . . . . . . . .
Anthrone Tests
Biting Rates of . mississipDiensis


4 FLORAL VOLATILE PHENOLOGY OF ILE VOMITORIA
Introduction . . . . . . . . . . . . . .
Materials and Methods . . . . . . . . . .
Volatile Collection . . . . . . . . .
GC/MS Analysis . . . . . . . . . . .
Data Analysis . . . . . . . . . . . .


Results


Sample Analysis . . . . . . . . . .
Blank and non-flowering samples Pollenkitt samples . . . . . .


Flowering season samples


Data Analysis . . . . . . . . . . .
Phenology of volatile emissions Cannonical correlation analysis Discussion . . . . . . . . . . . . . .
Sample Analysis . . . . . . . . . .
Blank and non-flowering samples Pollenkitt samples . . . . . .


102 102 108 108
112 113 115 115 . . 115
115 . . 115
121 . . 121
122 130 . . 130
130 130


.
.
.


. . .
. . .


. . . . . 55
. . . . . 55
55 58
. . . . . 62
. . . . . 64
. . . . . 65
. . . . . 65
. . . . . 73
78
. . . . . 82
. . . . . 85
88 88
. . . . . 89
89
91
. . . . . 92
93
. . . . . 93
. . . . . 94
. . . . . 96
. . . . . 97
100


. .


. .
. .
.
. .
. .
. .
.


. .










Data Analysis . . . . . . . . . . . .
Phenology of volatile emissions Cannonical correlation analysis


5 SUMMARY AND CONCLUSIONS . . . . . . . . . .
Research Results . . . . . . . . . . . .
Yaupon Holly . . . . . . . . . . . .
Sex ratio . . . . . . . . . . .
Size . . . . . . . . . . . . . .
Start of flowering . . . . . . .
Percent bloom . . . . . . . . .
Yaupon Feeding Guild . . . . . . . .
Profile of arthropod visitors
Visitation rhythms . . . . . . .
Sugar and blood feeding behavior
Floral Volatiles . . . . . . . . . .
Identification . . . . . . . . .
Phenology . . . . . . . . . . '
Canonical correlation analysis
Conclusions . . . . . . . . . . . . .


APPENDICES

A FLORAL PHENOLOGY OF IE VOMITORIA . . . .

B PERCENT BLOOM DATA . . . . . . . . . . . .

C GC TRACES OF POLLEN AND FLORAL VOLATILES

D CALCULATION OF SAMPLE CHEMICAL MASSES . . .

REFERENCES . . . . . . . . . . . . . . . . . . .

BIOGRAPHICAL SKETCH . . . . . . . . . . . . . . .


. . . 132 . . . 132 . . . 133


135 135 135 136 136 136 137 137 138 138 139
140 140 141 142 143


. . . 146 . . . 152 . . . 158 . . . 161 . . . 162 . . . 173


vii


. .
. .
. .
. .
. .
. .
. .
. .
. .
. .
. .
. .
. .
. .














L:ST OF ?ABLES


Table 2A=

2-1. Key to map of 1&A vomitoria positions on Hammock A . . . . . . . . . . . . . . . . . . . 23

2-2. Number of 1. vomitoria plants on Hammock A by size and sex . . . . . . . . . . . . . . . . 24

2-3. JIM vomitoria plants included in study . . . . 35 2-4. Mean bud counts for L. vomitoria, March 1995 . . 35 2-5. Percent bloom groups . . . . . . . . . . . . . . 40

2-6. Mean peak bloom . . . . . . . . . . . . . . . . 40

3-1. Invertebrates collected from I. vomitoria prior
to the start of flowering . . . . . . . . . . . 56

3-2. Invertebrates collected from 1. vomitoria during
the entire flowering season . . . . . . . . . . 57 3-3. Arthropods identified from AFS sweep samples
of 1. vomitoria, spring 1995 . . . . . . . . . 59 3-4. Mean number of principal insects collected
by diel period, spring 1995 . . . . . . . . . . 69 3-5. Mean number of principal insects collected
by plant, spring 1995 . . . . . . . . . . . . . 69

3-6. Mean number of principal insects collected
by bloom group, spring 1995 . . . . . . . . . . 72


viii













4-1. Essential pollen oils identified by GC/MS


4-2. Essential oils of flowering male J. vomitoria
identified by GC/MS . . . . . . . . . . . . . . 117

4-3. Mean mass by diel period for each essential
oil identified by GC/MS from flowering male
I. vomitoria . . . . . . . . . . . . . . . . . . 120

4-4. Standardized cannonical coefficients of essential
oils identified by GC/MS for correlation with
capture of !. mississio.iensis . . . . . . . . . 123

4-5. Probability values of essential oils for
inclusion in a model predicting high capture
of C. mississi .iensis . . . . . . . . . . . . . 125

4-6. Probability values for the contributions of
essential oils to a regression model predicting
capture of Q. mississiDiensis . . . . . . . . . 125

4-7. Standardized cannonical coefficients of essential
oils identified by GC/MS for correlation with
capture of 2. mutabilis . . . . . . . . . . . . 128

4-8. Probability values of essential oils for
inclusion in a model predicting high capture
of ]. mutabilis . . . . . . . . . . . . . . . . 129

4-9. Probability values for the contributions of
essential oils to a regression model predicting
capture of 2. mutabilis . . . . . . . . . . . . 129


ix


. . 116














LiST OF FIGURES


Figure pace

2-1. Distribution of Ilex vomitoria . . . . . . . . . 8

2-2. Leaves and flowers of J. vomitoria . . . . . . 10 2-3. Aerial view of salt marsh, Allen Park Road, Yankeetown, Florida . . . . . . . . . . . . . . 17

2-4. Enlargement of Hammock A, Allen Park Road, Yankeetown, Florida . . . . . . . . . . . . . . 18

2-5. Hammock A showing positions of j. vomitoria plants . . . . . . . . . . . . . . . . . . . . 22

2-6. Phenology of flower start for . vomitoria . . 26 2-7. Start of flowering by sex for . vomitoria 33

2-8. Percent bloom as a function of time, throughout the flowering season . . . . . . . . . . . . . 36

3-1. Number of r. mississiDoiensis collected from I.
vomitoria by diel period . . . . . . . . . . . 66

3-2. Daily totals for Q. mississioriensis collected from L.vomitoria by diel period, spring 1995 . 71 3-3. Number of Q. mutabilis collected from I. vomitoria by diel period . . . . . . . . . . . . . . . . 74

3-4. Daily totals for Q. mutabilis collected from .1. vomitoria, spring 1995 . . . . . . . . 77


x











Figure page

3-5. Number of thrips collected from I.
vomitorisa by diel period . . . . . . . . . . . 79

3-6. Daily totals for thrips collected from I.
Yomitoria, spring 1995 . . . . . . . . . . . . 83

3-7. Mean 1 min vacuum samples for three principal insects relative to percent bloom . . . . . . . 84

3-8. Percent sugar feeding of C. mississiiensis by diel period . . . . . . . . . . . . . . . . . . 86

3-9. Comparison of Q. mississiDiensis biting densities by diel period to male and female
visitations on flowering 1. vomitoria ..... 87


xi








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 ASSOCIATION OF CULICOIDES MISSISSIPPIENSIS HOFFMAN
WITH ILEX VOMITORIA AITON, THE YAUPON HOLLY By

ROBERT GORDON STEWART

AUGUST 1996


Chairperson: D. L. Kline
Major Department: Entomology and Nematology

The association between the salt marsh biting midge, Culicoides mississippiensis Hoffman, and flowering male yaupon holly, Ilex vomitoria Aiton, was investigated in field and laboratory experiments in 1995 and 1996. Over 200 insect and

floral volatile samples collected in spring 1995 were analyzed to identify phenological patterns of insect visitation in relationship to floral essential oil emissions. Insects were collected during one minute intervals, four times a day by AFS sweeper. Samples consisted mainly of thrips species (59.5%) and two ceratopogonids, Dasyhelea mutabilis Kieffer (30.6%) and C. mississippiensis (3.4%).

Adult C. mississippiensis were shown by anthrone tests to feed on nectar of male I. vomitoria flowers. Each of the three main insects visited flowers most frequently in the late xii








afternoon. Ceratopogonid visits were most abundant in the first half of the flowering season and during peak flowering. Thrips visits were most abundant in the second half of the flowering season during early and late flowering.

Twenty essential oils were identified from volatile samples of I. vomitoria by GC/MS. Two different patterns of

emission were detected. One pattern was characterized by increasing output from early morning to mid-afternoon, followed by a decrease towards sunset. The second pattern showed increased output throughout the day up to sunset. Essential oils most strongly correlated with high sample numbers of C. mississippiensis were dimethyl benzene, apinene, P-pinene, 1,4-dimethyl benzene, 4,8-dimethyl-1,3,7nonatriene, ethyl benzaldehyde, and ethenyl benzaldehyde. Those most strongly correlated with high sample numbers of D. mutabilis were dimethyl benzene, a-pinene, 1,4-dimethyl benzene, ethyl benzaldehyde, linalool oxide, 1-(2,4-dimethyl phenyl) ethanone, and phenyl acetonitrile. Future studies should examine attraction of C. mississippiensis to a blend in

which those essential oils not highly correlated to capture of D. mutabilis are manipulated.


xiii















CHAPTER 1
INTRODUCTION






Culicoides mississippiensis Hoffman, the salt marsh biting midge, is a major pest along the Florida Gulf Coast. Breeding and foraging zones of C. mississippiensis cover huge tracts of coast land, overlapping with areas of human habitation and outdoor activities. Even at lower than peak

population levels, biting rates frequently exceed several hundred per hour and can surpass two thousand per hour at peak levels (Lillie et al. 1988). The resulting interaction

between midges and humans has negatively impacted local economies, recreational activities and public health.

No control program is currently being administered in Florida. Insecticidal mists and fogging provide only short term population reduction because of repopulation from

untreated areas and development of resistance (Blanton and Wirth 1979). Moreover, salt marshes are considered

ecologically sensitive areas and broadcast spraying of


1








2


pesticides is no longer an acceptable option. Impoundment, the flooding of breeding zones by construction of dikes with tidal gates, was implemented in the 1930s and 1940s with some

success, but was stopped in the 1950s because of environmental concerns (Blanton and Wirth 1979).

Residents of severely infested localities such as Yankeetown, Levy County, Florida have mixed feelings about the need for control. During peak season, when Q.

mississippiensis is at extremely high densities in early morning and late afternoon, yard work is done in the middle of the day. While some people feel that control of their sand

gnats would help improve property values, others fear the attraction their region would gain to developers and tourists (personal communications).

Nevertheless, reduction of Culicoides spp. populations should be a public health priority in part because of allergic reaction to bites, but more importantly because of the potential disease vectoring. Worldwide, Culicoides spp. are known vectors for many viral and filarial diseases (Linley et al. 1983) . In the Caribbean, Q. furens (Poey) transmits Mansonella ozzardi Manson, causative agent of mansonellosis, a filarial disease. Interaction between active Q. furens









3

populations, found on both of Florida's coasts, and infected immigrants might result in establishment of mansonellosis in Florida. Bluetongue virus, a pathogen of livestock, is vectored by C. variipennis (Coquillett) in the United States and the eastern equine encephalitis virus has been isolated

from pooled Culicoides in Georgia (Karstad et al. 1957). Little attention has been given to the vector potential of C. mississippiensis. An ecologically acceptable control program

would not only bring relief in recreational and economic terms, but would serve to reduce the potential medical and veterinary impact of this pest as well as provide a model for control of other important Culicoides spp.

Culicoides mississippiensis has been reported along most of the U. S. Gulf Coast (Blanton and Wirth 1979). Culicoides

mississippiensis was first described by Hoffman (1926) and female cotypes from Pass Christian, Mississippi deposited in

the U. S. National Museum. Its appearance and biology are similar to those of C. hollensis (Melander and Brues), and until the late 1960s was thought to coexist with C. hollensis throughout its Gulf Coast range. Previous C. hollensis

identifications for the Gulf Coast have been reclassified as C. mississippiensis (Blanton and Wirth 1979).








4

Extensive studies on the biology of C. mississippiensis have been carried out in the salt marshes of Yankeetown, Levy County, Florida. Culicoides mississippiensis has two annual population peaks, one in the spring and the other in the fall

(Kline 1986). Emergent males and females feed on flower nectar. After their first oviposition event females convert to a haematophagous feeding behavior, showing preference for mammals as blood sources (Blanton and Wirth 1979) . Host searching is highest at dawn and dusk but may occur at reduced levels throughout the daytime (Lillie et al. 1988).

Oviposition occurs in salt marsh mud, mainly in stands of cordgrass, Spartina alterniflora Loisel (Kline and Wood 1988,

Kline 1989), which is alternately covered by salt water or exposed to the air due to tidal patterns.

Oviposition sites are in proximity to scattered islets of land called hammocks. Hammock vegetation is characterized by white cedar, southern pine, live oak, cabbage palm, saw palmetto and holly. Yaupon holly, Ilex vomitoria Aiton, is the most common holly species in Yankeetown. It blooms from mid-March to mid-April when C. mississippiensis populations are at their highest.








5

Culicoides mississippiensis has been shown to be

associated with flowering I. vomitoria (Lillie and Kline 1986). Field tests of attraction to detached flowers of I.

vomitoria and seven other plants were carried out using a portable plexiglass olfactometer by the author in April 1993.

In the absence of visual cues, significant attraction was demonstrated for both C. mississippiensis and Dasyhelea mutabilis Kieffer, a common gnat, only to flowers of I. vomitoria (unpublished personal data).

A follow-up study of C. mississippiensis visitation to flowering I. vomitoria and two other flowering plants, the saw palmetto Serenoa repens (Bartram) Small and the cabbage palm Sabal palmetto (Walter), was carried out in spring 1994. As

flowering of I. vomitoria ended in mid-April that of S. repens began. Similarly, the flowering season of S. palmetto started in late May as that of S. repens was ending. Each of those plants is the dominant flowering plant in Yankeetown hammocks during its own flowering season. Visitations of C.

mississippiensis to I. vomitoria were consistently high, whereas they were low to S. repens and non-existent to S. palmetto. Attraction to flowering 1. vomitoria was shown to be a fact, but the nature of the attractant system remained








6


unknown. An examination of the association between C. mississippiensis and I. vomitoria is needed to identify possible candidate floral attractants.

The main research objective of the present study was to investigate the association of C. mississippiensis with I. vomitoria. Specific objectives of the study were 1) to characterize the floral phenology of an I. vomitoria population in a salt marsh hammock of Yankeetown, Levy County,

Florida; 2) identify visitation patterns of the arthropod community associated with flowering male I. vomitoria to aid assessment of possible control on non-target arthropods; 3) establish the basis of primary attractancy of C.

mississippiensis to flowering male I. vomitoria; and 4) identify the phenological patterns of chemicals emitted by flowers of male I. vomitoria during the entire flowering season and determine correlation between specific chemicals and high C. mississippiensis visitations.
















CHAPTER 2
ILEX VOMITORIA AITON


Introduction

Ilex vomitoria Aiton, classified in the angiosperm family Aquifoliaceae, is a dioecious shrub or small tree growing to eight meters, found throughout the southeastern region of the United States (Figure 2-la) (Hudson 1979) . In Florida its distribution is coastal north of Brevard and Sarasota counties, inland in Polk, Marion, Alachua and Baker counties, and in all Panhandle counties except Leon and Calhoun (Figure 2-lb) (Wunderlin and Poppleton 1977).

The Aquifoliaceae are distinguished from other

angiosperms by being oligostemonous (i.e., having the same number or fewer stamens as perianth members) , syncarpous (i.e., having fused carpels) and by bearing fruit with few seeds and only one per carpel (Stebbins 1951). Family genera

are divided into two groups, those that are deciduous and those that are evergreen, including Ilex (Wunderlin and Poppleton 1977). Ilex vomitoria differs from other species in


7









8


A





--------- --------N"
*
Al* ** \
















B























Figure 2-1. Distribution of hIex vomitoria. A. United
States (reprinted with permission of C. M. Hudson).
B. Florida (reprinted with permission of R. P. Wunderlin
and J. E. Poppleton.)









9


the genus in having completely crenate leaves (Figure 2-2a). Hume (1953) grouped American hollies according to persistence

of leaves and fruit color. Species may be deciduous with black fruit, deciduous with red fruit, evergreen with black fruit, or evergreen with red fruit. Ilex vomitoria falls into

the last category. It is also the only fasciculate holly native to the U. S. (Hume 1953), producing flowers on the previous season's stem growth.

In the Aquifoliaceae, nectar is produced at the base of petals or occasionally on the upper surface near the middle of

petals (Loesener 1942). The flowers of I. vomitoria are tetramerous, with male flowers having four stamens and a rudimentary ovary while female flowers have four nonfunctional stamens and a pistil ending in a four-lobed stigma (Figure 2-2b,c). Bawa and Opler (1975) found that flowers of

most dioecious plants in a lowland tropical semideciduous forest of Costa Rica are small and white to yellow or green. Male I. vomitoria flowers are small, cream-colored and more numerous than female flowers (Hu 1979). Flowers are borne in fascicles. Fascicles of male plants have several secondary flower clusters, called cymes, with up to five flowers per









10


A


U


B


C


Figure 2-2. A. Leaf of Ilex vomitoria. B. Male flower
of I. vomitoria. C. Female flower of I. vomitoria.
(reprinted with permission of C. M. Hudson.)








11

cyme while those of female plants have one reduced cyme with a maximum of seven flowers (Hu 1979).

Although I. vomitoria is related in geological time to Asian holly species, more recent lineage can be traced to an original stock in the Ouachita Mountains of western Arkansas and eastern Oklahoma (Hu 1979). It has thrived in the post oak savannah of east central Texas, an area intensively exploited for cattle grazing. Because its dense growth interferes with grazing and movement of equipment, I. vomitoria is controlled by herbicidal applications (Bovey et

al. 1972, Duncan and Scifres 1983, Meyer and Bovey 1985). Adaptation to the drier conditions of seashore and hammocks has enabled I. vomitoria to establish itself in coastal areas

(Hu 1979) . The salt marshes invaded by I. vomitoria are protected wetlands and it is unlikely that the plants' growth will be controlled there as has been the case in Texas.

Historically, I. vomitoria has been important in both Native American and colonial cultures. The leaves when

roasted and boiled twice in water yield a black tea (Hume 1953). For centuries before colonial invasion, Native Americans used what they called yaupon tea as a social

beverage, medicine, and emetic (Hudson 1979). It was the








12

third use, to induce vomiting in order to rid the body of contamination before council meetings, that inspired the

species name of the plant. For much of the 17th century, prior to the advent of tea and coffee as important beverages

in Europe, yaupon tea was an export to Spain, France and England (Hudson 1979). Its use was again popularized in the South during the Civil War when coffee and tea supplies were cut off.

Florida is in a sub-tropical zone of vegetation. The relative abundance of dioecious plant species compared to monoecious species in South Florida (27% of all plant species) approaches that of tropical forests of Costa Rica (21-29%) but is twice the number of temperate forests in the United States (6-17%) (Bawa and Opler 1975). Ilex vomitoria and possibly

other hollies share two attributes of tropical dioecious plants relative to pollination.

While temperate dioecious plants are primarily wind

pollinated (Stebbins 1951), tropical dioecious plants have been shown to be insect pollinated (Bawa and Opler 1975). The

Aquifoliaceae are in the order Celestrales, which includes insect and wind pollinated families (Brizicky 1964) . Hollies are found both in temperate and sub-tropical zones of the








13

United States and are probably insect pollinated throughout the family range. Cavigelli et al. (1986) argue for insect

pollination of I. montana, a holly found from New York to Louisiana. Ilex cassine L. and I. krugiana Loess., both found in Florida, are insect pollinated (Tomlinson 1974). High numbers of insect visitations to I. vomitoria in Florida, especially Nematocera and Hymenoptera (personal observation), also implicate insect pollination for this holly species. The

small size and abundance of I. vomitoria pollen, however, suggest the possibility of wind dispersal. Such dispersal may have no value for pollination, but may result in a pollen plume whose volatiles serve as attractants for insect visitors.

Temperate dioecious plants produce larger female flowers

than male flowers (Baker 1948). The flowers of male and female I. vomitoria plants, however, are similar in size, a characteristic more typical of tropical dioecious plants (Bawa and Opler 1975).

Bawa and Opler (1975) observed a strong correlation between the timing of flowering in dioecious plants in Costa Rica forests and seasonal emergences of insect pollinators. In that study most insect visitors were small, with Hymen-









14

optera being the most likely pollinators. The flowering of I.

vomitoria coincides with the spring population peak of C. mississippiensis. Few other flowering plants are as ubiquitous as I. vomitoria in March and April. Poison ivy, Toxicodendron radicans L., is the only other plant in flower at that time from which C. mississippiensis adults have been collected in vacuum samples.

Fine tuning of the relationship between pollinator and plant can be effected through the timing of release of nectar and pollen rewards (Janzen 1966). Prior to flowering, plants may encourage insect visitations by secretions of extrafloral nectaries. The Aquifoliaceae are not listed by Elias (1983) among plant families with extrafloral nectaries and none were seen on I. vomitoria by the investigator. Most flowers begin nectar production before pollinators arrive and stop once a maximum volume nectar pool is achieved (Cruden et al. 1983).

Stiles (1975) suggests early morning as a good time for diurnally pollinated plants to present large amounts of nectar because production could occur throughout the night. However, cool nighttime temperatures may inhibit or delay nectar secretion in some plants (Cruden et al. 1983).








15

Red fruits develop on I. vomitoria and yield a maximum

of four seeds. Diaz-Colon et al. (1970) determined that aqueous extracts of the fruit can inhibit root growth of other

plant species' germinated seeds. Perhaps this capability gives I. vomitoria a competitive advantage over other plant species in colonization of new habitats. The fruit may be eaten by birds and seeds dispersed in their droppings. Germination and growth are slow, but more rapid population expansion can be achieved asexually by root-sprouting (Hume 1953).

The salt marsh habitat of Florida's Big Bend Gulf Coast

region, from St. Petersburg to Pensacola, is an important breeding ground for C. mississippiensis. Preliminary vacuum samples of flowering I. vomitoria made in spring 1994 (Chapter 4) contained both adult male and female C. mississippiensis. Nectar sugars are a significant reward to C. mississippiensis adults for I. vomitoria floral visitations (Chapter 4). Attraction to flowers is mediated at least in part by floral volatile emissions (Metcalf and Metcalf 1992).

The purpose of this study was to characterize the I. vomitoria population of a salt marsh hammock in Yankeetown, Levy county, Florida to facilitate volatile sampling.









16

Specific objectives of the study were to map distribution of plants, determine spatial and temporal patterns of flowering,

and select a representative group of plants for volatile collections.



Materials and Methods



Distribution Mapping

An aerial photograph taken by U. S. Air Force Tactical Reconnaissance Wing in 1981 (Figure 2-3) was used as a template for producing a scale map of the study site hammock, referred

to as hammock A (Figure 2-4). Four other neighboring hammocks were periodically visited for comparison to hammock A and were

designated as hammocks B-E. Quadrats 10m on each side were assessed and plant positions plotted on the map of hammock A.

Plants were considered as individuals if separated by at least one meter or if of different sex. Plants were tagged with pink surveyor's tape on which the hammock and code number were written. Each plant was sized as small (2m high).








17






4-) 4-i Sr* H 4-4


O 4r14 -Hq (0
A v M

4


2












f0 W,~ 4 4J






























0 -H
4-)

1k ( -d 0 -H



-Hi
4

U)


















tsr
r


atN


r ~ Jr


*


4


Figure 2-4. Enlargement of hammock A, Allen Park Road, Yankeetown,
Florida.


AM;








19


Spatial and Temporal Patterns of Flowering

Observations made in the spring of 1993 and 1994 indicated that the I. vomitoria flowering season begins in March and ends in April. In spring 1995 weekly observation of bud development was made starting in February. As soon as the first plant began flowering, complete surveys of hammock A I. vomitoria plants were made on an average of 2-3 day intervals. Date of first flowering and sex of each plant were noted. The data obtained were evaluated for sex ratio as well as spatial

and temporal variations in bloom by sex across hammock A. Daily temperature readings were made with a Taylor alcohol thermometer during each of the four diel sample periods, corresponding to hours at sunrise, late morning, early afternoon, and sunset.

Sample Group Selection and Evaluation

The cost of materials and time required for movement between plants resulted in focusing the study on male I. vomitoria. Large male plants, up to a maximum of three plants at any given time, were included in volatile sampling. Candidate sample plants were identified before the start of flowering. Plants were brought into the study at the first sign of flowering and replaced when fully senesced. In some








20

cases it was necessary to replace a plant before flowering had

completely stopped in order to accommodate inclusion of a plant just starting to flower.

Prior to the start of the flowering season the mean number of buds per centimeter of male stem was determined. Ten 5-cm-stem samples were randomly selected from each of three male plants with well-developed buds. All buds were counted using a dissecting microscope. Mean bud counts per centimeter of stem were determined for each plant and plant means examined by ANOVA and the REG-WQ (Ryan-Einot-GabrielWelsch Multiple Range Test) (SAS Institute, 1989).

Male bloom was quantified to make correlations between flowering phenology and volatile phenology. A male budded branch (sample segment) was enclosed in a tomato support cage.

During each sample period, the number of open flowers of a sample segment was counted. Counting was facilitated by marking counted cymes with a small piece of masking tape. At the end of sampling, the total length of each plant's budded stems was measured. That length multiplied by the mean flowers per centimeter of stem gave the total flowers per sample segment. Percent bloom was calculated by dividing the number of open flowers in any given sample period by the total








21

number of flowers on the sample segment. Mean values for three day peak bloom were compared for the five plants by ANOVA and REG-WQ. Data for all plants included in the study were combined in a group bloom profile to statistically compare peak insect activity and percent bloom (Chapter 3).



Results



Distribution Mapping

Ilex vomitoria plant positions are shown on the map in Figure 2-5. A key to specific plant positions is given in Table 2-1. There are 145 I. vomitoria plants on hammock A, in an area of approximately 5600 M2. Fifty-nine plants are

female, 71 are male and 15 are of undetermined sex (did not flower) . The sex ratio for Hammock A is 1.20:1. The data are presented in size groups in Table 2-2. Spatial and Temporal Patterns of Flowering

The first plant to flower in Hammock A, starting on March 14, 1995, was a male coded A36. The last bloom start date, on April 15, was for a female plant coded A104. Bloom start dates covered a span of 33 days. Dates for the start of flowering of individual plants are given in Appendix A.




















6







1


A B IC D EF G H IJK L M N










~~zr 4
IIF




K' 37


F 4 ft.
N 4 xi-


Krrtrr-


Figure 2-5. Hammock A showing positions of Ilex vomitoria in quadrats with sides of 10 m. Key to plants in each quadrat given in Table 3-1.


tQ tNi


O









23


Table 2-1. Key to map of Figure 2-5, relating Ilex vomitoria
plant designations to plotted positions on map.

Plant Designation Map Location

1 C6/7
2,3 D6/7
4-11 E6
12 D6
13 E5
14 E6/7
15-21 F7
22,29-31 G6
23 F6
24-26 EF5
27,28 FG4/5
32,33 G6/7
34-36 G7/8
37 FG8
38,62-64 H7
39,40 H6/7
41, 48, 142 H5
42-49,135 15
50-52,139 13/4
53-55,143 IJ3
56-58,132 IJ4/5
59, 61,145 J4
60,140 15/6
65-69 H7
70-75 IJ8
76-78,144 19
79,80 18
81-83,85-87 J8/9
84 K9
88,89 K8
90 L8
91,92 M8/9
93-95 M7
96,97,141 N7
98-100 LM6
101-105 M5
106-110 L4








24


Table 2-1. continued


Plant Designation


Map Location


111-117 L5
118,119,124-126 K4
120-123 K3/4
127-130 K5
131 JK5
133,134 IJ5
135-138 G8


Table 2-2. Number of Ilex vomitoria plants on Hammock A in
Yankeetown, Florida classed by size.

Size
Sex Small Medium Large Total

M 26 25 20 71

F 26 14 19 59

Unknown 14 0 1 15

Total 66 39 40 145








25

Phenology of start of flowering is shown in Figures 2-6.A-G. The mean start date for flowering of male plants was day 11,

while that of female plants was day 18. The difference

between male and female start dates was significant (ANOVA, F=109.0, df=l1,118, p<0.0001).

When size groups are considered individually, large plants are shown to flower earlier than medium plants, which flower earlier than small plants. A statistically significant difference for start of flowering exists between large and small male plants, and between all male plants, except small, and female plants (ANOVA, F=13.51, df=5,124, p<0.0001). Mean start dates for large, medium and small male plants were 9, 11, and 13 days, respectively. For large, medium and small

female plants mean start dates were 16, 17 and 19 days, respectively. No statistical differences were observed between female size groups relative to the date for start of flowering. A significant interaction existed for start of flowering with the mean daily temperature calculated from daily high and low temperature readings. A graph for the start of flowering by sex is shown in Fig. 2-7.







26


12


10

9




j6:-O,



3







12


12

9






A M-~
6V M
4 U



TjII Figure 2-6A. Phenology of flower start for _I. vomitoria on hammock A. Start of bloom indicated by white square. Top is for 3/14/95,
bottom for 3/16/95.







27


A B C D EF G H I JK L IM N 0



10 ;
9




6



40'
3KK
i2Z"
1M


12
11 A
10 AM

99







5 -4


Figure 2-6B. Phenology of flower start for I. vomitoria on hammock A. Start of bloom indicated by white square. Top is for 3/18/95, bottom for 3/21/95.







A B C D E 12


28


F G H IT JIK

10


10
9
8



6 7
U5
itt


2w

1

11

10




7 I

6
5

44


2
1

Figure 2-6C. Phenology of flower start for I. vomitoria on hammock A. Start of bloom indicated by white square. Top is for 3/24/95,
bottom for 3/27/95.


I

r


I
it









12


10,

9
8
7

6

5


A B C D E F G H I J K L M N







A--


A


12,

101


1074T
9r


8/A72EL


!AA

3 AW


1


a
1K


I? I


Figure 2-6D. Phenology of flower start for I. vomitoria on hammock A. Start of bloom indicated by white square. Top is for 3/29/95, bottom for 3/31/95.


29






30


A B IC ID IE F G 3 H I IIJI K IL IM IN 12
1

190 4 1
104





6 A
54


2



12




9
7'


6
5'
W&







Figure 2-6E. Phenology of flower start for I. vomitoria on hammock A. Start of bloom indicated by white square. Top is for 4/3/95,
bottom for 4/7/95.


01







31


_A 1B IC ID EI F IGI HI TJ K IL IM IN 1
12


10 r S


f -s 9"


A 4 - Ik


5








12


10 4

9




6








1a

Figure 2-6F. Phenology of flower start for 1. vomitoria on hammock A. Start of bloom indicated by white square. Top is for 4/9/95,
bottom for 4/12/95.








32


A JB C ID IE IF IGI HI I JI K IL IM IN 11




8
7

6


1 ~ T


Figure 2-6G. Phenology of flower start for I. vomitoria on hammock A. Start of bloom indicated by white square. Above is for 4/15/95.


0!








33


35

- male 30 - - female


25


420


0
Q)
15


10


5


0
March I I -I I ; r
12 14 16 18 21 24 27 29 31 3 7 9 12 15
March April
Date


Figure 2-7. First day of flowering by sex.
Ilex vomitoria on Hammock A, Allen Park Road, Yankeetown, Levy Co., March-April
1996.








34


Sample Group Selection and Evaluation

Plants included in the study are listed with dates of flowering in Table 2-3. A total of five large male plants were sampled.

Results of bud count data are given in Table 2-4. All means were evaluated by ANOVA and REG-WQ multiple comparison test. Male means are not significantly different from each

other (ANOVA, F=0.17, df=2,27, p=0.85), and so the overall mean for males is valid. Female means do differ significantly (ANOVA, F=25.8, df=2,27, p<0.0001), with each mean shown by multiple comparison to be significantly different. Male mean bud count per centimeter stem is significantly greater than female mean bud count (ANOVA, F=100,000, df=6,53, p<0.0001). The mean value of 20.4 buds/cm was used to evaluate percent bloom of male plants included in the study.

Data for calculation of percent bloom is found in Appendix B. Percent bloom for each of the five study plants is shown in Figure 2-8 A to E. Combined percent bloom data

for the study group was stratified into 8 groups for later statistical examination of the relationship between insect visitations and flowering status (Chapter 3). The 8 groups were constructed to include more than 15 data values each









35


Table 2-3. Ilex vomitoria plants included in the study by hammock code designation.

Plant Sex Flowering dates

A36 M 3/16-3/31

A14 M 3/24-4/7

A62 M 3/25-4/7

A35 M 4/1-4/15

A129 M 4/7-4/15










Table 2-4. Mean bud counts cm' (SE) for I. vomitoria in March
1995. Ten 5-cm-segments of budded stem from each of 3
male and 3 female plants were evaluated.

Sex
Plant Male Female

1 21.6 (2.61) 2.5 (0.40)

2 20.1 (3.56) 1.5 (0.28)

3 19.3 (1.85) 4.8 (0.31)


Overall mean 20.4 (1.55) 2.9 (0.33)


I










25



20



15



10



5

0
0


U
4' 25 Q)


20 15 10 5



0-


I I I I I 1 1 24 25 26 27 28 29 31 March
Date


1 2 3 4 7 April


Figure 2-8. Graph of percent bloom as a function of
time, throughout the flowering season.
A. Profile for plant A36.
B. Profile for plant A14.


36


A














I I F I I I I I I I





16 17 18 19 20 21 24 25 26 27 28 29 31
March


B








37


25 20 15 10 5 -


D


1 1 2 3 41 7 8 91 101 11 121 141 15
Api.il

Date

Figure 2-8. Graph of percent bloom as a function of
time, throughout the flowering season. C. Profile for plant A62. D. Profile for plant A35.


C



















25 26 127 28 1 29 311 1 2 3 4 7 8 9 March April


0 -


E
0
0




0)
a4


Ir


20 15 10



5



0


25 n








38


25

E

20



o 15


0
4J
a)
8 10a)

5



0

7 8 9 10 11 12 14 15
April
Date Figure 2-8. Graph of percent bloom as a function of
time, throughout the flowering season.
E. Profile for plant A129.








39

and to reflect the flowering cycle (increase to peak followed by decline). Those groups are shown in Table 2-5.

Statistical comparison of the five plants of the study group was made relative to the three days of highest percent bloom for each (peak bloom). Peak bloom declined significantly over the flowering season. Plant A36, the first plant to flower, had a significantly higher peak bloom than plant A14, and both A36 and A14 were significantly higher than the

other three plants in the study (ANOVA, F=271.1, df=14,45, p<0.0001). Peak bloom means are summarized in Table 2-6.



Discussion



Distribution MaDDing

The positions of I. vomitoria as shown on the map in Figure 2-5 are evenly scattered across hammock A. Although

male plants outnumber female plants, no irregular distribution patterns relative to sex were detected. It should be noted, however, that formal analysis of spatial association was not carried out.

Sex ratio for hollies may be skewed in favor of males. Cavigelli et al. (1986) found a male:female ratio of 1.35:1









40

Table 2-5. Percent bloom (B) groups. Eight groups were
constructed to reflect natural flowering cycle dynamics.


Group B(% bloom) n*
1 B 2 1 B<2 17
3 2 B<5 27
4 5 B<22 28
5 22>B25 18
6 5>B22 32
7 2>B2l 27
8 l>B 33


* Number of samples included in each bloom group.






Table 2-6. Mean peak bloom. Means are calculated for highest
percent bloom over three days for each plant included in the volatile study. Values of all four diel periods are
used for each day.

Plant N Mean SE

A36 12 18.5A* 0.38

A14 12 10.1B 0.94

A35 12 3.02C 0.18

A62 12 2.88C 0.18

A129 12 1.92D 0.13


*Means followed by different letters different at the 0.05 probability level comparison test.


are significantly by REG-WQ multiple









41

for I. montana Torr. & Gray. Richards' (1988) study of four natural or seminatural plots of I. aquifolium L. in Britain showed male:female sex ratios ranging from 1.11:1 to 2.16:1. The male:female sex ratio of I. opaca Aiton was determined to

be 1.03:1, seeming to contradict the male-skewed trend of other hollies (Clark and Orton, 1967). However, the study of Clark and Orton was done on seedlings which were followed over

a seven-year period in controlled conditions and does not reflect differential survival that may exist in a natural stand of I. opaca.

Two characteristics of hollies tend to complicate sex ratio determination. Their ability to reproduce vegetatively

from roots makes it difficult to distinguish close growing same sex plants as truly unique individuals. The sex ratio of I. vomitoria is lower than that found for I. montana (Cavigelli et al. 1986), but is consistent with a seemingly male-skewed sex ratio for holly species. Several hypotheses to explain male-biased sex ratios in hollies are discussed by

Cavigelli et al. (1986) . The hypothesis favored is that female energetic costs are higher than male costs, resulting in reduced frequency of flowering. Therefore long-term

studies should reveal a greater number of any year's non-








42

flowering plants to be female and the overall sex ratio to be close to 1:1. Studies carried out over several years in Hammock A could determine if this is the case for I. vomitoria.

Spatial and Temporal Patterns of Flowering

In their study of tropical dioecious plants, Bawa and Opler (1975) determined that male flowers open earlier in the

day than female flowers. The self-incompatibility of such plants would favor out crossing, resulting in a population with greater genetic variation. Ilex vomitoria also shows temporal segregation of male and female floral rewards, with

male plants producing flowers earlier in the season than female plants. This arrangement results in pollinator visitation of male plants before female plants. The advantage

for the plant of such a system is to increase the chance a pollinator will visit a female flower bearing pollen from an unrelated plant.

Sample Group Selection and Evaluation

The choice of conducting the study on male rather than female plants was based in part on the potential for greater levels of flowering. Plants with more flowers would provide

greater quantities of floral volatiles as well as insects.









43

Bud counts that showed significantly greater numbers in male than female plants supported the choice of male plants. Male plants were also shown to begin flowering significantly earlier than female plants. Choice of male plants therefore

allowed earliest start of insect sampling, yielding a more complete profile of insect visitation throughout the flowering season.

The level of percent bloom achieved by plants through the flowering season varied from high early in the season to low late in the season. It is difficult to judge how much the low bloom level achieved by A62, A35 and A129 was the product of thrips infestations and how much was a normal end of season phenomenon. Damage to buds caused by high thrips populations may be at least in part responsible for considerably lower percent blooms of plants. This explanation is supported by data presented in Chapter 3 showing that highest numbers of thrips in collections were significantly associated with lowest percent bloom.















CHAPTER 3
ILEX VOMITORIA FEEDING GUILD





Introduction



The Diptera genus Culicoides Latreille, 1909 includes a large number of species of extremely small biting flies with worldwide distribution. They are classified in the suborder Nematocera, family Ceratopogonidae, subfamily Ceratopogoninae and tribe Culicoidini (Wirth et al. 1980). Culicoides spp. are important pests and disease vectors. Males feed on the nectar of flowers, while females feed on blood. The genus shows great diversity of blood source preference, but most

feed on mammals and birds (Downes 1970) . In many species autogeny, that is the production of the first batch of eggs

following a sugar meal, provides an adaptive advantage in habitats where blood meals may not always be available (Downes 1970).

Worldwide there are more than 1,000 species of Culicoides (Wirth et al. 1980). In North America there are 144 described 44








45


Culicoides species, of which 48 can be found in Florida (Blanton and Wirth 1979). As many as 21 of those species have been found in Yankeetown, Florida (Kline 1986).

Adult Culicoides can be distinguished from other ceratopogonids by having a 3-segmented antenna, including the flattened scapes (segment 1), enlarged pedicel (segment 2) and flagellum (segment 3), which is subdivided into 13 flagellomeres. Wings are characteristically patterned with light

spots on a dark background. These patterns, as well as

antennal ratios, the number and type of antennal sensilla, and male terminalia are important taxonomic characters for species identification (Wirth, personal communication) . Species identifying criteria for Florida species are outlined by Blanton and Wirth (1979).

The four most common anthropophilic species in Yankeetown are C. mississippiensis Hoffman, C. furens Poey, C. barbosai Wirth & Blanton, and C. floridensis Beck (Lillie et al. 1988). While C. mississippiensis is bivoltine, with spring and fall population peaks, the other three species are most often found

from April to October when C. mississippiensis begin to decline and reach their lowest levels. Host seeking is

characterized as crepuscular in C. mississippiensis, C.








46

barbosai and C. floridensis but nocturnal in C. furens (Lillie et al. 1988).

Culicoides mississipiensis has been shown to be autogenous (Davis 1981). No sugar sources have been identified prior to the present study although midge vacuum samples by Lillie and Kline (1986) suggested flowering Ilex vomitoria Aiton, the yaupon holly, as a possible sugar source in spring. An important first step in studying the association between C. mississippiensis and I. vomitoria is to

establish the nature of the reward of attraction to the flowering plants. It may seem reasonable to assume that the insect visits are linked to sugar feeding, but determination of such a link should be made by conducting anthrone tests.

The success of many nematoceran pests and disease vectors is linked to autogeny and sugar feeding. Sugar feeding not only fuels initial egg production but also provides energy for flight (Van Handel 1984), enabling search for blood meal hosts and oviposition sites, and increases longevity (Hunter

1977, Jamnback 1961) . Lack of contact with humans and alternate hosts may favor higher sugar feeding rates (Edman et al. 1992, Van Handel et al. 1994) . Implications for control of protozoan vectors in sylvan cycles are significant given









47

that sugar meals increase transmission rates, possibly by nourishing developing parasites (Young et al. 1980).

The cold anthrone test (Van Handel 1972) is the primary

tool used in studying sugar feeding behavior of insects. Crushing a sugar fed insect in yellow anthrone solution will cause the reagent to change to color shades ranging from light

green to dark blue. Mosquitoes (Bidlingmayer and Hem 1973, Edman et al. 1992, Reisen et al. 1986, Smith and Kurtz 1994, Van Handel et al. 1994), black flies (McCreadie et al. 1994, Walsh and Garms 1980, Young et al. 1980), and biting midges (Magnarelli 1981, Mullens 1985) are the principal nematocerans on which anthrone testing has been used.

Nectar is considered a primary attractant of insects to flowers (Faegri and Pijl 1979) . It contains sugars, amino acids, lipids, anti-oxidants, and vitamins (Inouye 1980) . In addition to being important components of the nectar reward, sugars and amino acids may also act as attractants themselves (Baker and Baker, 1986). Nectar glycerides that fluoresce may also attract insects (Kevan 1976).

Flowers have been shown to emit a large number of volatile compounds which serve to attract insects. In some cases the volatiles act as mimics of insect glandular








48

secretions (Borg-Karlson and Groth 1986). Some pollens

produce their own set of volatiles that are important insect attractants (Pijl 1960, Dobson et al. 1987) . Perfumes, UVreflecting nectar guides, and visible color guides act as close range attractants to insects (Snyder and Miller 1972). Establishment of the nectar of I. vomitoria as a sugar source

for male and female C. mississippiensis would provide a rationale for examining floral volatiles as possible attractants.

Anthrone tests made on other Culicoides species (Magnarelli 1981, Mullens 1985) confirmed prior sugar feeding by host-seeking females, but did not identify sugar sources. Such data may be of limited value for determining circadian feeding rhythms and population feeding rates because anthrone negatives and weak positives could include insects that fed some time before sampling (Walsh and Garms 1980, Reisen et al. 1986, Smith and Kurtz 1994) . Insect samples taken directly from flowering plants would provide more reliable information since the time of collection is more directly related to the opportunity for sugar acquisition.

An ecologically important concern that must be addressed

before control of a pest is attempted is the impact such control may have on non-target arthropods. If control of C.









49

mississippiensis is to be based on its association with flowering male I. vomitoria, then other insects that visit the flowering plants must be identified and temporal patterns of floral visitations evaluated. Identification of floral

attractants that correlate to peak foraging by target insects,

but not to non-target insects, may facilitate selection of chemicals to be tested for response by insects.

The objectives of the following study were to identify the principal insects present on flowering male I. vomitoria, both before and during the flowering season, characterize the periodicity of their activity relative to the time of day and

time within the flowering season and establish the nectar feeding behavior of C. mississippiensis. A comparison of the

nectar and blood feeding of C. mississippiensis was also made.



Materials and Methods



Non-flowerinq Samples

Samples of insects were collected from three male I. vomitoria prior to the start of the flowering season at the

Yankeetown study site using an AFS sweeper (Meyer et al. 1983) . The sweeper is a backpack vacuum device with a









50

flexible extension arm that holds a collection container. Insects were collected in cardboard pint containers fitted with 100 mesh screen hot-glued into top and bottom covers.

Plants sampled were A14, A36 and A39. Samples were made by vacuuming vegetation for one minute. Vacuum sweeps were

made away from the collector to avoid influencing sample contents. Samples were collected on 5 March, 11 days before the start of the flowering season, at noon and sunset, and on 14 March, 2 days before flowering, at sunset only.

At completion of sampling, containers were closed and placed inside zipper-top plastic bags for insect anaesthetization by C02. Insects were then transferred to vials and

held on wet ice until return to Gainesville. Cooling of

samples before processing was needed to prevent spiders from

feeding on insects. All collected arthropods were freezekilled and preserved in 70% ethanol for later processing.

Totals were recorded for major taxa. Mean number per minute sample and standard error were calculated. Flowering Season Samples

Five flowering male I. vomitoria were vacuum-sampled with

an AFS sweeper four times a day throughout the flowering season of the holly. Plants sampled (dates of sampling)








51

included A36 (16-31 March), A14 (24 March-7 April), A62 (25 March-9 April), A35 (1-15 April) and A129 (7-15 April) . Oneminute AS sweeper samples were collected and handled according to the protocol for non-flowering season samples. The diurnal

portion of the diel cycle was divided into ten periods and samples were made during the first, fourth, seventh and tenth periods, corresponding to sunrise, late morning, early afternoon and sunset. Sampling was limited to the diurnal portion of the diel cycle because C. mississippiensis does not feed at night (Lillie et al. 1988) . Nighttime sampling may have yielded a more complete picture of the I. vomitoria feeding guild as well as the phenology of essential oil emission, but was not carried out.

Most male and female C. mississippiensis were removed from samples and stored separately at -600C until processed by the cold anthrone test. All other sample contents,

including some C. mississippiensis, were preserved in 70% ethanol and processed at a later date.

Preserved sample contents were sorted according to general taxa and subsequently as many identified to species as

possible. The total number of all invertebrates collected was recorded for each plant by diel period for each sample day of








52

the flowering season. Mean number per minute sample and standard error were calculated and general evaluations made for all taxa except C. mississippiensis and the two dominant insects, Dasyhelea mutabilis Kieffer and thrips. Those three

taxa were submitted to more detailed analysis of temporal visitation rhythms.

Visitation Rhythms

Sample data were analyzed to identify any daily or

seasonal rhythmic patterns of visitation by C. mississippiensis to flowering yaupon. Detailed analysis of visitations by D. mutabilis and thrips was also made because they would

probably be the most abundant non-target victims. Graphs were made for each of those insects of the numbers collected by plant and for the entire flowering season. Mean (standard error) one-minute samples by diel period were calculated to identify diurnal visitation patterns. Those means were

examined by ANOVA and significant differences identified by REG-WQ.

Visitation relative to flowering season was also analyzed. Insect means for each plant were compared by ANOVA

and REG-WQ to identify visitation patterns relating to the entire flowering season. Plants were sampled sequentially in









53

an order that can be considered to represent early (A36), mid (A14, A62), and late (A35, A129) flowering season.

Bloom groups were described for I. vomitoria male plants in Chapter 2 (Table 2-5). These bloom groups (BG) correspond to stages in flowering that may be referred to as initiation

(BGl,2), expansion (BG3,4), decline (BG5,6) and senescence (BG7,8). An analysis of visitations by bloom group was made comparing the mean number of insects captured for each of the eight bloom periods described. Means were examined by ANOVA

and significant differences identified by REG-WQ multiple comparison.

Anthrone tests

Anthrone tests were run on a stratified random sub-sample of males and gravid females. Gravid females could be easily distinguished from the few non-gravid females in samples by their swollen, lighter pigmented abdomens. Midges were

assigned chronological numbers according to the date and time of sampling. Random number tables were used to select midges

for anthrone testing, with numbers proportionately allotted to the four sample time periods for each sex. The estimate of nectar feeding rate was made using a 90% confidence interval

and sample size calculated by the confidence interval formula.









54

Because - was unknown, a value of 0.5 was used to calculate the sample size. By this method, 384 males and 384 females were tested to obtain a 90% confidence interval. Parous

females taken in samples were also tested.

Fresh anthrone reagent was made following the method of Van Handel (1972) every two to three days . Selected insects from each cohort were rinsed in distilled water to remove any nectar, which may have adhered to them during vacuuming, and placed individually in 10 x 75 mm test tubes. Midges were then crushed with a glass stirring rod in 0.2 ml anthrone reagent and evaluated one hour later for color change. Results were reported as positive or negative. The stirring rod was rinsed in two beakers of water and then dried with a paper towel. Controls were made by mixing the stirring rod in anthrone reagent alone after every six insect tests. Lack of available known negative insects, due to low field emergence levels at the time of testing, precluded better experimental control. Data was analyzed by chi-squared tests. Biting rates of C. mississippiensis

Midge biting rates were determined for the same four sample periods from 16 March to 15 April. Midges were

aspirated from the left forearm and numbers recorded every 60









55

seconds for 20 consecutive 60 second periods. When densities exceeded 20 bites/60sec, a 5 min aspiration sample was made. During the next fifteen minutes midges were killed, but not collected, as they were counted. Samples were later

identified to species.




Results




Non-flowering Samples

A total of nine AS sweeper samples were collected from non-flowering male I. vomitoria. The composition of the samples is given by summary taxa in Table 3-1.

Very low numbers of insects for major taxa were present on plants that had not yet flowered. In particular, the mean numbers of thrips, D. mutabilis, and C. mississippiensis were less than 2 each per sample. No taxon dominated samples and most taxa had less than 10% of the total sample contents. Flowering Season Samples

A total of 218 AS sweeper samples were collected from flowering I. vomitoria between 16 March and 15 April 1995. Sample contents are given by summary taxa in Table 3-2.









56


Table 3-1. Invertebrates collected from Ilex vomitoria prior
to the start of flowering. Samples (n= 9) from plants A36, A14 and A39 collected by AFS sweeper on 5 March and
14 March 1996.

Taxa Non-flowering Non-


totals (%)


Thysanoptera Dasyhelea mutabilis (male) D. mutabilis (female) Culicoides mississippiensis
(male)

C. mississippiensis (female) Other Nematocera Hymenoptera Collembola Hemiptera Coleoptera Other Diptera Acari

Lepidoptera Homoptera Araneida Psocoptera Mollusca


15 15

14

0


12 19 18 15

3

2

9

3

6 10

9

2

3


(9.7) (9.7) (9.0) (0.0)


(7.7) (12.3) (11.6) (9.7) (1.9) (1.3) (5.8) (1.9) (3.9) (6.5) (5.8) (1.3) (1.9)


flowering means (SE*)

1.7 (1.1) 1.7 (1.2) 1.6 (1.1) 0.0 (0.0)


1.3

2.1 2.0 1.7 0.3

0.2 1.0 0.3 0.7 1.1

1.0

0.2 0.3


(0.5) (0.7) (0.7) (0.5) (0.3)

(0.2) (0.7)

(0.2) (0.2) (0.8) (0.3)

(0.2) (0.2)


* Standard error









57


Table 3-2. Invertebrates collected from Ilex vomitoria
during the entire flowering season. Samples (n= 218) from plants A36, A14, A62, A35 and A129 collected by AFS
sweeper from 16 March and 15 April 1996.

Taxa From Flower From Flower
totals (%) means (SE*)


Thysanoptera Dasyhelea mutabilis (male) D. mutabilis (female) Culicoides mississippiensis
(male)

C. mississippiensis (female) Other Nematocera Hymenoptera Collembola Hemiptera Coleoptera Other Diptera Acari

Lepidoptera Homoptera Araneida Psocoptera Mollusca


65504 (59.5) 16223 (14.7) 17469 (15.9)

2459 (2.2)


1332

2394 1902 796 697 323 287

222 143 137

122 23

22


(1.2) (2.2) (1.7) (0.7)

(0.6) (0.3) (0.3)

(0.2) (0.1) (0.1) (0.1) (<0.1) (<0.1)


* Standard error


299.1

74.1 79.8

11.2


6.1 10.9 8.7 3.6 3.2 1.5 1.3 1.0 0.7 0.6 0.6 0.1 0.1


(30.7) (8.0) (7.8)

(1.2)


(0.5) (0.9) (0.6)

(0.4) (0.2) (0.1) (0.1) (0.1) (0.07) (0.07) (0.07)

(0.02) (0.02)









58

Specific identifications and relative occurrence are shown in Table 3-3.

More than 90% of all arthropods collected were either thrips or D. mutabilis. Culicoides mississippiensis comprised 3.4% of the sample total and was the third most abundant arthropod on flowering yaupon. Non-arthropods were very uncommon and included anoles, observed feeding on insects, and small birds. Arthropod visitors may be generally grouped into three categories: uncommon (less than 1% of the sample total), common (between 1 and 10% of the sample total) and very common (more than 10% of the sample total) . General analysis for uncommon and common visitors, except C. mississippiensis, follow. More detailed analysis of data for C.

mississippiensis and the two very common insect taxa is given separately in the section on temporal visitation patterns. Uncommon visitors

Psocoptera (<0.1%) . No sample had more than two psocids. Samples containing psocids were few and scattered throughout the flowering season.

Araneida (0.1%). Spiders could be found at low levels throughout the flowering season, with an unusually high number

(12) taken from A35 during diel period 1 on April 2.









59


Table 3-3. Arthropods identified from AS sweep samples of Ilex vomitoria, spring 1995. Occurrence is indicated by +++ (very common), ++ (common), and + (uncommon).

Taxa Occurrence

Arachnida
Acaridae
Trombiculidae +
Araneae
Anyphaenidae
Habana sp. +
Araneidae
Argiope sp. +
Metazygia sp. +
Neoscona sp. +
Salticidae
Hentzia sp. +
Theridiidae
Anelosinus sp. +
Coleosoma sp. +

Insecta
Collembola
Entomobriidae +
Coleoptera
Curculionidae +
Oedemeridae +
Diptera
Ceratopogonidae
Culicoides barbosai Wirth & +
Blanton
C. furens (Poey) +
C. mississippiensis Hoffman ++ Dasyhelea mutabilis Kieffer +++ Dasyhelea sp. +
Forcipomyia sp. +
Cecidomyidae +
Chironomidae +
Mycetophilidae +
Sciaridae +









60


Table 3-3 (continued)


Taxa


Occurrence


Hemiptera
Lygaeidae Homoptera
Aphididae
Toxoptera aurantii (Boyer de
Fonscolombe)
Cixiidae
Pintalia vibex Kramer
Myndus sp.
Psyllidae
Gyropsylla ilicis (Ashmead)
Aleyrodidae Hymenoptera
Eulophidae (4 spp.)
Mymaridae (1 sp.) Torymidae (2 spp.) Lepidoptera
Geometridae Lycaenidae
Callophrys gryneus (Hibner)
Nymphalidae
Agraulis vanillae (L.) Thysanoptera
Thripidae
Frankliniella bispinosa
(Morgan)
Leptothrips sp.
Heterothrips sp.


+



+

+

+


++ ++ ++


+

+


+
+








61

Homoptera (0.1%). These insects were few and scattered throughout the flowering season. The most often encountered Homoptera included Toxoptera aurantii (Boyer de Fonscolombe) and Gyropsylla ilicis (Ashmead) , the yaupon psyllid. All whiteflies in samples were adults.

Lepidoptera (0.1%). All lepidoptera collected from

flowering yaupon were Geometrid larvae. Damage to leaves caused by grazing larvae, though not extensive, could be seen on many of the yaupon plants in the study site. Cumulative

totals were greater than 50% at the end of two weeks and greater than 90% at the end of three weeks. More than 85% of all larvae were collected between 24 March and 8 April. The only adult Lepidoptera observed infrequently feeding on male yaupon flowers were the cedar hairstreak, Callophrys gryneus (Habner), and the Gulf fritillary, Agraulis vanillae (L.).

Acari (0.2%). Mites found on flowering yaupon were of two major groups, detrivores and parasites. Several

trombiculiids were found clinging to the abdomens of D. mutabilis collected on flowering yaupon.

Diptera (Brachycera and Cyclorrhapha) (0.3%). Larger Diptera were rarely trapped in AS sweeps but were occasionally observed at low numbers to be feeding from flowers. Most








62

frequently observed were syrphids (Syrphidae) and deer flies (Tabanidae). Most non-ceratopogonid Diptera visiting flowers were small cyclorrhaphan flies. Highest numbers occurred in two distinct phases, two days in late March (26-27 March) and nine days in April (2-12 April). Specimens collected on those eleven days accounted for more than 80% of all nonceratopogonid Diptera.

Coleoptera (0.3%). Coleoptera were collected in relatively low numbers throughout the flowering season. Some, especially the oedemerids, were observed feeding on pollen.

Hemiptera (0.6%). The majority of bugs collected from flowering yaupon were lygaeids (Lygaeidae). Lygaeid numbers increased steadily from 25 March to 7 April. Both adults and immatures were found at all times of the day on plants.

Collembola (0.7%). Nearly all Collembola collected were

of the family Entomobryidae. Their numbers increased to a peak of 103 on 27 March and then remained at stable but lower levels to the end of the flowering season. Higher numbers were found in diel period 1 samples than all other samples. Common visitors

Hymenoptera (1.7%). Bees and wasps are important visitors of flowering yaupon holly. Although relatively few in








63

number, megachilids and vespids could be seen carrying copious amounts of pollen as they visited both male and female plants. Vespids were also frequently seen searching in yaupon foliage.

The majority of Hymenoptera in AS sweep samples were very small members of Chalcidoidea families. Species most commonly collected from flowering yaupon were of the Eulophidae, Torymidae, and Mymaridae. Over the entire flowering season significantly fewer Hymenoptera visits occurred during diel period 1 than the other three collection periods. Collections

made during diel periods 7 and 10 had significantly higher numbers of Hymenoptera than earlier collections. Nearly two-thirds of all Hymenoptera collected were in samples of the last 11 days of the flowering season.

Nematocera (excluding C. mississippiensis and D. mutabilis) (2.2%). Nematocera were the second most abundant group of insects on flowering yaupon. Separate consideration is given to C. mississippiensis because of their relatively high numbers in samples. Other ceratopogonids were also found in important numbers, including Dasyhelea species and Forcipomvia species. Only two other species of Culicoides were collected, those being C. barbosai Wirth & Blanton and C. furens (Poey), both in very low numbers. Additional families








64

significantly represented in samples included Cecidomyidae, Chironomidae, Mycetophilidae, and Sciaridae.

Culicoides mississippiensis (3.4%). These ceratopogonid

insects were the third most abundant insects in AS sweep samples of flowering I. vomitoria in 1995. Culicoides

mississippiensis made up 95.7% of all biting midges collected, while 4.1% were C. barbosai and 0.2% C. furens. A total of 3,791 C. mississippiensis adults were collected, of which 2,459 (2.2%) were male and 1,332 (1.2%) female. Higher

numbers of males than females is in agreement with expected

feeding differences. Detailed analysis of C. mississippiensis data follows under visitation rhythms. Very common visitors

Dasyhelea mutabilis (30.6%). This common ceratopogonid gnat was the second most abundant insect in 1995 samples. A total of 33,692 D. mutabilis adults, of which 16,223 (14.7%) were male and 17,469 (15.9%) female, were collected. Detailed analysis of D. mutabilis data follows under visitation rhythms.

Thysanoptera (59.5%). Thrips were the most abundant insects in samples of flowering I. vomitoria. Holes, made by thrips, could be seen in numerous buds of plants A62, A35 and









65

A129. Total bloom of those three plants was far less than of A36 and A14, which were not infested. The vast majority of thrips collected were Frankliniella bispinosa (Morgan). Both adults and immatures were collected from flowering yaupon.

Detailed analysis of thrips data follows under visitation rhythms.

Visitation Rhythms

Culicoides mississippiensis. The combined total of males and females collected for each plant by diel period are shown in graphs of Figure 3-1. Fluctuating numbers reflect both daily and seasonal rhythms.

Daily rhythms. Combined data from all plants were

analyzed by GLM and REG-WQ multiple comparison to identify diel foraging preferences. Mean (standard error) values for each diel period are shown in Table 3-4. Diel 10 had a significantly greater number of biting midges in collections

than either diel 7 or diel 4, but was not significantly greater than diel 1.

Seasonal rhythms. Combined data was analyzed by GLM and REG-WQ multiple comparison to identify seasonal patterns of C. mississippiensis visitations to flowering I. vomitoria. Results of this analysis are given in Table 3-5. Mean








66


200 180 160 140 120 100 80 60 40 20

0


200


119 T20 121 '24 125 1 261 27 '28 129 131


-H











4-4
0 a;
z


March


April


Date


Figure 3-1. Number of Culicoides mississippiensis
collected from .lex vomitoria by diel period
(sunrise, late morning, early afternoon, sunset). A. Samples from plant A36. B. Samples from plant A14.


A


B

















24 25 26 27 28 29 31 1 2 3 4 7


16 17 T18
March


180 160

140 120 100 80 60

40 20


0


.








67


I 25' 26 '27 28 ' 291 311 1 2 3 4 7 8 9 March April


UH











~44
0
w4


Figure 3-1. Number of Culicoides mississippiensis
collected from Ilex vomitoria by diel period
(sunrise, late morning, early afternoon, sunset). C. Samples from plant A62. D. Samples from plant A35.


2? n


180 160 140 120 100 80 -


C


60

40 20 0


200 180 160

140 120


D


















1 TT2 3 -4 7- 8 9 10T- 1-1 '12 1 5T
April Date


100 80 60 40 20 0 -


-








-








-








68












200

U 180 - E

2 160

140 H 120

-H 100

80

o 60 ci 40 z 20

0
7 8 9 110 11 12 14 15
April

Figure 3-1. Number of Culicoides mississippiensis
collected from Ilex vomitoria by diel period
(sunrise, late morning, early afternoon, sunset).
E. Samples from plant A129.









69


Table 3-4. Mean number of principal insects collected
by diel period, spring 1995.

Diel Period C. mississipp. D. mutabilis Thrips


1 19.2 AB* 29.4D 321.0AB

4 7.4 B 103.0B 201.6B

7 11.6 B 189.8C 307.4AB

10 28.0 A 279.2A 394.3A


*Means within columns followed by different letters are significantly different at the 0.05 probability level by REG-WQ multiple comparison test.





Table 3-5. Mean number of principal insects collected by plant, spring 1995.


Plant


C. mississipp. D. mutabilis Thrips


A36 29.8 A* 204.5B 59.4C


A14 24.0 A 287.3A 83.2C

A62 16.0 AB 93.9C 378.9B

A35 7.1 B 95.2C 334.8B

A129 5.8 B 58.2C 806.OA



*Means within columns followed by different letters are significantly different at the 0.05 probability level by REG-WQ multiple comparison test.









70

numbers in descending order correspond to the order in which plants were sampled during the flowering season.

Significantly higher numbers of C. mississippiensis were collected from plants A36 and A14 than from A35 or A129 (ANOVA, F=2.99, df=27,25, p<0.004) . The mean for A62 was intermediary to others, though not significantly different from the means for A36 and A14. More than 50 adult C.

mississippiensis were collected on ten occasions from A36, on six from A14 and on five from A62, but only once from A35 and never from A14.

When data for all plants are combined, major collection peaks for C. mississippiensis can be identified every 3-4 days (Figure 3-2). The last such peak occurs on 4 April with two lesser peaks occurring on 8 and 12 April.

Examination of C. mississippiensis capture by bloom

group, using the eight bloom groups defined in Chapter 2 (Table 2-5), showed that significant differences existed among bloom group means (ANOVA, F=6.43, df=7,204, p<0.0001) . REG-WQ evaluation of bloom group means is given in Table 3-6. The mean for bloom group 5, corresponding to peak bloom and the

start of bloom decline, was significantly greater than all other bloom group means except those of bloom groups 3 and 4.










500





400





300





200
0




100





0
SI I I I I I I I I I I I
16 17 18 19 20 21 24 25 26 27 28 29 31 1 2 3 4 7 8 9 10 11 12 14 15
March April

Figure 3-2. Daily totals for Culicoides mississiopiensis
collected from Ilex vomitoria, spring 1995.









72


Table 3-6. Mean number of principal insects collected
per sample by bloom group, spring 1995. Bloom groups
correspond to the indicated range of percent bloom.


Bloom N C. mississip. D. mutabilis Thrips
group

1 (0-1%) 30 5.9 D* 67.5 CD 205.2 BC

2 (1-2%) 17 9.1 CD 129.9 BCD 415.8 ABC

3 (2-5%) 27 26.0 ABC 253.1 AB 201.6 BC

4 (5-22%) 28 28.1 AB 306.2 A 57.0 C

5 (22-5%) 18 37.3 A 203.0 ABC 123.3 C

6 (5-2%) 32 17.1 BCD 151.1 ABC 288.9 ABC

7 (2-1%) 27 11.6 BCD 100.1 BCD 590.9 A

8 (1-0%) 33 9.3 CD 31.5 D 521.4 AB



*Means within columns followed by different letters are significantly different at the 0.05 probability level by REG-WQ multiple comparison test.








73

Dasyhelea mutabilis. These ceratopogonids were the second most abundant insects in AS sweeper samples of flowering I. vomitoria in 1995. The combined totals of males and females collected for each plant by diel period are shown in graphs of Figure 3-3 A-E.

Daily rhythms. Combined data were analyzed by GLM and REG-WQ multiple comparison to identify diel foraging preferences. Mean (standard error) values for each diel period are shown in Table 3-4. Significant differences were

found for all diel periods, and mean visitations increased from sunrise to sunset.

Seasonal rhythms. Combined data relative to source plants were analyzed by ANOVA and REG-WQ to identify seasonal

visitation patterns. Plant samples represent early (A36, A14), middle (A62, A35) and late (A129) flowering season. Results, given in Table 3-5, show a significantly higher number of visitations made to early flowering plants than to middle and late flowering plants.

Peak numbers in samples are shown for 3-4 day intervals (Figure 3-4). A sharp decline in the number of D. mutabilis visiting I. vomitoria can be seen in the last third of the flowering season.










1500
1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200 100 0

1500 1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200 100
0


16 17 18 19 120 121 24 25 126 727 128 129 13 March


24 25 26 27 28 29 31 1 2 3 4 7
March April


Date
Figure 3-3. Number of D. mutabilis
collected from I. vomitoria by diel period.
(sunrise, late morning, early afternoon, sunset). A. Samples from plant A36. B. Samples from plant A14.


74


A


0
a)


B








75


1500
1400 1300
1200 1100 1000 900 800 700 600 500
400 300
200 100
0

1500
1400 1300
1200 1100 1000 900 800 700 600 500
400 300
200 100
0


1 2 3 4 7 8 9 10 11 1 r 1


April
Date
Figure 3-3. Number of D. mutabilis collected from I. vomitoria by diel period. (sunrise, late morning, early afternoon, sunset). C. Samples from plant A62. D. Samples from plant A35.


-T


-w


0
w4
a)


'25 '26 '27 28 1 29' 3T1 1 '2 1 3 '4 '7 8 T9 March April


D


-








76


1500
H 1400
1300 1200 1100 1000
900 0 800 4 700 1 600 z 500
400 300 200 100
0
7 9 10 1 12 14 715
April

Figure 3-3. Number of _D. mutabilis
collected from I. vomitoria by diel period
(sunrise, late morning, early afternoon, sunset).
E. Samples from plant A129.










4000 3500 u 3000 2500 2000

0
1500 1000 500



0 -


16 17 18 19 20 21 24 25 26 27 28 29 March


I I I I I I I I I I I I I 31 1 2 3 4 7 8 9 10 11 12 14 15
April


Figure 3-4. Daily totals for D. mutabilis collected from
I. vomitoria,spring 1995.


-2
-2








78

Analysis of the mean number of D. mutabilis collected relative to the level of bloom of plants showed that D. mutabilis is attracted in higher numbers in bloom group 4, corresponding to late rise to peak bloom, followed by bloom groups 3, 5, and 6, in that order (ANOVA, F=6.46, df=7,204, p<0.0001). Mean capture for bloom group 4 was significantly greater than mean captures of bloom groups 1, 2, 7, and 8, corresponding to beginning and end stages of flowering. Results of REG-WQ evaluation of bloom group means is given in Table 3-6.

Thysanoptera. These were the most abundant insects in samples, though not evenly distributed throughout the flowering season. The majority of adult thrips were F. bispinosa. Many immatures were collected throughout the flowering season. More than 65,000 thrips were collected from flowering I. vomitoria, but no effort was made to distinguish the numbers of immatures from adults. The totals of thrips for each plant are represented in graphs of Figure 3-5 A-E.

Daily rhythms. Combined data was analyzed by GLM and REG-WQ multiple comparison to identify diel foraging preferences. Mean (standard error) values for each diel










500

450 400 350 300 250

200 150 100 50

0

500

450


24 25
March


26 271 28 291 311 1 2 1 3 April Date


1 4 7


Figure 3-5. Number of thrips collected from
I. vomitoria by diel period. A. Samples from plant A36. B. Samples from plant A14.


79


A


16 17 18 19' 20 21 '24 125 26 27 '28 129 131
March


U)
04

4

0
w4


B


400 350 300 250 200 150 100 50

0


-








80


2000

1800 -C

1600 1400 1200 1000 800 600 400
04
-H 200

0
25 26 27 28 29 31 1 2 3 4 7 8 9 o March April
2000

1800 D

1600

1400

1200 1000 800 600

400 200

0 i 1 1 2 3 4 7 8 9 10 11 J -A 1T
April
Date Figure 3-5. Number of thrips collected from
.I. vomitoria by diel period.
C. Samples from plant A62.
D. Samples from plant A35.






















3400 3200 3000 2800 2600 2400 2200 2000 1800 1600 1400 1200 1000 800 600 400 200 0-


April

Figure 3-5. Number of thrips collected from
I. vomitoria by diel period. E. Samples from plant A129.


81


U)


04
-H



0


E


12 14 15


1 7 1 8 9 10 1








82

period are shown in Table 3-4. Relatively high numbers are collected in all diel periods, but the mean for diel period 10 is significantly higher than that of diel period 4.

Seasonal rhythms. Data analysis for thrips shows low population levels during the first two-thirds and a population

explosion during the last third of the flowering season. Combined plant totals for the entire season are shown in Figure 3-6.

Examination of bloom group data for thrips showed highest

mean numbers of thrips are associated with early and late phases of flowering, when percent bloom is at its lowest (ANOVA, F=5.27, df=7,204, p<0.0001.) Results of REG-WQ evaluation of bloom group data are given in Table 3-6. Graphs

of bloom group data reflect a very different pattern of visitation for thrips to flowering I. vomitoria from those of the two major ceratopogonids (Figure 3-7). Anthrone Tests

Positive anthrone tests were obtained for 427 of 768 (55.6%) midges tested. The number of male positives were 67

(53.2%), 23 (53.5%), 33 (55.0%), and 68 (43.9%) for diel periods 1,4, 7 and 10 respectively. Gravid female positives for the same periods were 57 (73.1%), 28 (65.1%), 51 (59.3%),









8000 7500 7000 6500 6000 8 5500 5000 4500 4000 43 3500
0
3000 2500 2000 1500 1000 500

0


16 17 18 19 20 21 24 25 26 27 28 29 31 1 2 3 4 7 8 9 10 11 12 14 15 March April

Figure 3-6. Daily totals for F. bisoinosa collected from
I. vomitoria,spring 1995.








84


50


40


30


20


10


0
350 300 250 200 150 100 50 0
700 600 500 400 300 200 100

0


bisvinosa


0-1 1-2 2-5 5-22 22-5 5-2


2-1 1-0


Percent bloom

Figure 3-7. Mean 1 min vacuum samples for three
insect species from flowering male I. vomitoria relative to percent bloom. Bloom increases to
a maximum of 22% and then decreases to <1%.


rd

4
0
flo

0

rd


_C. mississiOiensis


f.


12. mutabilis









85

and 100 (56.5%) (Fig. 3-8) . An overall significant difference (p 0.10) between male and gravid female anthrone positivity was found, with 236 of 384 (61.5%) females having sugar fed compared to 191 of 384 (49.7%) males. Male feeding rates were significantly higher during diel periods 1 (X2=14.426, df=l), 4 (X2=7. 688, df=1) , and 7 (2, =12. 670, df=1) than during diel period 10. Gravid female feeding rates decreased steadily from dawn to dusk, and those of periods 7 (X2=10.560, df=l) and 10 (X2=19.735, df=l) were significantly less than the rate for period 1.

Of 25 non-gravid, parous females collected and tested, 10 (40%) yielded positive anthrone results. The low number of parous females in samples indicates that the method of collection succeeded in excluding blood-meal seeking females. All controls were negative.

Biting rates of C. mississippiensis

Mean biting rates (range) were 140 (3-321), 79 (3-525), 117 (0-510), and 689 (51-2484) bites/h for diel periods 1, 4, 7 and 10, respectively. A comparison of mean biting rates and male/female floral samples for the four collection periods is shown in Figure 3-9. The biting rate for diel period 10 was significantly greater than rates for all other sample periods









86


=II males
females


4 7
Diel period


10


Figure 3-8. Percent sugar feeding of C.
mississippiensis by diel period determined by the cold anthrone test.


100


90 80 -


4)




0



0


70 60 50 40 30 20 10 0 -


1



















1000 900 800 0 700 4 600 500
0

400 300 200 100 0-


female BR








/


I -0- male I


--M
-A-


z


A


1


4


7/I
/
-4'


7


1000


- 900


- 800k,


- 700-.
4IJ . H

- 600
0

- 500 (0


- 400


- 300


- 200


- 100


0


10


Diel period


Figure 3-9. Comparison of the number of male and
female _Q. mississipiensis collected from flowering male I. vomitoria with the mean number of midges biting per hour (BR)for corresponding diel periods.


Al


87


. . '




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 ER9QH11Q0_XONC0H INGEST_TIME 2015-04-13T20:50:50Z PACKAGE AA00029887_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

ftjJIUIWI P'" " '^^^ ASSOCIATION OF CULICOIDES MISSISSIPPIENSIS HOFFMAN WITH ILEX VOMITORIA AITON, THE YAUPON HOLLY By ROBERT GORDON STEWART A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1996 UNIVERSITY OF FLORIDA LIBRARIES

PAGE 2

This work is dedicated to the memory of my mother, Margaret K. Stewart, whose warmth, encouragement, simplicity and unwavering love are the cornerstones of my life.

PAGE 3

ACKNOWLEDGMENTS Appreciation is extended to my major professor, Dr. Daniel L. Kline, for providing an exceptional opportunity to learn and apply diverse methods used in vector control, for his guidance and financial support. Many thanks also to my committee members. Dr. E. Greiner, Dr. D. Hall, Dr. H. McAuslane, and Dr. W. Wirth (posthumously) for direction and helpful advice. Thanks to the United States Department of Agriculture-ARS CMAVE and especially Dr. D. Barnard for the use of facilities and for providing funds for the duration of my program. Heartfelt thanks to H. T. McKeithen for his "main man" technical advice on mud, pesticides and computers, and to E. Rountree for administrative assistance, humorous barbs and weather reports . To J. Harrison thanks for many hours of statistical consultation. Thanks also to the many who helped in my work, including for gas chromatography D. Milne, Dr. M. Whitten, R. Heath and B. Deuben; for scanning electron microscopy D. Duzak; for arthropod identification D. Amalin, Dr. W. Grogan,

PAGE 4

Dr. V. Gupta, Dr. S. Halbert, Dr. A. Hamon, C. Tipping, and Dr. C. Welborn. For excellent technical advice and much appreciated support thanks to Dr . S. Allan and Dr. G. Hu. To all my friends in Gainesville, especially James and Grace Okine, Dini Miller, Likui Yang, Mbulaheni Nthangeni, Hilary George, Alicia Daniel, and Christine Masson, words cannot express enough the appreciation for bringing normalcy to my life in hectic and often stressful times. My mother and father were a constant source of moral support. To them and to my sisters, Diana and Carol, my warm appreciation is extended for their encouragement throughout my academic and professional career. , iv

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS iii LIST OF TABLES viii LIST OF FIGURES X ABSTRACT xii CHAPTERS < ,; . . 1 INTRODUCTION 1 2 ILEX VOMITORIA AITON . 7 Introduction 7 Materials and Methods 16 Distribution Mapping 16 Spatial and Temporal Patterns of Flowering . 19 Sample Group Selection and Evaluation .... 19 Results 21 Distribution Mapping 21 Spatial and Temporal Patterns of Flowering ... 21 Sample Group Selection and Evaluation 34 Discussion 39 Distribution Mapping 39 Spatial and Temporal Patterns of Flowering ... 42 Sample Group Selection and Evaluation 42 3 ILEX VOMITORIA FEEDING GUILD 44 Introduction 44 Materials and Methods 49 Pre-f lowering Samples 49 Flowering Season Samples 50 Visitation Rhythms 52 Anthrone Tests 53 V

PAGE 6

Biting Rates of C. mississippiensis .... 54 Results 55 Preflowering Samples 55 Flowering Season Samples 5 5 Uncommon visitors 58 Common visitors 62 Very common visitors 64 Visitation Rhythms 65 Culicoides mississippiensis 65 Dasvhelea mutabilis 73 Thysanoptera 78 Anthrone Tests 82 Biting Rates of C. mississippiensis 85 Discussion 88 Pre-f lowering Samples 88 Flowering Season Samples 89 Uncommon visitors 89 Common visitors 91 Very common visitors 92 Visitation Rhythms 93 Culicoides mississippiensis 93 Dasvhelea mutabilis 94 Thysanoptera 96 Anthrone Tests 97 Biting Rates of C. mississippiensis .... 100 4 FLORAL VOLATILE PHENOLOGY OF ILEX VOMITORIA . . 102 Introduction 102 Materials and Methods 108 Volatile Collection 108 GC/MS Analysis 112 Data Analysis II3 Results 215 Sample Analysis II5 Blank and non-flowering samples . . . 115 Pollenkitt samples 115 Flowering season samples 115 Data Analysis 12i Phenology of volatile emissions . . . 121 Cannonical correlation analysis . . . 122 Discussion 23 0 Sample Analysis 13 0 Blank and nonflowering samples ... 130 Pollenkitt samples 13 0 vi

PAGE 7

Flowering season samples 131 Data Analysis 132 Phenology of volatile emissions . . . 132 Cannonical correlation analysis . . . 133 5 SUMMARY AND CONCLUSIONS 135 Research Results 135 Yaupon Holly 135 Sex ratio 136 Size 136 Start of flowering 136 Percent bloom 137 Yaupon Feeding Guild 137 Profile of arthropod visitors .... 138 Visitation rhythms 138 Sugar and blood feeding behavior ... 139 Floral Volatiles 140 Identification 140 Phenology 141 Canonical correlation analysis .... 142 Conclusions 143 APPENDICES A FLORAL PHENOLOGY OF ILEX VOMITORIA 146 B PERCENT BLOOM DATA 152 C GC TRACES OF POLLEN AND FLORAL VOLATILES .... 158 D CALCULATION OF SAMPLE CHEMICAL MASSES 161 REFERENCES 162 BIOGRAPHICAL SKETCH 173 vii

PAGE 8

LIST OF TABLES Table page 2-1. Key to map of Ilex vomitoria positions on Hammock A 23 2-2. Number of I., vomitoria plants on Hammock A by size and sex 24 2-3. Ilex vomitoria plants included in study .... 35 2-4. Mean bud counts for I. vomitoria . March 1995 . . 35 2-5. Percent bloom groups 40 26. Mean peak bloom 40 31. Invertebrates collected from 1. vomitoria prior to the start of flowering 56 3-2. Invertebrates collected from I. vomitoria during the entire flowering season 57 3-3. Arthropods identified from AFS sweep samples of I. vomitoria , spring 1995 59 3-4. Mean number of principal insects collected by diel period, spring 1995 69 3-5. Mean number of principal insects collected by plant, spring 1995 69 3-6. Mean number of principal insects collected by bloom group, spring 1995 72 viii

PAGE 9

Table Page 4-1. Essential pollen oils identified by GC/MS . . 116 4-2. Essential oils of flowering male 1. vomitoria identified by GC/MS 117 4-3. Mean mass by diel period for each essential oil identified by GC/MS from flowering male 1. vomitoria 120 4-4. Standardized cannonical coefficients of essential oils identified by GC/MS for correlation with capture of C. mississippiensis 123 4-5. Probability values of essential oils for . . inclusion in a model predicting high capture of C. mississippiensis 125 .'."4-^1 Probability values for the contributions of essential oils to a regression model predicting capture of C. mississippiensis 125 4-7. Standardized cannonical coefficients of essential oils identified by GC/MS for correlation with capture of D. mutabilis 128 4-8. Probability values of essential oils for inclusion in a model predicting high capture of D. mutabilis 129 4-9. Probability values for the contributions of essential oils to a regression model predicting capture of D. mutabilis 129 ii ix

PAGE 10

LIST OF FIGURES Figure page 2-1. Distribution of Ilex vomitoria 8 2-2. Leaves and flowers of 1. vomitoria 10 2-3. Aerial view of salt marsh, Allen Park Road, Yankeetown, Florida 17 2-4. Enlargement of Hammock A, Allen Park Road, Yankeetown, Florida 18 2-5. Hammock A showing positions of I., vomitoria plants 22 2-6. Phenology of flower start for I. vomitoria . . 26 2-7. Start of flowering by sex for I. vomitoria . . 33 28. Percent bloom as a function of time, throughout the flowering season 36 31. Number of C. mississippiensis collected from I. vomitoria by diel period 66 3-2. Daily totals for C. mississippiensis collected from I. vomitoria by diel period, spring 1995 . 71 3-3. Number of D. mutabilis collected from I. vomitoria by diel period 74 3-4. Daily totals for D. mutabilis collected from I. vomitoria . spring 1995 77 X

PAGE 11

Figure , J '. ] page 3-5. Number of thrips collected from I. vomitoria by diel period 79 3-6. Daily totals for thrips collected from I. vomitoria , spring 1995 83 3-7. Mean 1 min vacuum samples for three principal insects relative to percent bloom 84 3-8. Percent sugar feeding of C. mississippiensis by diel period 86 3-9. Comparison of C. mississippiensis biting densities by diel period to male and female visitations on flowering I. vomitoria 87 xi

PAGE 12

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 ASSOCIATION OF CULICOIDES MISSISSIPPIENSIS HOFFMAN WITH ILEX VOMITORIA AITON, THE YAUPON HOLLY By ROBERT GORDON STEWART AUGUST 1996 Chairperson: D. L. Kline Major Department: Entomology and Nematology The association between the salt marsh biting midge, Culicoides mississippiensis Hoffman, and flowering male yaupon holly, Ilex vomitoria Alton, was investigated in field and laboratory experiments in 1995 and 1996. Over 200 insect and floral volatile samples collected in spring 1995 were analyzed to identify phenological patterns of insect visitation in relationship to floral essential oil emissions. Insects were collected during one minute intervals, four times a day by AFS sweeper. Samples consisted mainly of thrips species (59.5%) and two ceratopogonids , Dasyhelea mutabilis Kieffer (30.6%) and C . mississippiensis (3.4%). " Adult C. mississippiensis were shown by anthrone tests to feed on nectar of male I. vomitoria flowers. Each of the three main insects visited flowers most frequently in the late xii

PAGE 13

afternoon. Ceratopogonid visits were most abundant in the first half of the flowering season and during peak flowering. Thrips visits were most abundant in the second half of the flowering season during early and late flowering. Twenty essential oils were identified from volatile samples of I. vomitoria by GC/MS. Two different patterns of emission were detected. One pattern was characterized by increasing output from early morning to mid-afternoon, followed by a decrease towards sunset. The second pattern showed increased output throughout the day up to sunset . Essential oils most strongly correlated with high sample numbers of C. mississippiensis were dimethyl benzene, ccpinene, P-pinene, 1,4-dimethyl benzene, 4 , 8-dimethyl-l , 3 , 7nonatriene, ethyl benzaldehyde, and ethenyl benz aldehyde . Those most strongly correlated with high sample numbers of D. mutabilis were dimethyl benzene, a-pinene, 1,4-dimethyl benzene, ethyl benzaldehyde, linalool oxide, 1(2 , 4-dimethyl phenyl) ethanone, and phenyl acetonitrile . Future studies should examine attraction of C. mississippiensis to a blend in which those essential oils not highly correlated to capture of D. mutabilis are manipulated. xiii

PAGE 14

CHAPTER 1 INTRODUCTION Culicoides mississippiensis Hoffman, the salt marsh biting midge, is a major pest along the Florida Gulf Coast. Breeding and foraging zones of C. mississippiensis cover huge tracts of coast land, overlapping with areas of human habitation and outdoor activities. Even at lower than peak population levels, biting rates frequently exceed several hundred per hour and can surpass two thousand per hour at peak levels (Lillie et al . 1988). The resulting interaction between midges and humans has negatively impacted local economies, recreational activities and public health. ' No control program is currently being administered in Florida. Insecticidal mists and fogging provide only short term population reduction because of repopulation from untreated areas and development of resistance (Blanton and Wirth 1979) . Moreover, salt marshes are considered ecologically sensitive areas and broadcast spraying of

PAGE 15

2 pesticides is no longer an acceptable option. Impoundment, the flooding of breeding zones by construction of dikes with tidal gates, was implemented in the 1930s and 1940s with some success, but was stopped in the 1950s because of environmental concerns (Blanton and Wirth 1979) . Residents of severely infested localities such as Yankee town. Levy County, Florida have mixed feelings about the need for control. During peak season, when C. mississippiensis is at extremely high densities in early morning and late afternoon, yard work is done in the middle of the day. While some people feel that control of their sand gnats would help improve property values, others fear the attraction their region would gain to developers and tourists (personal communications). Nevertheless, reduction of Culicoides spp. populations should be a public health priority in part because of allergic reaction to bites, but more importantly because of the potential disease vectoring. Worldwide, Culicoides spp. are known vectors for many viral and filarial diseases (Linley et al . 1983). In the Caribbean, C. furens (Poey) transmits Mansonella ozzardi Manson, causative agent of mansonellosis , a filarial disease. Interaction between active C. furens

PAGE 16

3 populations, found on both of Florida's coasts, and infected immigrants might result in establishment of mansonellosis in Florida. Bluetongue virus, a pathogen of livestock, is vectored by C. variipennis (Coquillett) in the United States and the eastern equine encephalitis virus has been isolated from pooled Culicoides in Georgia (Karstad et al. 1957). Little attention has been given to the vector potential of C. mississippiensis . An ecologically acceptable control program would not only bring relief in recreational and economic terms, but would serve to reduce the potential medical and veterinary impact of this pest as well as provide a model for control of other important Culicoides spp. Culicoides mississippiensis has been reported along most of the U. S. Gulf Coast (Blanton and Wirth 1979) . Culicoides mississippiensis was first described by Hoffman (192 6) and female cotypes from Pass Christian, Mississippi deposited in the U. S. National Museum. Its appearance and biology are similar to those of C. hollensis (Melander and Brues) , and until the late 1960s was thought to coexist with C. hollensis throughout its Gulf Coast range. Previous C. hollensis identifications for the Gulf Coast have been reclassified as C. mississippiensis (Blanton and Wirth 1979)..

PAGE 17

Extensive studies on the biology of C. mississippiensis have been carried out in the salt marshes of Yankeetown, Levy 4 County, Florida. Culicoides mississippiensis has two annual population peaks, one in the spring and the other in the fall (Kline 1986) . Emergent males and females feed on flower nectar. After their first oviposition event females convert to a haematophagous feeding behavior, showing preference for mammals as blood sources (Blanton and Wirth 1979) . Host searching is highest at dawn and dusk but may occur at reduced levels throughout the daytime (Lillie et al . 1988). Oviposition occurs in salt marsh mud, mainly in stands of cordgrass, Spartina alternif lora Loisel (Kline and Wood 1988, Kline 1989), which is alternately covered by salt water or exposed to the air due to tidal patterns. Oviposition sites are in proximity to scattered islets of land called hammocks. Hammock vegetation is characterized by white cedar, southern pine, live oak, cabbage palm, saw palmetto and holly. Yaupon holly. Ilex vomitoria Alton, is the most common holly species in Yankeetown. It blooms from mid-March to mid-April when mississippiensis populations are at their highest.

PAGE 18

Culicoides mississippiensis has been shown to be associated with flowering I. vomitoria (Lillie and Kline 1986) . Field tests of attraction to detached flowers of 1. vomitoria and seven other plants were carried out using a portable plexiglass olfactometer by the author in April 1993. In the absence of visual cues, significant attraction was demonstrated for both C. mississippiensis and Dasvhelea mutabilis Kieffer, a common gnat, only to flowers of I. vomitoria (unpublished personal data) . A follow-up study of C. mississippiensis visitation to flowering I. vomitoria and two other flowering plants, the saw palmetto Serenoa repens (Bartram) Small and the cabbage palm Sabal palmetto (Walter), was carried out in spring 1994. As flowering of I. vomitoria ended in mid-April that of S. repens began. Similarly, the flowering season of S. palmetto started in late May as that of S. repens was ending. Each of those plants is the dominant flowering plant in Yankeetown hammocks during its own flowering season. Visitations of C. mississippiensis to I. vomitoria were consistently high, whereas they were low to S. repens and non-existent to S. palmetto . Attraction to flowering I. vomitoria was shown to be a fact, but the nature of the attractant system remained

PAGE 19

unknown. An examination of the association between C. mississippiensis and 1. vomitoria is needed to identify possible candidate floral attractants . The main research objective of the present study was to investigate the association of C. mississippiensis with 1. vomitoria. Specific objectives of the study were 1) to characterize the floral phenology of an 1. vomitoria population in a salt marsh hammock of Yankeetown, Levy County, Florida; 2) identify visitation patterns of the arthropod community associated with flowering male I. vomitoria to aid assessment of possible control on non-target arthropods; 3) establish the basis of primary attractancy of C. mississippiensis to flowering male I. vomitoria ; and 4) identify the phenological patterns of chemicals emitted by flowers of male I. vomitoria during the entire flowering season and determine correlation between specific chemicals and high C. mississippiensis visitations.

PAGE 20

CHAPTER 2 ILEX VOMITORIA AITON Introduction Ilex vomitoria Aiton, classified in the angiosperro family Aquif oliaceae , is a dioecious shrub or small tree growing to eight meters, found throughout the southeastern region of the United States (Figure 2-la) (Hudson 1979) . In Florida its distribution is coastal north of Brevard and Sarasota counties, inland in Polk, Marion, Alachua and Baker counties, and in all Panhandle counties except Leon and Calhoun (Figure 2-lb) (Wunderlin and Poppleton 1977) . The Aquifoliaceae are distinguished from other angiosperms by being oligostemonous (i.e., having the same number or fewer stamens as perianth members) , syncarpous (i.e., having fused carpels) and by bearing fruit with few seeds and only one per carpel (Stebbins 1951) . Family genera are divided into two groups, those that are deciduous and those that are evergreen, including Ilex (Wunderlin and Poppleton 1977) . Ilex vomitoria differs from other species in

PAGE 21

8 Figure 2-1. Distribution of Ilex vomitoria . A. United States (reprinted with permission of C. M. Hudson) . B. Florida (reprinted with permission of R. P. Wunderlin and J. E. Poppleton.)

PAGE 22

the genus in having completely crenate leaves (Figure 2-2a) . Hume (1953) grouped American hollies according to persistence of leaves and fruit color. Species may be deciduous with black fruit, deciduous with red fruit, evergreen with black fruit, or evergreen with red fruit. Ilex vomitoria falls into the last category. It is also the only fasciculate holly native to the U. S. (Hume 1953), producing flowers on the previous season's stem growth. In the Aquifoliaceae, nectar is produced at the base of petals or occasionally on the upper surface near the middle of petals (Loesener 1942) . The flowers of I. vomitoria are tetramerous, with male flowers having four stamens and a rudimentary ovary while female flowers have four nonfunctional stamens and a pistil ending in a four-lobed stigma (Figure 2-2b,c). Bawa and Opler (1975) found that flowers of most dioecious plants in a lowland tropical semideciduous forest of Costa Rica are small and white to yellow or green. Male I. vomitoria flowers are small, cream-colored and more numerous than female flowers (Hu 1979). Flowers are borne in fascicles. Fascicles of male plants have several secondary flower clusters, called cymes, with up to five flowers per

PAGE 23

r 1 10

PAGE 24

11 cyme while those of female plants have one reduced cyme with a maximum of seven flowers (Hu 1979) . Although 1. vomitoria is related in geological time to Asian holly species, more recent lineage can be traced to an original stock in the Ouachita Mountains of western Arkansas and eastern Oklahoma (Hu 1979) . It has thrived in the post oak savannah of east central Texas, an area intensively exploited for cattle grazing. Because its dense growth interferes with grazing and movement of equipment, Xvomitoria is controlled by herbicidal applications (Bovey et al. 1972, Duncan and Scifres 1983, Meyer and Bovey 1985). Adaptation to the drier conditions of seashore and hammocks has enabled I. vomitoria to establish itself in coastal areas (Hu 1979) . The salt marshes invaded by I. vomitoria are protected wetlands and it is unlikely that the plants' growth will be controlled there as has been the case in Texas. Historically, 1. vomitoria has been important in both Native American and colonial cultures. The leaves when roasted and boiled twice in water yield a black tea (Hume 1953). For centuries before colonial invasion. Native Americans used what they called yaupon tea as a social beverage, medicine, and emetic (Hudson 1979) . It was the

PAGE 25

12 third use, to induce vomiting in order to rid the body of contamination before council meetings, that inspired the species name of the plant. For much of the 17th century, prior to the advent of tea and coffee as important beverages in Europe, yaupon tea was an export to Spain, France and England (Hudson 1979) . Its use was again popularized in the South during the Civil War when coffee and tea supplies were cut off. Florida is in a subtropical zone of vegetation. The relative abundance of dioecious plant species compared to monoecious species in South Florida (27% of all plant species) approaches that of tropical forests of Costa Rica (21-29%) but is twice the number of temperate forests in the United States (6-17%) (Bawa and Opler 1975) . Ilex vomitoria and possibly other hollies share two attributes of tropical dioecious plants relative to pollination. While temperate dioecious plants are primarily wind pollinated (Stebbins 1951), tropical dioecious plants have been shown to be insect pollinated (Bawa and Opler 1975) . The Aquif oliaceae are in the order Celestrales, which includes insect and wind pollinated families (Brizicky 1964) . Hollies are found both in temperate and sub-tropical zones of the

PAGE 26

13 United States and are probably insect pollinated throughout the family range. Cavigelli et al . (1986) argue for insect pollination of I. montana , a holly found from New York to Louisiana. Ilex cassine L. and I. krugiana Loess., both found in Florida, are insect pollinated (Tomlinson 1974) . High numbers of insect visitations to I., vomitoria in Florida, especially Nematocera and Hymenoptera (personal observation) , also implicate insect pollination for this holly species. The small size and abundance of I.vomitoria pollen, however, suggest the possibility of wind dispersal . Such dispersal may have no value for pollination, but may result in a pollen plume whose volatiles serve as attractants for insect visitors. ' .• , 'i'' Temperate dioecious plants produce larger female flowers than male flowers (Baker 1948) . The flowers of male and female I. vomitoria plants, however, are similar in size, a characteristic more typical of tropical dioecious plants (Bawa and Opler 1975) . Bawa and Opler (1975) observed a strong correlation between the timing of flowering in dioecious plants in Costa Rica forests and seasonal emergences of insect pollinators. In that study most insect visitors were small, with Hymen-

PAGE 27

optera being the most likely pollinators. The flowering of 1. vomitoria coincides with the spring population peak of C. mississippiensis . Few other flowering plants are as ubiquitous as I. vomitoria in March and April. Poison ivy, Toxicodendron radicans L., is the only other plant in flower at that time from which mississippiensis adults have been collected in vacuum samples . Fine tuning of the relationship between pollinator and plant can be effected through the timing of release of nectar and pollen rewards (Janzen 1966) . Prior to flowering, plants may encourage insect visitations by secretions of extrafloral nectaries. The Aquif oliaceae are not listed by Elias (1983) among plant families with extrafloral nectaries and none were seen on 1. vomitoria by the investigator. Most flowers begin nectar production before pollinators arrive and stop once a maximum volume nectar pool is achieved (Cruden et al . 1983) . Stiles (1975) suggests early morning as a good time for diurnally pollinated plants to present large amounts of nectar because production could occur throughout the night. However, cool nighttime temperatures may inhibit or delay nectar secretion in some plants (Cruden et al . 1983).

PAGE 28

^ : 15 Red fruits develop on I. vomitoria and yield a maximum of four seeds. Diaz-Colon et al . (1970) determined that aqueous extracts of the fruit can inhibit root growth of other plant species' germinated seeds. Perhaps this capability gives 1. vomitoria a competitive advantage over other plant species in colonization of new habitats. The fruit may be eaten by birds and seeds dispersed in their droppings. Germination and growth are slow, but more rapid population expansion can be achieved asexually by root-sprouting (Hume 1953) . The salt marsh habitat of Florida's Big Bend Gulf Coast region, from St. Petersburg to Pensacola, is an important breeding ground for C. mississippiensis . Preliminary vacuum samples of flowering I. vomitoria made in spring 1994 (Chapter 4) contained both adult male and female C. mississippiensis . Nectar sugars are a significant reward to C. mississippiensis adults for I. vomitoria floral visitations (Chapter 4) . Attraction to flowers is mediated at least in part by floral volatile emissions (Metcalf and Metcalf 1992). The purpose of this study was to characterize the I. vomitoria population of a salt marsh hammock in Yankee town. Levy county, Florida to facilitate volatile sampling.

PAGE 29

Specific objectives of the study were to map distribution of plants, determine spatial and temporal patterns of flowering, and select a representative group of plants for volatile collections . Materials and Methods Distribution Mapping An aerial photograph taken by U. S. Air Force Tactical Reconnaissance Wing in 1981 (Figure 2-3) was used as a template for producing a scale map of the study site hammock, referred to as hammock A (Figure 2-4) . Four other neighboring hammocks were periodically visited for comparison to hammock A and were designated as hammocks B-E. Quadrats 10m on each side were assessed and plant positions plotted on the map of hammock A. Plants were considered as individuals if separated by at least one meter or if of different sex. Plants were tagged with pink surveyor ' s tape on which the hammock and code number were written. Each plant was sized as small (2m high) .

PAGE 30

17 >i 4J -H g rH •H 4-1 (C X -H o a to •H o o 4-1 0) a T3 (0 o ^; u (13 G (U to to jj M (0 U Q 0 W 1 ^ (0 -U K -H 0) 4-1 4J (0 fO -H U U H O H (0 < PQ CO to I Ti (U o u o CO to O CO Cn Hi a lO H (U M OJ (U 0) g c to -H X! 4-1 0 0 CO W 0) •H 4-1 CO Ui 0) Q) a -H U 0) u > H c T5 D rH Q IC (0 -H •H CO T! < CO C 0) •H to CO u CO 0 •H "to e 1 to CN o ^ 0) 0) CO 4J CJ •H H H 0 w a •H A* fa

PAGE 31

18

PAGE 32

Spatial and Temporal Patterns of Flowering Observations made in the spring of 1993 and 1994 indicated that the 1. vomitoria flowering season begins in March and ends in April. In spring 1995 weekly observation of bud development was made starting in February. As soon as the first plant began flowering, complete surveys of hammock A Xvomitoria plants were made on an average of 2-3 day intervals. Date of first flowering and sex of each plant were noted. The data obtained were evaluated for sex ratio as well as spatial and temporal variations in bloom by sex across hammock A. Daily temperature readings were made with a Taylor alcohol thermometer during each of the four diel sample periods, corresponding to hours at sunrise, late morning, early afternoon, and sunset. Sample Group Selection and Evaluation The cost of materials and time required for movement between plants resulted in focusing the study on male I., vomitoria. Large male plants, up to a maximum of three plants at any given time, were included in volatile sampling. Candidate sample plants were identified before the start of flowering. Plants were brought into the study at the first sign of flowering and replaced when fully senesced. In some

PAGE 33

20 cases it was necessary to replace a plant before flowering had completely stopped in order to accommodate inclusion of a plant just starting to flower. Prior to the start of the flowering season the mean number of buds per centimeter of male stem was determined. Ten 5-cm-stem samples were randomly selected from each of three male plants with well -developed buds. All buds were counted using a dissecting microscope. Mean bud counts per centimeter of stem were determined for each plant and plant means examined by ANOVA and the REG-WQ (Ryan-Einot-GabrielWelsch Multiple Range Test) (SAS Institute, 1989). Male bloom was quantified to make correlations between flowering phenology and volatile phenology. A male budded branch (sample segment) was enclosed in a tomato support cage. During each sample period, the number of open flowers of a sample segment was counted. Counting was facilitated by marking counted cymes with a small piece of masking tape. At the end of sampling, the total length of each plant's budded stems was measured. That length multiplied by the mean flowers per centimeter of stem gave the total flowers per sample segment. Percent bloom was calculated by dividing the number of open flowers in any given sample period by the total

PAGE 34

21 number of flowers on the sample segment. Mean values for three day peak bloom were compared for the five plants by ANOVA and REG-WQ . Data for all plants included in the study were combined in a group bloom profile to statistically compare peak insect activity and percent bloom (Chapter 3). Results . • Distribution Mapping Ilex vomitoria plant positions are shown on the map in Figure 2-5. A key to specific plant positions is given in Table 2-1. There are 145 I. vomitoria plants on hammock A, in an area of approximately 5600 m^ . Fifty-nine plants are female, 71 are male and 15 are of undetermined sex (did not flower) . The sex ratio for Hammock A is 1.20:1. The data are presented in size groups in Table 2-2. Spatial and Temporal Patterns of Flowering • The first plant to flower in Hammock A, starting on March 14, 1995, was a male coded A3 6. The last bloom start date, on April 15, was for a f emale plant coded A104 . Bloom start dates covered a span of 33 days. Dates for the start of flowering of individual plants are given in Appendix A.

PAGE 35

rH O 00 [> LD en ^1

PAGE 36

23 Table 2-1. Key to map of Figure 2-5, relating Ilex vomitoria plant designations to plotted positions on map. Plant Designation Map Location 1 C6/7 2,3 D6/7 4-11 E6 12 D6 13 E5 14 E6/7 15-21 F7 22,29-31 G6 23 F6 24-26 EF5 27,28 FG4/5 32,33 G6/7 34-36 07/ 8 37 FG8 38, 62-64 H7 39,40 H6/7 41,48, 142 H5 42-49, 135 15 50-52,139 13/4 53-55, 143 IJ3 56-58, 132 IJ4/5 59, 61, 145 J4 60, 140 15/6 65-69 H7 70-75 IJ8 76-78 144 T9 79, 80 18 81-83,85-87 J8/9 84 K9 88,89 K8 90 L8 91,92 M8/9 93-95 M7 96, 97, 141 N7 98-100 LM6 101-105 M5 106-110 L4

PAGE 37

24 Table 2-1. continued Plant Designation Map Location 111-117 L5 118,119,124-126 K4 120-123 .. K3/4 127-130 K5 131 JK5 133,134 IJ5 135-138 G8 Table 2-2. Number of Ilex vomitoria plants Yankeetown, Florida classed by size. on Hammock A in Sex Small Size Medium Large Total M 26 25 20 71 F 26 14 19 59 Unknown 14 0 1 15 Total 66 39 40 145

PAGE 38

25 Phenology of start of flowering is shown in Figures 2-6. A-G. The mean start date for flowering of male plants was day 11, while that of female plants was day 18. The difference between male and female start dates was significant (ANOVA, F=109.0, df=ll,118, p<0.0001). When size groups are considered individually, large plants are shown to flower earlier than medium plants, which flower earlier than small plants. A statistically significant difference for start of flowering exists between large and small male plants, and between all male plants, except small, and female plants (ANOVA, F=13.51, df=5,124, p<0.0001). Mean start dates for large, medium and small male plants were 9, 11, and 13 days, respectively. For large, medium and small female plants mean start dates were 16, 17 and 19 days, respectively. No statistical differences were observed between female size groups relative to the date for start of flowering. A significant interaction existed for start of flowering with the mean daily temperature calculated from daily high and low temperature readings. A graph for the start of flowering by sex is shown in Fig. 2-7.

PAGE 39

Figure 2-6A. Phenology of flower start for I. vomitoria on hammock A. Start of bloom indicated by white square. Top is for 3/14/95, bottom for 3/16/95.

PAGE 40

Figure 2-6B. Phenology of flower start for I. vomitoria on hammock A. Start of bloom indicated by white square. Top is for 3/18/95, bottom for 3/21/95.

PAGE 41

B C D E F G H I J K L M N O Figure 2-6C. Phenology of flower start for I. vomitoria on hammock A. Start of bloom indicated by white square. Top is for 3/24/95, bottom for 3/27/95.

PAGE 42

Figure 2-6D. Phenology of flower start for I. vomitoria on hammock A. Start of bloom indicated by white square. Top is for 3/29/95, bottom for 3/31/95.

PAGE 43

Figure 2-6E. Phenology of flower start for 1. vomitoria on hammock A. Start of bloom indicated by white square. Top is for 4/3/95, bottom for 4/7/95.

PAGE 44

Figure 2-6F. Phenology of flower start for I. vomitoria on hammock A. Start of bloom indicated by white square. Top is for 4/9/95, bottom for 4/12/95.

PAGE 45

Figure 2-6G. Phenology of flower start for I. vomitoria on hammock A. Start of bloom indicated by white square. Above is for 4/15/95.

PAGE 46

Figure 2-7. First day of flowering by sex. Ilex vomitoria on Hainmock A, Allen Park Road, Yankeetown, Levy Co., March-April 1996.

PAGE 47

Sample Group Selection and Evaluation Plants included in the study are listed with dates of flowering in Table 2-3. A total of five large male plants were sampled. Results of bud count data are given in Table 2-4. All means were evaluated by ANOVA and REG-WQ multiple comparison test. Male means are not significantly different from each other (ANOVA, F=0.17, df=2,27, p=0.85), and so the overall mean for males is valid. Female means do differ significantly (ANOVA, F=25.8, df=2,27, p<0.0001), with each mean shown by multiple comparison to be significantly different. Male mean bud count per centimeter stem is significantly greater than female mean bud count (ANOVA, F=100,000, df=6,53, p<0.0001). The mean value of 20.4 buds/cm was used to evaluate percent bloom of male plants included in the study. Data for calculation of percent bloom is found in Appendix B. Percent bloom for each of the five study plants is shown in Figure 2-8 A to E. Combined percent bloom data for the study group was stratified into 8 groups for later statistical examination of the relationship between insect visitations and flowering status (Chapter 3). The 8 groups were constructed to include more than 15 data values each

PAGE 48

35 Table 2-3. Vi 3 TTim r~* "L^ lictllUlUJOrL Ilex /~\ /—\ vomitoria Dlants Q.t=t> j.yriaLiori . included in the study by Plant Sex Flowering dates A3 6 M 3/16-3/31 A14 M 3/24-4/7 A62 M 3/25-4/7 A3 5 M 4/1-4/15 A129 M 4/7-4/15 Table 2-4. Mean bud counts cm"^ (SE) for I. vomitoria in March 1995. Ten 5 -cmsegments of budded stem from each of 3 male and 3 female plants were evaluated. Sex Plant Male Female 1 21.6 (2.61) 2.5 (0.40) 2 20.1 (3.56) 1.5 (0.28) 3 19.3 (1.85) 4.8 (0.31) Overall mean 20.4 (1.55) 2.9 (0.33)

PAGE 49

36 25 20 15 10 5 i o 4-> 0) o 04 16 ' 17 ' 18 ' 19 ' 20' 21 ' 24 ' 25 ' 26 ' 2?' 28 ' 29 ' 31 March 25 20 15 10 B 1 r I \ \ \ \ \ 1 \ \ r 24 25 26 27 28 29 31 1 2 3 4 7 March April Date Figure 2-8. Graph of percent bloom as a function of time, throughout the flowering season. A. Profile for plant A36. B. Profile for plant A14.

PAGE 50

25 § o 4J (D U C 20 15 10 5 0 25 20 15 10 5 0 ' 25 ' 26 ' 27 ' 28 ' 29 ' 3l' 1 ' 2 March April 3 I 4 I 7 I April 1 9' 10' 11' 12' 14' 15^ Date Figure 2-8. Graph of percent bloom as a function of time, throughout the flowering season. C. Profile for plant A62. D. Profile for plant A35.

PAGE 51

I I \ I I 1 1 [ 7 8 9 10 11 12 14 15 April Date Figure 2-8. Graph of percent bloom as a function of time, throughout the flowering season. E. Profile for plant A129.

PAGE 52

3t and to reflect the flowering cycle (increase to peak followed by decline). Those groups are shown in Table 2-5. Statistical comparison of the five plants of the study group was made relative to the three days of highest percent bloom for each (peak bloom) . Peak bloom declined significantly over the flowering season. Plant A3 6, the first plant to flower, had a significantly higher peak bloom than plant A14, and both A3 6 and A14 were significantly higher than the other three plants in the study (ANOVA, F=271.1, df=14,45, p<0.0001). Peak bloom means are summarized in Table 2-6. Discussion Distribution Mapping The positions of I. vomitoria as shown on the map in Figure 2-5 are evenly scattered across hammock A. Although male plants outnumber female plants, no irregular distribution patterns relative to sex were detected. It should be noted, however, that formal analysis of spatial association was not carried out. 'Sex ratio for hollies may be skewed in favor of males. Cavigelli et al . (1986) found a male: female ratio of 1.35:1

PAGE 53

40 Table 2-5. Percent bloom (B) groups. Eight groups were constructed to reflect natural flowering cycle dynamics. Group B(% bloom) n^ 1 B<1 30 2 1B^5 18 6 5>B>2 32 7 2>B>1 27 8 1>B 33 * Number of samples included in each bloom group. Table 2-6. Mean peak bloom. Means are calculated for highest percent bloom over three days for each plant included in the volatile study. Values of all four diel periods are used for each day. Plant N Mean SE A36 12 18. 5A* 0.38 A14 12 10. IB 0.94 A35 12 3.02C 0.18 A62 12 2.88C 0.18 A129 12 1.92D 0.13 *Means followed by different letters are significantly different at the 0.05 probability level by REG-WQ multiple comparison test.

PAGE 54

41 for I. montana Torr . & Gray. Richards' (1988) study of four natural or seminatural plots of I. agui folium L. in Britain showed male : female sex ratios ranging from 1.11:1 to 2.16:1. The male: female sex ratio of I. opaca Alton was determined to be 1.03:1, seeming to contradict the male-skewed trend of other hollies (Clark and Orton, 1967) . However, the study of Clark and Orton was done on seedlings which were followed over a seven-year period in controlled conditions and does not reflect differential survival that may exist in a natural stand of 1. opaca . Two characteristics of hollies tend to complicate sex ratio determination. Their ability to reproduce vegetatively from roots makes it difficult to distinguish close growing same sex plants as truly unique individuals. The sex ratio of 1. vomitoria is lower than that found for J. montana (Cavigelli et al . 1986), but is consistent with a seemingly male-skewed sex ratio for holly species. Several hypotheses to explain male-biased sex ratios in hollies are discussed by Cavigelli et al . (1986). The hypothesis favored is that female energetic costs are higher than male costs, resulting in reduced frequency of flowering. Therefore long-term studies should reveal a greater number of any year's non-

PAGE 55

42 flowering plants to be female and the overall sex ratio to be close to 1:1. Studies carried out over several years in Hammock A could determine if this is the case for I. vomitoria . STPatial and Temporal Patterns of Flowering In their study of tropical dioecious plants, Bawa and Opler (1975) determined that male flowers open earlier in the day than female flowers. The self -incompatibility of such plants would favor out crossing, resulting in a population with greater genetic variation. Ilex vomitoria also shows temporal segregation of male and female floral rewards, with male plants producing flowers earlier in the season than female plants. This arrangement results in pollinator visitation of male plants before female plants. The advantage for the plant of such a system is to increase the chance a pollinator will visit a female flower bearing pollen from an unrelated plant. Sample Group Selection and Evaluation The choice of conducting the study on male rather than female plants was based in part on the potential for greater levels of flowering. Plants with more flowers would provide greater quantities of floral volatiles as well as insects.

PAGE 56

43 Bud counts that showed significantly greater numbers in male than female plants supported the choice of male plants . Male plants were also shown to begin flowering significantly earlier than female plants . Choice of male plants therefore allowed earliest start of insect sampling, yielding a more complete profile of insect visitation throughout the flowering season . The level of percent bloom achieved by plants through the flowering season varied from high early in the season to low late in the season. It is difficult to judge how much the low bloom level achieved by A62, A35 and A129 was the product of thrips infestations and how much was a normal end of season phenomenon. Damage to buds caused by high thrips populations may be at least in part responsible for considerably lower percent blooms of plants. This explanation is supported by data presented in Chapter 3 showing that highest numbers of thrips in collections were significantly associated with lowest percent bloom.

PAGE 57

CHAPTER 3 ILEX VOMITORIA FEEDING GUILD Introduction The Diptera genus Culicoides Latreille, 1909 includes a large number of species of extremely small biting flies with worldwide distribution. They are classified in the suborder Nematocera, family Ceratopogonidae, subfamily Ceratopogoninae and tribe Culicoidini (Wirth et al . 1980). Culicoides spp. are important pests and disease vectors. Males feed on the nectar of flowers, while females feed on blood. The genus shows great diversity of blood source preference, but most feed on mammals and birds (Downes 197 0) . In many species autogeny, that is the production of the first batch of eggs following a sugar meal, provides an adaptive advantage in habitats where blood meals may not always be available (Downes 1970). Worldwide there are more than 1,000 species of Culicoides (Wirth et al . 1980) . In North America there are 144 described 44

PAGE 58

45 Culicoides species, of which 48 can be found in Florida (Blanton and Wirth 1979) . As many as 21 of those species have been found in Yankeetown, Florida (Kline 1986) . Adult Culicoides can be distinguished from other ceratopogonids by having a 3 -segmented antenna, including the flattened scapes (segment 1), enlarged pedicel (segment 2) and flagellum (segment 3), which is subdivided into 13 flagellomeres. Wings are characteristically patterned with light spots on a dark background. These patterns, as well as antennal ratios, the number and type of antennal sens ilia, and male terminalia are important taxonomic characters for species identification (Wirth, personal communication) . Species identifying criteria for Florida species are outlined by Blanton and Wirth (1979) . The four most common anthropophilic species in Yankeetown are C. mississippiensis Hoffman, C. fur ens Poey, C. barbosai Wirth & Blanton, and C. f loridensis Beck (Lillie et al . 1988) . While C. mississippiensis is bivoltine, with spring and fall population peaks, the other three species are most often found from April to October when C. mississippiensis begin to decline and reach their lowest levels. Host seeking is characterized as crepuscular in C. mississippiensis , C.

PAGE 59

46 barbosai and C. f loridensis but nocturnal in C. furens (Lillie et al. 1988) . Culicoides mississippiensis has been shown to be autogenous (Davis 1981) . No sugar sources have been identified prior to the present study although midge vacuum samples by Lillie and Kline (1986) suggested flowering Ilex vomitoria Alton, the yaupon holly, as a possible sugar source in spring. An important first step in studying the association between C. mississippiensis and I. vomitoria is to establish the nature of the reward of attraction to the flowering plants. It may seem reasonable to assume that the insect visits are linked to sugar feeding, but determination of such a link should be made by conducting anthrone tests. The success of many nematoceran pests and disease vectors is linked to autogeny and sugar feeding. Sugar feeding not only fuels initial egg production but also provides energy for flight (Van Handel 1984) , enabling search for blood meal hosts and oviposition sites, and increases longevity (Hunter 1977, Jamnback 1961) . Lack of contact with humans and alternate hosts may favor higher sugar feeding rates (Edman et al . 1992, Van Handel et al . 1994). Implications for control of protozoan vectors in sylvan cycles are significant given

PAGE 60

47 that sugar meals increase transmission rates, possibly bynourishing developing parasites (Young et al . 1980). The cold anthrone test (Van Handel 1972) is the primary tool used in studying sugar feeding behavior of insects. Crushing a sugar fed insect in yellow anthrone solution will cause the reagent to change to color shades ranging from light green to dark blue. Mosquitoes (Bidlingmayer and Hem 1973, Edman et al . 1992, Reisen et al . 1986, Smith and Kurtz 1994, Van Handel et al . 1994), black flies (McCreadie et al . 1994, Walsh and Garms 1980, Young et al . 1980), and biting midges (Magnarelli 1981, Mullens 1985) are the principal nematocerans on which anthrone testing has been used. Nectar is considered a primary attractant of insects to flowers (Faegri and Pijl 1979) . It contains sugars, amino acids, lipids, anti-oxidants , and vitamins (Inouye 1980) . In addition to being important components of the nectar reward, sugars and amino acids may also act as attractants themselves (Baker and Baker, 1986) . Nectar glycerides that fluoresce may also attract insects (Kevan 1976) . Flowers have been shown to emit a large number of volatile compounds which serve to attract insects. In some cases the volatiles act as mimics of insect glandular

PAGE 61

48 secretions (Borg-Karlson and Groth 1986) . Some pollens produce their own set of volatiles that are important insect attractants (Pijl 1960, Dobson et al . 1987). Perfumes, UVreflecting nectar guides, and visible color guides act as close range attractants to insects (Snyder and Miller 1972). Establishment of the nectar of 1. vomitoria as a sugar source for male and female C. mississippiensis would provide a rationale for examining floral volatiles as possible attractants. Anthrone tests made on other Culicoides species (Magnarelli 1981, Mullens 1985) confirmed prior sugar feeding by host-seeking females, but did not identify sugar sources. Such data may be of limited value for determining circadian feeding rhythms and population feeding rates because anthrone negatives and weak positives could include insects that fed some time before sampling (Walsh and Garms 1980, Reisen et al . 1986, Smith and Kurtz 1994). Insect samples taken directly from flowering plants would provide more reliable information since the time of collection is more directly related to the opportunity for sugar acquisition. An ecologically important concern that must be addressed before control of a pest is attempted is the impact such control may have on non-target arthropods. If control of C.

PAGE 62

49 mississippiensis is to be based on its association with flowering male I., vomitoria , then other insects that visit the flowering plants must be identified and temporal patterns of floral visitations evaluated. Identification of floral attractants that correlate to peak foraging by target insects, but not to nontarget insects, may facilitate selection of chemicals to be tested for response by insects. The objectives of the following study were to identify the principal insects present on flowering male I. vomitoria, both before and during the flowering season, characterize the periodicity of their activity relative to the time of day and time within the flowering season and establish the nectar feeding behavior of C. mississippiensis . A comparison of the nectar and blood feeding of C. mississippiensis was also made. Materials and Methods Nonflowering Samples Samples of insects were collected from three male I. vomitoria prior to the start of the flowering season at the Yankeetown study site using an AFS sweeper (Meyer et al . 1983). The sweeper is a backpack vacuum device with a

PAGE 63

50 flexible extension arm that holds a collection container. Insects were collected in cardboard pint containers fitted with 100 mesh screen hot-glued into top and bottom covers. Plants sampled were A14, A36 and A39. Samples were made by vacuuming vegetation for one minute. Vacuum sweeps were made away from the collector to avoid influencing sample contents. Samples were collected on 5 March, 11 days before the start of the flowering season, at noon and sunset, and on 14 March, 2 days before flowering, at sunset only. At completion of sampling, containers were closed and placed inside zipper-top plastic bags for insect anaesthetization by CO2 . Insects were then transferred to vials and held on wet ice until return to Gainesville. Cooling of samples before processing was needed to prevent spiders from feeding on insects. All collected arthropods were freezekilled and preserved in 7 0% ethanol for later processing. Totals were recorded for major taxa. Mean number per minute sample and standard error were calculated. Flowering Season Samples Five flowering male I. vomitoria were vacuumsampled with an AFS sweeper four times a day throughout the flowering season of the holly. Plants sampled (dates of sampling)

PAGE 64

51 included A36 (16-31 March), A14 (24 March-7 April), A62 (25 March-9 April), A35 (1-15 April) and A129 (7-15 April). Oneminute AS sweeper samples were collected and handled according to the protocol for non-flowering season samples. The diurnal portion of the diel cycle was divided into ten periods and samples were made during the first, fourth, seventh and tenth periods, corresponding to sunrise, late morning, early afternoon and sunset. Sampling was limited to the diurnal portion of the diel cycle because C. mississippiensis does not feed at night (Lillie et al . 1988) . Nighttime sampling may have yielded a more complete picture of the I. vomitoria feeding guild as well as the phenology of essential oil emission, but was not carried out. Most male and female C. mississippiensis were removed from samples and stored separately at -60°C until processed by the cold anthrone test. All other sample contents, including some C. mississippiensis , were preserved in 70% ethanol and processed at a later date. Preserved sample contents were sorted according to general taxa and subsequently as many identified to species as possible. The total number of all invertebrates collected was recorded for each plant by diel period for each sample day of

PAGE 65

' 52 the flowering season. Mean number per minute sample and standard error were calculated and general evaluations made for all taxa except C. mississippiensis and the two dominant insects, Dasyhelea mutabilis Kieffer and thrips . Those three taxa were submitted to more detailed analysis of temporal visitation rhythms. . Visitation Rhythms Sample data were analyzed to identify any daily or seasonal rhythmic patterns of visitation by C. mississippiensis to flowering yaupon. Detailed analysis of visitations by D. mutabilis and thrips was also made because they would probably be the most abundant nontarget victims. Graphs were made for each of those insects of the numbers collected by plant and for the entire flowering season. Mean (standard error) one-minute samples by diel period were calculated to identify diurnal visitation patterns. Those means were examined by ANOVA and significant differences identified by REG-WQ. Visitation relative to flowering season was also analyzed. Insect means for each plant were compared by ANOVA and REG-WQ to identify visitation patterns relating to the entire flowering season. Plants were sampled sequentially in

PAGE 66

53 an order that can be considered to represent early (A3 6) , mid (A14, A62), and late (A35, A129) flowering season. Bloom groups were described for I., vomitoria male plants in Chapter 2 (Table 2-5) . These bloom groups (BG) correspond to stages in flowering that may be referred to as initiation (BG1,2), expansion (BG3,4), decline (BG5,6) and senescence (BG7,8) . An analysis of visitations by bloom group was made comparing the mean number of insects captured for each of the eight bloom periods described. Means were examined by ANOVA and significant differences identified by REG-WQ multiple comparison . Anthrone tests Anthrone tests were run on a stratified random sub-sample of males and gravid females. Gravid females could be easily distinguished from the few non-gravid females in samples by their swollen, lighter pigmented abdomens. Midges were assigned chronological numbers according to the date and time of sampling. Random number tables were used to select midges for anthrone testing, with numbers proportionately allotted to the four sample time periods for each sex. The estimate of nectar feeding rate was made using a 90% confidence interval and sample size calculated by the confidence interval formula.

PAGE 67

54 Because ^ was unknown, a value of 0.5 was used to calculate the sample size. By this method, 384 males and 384 females were tested to obtain a 90% confidence interval. Parous females taken in samples were also tested. Fresh anthrone reagent was made following the method of Van Handel (1972) every two to three days . Selected insects from each cohort were rinsed in distilled water to remove any nectar, which may have adhered to them during vacuuming, and placed individually in 10 x 75 mm test tubes. Midges were then crushed with a glass stirring rod in 0.2 ml anthrone reagent and evaluated one hour later for color change . Results were reported as positive or negative. The stirring rod was rinsed in two beakers of water and then dried with a paper towel. Controls were made by mixing the stirring rod in anthrone reagent alone after every six insect tests. Lack of available known negative insects, due to low field emergence levels at the time of testing, precluded better experimental control. Data was analyzed by chi-squared tests. Biting rates of C. mississippiensis Midge biting rates were determined for the same four sample periods from 16 March to 15 April. Midges were aspirated from the left forearm and numbers recorded every 60

PAGE 68

55 seconds for 20 consecutive 60 second periods. When densities exceeded 20 bites/60sec, a 5 min aspiration sample was made. During the next fifteen minutes midges were killed, but not collected, as they were counted. Samples were later identified to species. Results Nonflowering Samples A total of nine AS sweeper samples were collected from non-flowering male I. vomitoria . The composition of the samples is given by summary taxa in Table 3-1. Very low numbers of insects for major taxa were present on plants that had not yet flowered. In particular, the mean numbers of thrips, D. mutabilis , and C. mississippiensis were less than 2 each per sample. No taxon dominated samples and most taxa had less than 10% of the total sample contents. Flowering Season Samples A total of 218 AS sweeper samples were collected from flowering I. vomitoria between 16 March and 15 April 1995. Sample contents are given by summary taxa in Table 3-2.

PAGE 69

56 Table 3-1. Invertebrates collected from Ilex vomitoria prior to the start of flowering. Samples (n= 9) from plants A3 6, A14 and A39 collected by AFS sweeper on 5 March and 14 March 1996. Taxa Nonflowering Nontotals (%) flowering means (SE*) Thysanoptera 1 R ( y . / ) 1 / ( ± 1 \ Dasyhelea mutabilis (male) Id (9.7) 1 1 ( 1 • 2 ) D. mutabilis (female) 1 A ( 9 . U ) 1 c D ( 1 . 1) Culicoides mississippiensis (male) U ( (J . U ) U u ( u 0 ) C. mississippiensis (female) 12 (7.7) 1 3 (0 .5) Other Nematocera 19 (12 .3) 2 1 (0 .7) Hymenoptera 18 (11.6) 2 0 (0 .7) Collembola 15 (9.7) 1 7 (0 .5) Hemiptera 3 (1.9) 0 3 (0 .3) Coleoptera 2 (1.3) 0 2 (0 .2) Other Diptera 9 (5.8) 1 0 (0 .7) Acari 3 (1.9) 0 3 (0 .2) Lepidoptera 6 (3.9) 0 7 (0 .2) Homoptera 10 (6.5) 1. 1 (0 .8) Araneida 9 (5.8) 1. 0 (0 .3) Psocoptera 2 (1.3) 0 . 2 (0 2) Mollusca 3 (1.9) 0 . 3 (0 2) * standard error

PAGE 70

57 Table 3-2. Invertebrates collected from Ilex vomitoria during the entire flowering season. Samples (n= 218) from plants A36, A14, A62, A35 and A129 collected by AFS sweeper from 16 March and 15 April 1996. Taxa From Flower totals (%) From means Flower ( SE* ) Thysanoptera U O u \J \Dy . D ) Z J7 _/ 1 . -L (30 .7) Dasvhelea mutabilis (male) _L D Z Z J 1 T A 1 \ 1 . X (8.0) D. mutabilis (female) _L / f± D ^ MR Q ^ p (7.8) Culicoides mississippiensis (male) 0 /l R Q Z 4i 3 y ( A . A I 1 1 . z (1.2) C. mississippiensis (female) 1332 (1.2) 6 . 1 (0.5) Other Nematocera 2394 (2.2) 10 .9 (0.9) Hymenoptera 1902 (1.7) 8 . 7 (0.6) Collembola 796 (0.7) 3 . 6 (0.4) Hemiptera 697 (0.6) 3 .2 (0.2) Coleoptera 323 (0.3) 1 .5 (0.1) Other Diptera 287 (0.3) 1 .3 (0.1) Acari 222 (0.2) 1 . 0 (0.1) Lepidoptera 143 (0.1) 0 .7 (0 . 07) Homoptera 137 (0.1) 0 . 6 (0.07) . Araneida 122 (0.1) 0 . 6 (0 . 07) Psocoptera 23 (<0.1) 0 1 (0.02) Mollusca 22 (<0 . 1) 0 1 (0.02) * standard error

PAGE 71

58 Specific identifications and relative occurrence are shown in Table 3-3. More than 90% of all arthropods collected were either thrips or D. mutabilis . Culicoides mississippiensis comprised 3.4% of the sample total and was the third most abundant arthropod on flowering yaupon. Non-arthropods were veryuncommon and included anoles, observed feeding on insects, and small birds. Arthropod visitors may be generally grouped into three categories: uncommon (less than 1% of the sample total), common (between 1 and 10% of the sample total) and very common (more than 10% of the sample total) . General analysis for uncommon and common visitors, except C. mississippiensis . follow. More detailed analysis of data for C. mississippiensis and the two very common insect taxa is given separately in the section on temporal visitation patterns. Uncommon visitors Psocoptera {<0.1%) . No sample had more than two psocids . Samples containing psocids were few and scattered throughout the flowering season. Araneida (0.1%). Spiders could be found at low levels throughout the flowering season, with an unusually high number (12) taken from A35 during diel period 1 on April 2.

PAGE 72

Table 3-3. Arthropods identified from AS sweep samples of Ilex vomitoria , spring 1995. Occurrence is indicated by +++ (very common) , ++ (common) , and + (uncommon) . Taxa Occurrence Arachziida Acaridae Trombiculidae + Araneae Anyphaenidae Habana sp. + Araneidae Argiope sp . + Metazyqia sp. + Neoscona sp . + Salticidae Hentzia sp. + Theridiidae Anelosinus sp . + Coleosoma sp. + Insecta Collembola Entomobriidae + Coleoptera Curculionidae + Oedemeridae + Diptera Ceratopogonidae Culicoides barbosai Wirth & + Blanton C. furens (Poey) + C. mississippiensis Hoffman ++ Dasyhelea mutabilis Kieffer +++ Dasvhelea sp . + Forcipomvia sp . + Cecidomyidae + Chironomidae + Mycetophilidae + Sciaridae +

PAGE 73

Table 3-3 (continued) Taxa Occurrence Hemiptera Lygaeidae + Homoptera Aphididae Toxoptera aurantii (Boyer de + Fonscolombe) Cixiidae Pintalia vibex Kramer + Myndus sp. + Psyllidae Gvropsvlla ilicis (Ashmead) + Aleyrodidae + Hymenoptera Eulophidae (4 spp . ) ++ Mymaridae (1 sp.) ++ Torymidae (2 spp.) ++ Lepidoptera Geometridae + Lycaenidae Callophrvs grvneus (Hiibner) + Nymphalidae Agraulis vanillae (L.) + Thysanoptera Thripidae C . Frankliniella bispinosa +++ (Morgan) Leptothrips sp . + Heterothrips sp . +

PAGE 74

61 Homoptera (0.1%) . These insects were few and scattered throughout the flowering season. The most often encountered Homoptera included Toxoptera aurantii (Boyer de Fonscolombe) and Gvropsvlla ilicis (Ashmead) , the yaupon psyllid. All whiteflies in samples were adults. Lepidoptera (0.1%). All lepidoptera collected from flowering yaupon were Geometrid larvae. Damage to leaves caused by grazing larvae, though not extensive, could be seen on many of the yaupon plants in the study site. Cumulative totals were greater than 5 0% at the end of two weeks and greater than 90% at the end of three weeks. More than 85% of all larvae were collected between 24 March and 8 April. The only adult Lepidoptera observed infrequently feeding on male yaupon flowers were the cedar hairstreak, Callophrvs qryneus (Hiibner) , and the Gulf fritillary, Aqraulis vanillae (L.). Acari (0.2%). Mites found on flowering yaupon were of two major groups, detrivores and parasites. Several trombiculiids were found clinging to the abdomens of D. mutabilis collected on flowering yaupon. Diptera (Brachycera and Cvclorrhapha) (0.3%). Larger Diptera were rarely trapped in AS sweeps but were occasionally observed at low numbers to be feeding from flowers. Most

PAGE 75

62 frequently observed were syrphids (Syrphidae) and deer flies (Tabanidae) . Most non-ceratopogonid Diptera visiting flowers were small cyclorrhaphan flies. Highest numbers occurred in two distinct phases, two days in late March (2 6-27 March) and nine days in April (2-12 April) . Specimens collected on those eleven days accounted for more than 80% of all nonceratopogonid Diptera. Coleoptera (0.3%). Coleoptera were collected in relatively low numbers throughout the flowering season. Some, especially the oedemerids, were observed feeding on pollen. Hemiptera (0.6%). The majority of bugs collected from flowering yaupon were lygaeids (Lygaeidae) . Lygaeid numbers increased steadily from 25 March to 7 April. Both adults and immatures were found at all times of the day on plants . Collembola (0.7%) . Nearly all Collembola collected were of the family Entomobryidae . Their numbers increased to a peak of 103 on 27 March and then remained at stable but lower levels to the end of the flowering season. Higher numbers were found in diel period 1 samples than all other samples. Common visitors Hymenoptera (1.7%). Bees and wasps are important visitors of flowering yaupon holly. Although relatively few in

PAGE 76

63 number, megachilids and vespids could be seen carrying copious amounts of pollen as they visited both male and female plants. Vespids were also frequently seen searching in yaupon foliage. The majority of Hymenoptera in AS sweep samples were very small members of Chalcidoidea families. Species most commonly collected from flowering yaupon were of the Eulophidae, Torymidae, and Mymaridae. Over the entire flowering season significantly fewer Hymenoptera visits occurred during diel period 1 than the other three collection periods. Collections made during diel periods 7 and 10 had significantly higher numbers of Hymenoptera than earlier collections. Nearly two-thirds of all Hymenoptera collected were in samples of the last 11 days of the flowering season. Nematocera (excluding C. mississippiensis and D. mutabilis ) (2.2%) . Nematocera were the second most abundant group of insects on flowering yaupon. Separate consideration is given to C. mississippiensis because of their relatively high numbers in samples. Other ceratopogonids were also found in important numbers, including Dasvhelea species and Forcipomvia species. Only two other species of Culicoides were collected, those being C. barbosai Wirth & Blanton and C. furens (Poey) , both in very low nijmbers . Additional families

PAGE 77

64 significantly represented in samples included Cecidomyidae, Chironomidae, Mycetophilidae, and Sciaridae. Culicoides mississippiensis (3.4%) . These ceratopogonid insects were the third most abundant insects in AS sweep samples of flowering I. vomitoria in 1995. Culicoides mississippiensis made up 95.7% of all biting midges collected, while 4.1% were C. barbosai and 0.2% C. fur ens . A total of 3,791 C. mississippiensis adults were collected, of which 2,459 (2.2%) were male and 1,332 (1.2%) female. Higher numbers of males than females is in agreement with expected feeding differences. Detailed analysis of C. mississippiensis data follows under visitation rhythms. Very common visitors ^ s ' Dasvhelea mutabilis (30.6%). This common ceratopogonid gnat was the second most abundant insect in 1995 samples. A total of 33,692 D. mutabilis adults, of which 16,223 (14.7%) were male and 17,469 (15.9%) female, were collected. Detailed analysis of D. mutabilis data follows under visitation rhythms. , ' -v Thvsanoptera (59.5%). Thrips were the most abundant insects in samples of flowering 1. vomitoria . Holes, made by thrips, could be seen in numerous buds of plants A62, A35 and

PAGE 78

65 A129. Total bloom of those three plants was far less than of A3 6 and A14, which were not infested. The vast majority of thrips collected were Frankliniella bispinosa (Morgan) . Both adults and immatures were collected from flowering yaupon. Detailed analysis of thrips data follows under visitation rhythms . Visitation Rhythms Culicoides mississippiensis . The combined total of males and females collected for each plant by diel period are shown in graphs of Figure 3-1. Fluctuating numbers reflect both daily and seasonal rhythms. " Daily rhythms . Combined data from all plants were analyzed by GLM and REG-WQ multiple comparison to identify diel foraging preferences. Mean (standard error) values for each diel period are shown in Table 3-4. Diel 10 had a significantly greater number of biting midges in collections than either diel 7 or diel 4, but was not significantly greater than diel 1. Seasonal rhythms . Combined data was analyzed by GLM and REG-WQ multiple comparison to identify seasonal patterns of C. mississippiensis visitations to flowering I. vomitoria. Results of this analysis are given in Table 3-5. Mean

PAGE 79

66 200 Date Figure 3-1. Number of Culicoides mississippiensis collected from Ilex vomitoria by diel period (sunrise, late morning, early afternoon, sunset) A. Samples from plant A3 6. B. Samples from plant A14 .

PAGE 80

67 200 180 160 140 H 120 100 80 H 25 ' 26 ' 27 ' 28 ' 29 ' 31 March April D 10 ' 11 ' 12 ' 14 ' 15 April Date Figure 3-1. Number of Culicoides mississippiensis collected from Ilex vomitoria by diel period (sunrise, late morning, early afternoon, sunset) C. Samples from plant A62. D. Samples from plant A3 5.

PAGE 81

68 200 180 160 140 120 100 80 60 40 20 0 Figure 3-1. Number of Culicoides mississippiensis collected from Ilex vomitoria by diel period (sunrise, late morning, early afternoon, sunset) , E. Samples from plant A129.

PAGE 82

69 Table 3-4. Mean number of principal insects collected by diel period, spring 1995. Diel Period C. mississipp. D. mutabilis Thrips 1 19.2 AB* 29 .4D 321. OAB 4 7.4 B 103 .OB 201. 6B 7 11.6 B 189 . 8C 3 07 .4AB 10 28.0 A 279 .2A 394. 3A *Means within coliomns followed by different letters are significantly different at the 0.05 probability level by REG-WQ multiple comparison test. Table 3-5. Mean number of principal insects collected by plant, spring 1995. Plant C. mississipp D. mutabilis Thrips A3 6 29 . ,8 A* 204. SB 59. 4C A14 24. .0 A 287. 3A 83 .2C A62 16. . 0 AB 93 .9C 378. 9B A3 5 7 . .1 B 95 .2C 334. 8B A129 5 . .8 B 58. 2C 806 . OA *Means within columns followed by different letters are significantly different at the 0.05 probability level by REG-WQ multiple comparison test.

PAGE 83

numbers in descending order correspond to the order in which plants were sampled during the flowering season. Significantly higher numbers of C. mississippiensis were collected from plants A3 6 and A14 than from A3 5 or A12 9 (ANOVA, F=2.99, df=27,25, p<0.004). The mean for A62 was intermediary to others, though not significantly different from the means for A36 and A14 . More than 50 adult C. mississippiensis were collected on ten occasions from A3 6, on six from A14 and on five from A62, but only once from A3 5 and never from A14 . When data for all plants are combined, major collection peaks for C_^ mississippiensis can be identified every 3-4 days (Figure 3-2) . The last such peak occurs on 4 April with two lesser peaks occurring on 8 and 12 April. Examination of C. mississippiensis capture by bloom group, using the eight bloom groups defined in Chapter 2 (Table 2-5), showed that significant differences existed among bloom group means (ANOVA, F=6.43, df=7,204, p<0.0001). REG-WQ evaluation of bloom group means is given in Table 3-6. The mean for bloom group 5, corresponding to peak bloom and the start of bloom decline, was significantly greater than all other bloom group means except those of bloom groups 3 and 4.

PAGE 84

71 ON CO o U en 00 M3 in o cn l> u rH S CO •ri CO •rH u Q -H CO CO -H CO • CO in •H ti CO ni C)l U ri 1 • r1 U t , H CJ •H CO ( — 1 w r \ CD •H i, H t, M o o -H CO g cH o (0 > JJ 0 X 0) rH > H rH •H g n3 o Q u 4-1 CM 0) 1 JJ ro u 0) 0) rH ^1 rH ;3 0 Oi u rH sTsuaTddxssTSSxui

PAGE 85

72 Table 3-6. Mean number of principal insects collected per sample by bloom group, spring 1995. Bloom groups correspond to the indicated range of percent bloom. Bloom N C. mississip . D. mutabilis Thrips group 1 (0-1%) 30 5 . .9 D* 67 . .5 CD 205 .2 BC 2 (1-2%) 17 9. .1 CD 129. .9 BCD 415 .8 ABC 3 (2-5%) 27 26 . . 0 ABC 253 . ,1 AB 201 .6 BC 4 (5-22%) 28 28. .1 AB 306 . .2 A 57 .0 C 5 (22-5%) 18 37 . .3 A 203 . , 0 ABC 123 .3 C 6 (5-2%) 32 17 , , 1 BCD 151. .1 ABC 288 .9 ABC 7 (2-1%) 27 11. . 6 BCD 100 . . 1 BCD 590 .9 A 8 (1-0%) 33 9 . .3 CD 31. .5 D 521 .4 AB *Means within columns followed by different letters are significantly different at the 0.05 probability level by REG-WQ multiple comparison test.

PAGE 86

73 Dasyhelea mutabilis . These ceratopogonids were the second most abundant insects in AS sweeper samples of flowering I. vomitoria in 1995. The combined totals of males and females collected for each plant by diel period are shown in graphs of Figure 3-3 A-E. Daily rhythms . Combined data were analyzed by GLM and REG-WQ multiple comparison to identify diel foraging preferences. Mean (standard error) values for each diel period are shown in Table 3-4. Significant differences were found for all diel periods, and mean visitations increased from sunrise to sunset. Seasonal rhythms . Combined data relative to source plants were analyzed by ANOVA and REG-WQ to identify seasonal visitation patterns. Plant samples represent early (A3 6, A14), middle (A62, A35) and late (A129) flowering season. Results, given in Table 3-5, show a significantly higher number of visitations made to early flowering plants than to middle and late flowering plants . Peak numbers in samples are shown for 3-4 day intervals (Figure 3-4) . A sharp decline in the number of D. mutabilis visiting I., vomitoria can be seen in the last third of the flowering season.

PAGE 87

74 m •H H X! 4-) PI u 1500 1400 1300 1200 1100 1000 900 800 700 600 500 — 400 300 200 I 100 0 1500 1400 1300 1200 _ 1100 _ 1000 900 800 700 600 500 400 300 200 100 0 A l6 'Tz ' 18 ' 19 ' 20 '21 ' 24 '25 '26 ' 27 '28 ' 29 '31 April Date Figure 3-3. Number of D. mutabilis collected from 1. vomitoria by diel period, (sunrise, late morning, early afternoon, sunset) A. Samples from plant A36. B. Samples from plant A14.

PAGE 88

75 April Date Figure 3-3. Number of D. mutabilis collected from I., vomitoria by diel period, (sunrise, late morning, early afternoon, sunset) C. Samples from plant A62. D. Samples from plant A3 5.

PAGE 89

76 H (0 e Ql 4-1 o u 0) 1500 -r 1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200 100 0 E ' 7 April 9 ' 10 ' ll '"'12 ' 14 ' 15 Figure 3-3. Number of D. mutabilis collected from 1. vomitoria by diel period (sunrise, late morning, early afternoon, sunset) . E. Samples from plant A12 9.

PAGE 90

77 in (N CM in CM tH CM O CM CTi U ' ' 5-1 1X> (fl g O TS (U u 0) o u CO H PI o in CO CTi iH CTi (0 H JJ O tn u a -H rH a H CO (0 P (0 -H >-l • o -IJ I e O (D > d • Di H| -H

PAGE 91

78 Analysis of the mean number of D. mutabilis collected relative to the level of bloom of plants showed that D. mutabilis is attracted in higher numbers in bloom group 4, corresponding to late rise to peak bloom, followed by bloom groups 3, 5, and 6, in that order (ANOVA, F=6.46, df=7,2 04, p<0.0001) . Mean capture for bloom group 4 was significantly greater than mean captures of bloom groups 1, 2, 7, and 8, corresponding to beginning and end stages of flowering. Results of REG-WQ evaluation of bloom group means is given in Table 3-6. Thysanoptera . These were the most abundant insects in samples, though not evenly distributed throughout the flowering season. The majority of adult thrips were F. bispinosa . Many immatures were collected throughout the flowering season. More than 65,000 thrips were collected from flowering I., vomitoria , but no effort was made to distinguish the numbers of immatures from adults. The totals of thrips for each plant are represented in graphs of Figure 3-5 A-E. Daily rhythms . Combined data was analyzed by GLM and REG-WQ multiple comparison to identify diel foraging preferences. Mean (standard error) values for each diel

PAGE 92

79 a -H 4-1 o M 0) March 500 B April Date Figure 3-5. Number of thrips collected from i. vomitoria by diel period. A. Samples from plant A3 6. B. Samples from plant A14.

PAGE 93

80 25 ' 26 ' 27 March 28 ' 29 2000 April April Date Figure 3-5. Number of thrips collected from Xvomitoria by diel period. C. Samples from plant A62. D. Samples from plant A3 5. „• -

PAGE 94

81 3400 3200 3000 2800 2600 2400 2200 2000 1800 1600 1400 1200 1000 800 600 400 200 0 -April Figure 3-5. Number of thrips collected from I., vomitoria by diel period. E. Samples from plant A129.

PAGE 95

82 period are shown in Table 3-4. Relatively high numbers are collected in all diel periods, but the mean for diel period 10 is significantly higher than that of diel period 4. Seasonal rhythms . Data analysis for thrips shows low population levels during the first two-thirds and a population explosion during the last third of the flowering season. Combined plant totals for the entire season are shown in Figure 3-6. Examination of bloom group data for thrips showed highest mean numbers of thrips are associated with early and late phases of flowering, when percent bloom is at its lowest (ANOVA, F=5.27, df=7,204, p<0.0001.) Results of REG-WQ evaluation of bloom group data are given in Table 3-6. Graphs of bloom group data reflect a very different pattern of visitation for thrips to flowering I. vomitoria from those of the two major ceratopogonids (Figure 3-7) . Anthrone Tests Positive anthrone tests were obtained for 427 of 768 (55.6%) midges tested. The number of male positives were 67 (53.2%), 23 (53.5%), 33 (55.0%), and 68 (43.9%) for diel periods 1,4, 7 and 10 respectively. Gravid female positives for the same periods were 57 (73.1%), 28 (65.1%), 51 (59.3%),

PAGE 96

83

PAGE 97

84 0-1 1-2 2-5 5-22 22-5 5-2 2-1 1-0 Percent bloom Figure 3-7. Mean 1 min vacuum samples for three insect species from flowering male 1. vomitoria relative to percent bloom. Bloom increases to a maximum of 22% and then decreases to <1%.

PAGE 98

85 and 100 (56.5%) (Fig. 3-8). An overall significant difference (p<0.10) between male and gravid female anthrone positivity was found, with 236 of 384 (61.5%) females having sugar fed compared to 191 of 384 (49.7%) males. Male feeding rates were significantly higher during diel periods 1 (X^=14.426, df=l) , 4 (X^=7.688, df=l), and 7 (J^ =12. 670, df = l) than during diel period 10. Gravid female feeding rates decreased steadily from dawn to dusk, and those of periods 7 (X^=10.560, df = l) and 10 (X^=19.735, df=l) were significantly less than the rate for period 1. Of 25 non-gravid, parous females collected and tested, 10 (40%) yielded positive anthrone results. The low number of parous females in samples indicates that the method of collection succeeded in excluding blood-meal seeking females. All controls were negative. Biting rates of C. mississippiensis Mean biting rates (range) were 140 (3-321), 79 (3-525), 117 (0-510), and 689 (51-2484) bites/h for diel periods 1, 4, 7 and 10, respectively. A comparison of mean biting rates and male/ female floral samples for the four collection periods is shown in Figure 3-9. The biting rate for diel period 10 was significantly greater than rates for all other sample periods

PAGE 99

100 90 80 70 60 50 40 30 20 10 4 7 Diel period ] males i females 10 Figure 3-8. Percent sugar feeding of C. mississippiensis by diel period determined by the cold anthrone test.

PAGE 100

87 1000 1000 Diel period Figure 3-9. Comparison of the number of male and female £. mississi ppiens-j s collected from flowering male 1. vomitoria with the mean number of midges biting per hour (BR) for corresponding diel periods. . ,>

PAGE 101

and the rate for diel period 1 was significantly greater than that of diel period 4 (ANOVA, F=23.11, df=3,92, p<0.001). The lowest temperature for which biting activity was recorded was 10.5°C (8:25 A.M., 3 April) and the highest was 31°C (midafternoon, 9-11 April). Discussion Non-flowering Samples Samples of nonflowering I. vomitoria taken in a study by Lillie and Kline (1986) contained very low numbers of C. mississippienisis compared to samples of flowering I. vomitoria. In the same study, higher numbers of C. mississippiensis were found in samples of non-flowering S. alternif lora , at emergence sites of C. mississippienisis , than non-flowering I. vomitoria , distant from emergence sites. The density of I. vomitoria is much lower in hammocks than is the density of S. alternif lora in the salt marsh. If C. mississippiensis had been attracted to non-flowering I. vomitoria then the numbers should have been at least as high as on S. alternif lora . Nonflowering I. vomitoria therefore seems to be an accidental resting site for C.

PAGE 102

89 mississippiensis . No other arthropod taxa were reported by that study. Low numbers of all taxa reported by the present study indicate that non-flowering I.vomitoria serves as a resting site randomly accessed by C. mississippiensis and other insects. Higher numbers on flowering plants would therefore indicate attraction to the plants related to the flowers . Flowering Season Samples Uncommon visitors Psocoptera These insects are incidental visitors, not intimately associated with flowering vomitoria . Araneida . Spiders are non-nectar feeders, present on plants to feed on other arthropods that were feeding on the plants . Homoptera . The yaupon psyllid, G. ilicis, is the only known gall-making psyllid frequenting yaupon holly (Mead 1983). Yaupon is the only plant on which this psyllid is known to breed. Numerous galls could be observed on male and female plants in the study site. Aphids were in such low numbers as to suggest these pests of citrus to be incidental visitors to yaupon holly. Immature whiteflies were neither present in samples nor on the undersurface of leaves.

PAGE 103

90 suggesting that the few adults were accidental visitors to yaupon . Lepidoptera . Since Geometrid adults are nocturnal feeders (Brantjes 1978) , it is possible that adults visited yaupon to ingest nectar and to oviposit but were not present during the four diel collection periods. • Acari . AS sweep samples of flowering false-willow, Baccharis angustif olia Mich., also contained D. mutabilis parasitized by trombiculiids (unpublished personal data, October 1995) . Trombiculiids were previously reported on D. mutabilis collected in California (Whitsel and Schoeppner 1967) . Diptera (Brachycera and Cvclorrhapha) . Most of those flies were probably feeding on nectar, though no anthrone tests were carried out for confirmation. Coleoptera . Beetles are commonly found on flowering plants feeding on pollen. It is such beetle transfer of pollen, incidental to feeding on the pollen, which Faegri and Pijl (1979) consider the most primitive form of pollination. Hemiptera . Lygaeids are often thought of as seed eaters. However, given the lack of seeds on the male yaupon it is more likely that observed lygaeids were feeding either

PAGE 104

91 on sap or other insects (Borror et al . 1992). The presence of iinmatures and adults suggests use of plants for breeding. Collembola . Collembola are found primarily in leaf litter, but the data of this study indicate migration from litter to flowering vegetation in the early morning. Common visitors Hymenoptera . Observation of pollen on the bodies of larger Hymenoptera suggests they may have a role in pollination. Vespids may have been searching through foliage for Geometrid larvae or spiders . Higher numbers of chalcids visiting flowering yaupon in the afternoon suggest a preference to forage during the warmth of the afternoon. Nematocera . It is likely that most nematocera visited yaupon flowers to obtain nectar. Culicoides mississippiensis . Males are sugar feeders for their entire life span, but females are thought to feed on sugar only prior to producing the first batch of eggs and then convert to blood feeding. Males would therefore be expected to return to plants for successive sugar meals while females would be expected to make only one visit. Greater numbers of males than females in samples is in agreement with the results of samples collected by Lillie and Kline (1986) .

PAGE 105

92 Very common visitors ' Dasvhelea mutabilis . These gnats are so small that they are able to obtain nectar from I. vomitoria flowers without making contact with anthers. What role they may play in pollination of other plants in the Gulf Coast region has not been evaluated. In Brazil Dasvhelea spp.are associated with flowering cocoa and may be cocoa pollinators (Winder 1977) . Although anthrone tests were not carried out on D. mutabilis specimens, it is very likely that they fed on nectar during visits to flowering 1. vomitoria. They were commonly observed at nectar pools at the same time C. mississippiensis adults were feeding. Both males and females are known to feed only on sugars. A similar number of males and females in samples therefore confirms the expectation that individuals of both sexes would make return visits to plants for sugar feeding. Male/ female sample sex ratio was nearly 1:1, in contrast to the male-biased ratio of C. mississippiensis . Dasvhelea spp. are the only ceratopogonids whose females do not take blood meals. Thvsanoptera . Although considered primarily pollen feeders, thrips have been shown to feed on nectar, petals.

PAGE 106

93 other flower parts and foliage of plants (Milne et al . 1996). In a study of thrips distribution and abundance in Florida, Childers et al . (1990) found that F. bispinosa made up 92% of all thrips collected from citrus and was the dominant thrips species on all non-citrus flowering plants sampled. Ilex vomitoria was not included in that study, but results of the present study indicate that it too is a major plant resource for this thrips species. It is the dominant species on flowering yaupon. The presence of great numbers of immatures as well as adults indicates use of flowering yaupon as a breeding site by F. bispinosa . For that reason 1. vomitoria should be evaluated as a host plant that may be important to maintenance of F. bispinosa populations. Visitation Rhythms Culicoides mississippiensis Daily rhythms . Preference for sugar feeding in crepuscular diel phases corresponds to crepuscular blood feeding preferences observed for this species and is discussed in more detail in the section on anthrone tests. Seasonal rhythms . This species shows a bivoltine population pattern (Kline 1986), with peaks in spring and fall. Analysis of insect data for collections from flowering yaupon

PAGE 107

94 provides a means for distinguishing important emergence phases within the overall spring population peak. Control based on attraction to floral volatiles aimed at biting midges at the time of expected emergence would most likely have greater impact on population reduction and be more cost effective than control applied indiscriminately throughout the spring. Two possible explanations for observed collection peaks need to be considered. The first is that those peaks represent major emergence events of C. mississippiensis . This possibility could have been corroborated had emergence trap collections also been made during the study. The second is that peak attraction to flowers is occurring on those days. This possibility will be considered in Chapter 5 by statistical comparison of qualitative and quantitative floral volatile emissions on peak days with off-peak days. Analysis of the mean number of C. mississippiensis collected relative to level of bloom of plants indicates that at least quantitative effects of floral volatile emissions are involved in increased insect attraction. Dasyhelea mutabilis . Daily rhythms . Dasyhelea mutabilis shows a marked preference for visiting flowering I. vomitoria in the late

PAGE 108

95 afternoon and early evening. Comparatively fewer visitations were made in early morning . Seasonal rhythms . Little is known of the population dynamics of D. mutabilis . Analysis of data for collections from flowering yaupon provides a window on spring emergence patterns for this species. Given the high numbers of D. mutabilis found on flowering vomitoria , it is likely that this species would be impacted by any control aimed at C. mississippiensis that is based on 1. vomitoria floral attractants. Differences between target and non-target species relative to volatile attractants and foraging times may help in constructing a program that minimizes the impact on the nontarget. Seasonal population peaks show very close similarity in timing to those of C. mississippiensis . The decline in population during the last third of the flowering season was also similar to the decline observed for C. mississippiensis . Bloom group data for both D. mutabilis and C. mississippiensis clearly show attraction to I. vomitoria greatest when bloom is maximal and least when bloom is minimal. As was the case with C. mississippiensis , quantitative effects of floral volatile

PAGE 109

96 production seem to be related to greater numbers of D. mutabilis visiting I. vomitoria. Thvsanoptera . Daily rhythms . Constant populations of immatures are maintained on plants. Transitory peaks most likely represent recently arrived migrating swarms of adults. On one occasion the researcher, while standing over A129, observed such a swarm flying away from that plant. Seasonal rhythms . The flowers of A62, A35 and A129 had heavy damage due to thrips, such as numerous unopened buds with holes eaten through petals. None of those three plants achieved the magnitude of bloom seen in A36 and A14 . It is very likely that thrips infestations were responsible for those blooming deficiencies. Bloom group data suggest arrival of swarms on plants early in flowering, causing immediate reduction of bloom potential through damage to buds, and development of new migratory swarms during the plants ' senescence. Low numbers of arthropods other than thrips in late season samples may reflect the impact of thrips injury to a plant's attraction for those arthropods. Attention should be given to the role played by I. vomitoria in maintenance and growth of F. bispinosa spring populations.

PAGE 110

97 Anthrone Tests : .j Positive anthrone test results, in a range similar to those found for other autogenous nematocera (Bidlingmayer and Hem 1973, Magnarelli 1981, Reisen et al . 1986), are strong evidence of I., vomitoria nectar sugar ingestion by C. mississippiensis . During the study, no other plants bloomed as long and prolifically or were as widespread in hammocks as I. vomitoria . There were 145 I. vomitoria plants in the study site. Of the 131 that bloomed in spring 1995, 71 were male. Other plants in bloom during the study period included southern wax myrtle, Myrica cerif era L., black needlerush, Juncus roemerianus Scheele, and poison ivy. Toxicodendron radicans L. Only vacuum samples of poison ivy contained adult C. mississippiensis . Male I. vomitoria plants are shown by this study to be an important, widely available nectar source for both male and female C. mississippiensis during spring emergence when midge populations reach their highest levels. A significant decrease in anthrone positives in the last sample period might be explained by reduced nectar availability. Increased competition with other insects, such as D. mutabilis and thrips species (90% of all insects in diel period 10) , might have caused such temporal responses in

PAGE 111

percent sugar feeding. However, diel period 10 samples with higher than mean anthrone positivity also had high numbers of D. mutabilis and thrips . Consequently, decreased feeding rates by midges cannot be explained by interspecific competition. • Decreased nectar production by flowers could also be reflected in lower number of anthrone positives during diel period 10 . Nectar availability has been shown to vary seasonally, daily and diurnally (Kato 1993, Wright 1988), and may also relate to water availability (Petanidou and Voukou 1993). No data are available on the phenology of nectar production by X. vomitoria, nor were data collected during this study. It is therefore impossible to evaluate flowerbased effects on nectar feeding. Significant changes in nectar sugar feeding might best be explained as a function of biting midge foraging behavior. Lillie et al . (1988) found that the host-seeking behavior of C. mississippiensis peaked during periods 1 and 10 of the diel cycle. Biting rates for C. mississippiensis parous females found in this study were in agreement with the results of Lillie et al . (1988) . Biting rate increases were consistently noted in diel period 10. Visitation of flowering I. vomitoria

PAGE 112

99 by C. mississippiensis reflects similar periodicity, with highest numbers of midges collected from flowers during diel period 10 followed by period 1 (Figure 4-9) . As increasing numbers of midges descend on I., vomitoria towards sunset, it would be expected that a greater proportion of samples would include recent arrivals that had not yet fed. This would be reflected by a lower frequency of anthrone positives for diel period 10. Higher feeding rates in earlier sample periods suggest that midges had been present and feeding for longer times than those collected in diel period 10. The results of this study bear resemblance to those found by Van Handel and Day (1990) in a study of nectar feeding by Aedes taeniorhynchus (Wiedemann) . In that study, low anthrone positive rates were found in Ae. taeniorhynchus collected before sunset and high rates for those collected at sunrise. It was determined that most sugar feeding took place soon after dark. If this were also the case with C. mississippiensis , high numbers of anthrone negatives would be expected before sunset. Biting midges collected in diel periods 4 and 7 would then be those that had fed during the night and rested on the plants through the following morning. However, this scenario seems unlikely due to the nature of

PAGE 113

sampling. Vacuuming a plant at sunrise should effectively remove any insects which had lingered through the night. Insects collected must have arrived between vacuum events. It can therefore be concluded that C. mississippiensis feeding occurs throughout the day and possibly during the night. The presence of sugarfed parous females in collections indicates that females will seek nectar beyond the first oviposition event. Mullens (1985) suggested that higher sugar feeding rates in parous females than males and nulliparous females of C. variipennis reflects the need for energy to seek blood hosts. The same may be true for C. mississippiensis . Biting rates . The biting rates determined by this study confirmed the crepuscular host-seeking of C. mississippienisis reported by Lillie et al . (1988). This study further establishes crepuscular nectar-seeking by C. mississippienisis . The significance of sugar feeding rates of midges collected directly from flowering plants is different from that of those collected from a blood-meal source. In the latter case males are absent from the sample and only parous females seeking blood to mature eggs are included. Flowerbased collections can reveal behavior patterns for males,

PAGE 114

101 nulliparous and parous females. It is likely that most males and females collected from flowers were there to feed on nectar. Negatives are those midges that had not yet ingested nectar at the time of collection, but may have tested positive if collected at some later time. Males may also use flowering I. vomitoria as a resting site more frequently than females, which disperse in search of blood meals following the first oviposition event. Little is known about the mating behavior of C. mississippiensis . Male swarms, over objects that contrast with the background, have been observed in Yankeetown (personal communications, D. Kline, P. Choate) . Females may enter the swarms to mate. Although matings on flowering I. vomitoria were not observed during field collections for the present study, such behavior cannot be ruled out.

PAGE 115

CHAPTER 4 •FLORAL VOLATILE PHENOLOGY OF ILEX VOMITORIA Introduction The fragrance of flowers is a secondary attractant whose function is to draw potential pollinators to the plant to obtain a primary attractant, such as nectar (Faegri and Pijl 197 9) . In the process, many of those insects may coincidentally pick up pollen grains and transfer them to the stigma of flowers of that plant or its conspecif ics . Most previous studies of floral volatiles relative to insect visitation have focused on euglossine bees (Cane and Tengo 1981, Whitten et al . 1988), fruit flies (Lewis et al . 1988), lepidoptera (Cantelo and Jacobson 1979, Haynes et al . 1991, Heath et al . 1992, Zhu et al . 1993) and weevils (Nielsen et al . 1995). Little has been done to identify chemicals from floral fragrance for the purposes of monitoring or control of medically important insects. 102

PAGE 116

103 Fragrances that attract insects to plants, resulting in pollination, are synomones because the response produced upon perception of the fragrances is adaptively favorable to both insects and plants (Dicke and Sabelis 1988) . The relationship between I., vomitoria and C. mississippiensis may not be that of a synomone induced partnership. Direct observation of insects taking nectar from male flowers during the study revealed that it is possible for C. mississippiensis adults to approach the nectar pool without ever contacting the anthers, perched relatively high above the pool. Although accidental pollination by biting midges may occur, it is unlikely that yaupon pollination depends on these or other equally small ceratopogonids . Much of yaupon pollination probably results from visits by larger Diptera and Hymenoptera. Nevertheless, C. mississippiensis is an important visitor to yaupon, depending on the flowers as a major source of nectar. It is tempting to conclude that yaupon floral fragrance is kairomonal in the case of C. mississippiensis since it would appear that only the insects benefit from their attraction to the flowers (Dicke and Sabelis 1988) . Floral fragrance usually results from emission of many different chemicals, often referred to as essential oils.

PAGE 117

104 blended in different amounts. Most essential oils are produced in the cytoplasm of the epidermis and mesophyll cells of sepals or osmophore and diffuse to the air at favorable temperatures (Fahn 1979) . Those not emitted may enter lipid biosynthesis pathways in roots and rhizomes or be oxidized for energy (Croteau 1988). Essential oils include terpenes, phenylpropanoids , alcohols, aldehydes, esters, acids and sulfur compounds (Metcalf and Metcalf 1992). The main fragrance of flowers results from the presence of terpenes (Fahn 1979), having high volatility due to low molecular weights . Terpenes are synthesized via the mevalonic acid pathway. Isoprene units, with five carbon atoms, form the structural base of terpenes. The most important terpenes of floral fragrance are monoterpenes (two isoprene units) and sesquiterpenes (three isoprene units) . Oxidation of terpenes by plants yields alcohols, aldehydes, ketones and esters (Metcalf and Metcalf 1992). A tremendous variety of essential oils has been reported and is summarized by Knudsen et al . (1993). Fragrance specifically associated with pollen grains is also thought to be a potential secondary attractant. In a study of the chemicals found in pollenkitt, the oily coating

PAGE 118

105 of angiosperm pollen, Dobson (1988) established a wide range of diversity of essential oils found on the pollen of 69 different plants. Such diversity points to species specific pollen-related olfactory cues for arthropod visitors to flowering plants. Ilex vomitoria produces copious amounts of pollen, and so the essential oils associated with a pollen plume should also be evaluated as potential attractants. Headspace extraction has been used extensively in analysis of floral volatiles (Heath et al . 1992, Patt et al . 1988, Tatsuka et al . 1990) and for preparation of extracts for use in bioassays (Healy and Jepson 1988) . The method is described in detail by Heath et al . (1992). The process involves collection of volatiles from flowers onto an absorbent, usually a porous polymer, by creating a flow of air across the flowers and then through the absorbent. The method allows for extraction of very volatile compounds that are often missed by other methods (Budde and Eichelberger 1979) . Choice of absorbent can be critical. Tenax has been indicated as a poor absorbent for carbonyl compounds (Tatsuka et al. 1990) and pinene (Patt et al . 1988). Super Q is the absorbent of choice, accommodating those not collected by Tenax as well as most other floral volatiles. Because of

PAGE 119

106 artifact production associated with solvent extraction and steam distillation, Knudsen et al . (1993) endorse headspace extraction as the preferred method for work with floral volatiles. A further advantage of headspace extraction over other extraction techniques is the ability to extract volatiles as they are produced by living flowers. In the identification of floral essential oils, a major consideration is whether flowers to be sampled are attached to plants or detached. Flowering shrubs and trees are subjected to highly variable conditions of humidity, temperature and lighting. Working with detached flowers theoretically allows for the study of intrinsic volatile emissions free of the influence of climate. Studies have been done in which only detached flowers were sampled (Pellmyr et al . 1987, Tatsuka et al . 1990), but the actual value of such an approach can only be evaluated in comparative studies. Loughrin et al . (1993) found no difference in volatile production by attached and detached flowers of Nicotiana suaveolens Lehman sampled over 24 hours. When extended diurnal rhythmicity is of interest, detached flowers may not always be useful . In a study carried out over four days, Matile and Altenburger (1988) took volatile samples from both detached and attached flowers of

PAGE 120

107 several different plants at 3 hour intervals. Their results show that while some plants maintain normal rhythmic fragrance production, whether detached or not, others lose diurnal rhythmicity when detached. Heath et al . (1992) also showed that release rates of chemicals may be altered in detached flowers. In that study, benzyl acetate increased significantly in detached flowers of night-blooming jessamine while phenyl acetaldehyde decreased significantly. Mookherjee et al. (1986) also found important differences between the chemicals emitted by intact versus excised flowers of jasmine, lilac and Easter lily plants. Gas chromatography has provided a window on the chemical world in which we and the organisms we study live. In entomology its uses are varied, including studies on the mechanisms of intraspecif ic recognition (Bonavita-Cougourdan 1987), mate selection (Peschke 1987), strategies for host selection for parasitoids (Phelan et al . 1991) and taxonomic relationships (Carlson and Service 1980, Wirth and Morris 1985) . Gas chromatography enables the quantitative characterization of the components of a mixture extracted from some source. The process depends on the differential dissolving of those components by the stationary phase of the

PAGE 121

108 column (Perry 1991) . GC may also be combined with mass spectrometry to provide a means of qualitative identification of the components of a mixture. On passing from GC into the mass spectrometer, molecules are ionized and then separated according to their mass-to-charge ratios, producing a spectriom that can be matched by computer to a library of such spectra (Clement and Karasek 1985) . The objectives of the present study were to collect and identify volatiles of intact flowering male I. vomitoria four times a day throughout the flowering season. Data was statistically analyzed to identify phenological patterns of essential oil production by plants and correlation of those patterns with C. mississippiensis visitations. Collection of pollen was also made to determine how much pollen contributes to floral fragrance. Materials and Methods Volatile Collection The sampling apparatus consisted of a Reynolds® oven bag (turkey size, 62 cm x 48 cm, made of nylon resin, exposure temperatures not to exceed 2 04°C) pulled over a segment of

PAGE 122

109 flowering branch and supported by a two ring tomato cage (distal diameter 24 cm, proximal diameter 18 cm) . The cage support was held onto the branch by duct tape, and the oven bag tightly tied around the branch with string. Each oven bag had two pipette ports reinforced externally by duct tape and centrally positioned on opposite sides of the bag, one placed at the closed end and the other halfway towards the open end. A Y-shaped wooden support was used to prevent the sampling apparatus from bending to the ground. Volatiles were collected by removal of air from the oven bag, using an air vacuum pump with DC motor connected to a 12 volt deep cycle marine battery, across a Pasteur pipette containing 4 mg Super Q (Alltech, Deerfield, IL) . The absorbent was sandwiched by glass wool, that had been washed in pentane, inside the pipette. Ambient air drawn into the oven bag was filtered through activated charcoal (20-60 mesh) (Sigma) held by glass wool inside a second pipette. Pipettes were inserted into the two pipette ports and held in place by two overlapping strips of duct tape. An air flow of 250 ml/min was monitored with an in-line air flow meter. A blank sample and sample of non-flowering vegetation was made prior to the start of the flowering season. Once

PAGE 123

110 flowering had started, one-hour samples were collected, four times a day. Differences between internal oven bag temperatures and the ambient air temperature were monitored by an indoor-outdoor thermometer. Shading of the oven bag by a campers foil emergency blanket was provided when needed and effectively maintained internal temperatures at or below that of the ambient temperature. On completion of volatile sampling, pipettes were wrapped in aluminum foil, placed in a jar and held on ice until return to Gainesville, where they were transferred to an ultra low (-70°C) freezer. Each oven bag and air filter pipette was used only once a day and heated at 95 °C for one hour before reuse. Pollen was collected from intact flowers on Whatman No. 1 filter paper attached to the exhaust outlet of an AFS backpack sweeper. It was then wrapped in aluminum foil and handled in the same way as volatile samples. Samples were prepared by extraction with methylene chloride. Only the first ten drops were collected, yielding a volume of approximately 3 0 ij1 solution. Nearly all of the extract can be found in the first ten drops of solvent passed through the adsorbent (C. M. Whitten, personal communication) . To this was added 2 0 ]il internal standard in solution

PAGE 124

Ill containing 100 ng standard/ial . Each 1 ]il of the final mixture injected into the GC therefore contained 40 ng of the internal standard. The internal standard used was tridecane, a 13 carbon alkane. Selection of tridecane was made following injection of a 1 ul mixture of Cg-C^g alkanes (prepared by the laboratory of Mr. R. Heath, USDA/CMAVE) with an extracted sample of volatiles from I. vomitoria. Since the C^j alkane showed no overlap, yielding a distinctly separate peak from those of I., vomitoria volatiles, it was chosen for use as the internal standard. Methylene chloride was used for extraction of pollenkitt following the method of Dobson (1988) . Approximately 1 mg of pollen was washed in 0.5 ml methylene chloride for 20 seconds and filtered through Whatman No. 1 filter paper fitted to a syringe. The short solvent extraction period allows for removal of pollenkitt, located in the exine, without removing other chemicals found in the intine of the pollen (Dobson et al. 1988). To 30 yl pollenkitt extract was added 20 ]il tridecane as a standard for quantification of identified compounds. Two pollenkitt samples were processed by GC/MS as previously described.

PAGE 125

112 GC/MS Analysis Access to GC/MS was limited. However, GC without coupled mass spectroscopy was available for the duration of the project. It was therefore decided to identify I., vomitoria volatiles by running selected samples on GC/MS and then, using the same protocol, extrapolate those results to chromatograms produced by GC alone for the majority of samples. Volatile identification was made on pollen and selected whole flower samples by a HP 5995A gas chromatograph/mass spectrometer with data analysis by Teknivent Vector/Two data analysis system. All other samples were processed for phenological data on a HP 5890 Series II Gas Chromatograph following the same GC protocol which had been used on the HP 5995A GC/MS. Hydrogen was used as the carrier gas at a flow rate of 7.7 ml min'^ in a fused silica capillary column of intermediate polarity (SPB5, 30m X 0.25mm i.d., 0.25um film). Injection was splitless, on-column using a HP Automatic Sampler 7 673A and detection by Flame Ionization Detector (FID) . The GC was temperature programmed from 60°C to 290°C at 3 °C min"\ injector set at 63°C and detector at 340°C. Blank samples, containing solvent mixed with standard, were run at the start of each series of samples .

PAGE 126

113 Peaks were matched by their relative distance from the retention time of the internal standard as well as their relative distance from each other. Mean retention times were calculated from weighted retention times for peaks found in two or more plants . Peaks recorded in blanks were produced bysolvent or impurities in solvent and were deleted from each plant's GC profile. Data Analysis In order to identify phenological changes in volatile production all data obtained by GC was converted to mass of chemical per microliter extract for a one hour sample (ng/ulhr) . To do this, a correction factor was first applied to each peak area to compensate for variation caused by procedure. The correction factor, the ratio of the peak area to peak area of the internal standard, is multiplied by the mass of the standard to give the mass of the chemical under consideration (Dean 1995) . Because of the large number of samples included in this project, division of each internal standard area by the sample's mean internal standard area was carried out. This was then used as the correction factor in the above mass calculation. The mass thereby obtained, after multiplication by 40ng (mass of internal standard per yl) ,

PAGE 127

114 was the mass of that particular chemical produced by the entire cluster of flowers sampled on an individual plant for one hour. A sample calculation is given in Appendix C. Data sets for each identified chemical were examined by ANOVA and REG-WQ in order to determine diel and seasonal patterns of emission. Combined data sets for chemicals were matched to percent bloom values and C. mississippiensis capture data. Analysis was made by Canonical Correlation and Regression (SAS Institute, 1989) . The goal of Canonical Correlation Analysis was to identify the combination of chemicals (identified by GC/MS) most strongly correlated with high numbers of biting midges in AFS sweep samples. Capture values for C. mississippiensis were transformed with square roots and analysis performed on the transformed values. Regression in reference to Canonical Correlation Analysis was used to develop an ANOVA model based on chemical data most strongly correlated with high capture of C^ mississippiensis . This process involved regression tests that evaluated the impact of removal of selected sets of chemicals. The model was refitted to include only those chemicals which were significant (p<0.05) or marginally significant ( 0 . 05
PAGE 128

115 Results Sample Analysis Blank and non-flowering vegetation samples Blank and nonflowering vegetation volatile samples were both negative for identifiable peaks. Pollenkitt samples The majority of peaks produced by GC/MS analysis of pollenkitt samples were found in a range of very high retention times. Chemicals that eluted at retention times below 40 minutes, the range of I. vomitoria floral volatiles, are listed in Table 4-1. Most are either aldehydes or acids. Those chemicals comprised less than 4 % by peak area of all pollenkitt volatiles. A sample GC trace of pollenkitt volatiles is included in Appendix C. Floral samples A total of 194 volatile samples were analyzed by GC/MS. A summary of peak retention times for chemicals found in two or more plants is given in Table 4-2. A sample GC trace of floral volatiles is shown in Appendix C. A total of 51 chemicals eluted, of which 20 were identified by GC/MS. Identified chemicals include (in order of elution) dimethyl

PAGE 129

116 Table 4-1. Essential oils identified by GC/MS from pollen of Ilex vomitoria , spring 1995. Only chemicals that eluted in a range of retention times corresponding to that for floral volatiles are listed. Retention Chemical Time 5.33 unidentified 6.50 unidentified 7 .78t heptanal 13 .45 2-octenal 14.57 heptanoic acid 15.50 nonanal 20.20 decanal 22.70 2-decenal 23 .32 nonanoic acid 33.38 pentadecane

PAGE 130

117 Table 4-2. Essential oils of flowering male I. vomitoria identified by GC/MS. Mean retention times are based on the number of samples (N) in which a chemical was found. Mean RT N Chemical 3 . 021 97 3 . 076 105 3 .492 98 3 . 823* 53 dimethyl benzene 4 .467 15 4 . 672 14 4 .930 10 5 .393 3 5 .562 6 5 . 678 11 5 .758 3 6 .245 3 6 .631* 140 (X-pinene 7 .896* 104 P-pinene 8 . 967* 34 (2— methvl dtddvI ) bp^n^^^ne* 9 .617* 22 1 imonene 9 .905* 98 cis-P-ocimene 10 .346* 145 trans "P-ocimene 10 .412* 55 1,2 -diethyl benzene 10 . 626* 141 1,3 -diethyl benzene 10 .846* 144 1,4-diethyl benzene 11 .712 4 11 . 844* 21 linalool oxide 12 .317 35 linalool t= tentative identifications *= confirmed identifications

PAGE 131

ible 4-2 (continued) Mean RT N Chemical 12 .461* 35 13 . 041* 135 4 , 8-dimethyl-l , 3 , 7-nonatriene 13 648 2 13 . 844* 56 T-jVipnvl acp t" oni t" r i 1 e 14 .918 27 15 . 259t 147 C • — i AJf -L. Jk^d ± i--> O. _L i.±_y ^ 15 .590 16 15 . 901t 153 ethenyl benzaldehyde 16 . 451 12 16 . 876t 49 dodecenal 17 .666 9 17 . 998 9 18 .727 94 19 .451* 161 1 — (9 4 — r] "1 m(=*1" Viv1 nlnf=i'nvl \ *=it~Vi^'no'n(=i 20 .282* 157 1 — ^4 — *=it'}ivl 'n}if^Ti\/l ^ *^ 1" Vi "n on *a 20 . 817 18 . r± -7 ^ Q . O O f± 1 9 J. ^ . ^ ^ u Q . u J. 3 25 . 656 15 26 .890 71 27 . 527 5 28 .261 10 29 .116 35 29 .930 12 30 .257* 76 E, E-a-f arnesene

PAGE 132

119 benzene, a-pinene, P-pinene, (2 -methyl propyl) benzene, limonene, cis-P-ocimene , trans-P-ocimene, 1 , 2-diethyl benzene , 1,3 -diethyl benzene, 1,4 -diethyl benzene, linalool oxide, linalool, 4 , 8-dimethyl-l , 3 , 7-nonatriene , phenyl acetonitrile, ethyl benzaldehyde , ethenyl benzaldehyde, dodecenal, l-(2,4dimethyl phenyl) ethanone, 1(4-ethyl phenyl) ethanone, and E, E-a-f arnesene . The GC program was unable to consistently discriminate between trans-p-ocimene and 1, 2-diethyl benzene. The latter peak always appeared as a shoulder on the trans-Pocimene peak. In 23 samples from plants A14, A62 and A129 discrimination between the two was sufficient for independent integration of their peak areas . The mass totals of all identified chemicals by diel period are 157.4 ng/ial-hr, 234.8 ng/yl-hr, 241.4 ng/pl-hr, and 23 0.3 ng/ial-hr for diel periods 1, 4, 7, and 10, respectively. Since each hour's volatile sample was extracted to 3 0 yl and mixed with 20 ]il internal standard/ solvent mixture, each ]il contains 1/50 th of the total hourly sample. The mean hourly samples by diel period for the 20 identified chemicals are therefore 7.87 yg, 11.74 pg, 12.07 \ig and 11.52 yg for diel periods 1, 4, 7 and 10, respectively. Data for individual chemicals are listed in Table 4-3.

PAGE 133

120 Table 4-3 . Mean mass of each essential oil identified by GC/MS from flowering male Ilex vomitoria spring 1995. Mean mass (ng) /^l-hr for diel: 1 A *i 7 10 dimethyl benzene 0 .7 1 . 4 1 . 5 1 .7 a-pinene 4 .4A* 9 . IBC 11 .7C 7 . 6BC P-pinene 1 .4A 3 . 6B 5 .4C 3 . IB (2 -methyl propyl) benz . 0 .2A 0 . 3AB 0 .3A 0 . 8B limonene 0 .2 0 .4 0 .9 0 .3 cis-p-ocimene 0 . 8A 2 , . 9B 4 . IC 2 .2B t . p-oc . /1 , 2 -diet .benz . 82 . . 8 127 . . 0 : L13 . .7 113 .9 1,3 -diethyl benzene 22 , .3 25 , .6 27 , .2 32 .0 1,4-diethyl benzene 5 , .7 6, .9 7 , .7 8 . 6 linalool oxide 0 , .4 1. .6 1, .7 0 .9 linalool 0 , .2 0. .6 0. ,7 1 .0 4 , 8-dimet .-1,3, 7-nonat . 1. .4A 4, .9B 9. .4C 5 . 2B phenyl acetonitrile 0 . .5 5 . ,8 8. .1 3 .2 ethyl benzaldehyde 3 . , lA 3 , , 9AB 4. ,3B 5 . OB ethenyl benzaldehyde 6, , 5 7 . ,2 8, ,5 9 .3 dodecenal 0, , lA 1. .IB 1. , 8B 1 .2B 1(2 , 4-dimet .phen . ) eth . 16 . ,1 19 . , 1 21. .5 24 . 0 1(4-eth.phen. ) eth. 8. ,2 9, ,5 9 . .3 10 . 8 E, E-a-f arnesene 2 . ,4 3 . ,9 5. ,1 1 .2 *Means within rows followed by different letters are significantly different at the 0.05 probability level by REG-WQ multiple comparison test.

PAGE 134

121 Data Analysis Phenology of volatile emissions In the case of all identified chemicals, emissions increase from diel 1 to diel 7. The patterns for l(4-ethyl phenyl) ethanone and combined trans-P-ocimene/1 , 2-diethyl benzene appear to differ, but there are no significant differences for their diel emission masses. The increase from diel 1 to diel 4 is statistically significant in a-pinene, Ppinene, cis-P-ocimene , 4 , 8-dimethyl-l , 3 , 7-nonatriene and dodecenal (ANOVA, P<0.001, REG-WQ) . Two general patterns of volatile production from diel 7 to diel 10 can be seen in the data. Reduction of volatile production from diel 7 to diel 10 can be seen in about half of all identified chemicals. Included in this group are apinene, P-pinene, limonene, cis-p-ocimene, linalool oxide, 4 , 8-dimethyl-l , 3 , 7-nonatriene, phenyl acetonitrile, dodecenal and E, E-a-farnesene. Of those, reduction from diel 7 to diel 10 is statistically significant in P-pinene (ANOVA, F=5.80, df=81,122, p<0.0Q01), cis-p-ocimene (ANOVA, F=6.30, df= 81,122, p<0.0001), 4, 8-dimethyl-l, 3, 7-nonatriene (ANOVA, F=3.69, df=81,122, p<0.0001).

PAGE 135

122 The second pattern seen in 8 of the remaining chemicals is that of continued increase in emissions from diel 7 to diel 10. Significant increases were found only for (2-methyl propyl) benzene. Cannonical correlation analysis Culicoides mississippiensis . Analysis of all chemical data obtained in the study by Canonical Correlation produced a correlation of 0.493. Standardized canonical coefficients (Wl) for each chemical are listed in Table 4-4. High captures of C. mississippiensis are correlated to high amounts of chemicals with positive coefficients and low amounts of chemicals with negative coefficients. The results show that high amounts of ethyl benzaldehyde and a-pinene combined with low amounts of 1,4-diethyl benzene are most strongly correlated with high captures of C. mississippiensis . These coefficients can be used to direct selection of chemicals to be removed from consideration in regression analysis. When bloom groups are compared by REG-WQ relative to Wl , bloom group 5 is shown to have a significantly higher mean Wl value than all other bloom groups. Bloom group 5 is followed by bloom groups 4, 6, and 3 in that order. These results confirm that the highest correlation between volatile

PAGE 136

123 Table 4-4. Standardized canonical coefficients of indicated chemicals for correlation with the square root of capture of C. mississippiensis . High amounts of chemicals with positive coefficients and low amounts of chemicals with negative coefficients correlate to high square root of capture. Chemical Wl dimethyl benzene -0 . 5015 a-pinene 1 .3847 P-pinene -0 . 6179 (2 -methyl propyl) benzene 0 .0699 limonene -0 .2563 cis-P-ocimene 0, .0814 t . -P-ocimene/1 , 2-diethyl benzene -0, .3269 1,3 -diethyl benzene 0 . 1856 1,4-diethyl benzene -2 , .3726 linalool oxide -0 , .0780 linalool 0 , .0025 4, 8 -dimethyl1, 3 , 7-nonatriene -0 , .4755 phenyl acetonitrile -0. . 1490 ethyl benzaldehyde 1. .4101 ethenyl benzaldehyde 0 . .7241 dodecenal 0. ,2890 1(2 , 4-dimethyl phenyl) ethanone 0. ,5979 l-(4-ethyl phenyl) ethanone -0, ,3292 E, E-a-f arnesene 0 . .2140

PAGE 137

124 composition and highest captures of C. mississippiensis also correlate to the periods of peak flowering of I., vomitoria . A similar exeimination of Wl values relative to diel periods shows highest mean Wl for diel 10 followed by 1, 4 and 7 in that order. However, none of the diel periods were significantly different from others. Analysis of variance for all chemicals yields an F value of 3.502 and p<0.0001, showing that at least one of the chemicals is a significant predictor for the response variable (capture^^^) . The model has an r^ value of 0.2656, indicating that other factors not considered in the analysis play an important role in predicting the response variable. A list of p values for each individual chemical included in the initial model is shown in Table 4-5. Test statements were used in regression analysis to identify chemicals that could be dropped from the model without significantly affecting the model. For example, when 1,4-diethyl benzene and dimethyl benzene are evaluated in a regression test, the p value for the group is 0.0001, showing that they need to be kept in the model. On the other hand, a test of the group linalool oxide, linalool and E, E-a-farnesene produces a p value of 0.9975, indicating that the group can be

PAGE 138

125 Table 4-5. Probability values of indicated chemicals for inclusion in a model predicting high square root of capture of C. mississippiensis . Chemical p intercept 0 . 0001 dimethyl benzene . 0 . 0620 a-pinene 0 . 0003 P-pinene 0 . 0228 (2-methyl propyl) benzene ' • 0 .6630 limonene 0 .2969 cis-P-ocimene 0 .7701 t . -P-ocimene/1 , 2-diethyl benzene 0 2263 1,3 -diethyl benzene 0 7681 1,4-diethyl benzene 0 0183 linalool oxide 0 8722 linalool 0 9915 4, 8 -dimethyl1, 3 , 7-nonatriene 0 0334 phenyl acetonitrile 0 7156 ethyl benzaldehyde 0 0139 ethenyl benzaldehyde 0 4826 dodecenal 0 3544 1(2 , 4-dimethyl phenyl) ethanone 0. 4622 l-(4-ethyl phenyl) ethanone 0 . 4468 E, E-a-f arnesene 0 . 4734 Table 4-6. Probability values of indicated chemicals for inclusion in a model predicting high square root of capture of C. mississippiensis following regression analysis . Chemical p Parameter Estimate intercept 0.0001 3.170 dimethyl benzene 0.0101 -0.269 a-pinene 0.0001 0.098 P-pinene 0.0584* -0.105 ' -vi 1,4-diethyl benzene 0.0001 -0.431 4, 8-dimethyl-l, 3 , 7-nonatriene 0.0354 -0.069 ethyl benzaldehyde 0.0029 0.484 ethenyl benzaldehyde 0.0390 0.207 * Indicates marginal significance

PAGE 139

126 dropped without significantly reducing the fit of the regression model. After running numerous tests of chemical groups, based on low absolute Wl values, a model for predicting optimal capture of C. mississippiensis was developed including a-pinene, Ppinene, 1,4-diethyl benzene, dimethyl benzene, 4 , 8-dimethyl1 , 3 , 7-nonatriene, ethyl benzaldehyde and ethenyl benzaldehyde . P values and parameter estimates, used in model construction, are shown in Table 4-6. The model based on those parameter estimates is: capture^^^ = 3.17 0 . 269 {dimethyl benzene} + 0 . 098 {a-pinene} 0 . 104{p-pinene} 0 . 431{1, 4-diethyl benzene} 0 . 0 6 9 { 4 , 8 -dime thyl -1,3,7 -nona tr iene } + 0.484 {ethyl benzaldehyde} + 0 . 207 {ethenyl benzaldehyde } Plots of residuals versus predictors were used to look for higher order polynomial trends, but none were apparent. Dasvhelea mutabilis. Capture data for D. mutabilis was handled in the same way for canonical correlation analysis as was the data for C. mississippienisis . This analysis produced a correlation of 0.530. Standardized canonical coefficients

PAGE 140

127 (Wl) for each chemical are listed in Table 4-7. Analysis of variance for all chemicals yields an F value of 3.781 and p<0.0001 (df=19,184), showing that at least one of the chemicals is a significant predictor for the response variable (capture^'^) . The model has an r^ value of 0.2808, indicating that other factors not considered in the analysis play an important role in predicting the response variable. A list of p values for each individual chemical included in the initial model is shown in Table 4-8. Essential oils most highly correlated with high capture^''^ of D. mutabilis included a-pinene, 1,4-diethyl benzene, linalool oxide, 1(2 , 4-dimethyl phenyl) ethanone, ethyl benzaldehyde, dimethyl benzene, and phenyl acetonitrile . Probability values and parameter estimates for those essential oils following removal of other chemicals from the model are shown in Table 4-9. Based on those values a model may be developed of the form: capture^'^ = 8.025 + 0.228 {a-pinene} -1.097 {1,4-diethyl benzene} 2 . 288 {linalool oxide} + 0 . 199 {1(2 , 4dimethyl phenyl) ethanone} +1.3 06 {ethyl benzaldehyde} -0 . 85 6 {dimethyl benzene} + 0.298 {phenyl acetonitrile) Plots of residuals versus predictors did not reveal higher order polynomial trends to be included in the model.

PAGE 141

128 Table 4-7. Standardized canonical coefficients of indicated chemicals for correlation with with the square root of capture of D. mutabilis . High amounts of chemicals with positive coefficients and low amounts of chemicals with negative coefficients correlate to high square root of capture. Chemical Wl dimethyl benzene -0 .4567 a-pinene 1 . 1385 P-pinene -0 .2020 (2 -methyl propyl) benzene 0 .2072 limonene -0 .4611 cis-P-ocimene 0 .4575 t . -P-ocimene/1 , 2-diethyl benzene -0, .4564 1,3-diethyl benzene > 0, .2590 1,4-diethyl benzene . -1, .6480 linalool oxide -1, . 0172 linalool -0, . 0619 4 , 8-dimethyl-l , 3 , 7-nonatriene -0 . . 0753 phenyl acetonitrile -0. ,6844 ethyl ben z aldehyde 0. .9143 ethenyl benzaldehyde 0 , .5333 dodecenal '"' 0 . .3862 1(2 , 4-dimethyl phenyl) ethanone 1. ,9740 l-(4-ethyl phenyl) ethanone -0 . ,8788 E, E-a-f arnesene -0 . 1695

PAGE 142

129 Table 4-8. Probability values of indicated chemicals for inclusion in a model predicting high square root of capture of D. mutabilis . Chemical p intercept 0 . 0001 dimethyl benzene 0 . 0607 a-pinene 0 .0009 p-pinene 0 .4083 (2 -methyl propyl) benzene 0 .1549 limonene 0 . 0392 cis-p-ocimene 0, .0710 t . -P-ocimene/1 , 2-diethyl benzene 0, . 0630 1,3 -diethyl benzene 0, .6497 1,4-diethyl benzene 0 , .0702 linalool oxide 0, .0216 linalool 0. .7731 4 , 8-dimethyl-l , 3 , 7-nonatriene 0 . .7085 phenyl acetonitrile 0, .0659 ethyl benzaldehyde 0. ,0772 ethenyl benzaldehyde 0. .5680 dodecenal 0, .1726 1(2 , 4-dimethyl phenyl) ethanone 0 . .0079 l-(4-ethyl phenyl) ethanone 0. .0259 E, E-a-f arnesene 0 . , 5308 Table 4-9. Probability values of indicated chemicals for inclusion in a model predicting high square root of capture of D. mutabilis following regression analysis. Parameter estimates are used for construction of the model . Chemical p Parameter Estimate intercept 0.0001 8.025 a-pinene 0.0001 0.228 1,4-diethyl benzene 0.0001 -1.097 linalool oxide 0.0006 -2.288 1(2 , 4-dimethyl phenyl) eth. 0.0149 0.199 ethyl benzaldehyde 0.0140 1.306 dimethyl benzene 0.0038 -0.856 phenyl acetonitrile 0.0102 0.298

PAGE 143

130 Discussion Sample Analysis , Blank and Nonflowering samples The lack of volatiles from blank and nonflowering samples indicates that the method of volatile sampling was not able to detect chemicals from either the oven bag or duct tape used for seals. Absence of detectable quantities of volatiles may have been due to the relatively short time of sample collection. In the case of nonflowering vegetation, lack of volatiles may also relate to the fact that all vegetation was from the previous season's growth. As noted in Chapter 3, new leaves are produced after flowering has occurred. Pollenkitt samples High retention times of pollenkitt peaks indicate that those chemicals have low volatility (W. M. Whitten, personal communication) . None of the pollenkitt volatiles with retention times less than 40 minutes was also found in floral volatiles. Pollenkitt volatile contributions to overall floral fragrance were therefore not detectable by the protocol of this study. In a study done on two Rosa species, Dobson et

PAGE 144

131 al . (1987) also found that the volatiles of pollenkitt were distinctly different from those of flowers. This does not mean that pollenkitt volatiles are necessarily specific to pollenkitt. Heptanal, octanal, nonanal, heptanoic acid and nonanoic acid, identified from pollenkitt in this study, have also been identified from flowers of the chestnut, Castanea creata (Yamaguchi and Shibamoto 1980) . Heptanal, decanal, and nonanal have been identified in the floral fragrance of orchids (Omata et al . 1990, Borg-Karlson and Groth 1986). Flowering season samples The absence of data for a chemical from any plant does not necessarily mean that the chemical was not emitted by the plant since emission rates may have been below the sensitivity threshold for the procedure. Many of the floral essential oils identified in this study are commonly found in emissions of other flowering plants (Chang et al . 1988, Nielsen et al . 1995, Patt et al . 1988, Pellmyr et al . 1987). Significant electroantennogram response has been identified to limonene by the European grapevine moth (Gabel et al . 1992) and to limonene, linalool and both cisand trans-P-ocimene by alfalfa seed chalcids. Both limonene and a-pinene have been identified as attractants for scolytid beetles and boll

PAGE 145

132 weevils and linalool as an attractant for silkworms (Berenbaum and Siegler 1992). Results such as these showing insect attraction to essential oils common to many flowering plants supports future investigation of possible attraction by C. mississippiensis to those chemicals. Data Analysis Phenolocrv of volatile emissions Diffusion of terpenes through plant cell walls and cuticle is a temperature dependent event (Fahn 1979) . The GC data of the present study support field observation of increased floral scent in late morning and early afternoon when higher temperatures would favor essential oil diffusion to the plant exterior. Quantities of monoterpene emissions by flowers have also been shown to vary relative to the age of plants and to time within the flowering season (Chang et al . 1988). Other studies have shown that some floral terpenes are released rhythmically while others are not (Matile and Altenburger 1988, Nielsen et al . 1995). Only l(4-ethyl phenyl) ethanone and combined trans-P-ocimene/1 , 2-diethyl data did not fit one of the two rhythmic emission patterns shown for other I. vomitoria essential oils. Since there were no significant

PAGE 146

133 differences shown for any of their diel period emissions, they may also fit one of the two emission patterns. Cannonical correlation analysis Culicoides mississippiensis . Occasionally attractants for insects are single compounds, such as methyl eugenol for tephritids (Fraenkel 1959) . However, synergistic interaction of essential oils in attraction systems is a more common phenomenon. Guerin et al . (1983) determined that carrot flies are more strongly attracted to a mixture of hexenal and trans asarone than to either of the chemicals alone. A study on the attraction of sunflower volatiles to the red sunflower seed weevil showed that a mixture of five chemicals produced attaction, but that removal of only one chemical from the blend reduced attraction (Roseland et al . 1992). The model for C. mississippiensis capture predicts the response of capture number to a change in the output of any one chemical. For example, if the release of a-pinene was increased by one unit, and all other chemicals held constant, the capture^'^ of C. mississippiensis trapped would be expected to increase by a factor of 0.098. If on the other hand, Ppinene output increased by one unit while all other chemicals

PAGE 147

134 were held constant, the capture^^^ of C. mississippiensis trapped would be expected to decrease by a factor of 0.104. The model does not necessarily predict response to simultaneous changes in more than one chemical. The results indicate that potential attractants for C. mississippiensis could include from one to all of the chemicals in the model. Chemical blends, in which levels of chemicals not highly correlated with capture of D. mutabilis are manipulated, would be of special interest. The results for D. mutabilis indicate that it may be possible to tailor a volatile blend that is attractant to C. mississippiensis but not to D. mutabilis . Test blends should avoid manipulation of a-pinene, 1,4-diethyl benzene, ethyl benzaldehyde, and dimethyl benzene and focus on chemicals not shared by the two models .

PAGE 148

CHAPTER 6 SUMMARY AND CONCLUSIONS Research Results It is unrealistic to expect to find quick solutions to an insect control problem as challenging as that of Culicoides mississippiensis Hoffman. The goal of the present study was to evaluate the association between this pest and flowering yaupon holly. The general outcomes of the research were an increased understanding of that association and several leads for future study in the field of electrophysiology . Specific results are summarized for the plant, the plant's feeding guild, and floral volatiles. Yaupon Holly The yaupon holly. Ilex vomitoria Alton, was intensively studied during its entire flowering season from 16 March to 15 April 1995 on a hammock in Yankeetown, Florida. All plant positions were plotted on a map of the hammock and data collected for sex ratio, size and start of flowering. Percent bloom as a function of time was determined for five male 135

PAGE 149

136 plants based on an estimate of flowering potential and actual flowering data. Sex ratio A total of 145 plants were mapped, of which 71 were male, 59 female, and 15 of undetermined sex. The sex ratio of 1.20:1 (male : female) for flowering plants is comparable to sex ratios for other hollies (Cavigelli et al . 1986, Richards 1988) . Size Of the 71 male plants, 26 (36.6%) were small, 25 (35.2%) were medium and 20 (28.2%) were large. Female size totals were 26 small (44.1%), 14 were medium (23.7%) and 19 were large (32.3%). All but one plant of undetermined sex were small. Start of flowering Plants started flowering over a 33 day span. The mean start day for males (day 11) was significantly different from that of females (day 18) . All small plants begin flowering significantly later than other plants. Medium male plants began flowering significantly later than large male plants, but there was no significant difference between medium and

PAGE 150

137 large female plants. There was significant interaction between the start of flowering and mean daily temperature. Percent bloom Five large male plants were included in a study of the visiting insect community and floral volatile emissions over the entire flowering season. Male plants were chosen for the study because they had significantly more buds per cm stem (20.4) than female plants (2.9) and would serve as better sources of insects and volatiles. Plants that started flowering early in the season reached significantly higher percent bloom than plants that began flowering late in the season. Mean peak bloom ranged from 1.9% (late season plant) to 18.5% (early season plant). Depression of flowering in late season plants is significantly associated with high thrips populations. Thrips damage to buds was common in late season plants and was most likely the cause of reduced flowering. Eight percent bloom classes were constructed to facilitate correlation of insect and volatile samples to flowering. Yaupon Feeding Guild One-minute vacuum samples of arthropods were collected by AFS sweeper before the start of the flowering season and four

PAGE 151

138 times daily throughout the flowering season. Pre-f lowering means for important insect taxa were less than means for taxa collected from flowering plants. Insect data for flowering season samples provided a profile of arthropod visitations. Profile of arthropod visitors More than 90% of all arthropods were either thrips (59.5%) or the non-biting ceratopogonid Dasvhelea mutabilis Kieffer (30.6%). Common visitors to flowering yaupon were C. mississippiensis (3.4%), other nematocera (2.2%) and chalcidoid Hymenoptera (1.7%). Uncommon visitors were Collembola (0.7%), Hemiptera (0.6%), Coleoptera (0.3%), other Diptera (0.3%), Acari (0.2%), Lepidoptera (0.1%), Homoptera (0.1%), Araneida (0.1%), Pscoptera (<0.1%) and Mollusca (<0.1%) . Visitation rhythms Dasvhelea mutabilis. These insects had similar dynamics to C. mississippiensis relative to their visits of flowering yaupon. Highest numbers were found in late afternoon, in the first half of the flowering season, and were significantly higher at periods of peak bloom. However, the number of visitations was significantly lower in the early morning than at any other time of the day.

PAGE 152

139 Thrips . Thrips visitations to flowering yaupon were highest during the second half of the flowering season. Their numbers were high throughout the day, but highest in the late afternoon. Their numbers were also significantly higher at periods of low bloom, during the initiation of flowering and flowering senescence. Culicoides mississippiensis . Preference for visiting flowering yaupon in crepuscular periods was shown, with the mean number of these biting midges collected before sunset significantly higher than any other time of the day except the early morning. Peak visitations occurred every three to four days in the first half of the flowering season (16 March to 4 April) dropping to lower levels after that time. Visitations by C. mississippiensis were significantly higher at periods of peak bloom. Anthrone tests were carried out on a subsample of C. mississippiensis to identify nectar feeding behavior. Nectar feeding was compared to blood-meal seeking behavior. Sugar and blood feeding behavior Positive anthrone tests were obtained for 55.6% of biting midges tested, indicating that C. mississippiensis is attracted to flowering yaupon to feed on nectar. There were significantly more female anthrone positives (61.5%) than male

PAGE 153

140 anthrone positives (49.7%). Males and females took nectar meals throughout the day. Lowest positive rates for both male and female were found for early evening possibly due to the high number of recent arrivals on plants. During the study period, the biting rate for female C. mississippiensis in the early evening (689 bites/h) was significantly greater than at any other time of the day, followed by early morning (140 bites/h) . The range of biting rates over the study period was 0-2484 bites/h. Host seeking was recorded over a temperature range of 10 . 5°C-31 . 0°C . Floral Volatiles Samples of blank, non-floral vegetation and pollen samples were analyzed by GC/MS to make comparison with floral volatiles. No identifiable peaks were found in either blank or non-floral vegetation samples. Pollen samples contained mostly high molecular weight (low volatility) compounds, with small amounts of heptanal and nonanal . Pollen was not shown to account for any of the chemicals present in floral samples. Identification More than 190 volatile samples were analyzed for chemical composition. Of 51 chemicals which eluted, 20 were at high enough levels to be identified. They were dimethyl benzene,

PAGE 154

141 a-pinene, p-pinene, (2 -methyl propyl benzene, limonene, cis-Pocimene, trans-P-ocimene , 1,2-diethyl benzene, 1,3-diethyl benzene, 1,4-diethyl benzene, linalool oxide, linalool, 4,8dimethyl-1, 3 , 7-nonatriene, phenyl acetonitrile , ethyl benzaldehyde, ethenyl benz aldehyde, dodecenal, 1(2 , 4-diinethyl phenyl) ethanone, l-(4-ethyl phenyl) ethanone, and E,E-afarnesene. Mean production of these chemicals was in a range of 7.87 iig/hr (early morning) to 12.07 ug/hr (early afternoon) . Microliter volumes of samples were analyzed by GC/MS and mean mass/ul-hr ranged from 0 . 1 ng (dodecenal, early morning) to 127.0 ng (combined trans -P-ocimene and 1,2-diethyl benzene) . Phenol ocrv Two patterns of volatile emission were detected. Emission of nine chemicals increased from early morning to early afternoon and then decreased to sunset. Of those the decrease from afternoon to sunset is significant for P-pinene, cis-p-ocimene, and 4 , 8-dimethyl-l , 3 , 7-nonatriene . The second pattern is that of increase from early morning up to sunset. Eight chemicals show this pattern, including dimethyl benzene, (2 -methyl propyl) benzene, 1,3-diethyl benzene, 1,4-diethyl benzene, linalool, ethyl benzaldehyde, ethenyl benzaldehyde

PAGE 155

142 and 1(2 , 4-dimethyl phenyl) ethanone. Only (2 -methyl propyl) benzene showed a significant increase from early afternoon to sunset . Canonical correlation analysis Data for all identified chemicals was analyzed by canonical correlation, yielding a correlation of 0.493. Standardized canonical correlation coefficients were used to direct selection of chemicals to be included in regression analysis for development of a linear model for optimal capture of C. mississippiensis . This analysis showed that seven of the twenty chemicals are significantly correlated to peak biting midge visitation of flowering yaupon. Those seven are dimethyl benzene, a-pinene, P-pinene, 1,4 -diethyl benzene, 4 , 8-dimethyl-l, 3 , 7-nonatriene, ethyl benzaldehyde and ethenyl benz aldehyde . Seven essential oils of I. vomitoria are significantly correlated to visitation of D. mutabilis to flowering male plants. Those chemicals are a-pinene , 1 , 4-diethyl benzene, linalool oxide, 1( 2 , 4-dimethyl phenyl) ethanone, ethyl benzaldehyde, dimethyl benzene and phenyl acetonitrile . These results indicate that D. mutabilis may be attracted to a different blend composition than C. mississippiensis .

PAGE 156

143 Conclusions The results of the research indicate that C. mississippiensis visits flowering I. vomitoria to obtain nectar. Canonical correlation identified seven essential oils emitted by male I. vomitoria flowers as possible secondary attractants for C. mississippiensis . Seven essential oils were also correlated to high capture of D. mutabilis . Four essential oils were highly correlated for both C. mississippiensis and D. mutabilis . but the remaining three were different. Yuval (1992) recommends studies to identify attractants that would concentrate mosquitoes at bait stations. The results of this study provide leads for formulating essential oil blends that may be capable of concentrating C. mississippiensis at bait stations. Although it is possible that other factors, such as flower color and weather conditions, play important roles in attracting biting midges to flowering yaupon, those seven chemicals merit further investigation. A behavioral bioassay should first be developed. The Y-tube model developed by Blackwell et al. (1993) should be considered for this purpose. Studies have shown attractant responses to l-octen-3-ol by

PAGE 157

144 other species of biting midges (Kline et al . 1994, Ritchie et al . 1994). Evaluation of the attractancy of C. mississippiensis to l-octen-3-ol should be made to determine its suitability for use in development of a Y-tube bioassay. Electroantennogram trials, following the method of Blackwell et al . (1993) should then be carried out, first on whole flower extract, then on discrete chemicals and blends. Comparison of the models for C. mississippiensis and D. mutabilis developed by canonical correlation analysis should guide selection of chemicals for candidate attractant blends. Discrete chemicals or blends that produce significant EAG responses should next be tested for possible attraction by a working bioassay. Any attractants for C. mississippiensis may also prove to be attractants for D. mutabilis , thrips and a variety of other arthropods. Timed release of attractants in a trapping system could prevent capture of non-targets. Release only in the early morning would both target C. mississippiensis at one of its most active times and avoid peak activity of Hymenoptera and D. mutabilis . the most likely non-target victims of control based on yaupon volatiles. Nevertheless, susceptibility of D. mutabilis to yaupon-based control must be

PAGE 158

145 thoroughly evaluated to prevent undesired reductions of its population. Such evaluation should also determine the feasibility of late afternoon release of attractants in order to target C. mississippiensis when it is most active.

PAGE 159

APPENDIX A FLORAL PHENOLOGY OF ILEX VOMITORIA The following table contains data relative to the start of flowering and size of each of the Ilex vomitoria plants included in the dissertation study. Plants are coded with the letter A, corresponding to the study site hammock, and number. All plant locations are indicated on the map shown in figure 3-5. Sex determination was made on the basis of flower structure. Plants that did not flower were considered to be of undetermined sex. Size was approximated visually as small, medium or large. Start of flowering was made by direct observation of plants every two to three days. 146

PAGE 160

147 Table A-1. Floral phenology of Ilex vomitoria on Hammock A in Yankeetown, Florida. Sex is listed as female (F) , male (M) , or undetermined (U) . Sizes are small (S) , medium (M) , or large (L) . Plants indicated by * are those included in dissertation sampling exercises. ID Sex Size Start Bloom Al F M 3 / 9 Q A2 F T. XJ A3 F s A4 M c A5 u C o IN \JVi Hj A6 u q A7 F A8 M c D A9 T. J / A'± AlO M r. All F s 3/29 A12 M L 3/21 A13 M M 3/18 A14* M L 3/21 A15 F S 3/29 A16 M L 3/18 A17 F M 4/9 A18 M L 3/24 A19 M L 3/24 A20 M L 3/24 A21 F L 3/27 A22 F L 4/9 A23 M M 4/3 A24 F S 3/29 A25 M M 3/27 A2 6 U S NONE A27 F S 3/18 A2 8 F L 3/24 A29 M S 3/24 A3 0 M M 3/18

PAGE 161

148 Table A-1. Continued ID Sex Size Start Bloom A31 F s 3/29 A3 2 M L 3/29 A3 3 F s 3/24 A3 4 F L 4/7 A35* M L 3/24 A36* M L 3/14 A3 7 F L 3/18 A3 8 M s 4/7 A3 9 M L 3/16 A40 M L 3/18 A41 F s 3/29 A42 F s NONE A43 F s 3/31 A44 F s 4/3 A45 F s 3/31 A46 F L 3 /29 A47 M s 3/21 A48 M s 3 /24 A49 M s 3/21 A50 F L 3/29 A51 F s 3/29 A52 u L NONF A53 M L 3/18 V--.,, A54 F L 3/24 A55 U S NONE A56 F L 3/31 A57 F L 3/29 ASS F L 3/27 A59 M M 3/27 A60 F S 3/31 A61 F S 4/7 A62* M L 3/24 A63 F L 4/7 A64 F S 4/7 A65 F L 3/29

PAGE 162

149 Table A-1. Continued ID Sex Size Start Bloom A66 M M 3/16 A67 M M 3 /21 A68 u s NONE A69 M s 3/27 A70 u s NONE A71 F s •4/3 A72 M s 4/7 A73 F s 3/31 A74 F s 4/3 A75 F M 3/31 A76 F s 4/12 A77 M M 3/27 A78 M M 3/24 A79 u s NONE A80 u s NONE A81 F M 4/3 A82 F M 4/3 A83 u s NONE A84 M M 3/24 A85 F M 4/3 A86* F J / ^ X A87 M M 3/94 A88 F M 3/21 A89 F M 3/24 A90 F M 3/31 A91 F M 3/24 A92 M S 3/24 A93 M S 3/21 A94 F S 4/7 A95 M S 3/24 A96 M S 3/29 A97 F S 3/24 A98 M S 3/21 A99 M S 3/27 AlOO M M 3/24

PAGE 163

150 Table A-1. Continued ID Sex Size Start Bloom AlOl M S 3/24 A102 u s NONE A103 U s NONE A104 F s 4/15 A105 F M 4/3 A106 M M 3/24 A107 F M 3/27 A108 M L 3 /24 A109 M M 3/27 Alio M M 3/27 Alll F M 3/29 A112 M M 3/24 A113 M s 3/24 A114 M s 3/27 A115 M s 3/24 A116 M s 3/27 A117 F s 4/7 A118 F s 4/7 A119 M M 3/24 A120 M L 3/21 A121 M M 3 /24 A122 M L 3/21 A123 M M 3/24 A124 M M 3/21 A125 F S 4/7 A126 F S 4/7 A127 M M 3/27 A128 M S 3/31 A129 M L ; 3/29 A130 F M '3/29 A131* F L 3/31 A132 M T S 3/24 A133 M S 3/29 A134 F . S 4/7 A13 5 U S NONE

PAGE 164

Table A-1. Continued ID Sex Size Start Bloom A13 6 F S 3/27 A13 7 M s 3/27 A138 F s 3/24 A139 M L 3/24 A140 M S 3/24 A141 M M 3/24 A142 M M 3/24 A143 M M 3/27 A144 M S 4/7 A145 M M 3/29

PAGE 165

APPENDIX B PERCENT BLOOM DATA . r . , ' The following tables contain data relative to the percent bloom of each plant included in the volatile study. Diel periods correspond to hour-long sample periods at sunrise, late morning, early afternoon and sunset. Percent bloom was determined by multiplying the total sample segment length (used in volatile sampling) by the average number of buds/cm (males: 20 buds/cm, females: 3 buds/cm) to yield a flower potential value (the total number of flowers that the sample segment could produce.) The number of flowers on the sample segment during any diel period may then be divided by the flower potential to give percent bloom. Sample Calculation (A36 during diel period 1 on 3/16/95) : Total Sample Segment Length = 106.5 cm Mean buds /cm = x 20 Flower potential = 2130 Percent Bloom = No. open flowers = 3 = 0.1% Flower potential 213 0 152

PAGE 166

153 Table B-1. Percent bloom data for plant A36. Total sample segment length of 106.5 cm. Flower potential of 2130. Numbers shown in each diel period are for the number of open flowers and percent bloom indicated in parentheses. Diel Period Date 1 4 7 10 3/16 3(0.1) 6(0.3) 4(0.2) 4(0.2) 3/17 2(0.1) 5(0.2) 6(0.3) 5(0.2) 3/18 6(0.3) 28(1.3) 30(1.4) 30(1.4) 3/19 54(2.5) 67(3.1) 70(3.3) 73(3.4) 3/20 62(2.9) 89(4.2) 123(6.2) 128(6.0) 3/21 163(7.7) 160(7.5) 169(7.9) 169(7.9) 3/24 414(19.4) 390(18.3) 382(17.9) 380(17.8) 3/25 464(21.8) 405(19.0) 374(17.6) 365(17.1) 3/26 409(19.6) 380(17.8) 376(17.7) 370(17.4) 3/27 436(20.5) 386(18.1) 344(16.2) 340(16.0) 3/28 369(17.3) 340(16.0) 294(13.8) 294(13.8) 3/29 143(6.7) 125(5.9) 109(5.1) 109(5.1) 3/31 104(4.9) 104(4.9) 104(4.9) Note: No data collection on 3/31, inclement weather. diel period 10 due to

PAGE 167

154 Table B-2 . Percent bloom data for plant A14 . Total sample segment length of 148.0 cm. Flower potential of 2960. Numbers shown in each diel period are for the number of open flowers and percent bloom indicated in parentheses . Diel Period Date 1 4 7 10 3/24 16 (0. 5) 29 (1. 0) 30(1. 0) 29 (1. 0) 3/25 70 (2 . 4) 70 (2 . 4) 79 (2 . 7) 77 (2 . 6) 3/26 130 (4. 4) 131(4. 4) 129 (4 . 4) 125 (4. 2) 3/27 157 (5 . 3) 155 (5 . 2) 149 (5 . 0) 147 (5 . 0) 3/28 243 (8. 2) 240 (8 . 1) 239 (8 . 1) 239 (8 . 1) 3/29 246 (8. 3) 240 (8. 1) 220 (7 . 4) 220 (7 . 4) 3/31 446 (15 '.1) 446 (15 .1) 446 (15 .1) 4/1 362 (12 .2) 362 (12 .2) 4/2 81 (2 . 7) 100 (3 . 4) 106 (3 . 6) 105 (3 . 5) 4/3 94 (3 . 2) 92 (3 . 1) 90 (3 . 0) 90 (3 . 0) 4/4 99 (3 . 3) 75 (2 . 5) 70(2. 4) 70 (2 . 4) 4/7 4(0. 2) Note: No data collection on 3/31, diel period 10 and 4/1, diel periods 1 and 4, due to inclement weather. No data collection after diel period 1 on 4/7 due to end of flowering.

PAGE 168

155 Table B-3. Percent bloom data for plant A62. Total sample segment length of 182.0 cm. Flower potential of 3640. .;, Numbers shown in each diel period are for the number of open flowers and percent bloom indicated in parentheses. Diel Period Date 1 4 7 10 3/25 25 (0 .7) 28 (0 .8) 34 (0 .9) 28(0 .8) 3/26 34 (0 .9) 28(0 .8) 28 (0 .8) 26(0 .7) 3/27 44 (1 .2) 46 (1 .3) 46 (1 .3) 45 (1 .2) 3/28 172 (4 .7) 150 (4 .1) 152 (4 .2) 152 (4 .2) 3/29 146 (4 .0) 95 (2 .6) 88 (2 .4) 88 (2 .4) 3/31 59 (1 .6) 59 (1 .6) 59 (1, .6) 4/1 44 (1, .2) 44 (1 .2) 4/2 7(0 .2) 30(0 .8) 35(1. .0) 35(1 .0) 4/3 32 (0 .9) 35(1 .0) 35 (1. .0) 35(1 .0) 4/4 68 (1 .9) 40 (1 .1) 33 (0 . .9) 33(0 .9) 4/7 6(0 .2) 13 (0 .4) 13 (0, .4) 13 (0 .4) 4/8 6(0 .2) 6(0 .2) 6(0. .2) 6(0 .2) 4/9 6(0 .2) 6(0 .2) 6(0. .2) 6(0 .2) Note: No data collection on 3/31, diel period 10 and 4/1, diel periods 1 and 4, due to inclement weather.

PAGE 169

156 Table B-4. Percent bloom data for plant A35. Total sample segment length of 213.0 cm. Flower potential of 4260. Numbers shown in each diel period are for the number of open flowers and percent bloom indicated in parentheses . Diel Period Date 1 4 7 10 4/1 64 (1 .5) 64(1. 5) 4/2 126 (3 0) 135 (3 2) 149 (3 .5) 149 (3 . 5) 4/3 151 (3 5) 110 (2 6) 110 (2 .6) 110 (2 . 6) 4/4 73 (1. 7) 45 (1 1) 31(0 7) 31(0. 7) 4/7 119 (2 . 8) 159 (3 . 7) 159 (3 7) 159 (3 . 7) 4/8 170 (4. 0) 126 (3 . 0) 125 (2 9) 125 (2 . 9) 4/9 126 (3 . 0) 96 (2 . 2) 96 (2 2) 96 (2 . 2) 4/10 73 (1. 7) 60 (1. 4) 60(1 4) 60 (1. 4) 4/11 51 (1. 2) 45 (1. 1) 35(0 8) 31(0. 7) 4/12 60 (1. 4) 51(1. 2) 42 (1 0) 42 (1. 0) 4/14 33 (0. 8) 33 (0. 8) 33 (0 8) 33 (0. 8) 4/15 18(0. 4) 18 (0. 4) 12 (0 3) 7(0. 2) Note: No data collection on 4/1, diel periods 1 and 4, due to inclement weather.

PAGE 170

157 Table B-5. Percent bloom data for plant A129. Total sample segment length of 181.0 cm. Flower potential of 3620. Numbers shown in each diel period are for the number of open flowers and percent bloom indicated in parentheses. Diel Period Date 1 4 7 10 4/7 33 (0 9) 33 (0 9) 33 (0 .9) 4/8 30(0 .8) 29 (0 8) 29(0 8) 29 (0 .8) 4/9 34 (0 .9) 20(0 6) 20(0 6) 20(0 .6) 4/10 39(1 .1) 49 (1 4) 49 (1 4) 49 (1 .4) 4/11 91 (2 .5) 70 (1 9) 58 (1 6) 53 (1 .5) 4/12 101 (2 .8) 83 (2 3) 81 (2 2) 81 (2 .2) 4/14 53 (1 .5) 53 (1 5) 53 (1 5) 53 (1 .5) 4/15 25(0 .7) 17 (0 5) 7 (0 2) 8(0 .2) Note: No data collection on 4/7, diel period 1 due to unexpected end of bloom by A14 as indicated by 4/7, diel period 1 sample of that plant and replacement by A129.

PAGE 171

' APPENDIX C SAMPLE GAS CHROMATOGRAPHY TRACES Sample gas chromatography traces of pollen and floral volatile samples analyzed by the protocol described in Chapter 4 are included in the following two pages. The floral volatile sample is that of diel period 2 on 31 March 1995 for plant A3 6. The tridecane internal standard peak is found at 21.2 min on each GC trace. Peaks below 3 minutes were produced by the solvent. 158

PAGE 172

159 = T O O "S © • © X T • O S — S S 3 o c •H E (1) g o c o •H U a u 0) Pi o CM in tn a H CO o •H o > X 0) 4-1 o 0) O (D u u u o 4-) -H 0) o 0, I o 0) tn H m m m (U vO -aCM o . — 1 1 — 1 >3(U O if OJ o
PAGE 173

160 r O Si5 o exJ ^ ^ Q <3 23 o e: 2: O = 3 • S T -Si u CM » 01 OJ o « o o o t— 1 00 OJ o o n c •H e 0) 6 O -H fvl H C O •H u C 0) u ID 0) •H a 03 -H

PAGE 174

APPENDIX D CALCULATION OF SAMPLE CHEMICAL MASSES The following is a sample calculation of the mass of a chemical emitted by a flowering male 1^ vomitoria in a one hour sample. The calculation uses data obtained by gas chromatographic analysis of a mixture of sample extract with internal standard so that each ]il of the mixture contains 40 ng of the internal standard. ' Source: A3 6 Date: March 19, 1995 Diel Period: 7 Data Peak area a-pinene iKp) = 16193 Peak area internal standard (A. J = 231660 Mean value for all internal standard peak areas (Aj^^J = 201305 Calculation 1. Standardization of a-pinene peak area Standardization Factor (SF)= A^^/A^^^ = 231660/201305 = 1.1508 Standardized a-pinene (Ag^^) = SF*A„„ = 1.1508*16193 = 18635 2. Mass calculation of standardized a-pinene peak area Kp = [Ag^p/A^.^] *40ng = [18635/201305] *40ng = 3.7 ng 161

PAGE 175

REFERENCES Baker, H. G. 1948. Corolla-size in gynodioecioius and gynomonoecious species of flowering plants. Proc . Leeds Philos. Lit. Soc . Sci. Sect. 1948: 136-139. Baker, H. G. And I. Baker. 1986. The occurrence and significance of amino acids in floral nectar. Plant. Syst. Evol. 151: 175-186. Bawa, K. S. and P. A. Opler. 1975. Dioecism in tropical forest trees. Evolution 29: 167-179. Berenbaum, M. and D. Siegler. 1992. Biochemicals : engineering problems for natural selection, pp 89-121. In B. D. Roitberg and M. B. Isman (eds) . Insect chemical ecology: an evolutionary approach. Chapman and Hall. New York. 3 59 pp. Bidlingmayer, W. L. and D. G. Hem. 1973. Sugar feeding by Florida mosquitoes. Mosq. News 33: 535-538. Blackwell, A., A. J. Mordue (Luntz), B. S. Hansson, L. J. Wadhams and A. J. Pickett. 1993. A behavioural and electrophysiological study of oviposition cues for Culex guinouef asciatus . Physiol. Entomol . 18: 343-348. Blanton, F. S. and W. W. Wirth. 1979. The sand flies (Culicoides) of Florida (Diptera: Ceratopogonidae) . Florida Department of Agriculture and Consumer Services. Gainesville, Florida. 204pp. Bonavita-Cougourdan, A., J. L. Clement and C. Lange . 1987. Nestmate recognition: the role of cuticular hydrocarbons in the ant Camponotus vagus Scop. J. Entomol. Sci. 22: 1-10 . 162

PAGE 176

163 Borg-Karlson, A. K. and I. Groth. 1986. Volatiles from four species in the sections Arachnitif ormes and Araneiferae of the genus Ophrys as insect mimetic attractants. Phytochemistry 25: 1297-1299. Borror, D. J., C. A. Triplehorn, and N. F. Johnson. 1992. An introduction to the study of insects. Harcourt Brace College Publishers. New York. 875 pp. Bovey, R. W. , H. L. Morton, R. E. Meyer, T. 0. Flynt and T. E. Riley. 1972. Control of yaupon and associated species. Weed Sci. 20: 246-249. Brantjes, N. B. M. 1978. Sensory responses to flowers in night-flying moths, pp 3-19. In A. J. Richards (ed. ) : The pollination of flowers by insects. Academic Press. London . Brizicky, G. K. 1964. The genera of Celestrales in the southeastern United States. J. Arnold Arbor. Harv. Univ. 45: 206-234. Budavari, S. (ed.) 1989. The Merck index. 11th ed. Merck and Co., Inc. Rahway, NJ. Budde, W. L. and J. W. Eichelberger . 1979. Organic analysis using gas chromatography /mass spectrometrya techniques and procedures manual. Ann Arbor Science Publishers, Inc. Ann Arbor, Michigan. 242 pp. Cane, J. H. and J. Tengo . 1981. Pheromonal cues direct mate-seeking behavior of male Colletes cunicularius (Hymenoptera: Colletidae) . J. Chem. Ecol . 7: 427436. Cantelo, W. W. and M. Jacobson. 1979. Phenacetaldehyde attracts moths to bladder flower, Arauiia sericofera and to blacklight traps. Environ. Entomol. 8: 444-447. Carlson, D. A. and M. W. Service. 1980. Identification of mosquitoes of Anopheles qambiae species complex A and B by analysis of cuticular components. Science 207: 10891091.

PAGE 177

164 Cavigelli, M. , M. Poulos, E. P. Lacey and G. Mellon. 1986. Sexual dimorphism in a temperate dioecious tree. Ilex montana (Aquifoliaceae) . Am. Midi. Nat. 115: 397-406. Chang, J. F., J. H. Benedict, T. L. Payne and B. J. Camp. 1988. Volatile monoterpenes collected from the air surrounding flower buds of seven cotton genotypes. Crop Sci. 28: 685-688. Childers, C. C, R. J. Beshear, J. R. Brushwein and H. A. Denmark. 1990. Thrips (Thysanoptera) species, their occurrence and seasonal abundance on developing buds and ' , flowers of Florida citrus. J. Entomol . Sci. 25: 601614. Clark, R. B. and E. R. Orton Jr. 1967. Sex ratio in Ilex opaca Ait. Hortscience 2: 115. ^ .. . Clement, R. E. and R. W. Karasek. 1985. Gas chromatography/ mass spectrometry/ computer instrumentation, pp 21-47. In F. W. Karasek, 0. Hutzinger and S. Safe (eds.): Mass spectrometry in environmental sciences. Plenum Press. New York. 578 pp. Croteau, R. 1988. Catabolism of monoterpenes in essential oil plants. Dev. Food Sci. 18: 65-84. Cruden, R. W. , S. M. Hermann and S. Peterson. 1983. Patterns , of nectar production and plant-pollinator coevolution. '. pp. 80-125. In B. Bentley and T. Elias (eds.): The . biology of nectaries. Columbia University Press. New York. 259 pp. Davis, E. L.1981. Laboratory studies on life cycle development and adult blood-feeding of Culicoides mississippiensis Hoffman (Diptera: Ceratopogonidae) . Unpubl . M.S. Thesis, Univ. of Florida. 107p. Dean, J. A. 1995. Analytical chemistry handbook. McGrawHill, Inc. New York. Diaz-Colon, J. D. , J. R. Baur and R. W. Bovey. 1970. Some characteristics of a growth inhibitory factor in yaupon fruit. Econ. Bot . 24: 403-407.

PAGE 178

165 Dicke, M. , and M. W. Sabelis. 1988. Inf ochemical terminology: should it be based on cost-benefit analysis rather than origin or compounds? Funct. Ecol . 2: 131139. » '^'p" Dobson, H. E. M. 1988. Survey of pollen and pollenkitt lipidschemical cues to flower visitors? Am. J. Bot. 75: 170-182. Dobson, H. E. M., J. Bergstrom, G. Bergstrom, and I. Groth. 1987. Pollen and flower volatiles in two Rosa species. Phytochemistry . 26:3171-3173. Downes, J. A. 1970. The ecology of blood-sucking Diptera: an evolutionary perspective. pp 232-258. In A. M. Fallis (ed.) Ecology and physiology of parasites, a symposium. Univ. Toronto Press. Toronto. Duncan, K. W. and C. J. Scifres. 1983. Yaupon and associated vegetation responses to seasonal tebuthiuron applications. J. Range Manage. 36: 568-571. Edman, J. D. , D. Strickman, P. Kittayapong and T. W. Scott. 1992. Female Aedes aeqypti (Diptera: Culicidae) in Thailand rarely feed on sugar. J. Med. Entomol . 29: 1035-1038 . Elias, T. S. 1983. Extrafloral nectaries: their structure and distribution. pp. 174-203. In B. Bentley and T. Elias (eds.): The biology of nectaries. Columbia University Press. New York. 259 pp. Faegri, K. and L. van der Pijl. 1979. The principles of pollination ecology. Pergamon Press. New York. 244pp. Fahn, A. 1979. Secretory tissues in plants. Academic Press. New York. 3 02pp. Fraenkel, G. 1959. The raison d'etre of secondary plant substances. Science 129: 1466-1470. Guerin, P. M. , E. Stadler and H. R. Buser. 1983. Identification of host plant attractants for the carrot fly, Psila rosae . J. Chem. Ecol. 9: 843-861.

PAGE 179

166 Haynes, K. F., J. Z. Zhao, and A. Latif. 1991. Identification of floral compounds from Abelia qrandi flora that stimulate upwind flight in cabbage looper moths. J. Chem. Ecol. 17: 637-646. Healy, T. P., and P. C. Jepson. 1988. The location of floral nectar sources by mosquitoes: the long-range responses of Anopheles arabiensis Patton (Diptera: Culicidae) to Achillea millefolium flowers and isolated floral odour. Bull. Entomol. Res. 78: 651-657. Heath, R. R., P. J. Landolt, B. Dueben and B. Lenczewski . 1992. Identification of floral compounds of nightblooming jessamine attractive to cabbage looper moths. Environ. Entomol. 21: 854-859. Hoffman, W. A. 1926. Notes on Ceratopogoninae (Diptera). Proc. Ent. Soc . Wash. 28: 156-159 Hu, S. Y. 1979. The botany of yaupon. pp. 10-39. In C. M. Hudson (ed. ) : Black drink: a native American tea. The University of Georgia Press. Athens, Georgia. 175 pp. Hudson, C. M. (ed.) 1979. Black drink: a native American tea. The University of Georgia Press. Athens, Georgia. 175 pp. Hume, H. H. 1953. Hollies . The Macmillan Company. New York . Hunter, D. M. 1977. Sugar-feeding in some Queensland black flies. J. Med. Entomol. 14: 229-232. Inouye, D. W. 1980. The terminology of floral larceny. Ecology 61: 1251-1253. Jamnback, H. 1961. Observations on Culicoides obsoletus (Meigen) in the laboratory (Diptera: Ceratopogonidae) . Mosq. News 21: 48-53. Janzen, D. H. 1967. Synchronization of sexual reproduction of trees within the dry season in Central America. Evolution 21: 620-637.

PAGE 180

167 Karstad, L.H., 0. K. Fletcher, J. Spalatin, R. Roberts and R. P. Hanson. 1957. Eastern equine encephalomyelitis virus isolated from three species of Diptera from Georgia. Science 125: 395-396. Kato, M. 1993. Floral biology of Nepenthes gracilis (Nepenthacene) in Sumatra. Am. J. Bot . 80: 924-927. Kevan, P. G. 197 6. Flourescent nectar. Science 194: 341-342. Kline, D. L. 1986. Seasonal abundance of adult Culicoides spp. (Diptera: Ceratopogonidae) in a salt marsh in i Florida, USA. J. Med. Entomol . 23: 16-22 Kline, D. L. 1989. Seasonal and spatial abundance of Culicoides spp. at Yankeetown, Florida. Fla. Entomol. 1. 72: 111-117. Kline, D. L. and J. R. Wood. 1988. Habitat preference of coastal Culicoides spp. at Yankeetown, Florida. J. Am. Mosq. Control Assoc. 4: 456-465. Kline, D. L., D. V. Hagan, and J. R. Wood. 1994. Med. Vet. Entomol. 8: 25-30. Knudsen, J. T., L. Tollsten and L. G. Bergstrom. 1993. Floral scentsa checklist of volatile compounds isolated by head-space techniques. Phytochemistry 33: 253-280. Lewis, J. A., C. J. Moore, M. T. Fletcher, R. A. Drew and W. Kitching. 1988. Volatile compounds from flowers of Spathiohvllum cannaef olium . Phytochemistry 27: 27552757. Lillie, T. H. and D. L. Kline. 1986. Occurrence of Culicoides mississippiensis on different types of vegetation. Fla. Entomol. 69: 672-677. Lillie, T. H., D. L. Kline and D. W. Hall. 1988. Hostseeking activity of Culicoides spp. (Diptera: Ceratopogonidae) near Yankeetown, Florida. J. Med. Entomol. 24: 503-511.

PAGE 181

168 Linley, J. R. , A. L. Hoch and F. P. Pinheiro. 1983. Biting midges (Diptera: Ceratopogonidae) and human health. J. Med. Entomol. 20: 347-364. Loesener, T. 1942. Aquif oliaceae . Nat. Pflanzenfam. ed. 2. 20b: 36-86. Loughrin, J. H., T. R. Kami It on -Kemp, H. R. Burton and R. A. Anderson. 1993. Effect of diurnal sampling on the headspace composition of detached Nicotiana suaveolens flowers. Phytochemistry 32: 1417-1419. Magnarelli, L. A. 1981. Parity, follicular development, and sugar feeding in Culicoides me Ileus and hollensis . Environ. Entomol. 10: 807-811. Matile, P., and R. Altenburger. 1988. Rhythms of fragrance emission in flowers. Planta 174: 242-247. McCreadie, J. W. , M. H. Colbo and F. F. Hunter. 1994. Notes on sugar feeding and selected wild mammalian hosts of black flies (Diptera: Simuliidae) in Newfoundland. J. Med. Entomol. 31: 566-570. Mead, F. W. 1983. Yaupon psyllid, Gyropsylla ilicis (Ashmead) (Homoptera: Psyllidae) . Florida Dept. Agric . , Div. Plant Ind. , Ent . Circ . 247: 1-2. Metcalf, R. L. and E. R. Metcalf. 1992. Plant kairomones in insect ecology and control. Chapman and Hall. New York. 168 pp. Meyer, R. P., w. K. Reisen, B. R. Hill and V. M Martrinez . 1983. The AFS Sweeper, a battery-powered backpack mechanical aspirator for collecting adult mosquitoes. Mosq. News 43: 346-5 0. Meyer, R. E. and R. W. Bovey. 1985. Response of yaupon ( ilex vomitoria) and understory vegetation to herbicides. Weed Sci. 33: 415-419.

PAGE 182

169 Milne, J. R., G. H. Walter, D. Kaonga and G. C. Sabio. 1996. The importance of non-pollen plant parts as food sources for the coirmon blossom thrips, Frankliniella schultzei . Entomol. Exp. Appl . 78: 271-281. Mookherjee, B. D. , R. W. Trenkle, R. A. Wilson, M. Zampino, K. P. Sands and C.J. Mussinan. 1986. Fruits and flowers: live vs deadwhich do we want? pp. 415-424. In B. M. Lawrence, B. D. Mookherjee and B. J. Willis (eds.) : Flavors and fragrances: a world perspective. Proceedings of the 10th international congress of essential oils, fragrances and flavors. Washington, DC. Mullens, B. A. 1985. Age-related adult activity and sugar feeding by Culicoides variipennis (Diptera: Ceratopogonidae) in southern California, USA. J. Med. Entomol. 22: 32-37. Nielsen, J. K. , H. B. Jakobsen, P. Friis, K. Hansen, J. M0ller and C. E. Olsen. 1995. Asynchronous rhythms in the emission of volatiles from Hesperis matronalis flowers. Phytochemistry 38: 847-851. Omata, A., S. Nakamura, K. Yomogida, K. Moriai, Y. Ichikawa, and I. Watanabe. 1990. Volatile components of TO-YORAN flowers ( Cymbidium faberi and C. virescens ) . Agric . Biol. Chem. 54: 1029-1033. Patt, J. M., D. F. Rhoades and J. A. Corkill. 1988. Analysis of the floral fragrance of Platanthera stricta . Phytochemistry 27: 91-95. Pellmyr, 0., G. Bergstrom, and I. Groth. 1987. Floral fragrances in Actaea . using differential chromatograms to discern between floral and vegetative volatiles. Phytochemistry 26: 1603-1606. Perry, J. A. 1991. Introduction to analytical gas chromatography: history, principles and practice. Dekker. New York. 42 6 pp.

PAGE 183

170 Peschke, K. 1987. Cuticular hydrocarbons regulate mate recognition, male aggression, and female choice of the rove beetle, Aleochara curtula . J. Chem. Ecol . 13: 1993-2008. Petanidou, T. and D. Voukou . 1993. Pollination ecology of Labiatae in a phryganic (east Mediterranean) ecosystem. Am. J. Bot. 80: 892-899. Phelan, P. L., A. W. Smith and G. R. Needham. 1991. Mediation of host selection by cuticular hydrocarbons in the honeybee tracheal mite Acarpis woodi (Rennie) . J. Chem. Ecol. 17: 463-473. Pijl, L. van der. 1960. Ecological aspects of flower evolution. I. Phyletic evolution. Evolution 14: 403-416. Reisen, W. K. , R. P. Meyer and M. M. Milby. 1986. Patterns of fructose feeding by Culex tarsalis (Diptera: Culicidae) . J. Med. Entomol . 23: 366-373. Richards, A. J. 1988. Male predominant sex ratios in holly (Ilex aguifolium L., Aquif oliaceae) and roseroot (Rhodiola rosea L., Crassulaceae) . Watsonia 17: 53-57. Ritchie, S. A., P. H. A. Van Essen, J. A. Kemme, B. H. Kay, and D. Allaway. 1994. Response of biting midges (Diptera: Ceratopogonidae) to carbon dioxide, octenol, and light in southeastern Queensland, Australia. J. Med. Entomol. 31: 645-648. . Roseland, C. R. , M. B. Bates, R. B. Carlson and C. Y. Oseto. 1992. Discrimination of sunflower volatiles by the red sunflower seed weevil. Entomol. Exp. Appl . 62: 99-106. SAS Institute. 1989. Statistical analysis system. SAS Institute Inc. Gary, North Carolina. Smith, S. M. and R. M. Kurtz. 1994. The use of an anthrone reagent to detect sugar meals and their persistence in the mosquito Aedes triseriatus (Diptera: Culicidae). Great Lakes Entomol. 27: 49-55.

PAGE 184

171 Snyder, A. W and W. H. Miller. 1972. Fly color vision. Vision Res. 12: 1389-1396. Stebbins, G. L. 1951. Natural selection and differentiation of angiosperm families. Evolution 5: 299-324. Stiles, F. G. 1975. Ecology, flowering phenology and hummingbird pollination of some Costa Rican Helliconia species. Ecology 56: 285-301. Tatsuka, K., S. Suekane, Y. Sakai and H. Sumitani . 1990 Volatile constituents of kiwi fruit flowers: simultaneous distillation and extraction versus headspace sampling. J. Agric . Food Chem. 38: 2176-2180. Tomlinson, P. B. 1974. Breeding mechanisms in trees native to tropical Florida; a morphological assessment. J. Arnold Arbor. Harv. Univ. 55: 269-290. Van Handel, E. 1972. The detection of nectar in mosquitoes. Mosq. News. 32: 458. Van Handel, E. 1984. Metabolism of nutrients in the adult mosquito. Mosq. News 44: 573-79. Van Handel, E., and J. F. Day. 1990. Nectar-feeding habits of Aedes taeniorhynchus . J. Am. Mosq. Control Assoc. 6: 270-273 . Van Handel, E., J. D. Edman, J. F. Day, T. W. Scott, G. G. Clark, P. Reiter and H. C. Lynn. 1994. Plant-sugar, glycogen, and lipid assay of Aedes aeqypti collected in urban Puerto Rico and rural Florida. J. Mosq. Control Assoc. 10: 149-153. Walsh, J. F. and R. Garms . 1980. The detection of plant sugars in Simulium damn os urn s. 1. by means of the cold anthrone test. Trans. R. Soc . Trop. Med. Hyg. 74: 811-813. Whitsel, R. H. and R. F. Schoeppner . 1967. Parasitic trombidiid mites on Dasvhelea mutabilis (Coquillett) (Diptera: Ceratopogonidae) . Proc . Ent . Soc. Wash. 69284286 .

PAGE 185

172 Whitten, W. M. , H. G. Hills, and N. H. Williams. 1988. Occurrence of ipsdienol in floral fragrances. Phytochemistry 27: 2759-2760. Winder, J. A. 1977. Field observations on Ceratopogonidae and other Diptera: Nematocera associated with cocoa flowers in Brazil. Bull. Ent . Res. 67: 57-63. Wirth, W. W., B. de Meillon, and E. Haeselbarth. 1980. Family Ceratopogonidae. pp 150-174. In R. W. Crosskey (ed.) Catalogue of the Diptera of the Afrotropical region. British Museum. London. 143 7 pp. Wirth, W. W., and C. Morris. 1985. The taxonomic complex, Culicoides variipenis . pp. 165-175. In T. L. Barber and M. M Jochim (eds.) : Bluetongue and related orbiviruses. Alan R. Liss, Inc. New York. 746 pp. Wright, D. H. 1988. Temporal changes in nectar availability and Bombus appositus (Hymenoptera : Apidae) foraging profits. Southwest Nat. 33: 219-227. Wunderlin, R. P. and J. E. Poppleton. 1977. The Florida species of Ilex (Aquifoliaceae) . Fla. Sci . 40: 7-21. Yamaguchi, K. and T. Shibamoto. 1980. Volatile constituents of the chestnut flower. J. Agric . Food Chem. 28: 82-84. Young, C. J., D. P. Turner, R. Killick-Kendrick, J. A. Rioux and A. J. Leary. 1980. Fructose in wild-caught Phlebotomus ariasi and the possible relevance of sugars taken by sand flies to the transmission of leishmaniasis. Trans. R. Soc . Trop . Med. Hyg. 74: 363-366. Yuval, B. 1992. The other habit: sugar feeding by mosquitoes. Bull. Soc. Vector Ecol . 17: 150-156. Zhu, Y., A. J. Keaster and K. 0. Gerhardt . 1993. Field observations on attractiveness of selected blooming plants to noctuid moths and electroantennogram responses of black cutworm (Lepidoptera : Noctuidae) moths to flower volatiles. Environ. Entomol . 22: 162-166.

PAGE 186

BIOGRAPHICAL SKETCH Robert Gordon Stewart was born in Winchester, Massachusetts, on June 7, 1949. He graduated from Reading Memorial High School in Reading, Massachusetts, in 1967 and began undergraduate study at Bowdoin College the same year. He received the A.B. degree with a major in biology in 1971. Following graduation Mr. Stewart entered Peace Corps service as a secondary science teacher in Zaire. Between 1971 and 1982 he taught in Zaire for a total of nine years. While in Zaire he also worked as a trainer of new Peace Corps science teachers for five training programs. From 1982 to 1986 Mr. Stewart taught biology at Our Lady of Nazareth Academy in Wakefield, Massachusetts. In 1986 he returned to Zaire under the joint auspices of the University of Lowell and the State University of New York at Stony Brook for one year of field research on the spatial and temporal distribution of Culicoides spp. in the Lomako Forest. Five more years of teaching secondary biology included one year at the EdCo Alternative High School of Boston, 173

PAGE 187

174 Massachusetts, and two years each at the Casablanca American School in Casablanca, Morocco, and the International School of Lusaka in Lusaka, Zambia. Mr. Stewart matriculated at the University of Florida in August 1992 to begin studies for attainment of the Doctor of Philosophy degree in entomology.

PAGE 188

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. Daniel L. Kline, Chair Assistant Professor of Entomology and Nematology 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 afid quality, as a dissertation for the degree of Doct0^^^'''^iiosophy . Ellis Greiner Professor of Veterinary Medicine 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. Donald Hall Professor of Entomology and Nematology 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. Heather McAuslane Assistant Professor of Entomology and Nematology

PAGE 189

This dissertation was submitted to the Graduate Faculty of the College of Agriculture and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August, 1996 Dean, College of Agriculture Dean, Graduate School