Citation
Mosquito and Sentinel Chicken Interactions with an Assessment of Experimental Cage Design and Flight Activity of Mosquitoes in Orange County, Florida: 2005-2006

Material Information

Title:
Mosquito and Sentinel Chicken Interactions with an Assessment of Experimental Cage Design and Flight Activity of Mosquitoes in Orange County, Florida: 2005-2006
Creator:
Kobylinski, Kevin C. ( Dissertant )
Rutledge-Connelly, Cynthia Roxanne ( Thesis advisor )
Breaud, Thomas ( Reviewer )
Day, Jon ( Reviewer )
Allan, Sandra ( Reviewer )
Kline, Dan ( Reviewer )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
2006
Language:
English

Subjects

Subjects / Keywords:
Antibodies ( jstor )
Arboviruses ( jstor )
Birds ( jstor )
Chickens ( jstor )
Encephalitis ( jstor )
Epidemics ( jstor )
Infections ( jstor )
Saint Louis encephalitis virus ( jstor )
Surveillance ( jstor )
West Nile virus ( jstor )
Dissertations, Academic -- UF -- Entomology and Nematology
Entomology and Nematology Thesis, Ph.D.
Orange County ( local )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( margct )

Notes

Abstract:
An experimental cage was created to attract and contain mosquitoes. Mosquitoes are funneled into the cage through a baffle. Mosquitoes are contained in the cage, in the exit trap, or they escape through the baffle slit. Trials were conducted in Gainesville, FL from May 26 through July 25, 2005 to determine mosquito escape rates from the experimental cage. Five colony-raised mosquito species were used: Culex nigripalpus Theobald, Culex quinquefasciatus Say, Anopheles quadrimaculatus Say, Aedes albopictus (Skuse), and Aedes aegypti (Linnaeus). Mosquito escape rates from the cage with the baffle slit open and two restrained chickens inside the cage were Cx. nigripalpus 77.7 ± 6.8%, Cx. quinquefasciatus 47.4 ± 10.9%, An. quadrimaculatus 65.8 ± 12.1%, Ae. albopictus 58.7 ± 4.6%, and Ae. aegypti 57.0 ± 5.5%. Chickens display a variety of defensive behaviors in response to host seeking mosquitoes. From July 26 through August 13, 2005, in Gainesville, Florida, adult chickens were exposed to known densities of mosquitoes. A restrained chicken and an unrestrained chicken were housed in two separate experimental cages and exposed to 300 Cx. nigripalpus or 300 Ae. albopictus. Blood-feeding success of Ae. albopictus varied significantly between the unrestrained chicken (6.1%) trials and the restrained chicken (22.2%) trials (Wilcoxon Rank Sum test, P < 0.025) . Groups of four chickens were exposed in the experimental cage once a week at two field sites, Tibet-Butler Preserve and Moss Park, in Orange County, FL from October 12, 2005 to July 5, 2006. All mosquitoes were aspirated from the cage the morning after exposure, collected from the exit trap, identified to species, pooled by species, and frozen at -80ºC. Twelve species of mosquitoes were captured in the experimental cage at Tibet-Butler Preserve and seven species at Moss Park. Coquillettida perturbans (Walker), Cx. nigripalpus, and Culex erraticus (Dyar and Knab) were frequently captured in the experimental cage at both sites. Mansonia titillans (Walker) was captured frequently at Tibet-Butler Preserve. Mosquitoes were more frequently captured in the cage than in the exit trap at Tibet-Butler Preserve (T-test, t = 14.86, P < 0.0001) and Moss Park (Wilcoxon Rank Sum test, Z = 2.80, P = 0.0052). Coquillettidia perturbans (Wilcoxon Rank Sum test, Z = 2.39, P = 0.0085), Cx. erraticus (Wilcoxon Rank Sum test, Z = 1.86, P = 0.0317), and Ma. titillans (Wilcoxon Rank Sum test, Z = 2.59, P = 0.0048) were more likely to fly upward and become captured in the exit trap when blood-fed. A rotator trap, baited with incandescent light and CO2, was operated on all nights that experimental chickens were exposed from November 7, 2005 through July 5, 2006 to determine periods of mosquito flight activity at night. Coquillettida perturbans (ANOVA, F = 4.74, P < 0.0001) and Cx. erraticus (Kruskal-Wallis, Chi-square = 25.2, P = 0.0007) exhibited crepuscular host-seeking patterns. Anopheles crucians Wiedemann (ANOVA, F = 4.14, P = 0.0003), Cx. nigripalpus (ANOVA, F=13.96, P < 0.0001), Cx. quinquefasciatus (Kruskal- Wallis, Chi-square = 22.4, P = 0.0022), and Culiseta melanura (Coquillett) (ANOVA, F = 6.55, P < 0.0001) exhibited nocturnal host-seeking patterns.
Thesis:
Thesis (Ph.D.)--University of Florida, 2006.
Bibliography:
Includes biographical references.
General Note:
Vita.
General Note:
Document formatted into pages; contains 166 p.
General Note:
Title from title page of document.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Kobylinski, Kevin C.. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
3/1/2007
Resource Identifier:
659561065 ( OCLC )

Downloads

This item has the following downloads:

kobylinski_k ( .pdf )

kobylinski_k_Page_157.txt

kobylinski_k_Page_008.txt

kobylinski_k_Page_076.txt

kobylinski_k_Page_017.txt

kobylinski_k_Page_025.txt

kobylinski_k_Page_112.txt

kobylinski_k_Page_138.txt

kobylinski_k_Page_063.txt

kobylinski_k_Page_113.txt

kobylinski_k_Page_083.txt

kobylinski_k_Page_148.txt

kobylinski_k_Page_102.txt

kobylinski_k_Page_140.txt

kobylinski_k_Page_077.txt

kobylinski_k_Page_133.txt

kobylinski_k_Page_110.txt

kobylinski_k_Page_016.txt

kobylinski_k_Page_014.txt

kobylinski_k_Page_109.txt

kobylinski_k_Page_027.txt

kobylinski_k_Page_015.txt

kobylinski_k_Page_106.txt

kobylinski_k_Page_093.txt

kobylinski_k_Page_136.txt

kobylinski_k_Page_018.txt

kobylinski_k_Page_135.txt

kobylinski_k_Page_004.txt

kobylinski_k_Page_010.txt

kobylinski_k_Page_069.txt

kobylinski_k_Page_043.txt

kobylinski_k_Page_145.txt

kobylinski_k_Page_073.txt

kobylinski_k_Page_061.txt

kobylinski_k_Page_036.txt

kobylinski_k_Page_022.txt

kobylinski_k_Page_056.txt

kobylinski_k_Page_123.txt

kobylinski_k_Page_118.txt

kobylinski_k_Page_129.txt

kobylinski_k_Page_057.txt

kobylinski_k_Page_067.txt

kobylinski_k_Page_105.txt

kobylinski_k_Page_150.txt

kobylinski_k_Page_044.txt

kobylinski_k_Page_122.txt

kobylinski_k_Page_128.txt

kobylinski_k_Page_101.txt

kobylinski_k_Page_108.txt

kobylinski_k_Page_049.txt

kobylinski_k_Page_037.txt

kobylinski_k_Page_091.txt

kobylinski_k_Page_107.txt

kobylinski_k_Page_130.txt

kobylinski_k_Page_064.txt

kobylinski_k_Page_120.txt

kobylinski_k_Page_116.txt

kobylinski_k_Page_131.txt

kobylinski_k_Page_088.txt

kobylinski_k_Page_162.txt

kobylinski_k_Page_029.txt

kobylinski_k_Page_047.txt

kobylinski_k_Page_072.txt

kobylinski_k_Page_098.txt

kobylinski_k_Page_070.txt

kobylinski_k_Page_121.txt

kobylinski_k_Page_065.txt

kobylinski_k_Page_001.txt

kobylinski_k_Page_079.txt

kobylinski_k_Page_117.txt

kobylinski_k_Page_007.txt

kobylinski_k_Page_040.txt

kobylinski_k_Page_137.txt

kobylinski_k_Page_163.txt

kobylinski_k_Page_160.txt

kobylinski_k_Page_126.txt

kobylinski_k_Page_096.txt

kobylinski_k_Page_165.txt

kobylinski_k_Page_114.txt

kobylinski_k_Page_033.txt

kobylinski_k_Page_034.txt

kobylinski_k_Page_124.txt

kobylinski_k_Page_087.txt

kobylinski_k_Page_144.txt

kobylinski_k_Page_154.txt

kobylinski_k_Page_142.txt

kobylinski_k_Page_050.txt

kobylinski_k_Page_090.txt

kobylinski_k_Page_038.txt

kobylinski_k_pdf.txt

kobylinski_k_Page_143.txt

kobylinski_k_Page_035.txt

kobylinski_k_Page_062.txt

kobylinski_k_Page_119.txt

kobylinski_k_Page_089.txt

kobylinski_k_Page_139.txt

kobylinski_k_Page_086.txt

kobylinski_k_Page_094.txt

kobylinski_k_Page_059.txt

kobylinski_k_Page_071.txt

kobylinski_k_Page_009.txt

kobylinski_k_Page_149.txt

kobylinski_k_Page_021.txt

kobylinski_k_Page_164.txt

kobylinski_k_Page_152.txt

kobylinski_k_Page_011.txt

kobylinski_k_Page_092.txt

kobylinski_k_Page_104.txt

kobylinski_k_Page_042.txt

kobylinski_k_Page_054.txt

kobylinski_k_Page_068.txt

kobylinski_k_Page_161.txt

kobylinski_k_Page_159.txt

kobylinski_k_Page_013.txt

kobylinski_k_Page_156.txt

kobylinski_k_Page_031.txt

kobylinski_k_Page_002.txt

kobylinski_k_Page_005.txt

kobylinski_k_Page_078.txt

kobylinski_k_Page_003.txt

kobylinski_k_Page_074.txt

kobylinski_k_Page_085.txt

kobylinski_k_Page_155.txt

kobylinski_k_Page_127.txt

kobylinski_k_Page_115.txt

kobylinski_k_Page_100.txt

kobylinski_k_Page_075.txt

kobylinski_k_Page_099.txt

kobylinski_k_Page_024.txt

kobylinski_k_Page_125.txt

kobylinski_k_Page_046.txt

kobylinski_k_Page_012.txt

kobylinski_k_Page_084.txt

kobylinski_k_Page_151.txt

kobylinski_k_Page_051.txt

kobylinski_k_Page_041.txt

kobylinski_k_Page_060.txt

kobylinski_k_Page_141.txt

kobylinski_k_Page_023.txt

kobylinski_k_Page_097.txt

kobylinski_k_Page_028.txt

kobylinski_k_Page_095.txt

kobylinski_k_Page_052.txt

kobylinski_k_Page_147.txt

kobylinski_k_Page_026.txt

kobylinski_k_Page_055.txt

kobylinski_k_Page_132.txt

kobylinski_k_Page_058.txt

kobylinski_k_Page_030.txt

kobylinski_k_Page_039.txt

kobylinski_k_Page_081.txt

kobylinski_k_Page_066.txt

kobylinski_k_Page_032.txt

kobylinski_k_Page_019.txt

kobylinski_k_Page_048.txt

kobylinski_k_Page_146.txt

kobylinski_k_Page_080.txt

kobylinski_k_Page_053.txt

kobylinski_k_Page_006.txt

kobylinski_k_Page_020.txt

kobylinski_k_Page_103.txt

kobylinski_k_Page_166.txt

kobylinski_k_Page_153.txt

kobylinski_k_Page_111.txt

kobylinski_k_Page_134.txt

kobylinski_k_Page_045.txt

kobylinski_k_Page_158.txt

kobylinski_k_Page_082.txt


Full Text












MOSQUITO AND SENTINEL CHICKEN INTERACTIONS WITH AN
ASSESSMENT OF EXPERIMENTAL CAGE DESIGN AND FLIGHT
ACTIVITY OF MOSQUITOES IN ORANGE COUNTY, FLORIDA: 2005-2006















By

KEVIN C. KOBYLINSKI


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2006

































Copyright 2006

by

Kevin C. Kobylinski










ACKNOWLEDGMENTS

I thank Robyn Raban for her countless efforts, suggestions, and help. I thank Sean

McCann, Erin Vrzal, Aissa Doumboya, Melissa Doyle, Aaron Lloyd, Joyce Urban, Haze Brown,

and David Hoel for their support, help, and suggestions concerning rearing and experiments in

Gainesville. I thank Gregg Ross, Amador Rodriguez, Columbus Holland, Terry Hughes, Sue

Durand, Armond Cross, and the Orange County Mosquito Control Division Staff for their

support, help, and suggestions concerning field research. I thank Dr. Don Shroyer, Dr. Jerry

Hogsette, Dr. Phil Koehler, Frank Wessels, Leslie Rios, and the Urban Entomology Laboratory

for their thoughtful suggestions and support. I thank Dr. Cynthia Lord, Dr. Ramon Littell,

Robyn Raban, and Alejandro Arevalo for their help with statistical analysis. I thank Dr. Sandra

Allan and Dr. Dan Kline for their support, help, suggestions, and use of laboratory supplies and

space for work done in Gainesville. I thank Dr. Cynthia Rutledge-Connelly, Dr. Thomas

Breaud, Dr. Jon Day, Dr. Sandra Allan, and Dr. Dan Kline for all their support, help, and

suggestions during the experiments and writing process. This research was funded in part by the

National Institutes of Health Grant AI042164, "Modeling and Empirical Studies of Arboviruses

in Florida." This research was funded in part by the Mosquito Control Division, Health and

Family Services Department, Orange County Florida.









TABLE OF CONTENTS


page

A CK N O W LED G M EN TS ................................................................. ........... ............. 3

T A B L E O F C O N TEN T S........................................................ ...............................................4

L IST O F T A B L E S ..................................................................................................... . 7

L IST O F FIG U R E S ............................................................................... 9

ABSTRAC T ................................................... ............... 10

CHAPTER

1 IN TR O D U C T IO N ................................................................................ 12

Arboviruses in Florida ..............................................................................13
St. Louis Encephalitis V irus ............................................................... ..............14
W e st N ile V iru s ......................................................................................................... 1 5
Eastern Equine Encephalitis V irus .......................................................... ...... 16
Western Equine Encephalitis Virus: Highlands J Virus ..............................................17
Culex nigripalpus and Virus Transmission in Florida................................. ...............18
St. Louis Encephalitis Virus ..................................... ......... .. ...... ................. 18
W est N ile V irus ..................................................... ............... .. ...... 19
Eastern Equine Encephalitis Virus ....................................... .......... ............20
Other Potential Vectors of Arboviruses in Florida ............... .................. 20
Potential Vectors of St. Louis Encephalitis Virus ....................................................21
Potential V sectors of W est N ile virus ..................................................................... ... 22
Potential Vectors of Eastern Equine Encephalitis Virus ..............................................23
Surveillance of A rboviruses .......................................................................................25
H u m an Su rveillance ....................................................................................................2 6
W weather P pattern s ................................. .................. ... ......................... ........................ 2 8
V sector Su rveillan ce ......................................................................................... 2 8
M monitoring of M mosquito Abundance..................................................... ...... ......... 29
P arity A nalysis................. ................................... 30
Minimum Infection Rates in Mosquitoes ............................................................... 30
Non-Human Vertebrate Surveillance .................................................32
Live, W ild V ertebrate Surveillance ............................................................ .....32
W ild bird surveillance ............................................................................................. 33
W ild m am m al surveillance.......... ..... .................................... ............ .............. 34
W ild reptile surveillance ........................................................ 36
D ead, W ild B ird Surveillance.................................................... ............... 36
D om estic A nim al Surveillance................................................ ............................ 38




4









Sentinel A nim al Surveillance .............................................. ....... ........................ 41
Sentinel m amm als .................................... ... .. .......... ....... .... 42
S en tin el b ird s ................................................................................ 4 2

2 C A G E E SC A PE TR IA L S............................................................. .....................................50

In tro d u ctio n ................... ...................5...................0..........
M materials and M methods ................................... .. ...... ..... .. ............51
M mosquito R hearing .............. .......................... .................... .......... 52
M a rk in g ...................................................................................................................... 5 3
C h ick e n s ................................................................5 3
E xperim mental C age D design ..................................................................... ..... ...............54
A sp irato r ............................................................................... 5 5
E sc a p e R a te s .......................................................................................................5 6
E n try R ate s ................................................................5 7
R e su lts ............... ... ................. ............................................................................................... 5 7
E scape T rial R results ................................................................57
C age E ntry R results ................................................................58
D iscu ssio n ................... ...................5...................9..........
C onclu sions..... ..........................................................62

3 SENTINEL CHICKEN DEFENSIVE BEHAVIOR ....... .......................72

In tro d u c tio n .............. ...... ......... ....................................................................................... 7 2
M materials an d M eth o d s ...................................................................................................... 7 3
Mosquito Rearing ............... ......... ......... ...............73
Mark-Release-Recapture Studies ................................................75
C h ick e n s .............................................................7 6
H o st D efen siv e B eh av ior............................................................................................ 76
Effect of Host Defensive Behavior................. ...... ......... .....................77
R e su lts .............. ..... ............ ....................................... ................................. ............ 7 8
D iscu ssion ........... .... ........... ....................................... ............................79
C onclu sions..... ..........................................................84

4 DESCRIPTION OF FIELD SITES: ARBOVIRUSES/VECTORS/HOSTS ........................88

In tro d u c tio n .............. ..... .......... ....................................................................................... 8 8
Site D description s ......... ...... ............ ...................................... ............................89
Tibet-Butler Preserve ........................................................................... ......... ..................89
M o ss P ark ......... ...... ............ ..................................... ............................9 1
A rboviruses ......... ............................................................................ ...... 92
Tibet-Butler Preserve ................. ........ .............. ............. 92
M oss Park ..................................................................................................... ............. 93
V ectors of A rboviruses ................................................................................................. ........ 94
Avian Hosts of Arboviruses ................................. ........................... ...........95
D iscu ssio n ................... ...................9...................8..........










5 EXPERIMENTAL ARBOVIRAL SURVEILLANCE IN ORANGE COUNTY, FL.........105

Introduction ................ .................. .............. ...............105
M materials an d M eth od s ............................................................................... ..................... 10 5
Sentinel Chickens ......... ............................................................ 105
M o sq u ito e s .............................................................................10 7
R results ........... ........ ..... ........... ........ ....................... ............... 108
D iscu ssio n ......... ...... ................................................. ............................10 8
C o n clu sio n ......... ..... ................................................ ...........................1 10

6 C A G E A N A L Y SIS .............. ..... ............ ...............................................................114

Introduction ................ .................. .............. ...............114
Materials and Methods ............... ......... ................. 115
Results ........... ................. ............................. ............... 116
Tibet-Butler Preserve ......................................................................... ........ ........ .... ........ 116
M oss Park .......... .... ..... .... ...... ................ ......... 117
D iscu ssio n ......... ...... ................................................. ............................1 18
C onclu sions..... .........................................................12 1

7 ROTATOR STUDIES .................. ......... ..................128

Introduction ................ .................. .............. ...............128
M materials and M methods ...............................................................129
Results ........... ................. ............................. ............... 130
Tibet-Butler Preserve ......................................................................... ........ ........ .... ........ 130
M oss Park ........... ......... ..... ........ ...... ..................... ........ 131
D iscu ssio n ......... ...... ................................................. ............................13 2
C onclu sions..... .........................................................134

8 C O N C L U S IO N S ............................................................................................................. 14 6

REFEREN CES CITED ................................................ .. 148

BIOGRAPHICAL SKETCH .................................... .............................. 166

















6









LIST OF TABLES


Table page

2-1 L arval feeding regime en ............................................................................. .....................63

3-1 Effect of host defensive behavior on recovery, blood feeding success, and location
re c o v ere d ..........................................................................8 5

3-2 Effect of Cx. nigripalpus density on recovery, blood feeding success, and location
recovered when one or four chickens were present .................................. ............... 85

3-3 Behaviors displayed by chickens in response to mosquitoes.................. ............... 86

3-4 Defensive behaviors displayed by chickens in the presence of mosquitoes with
com m entry ...................... .. .. ......... .. .. ....................................................87

4-1 Seroconversions for EEEV at sites operated by OCMCD from 1978-2004 ................100

4-2 Seroconversions for SLEV at sites operated by OCMCD from 1978-2004 ....................100

4-3 Seroconversions for WNV at sites operated by OCMCD from 1978-2004.......................101

5 -1 R ain (m m ) ...................................... ..................................... ................. 1 1 1

5-2 Relative humidity (% ) .................................. .. .............. .. .. ............ 111

5-3 Tem perature (oC) ............... ................. ........... ................ ............. .. 111

5-4 W ind speed (km /h) .................................................... ............ .. ............ 112

5-5 B arom etric pressure (m m -H g) ..................................................................................... 112

6-1 M osquitoes captured in experimental cage at TBP...........................................................122

6-2 Other Diptera captured in experimental cage at TBP ................................ ............... 122

6-3 Mosquitoes captured in experimental cage at MP... ................ ...............123

6-4 Other Diptera captured in experimental cage at MP ...................................123

6-5 Mean number of mosquitoes caught in the exit trap and the cage at TBP .......................124

6-6 Percent of blood-fed mosquitoes caught in the exit trap and the cage at TBP................... 124

6-7 Mean number of blood-fed and non blood-fed mosquitoes by location at TBP................ 125

6-8 Mean number of mosquitoes caught in the exit trap and the cage at MP ..........................126









6-9 Percent of blood-fed mosquitoes caught in the exit trap and cage at MP ........................126

6-10 Mean number of blood-fed and non blood-fed mosquitoes by location at MP ..................127

7-1 Rotator time schedule ........ .. ................................ .. .. ....... ................. 135

7-2 Tibet-Butler Preserve rotator catch totals 11/7/05 to 7/4/06 .................. ..................136

7-3 M oss Park rotator catch totals 11/8/05 to 7/5/06..................................... ............... 137









LIST OF FIGURES


Figure page

2-1 Original sentinel chicken cage used by OCM CD ..................................... .................63

2-2 Experimental sentinel chicken cage used for research........................ ................... 64

2-3 Line draw ing of experim ental cage......... ................. ................................. ............... 65

2-4 Line drawing of experimental inner cage.................. .............. .................. 66

2-5 Sentinel chicken cage exit trap......... .................. ........ ............................... ............... 67

2-6 A ssem bled aspirator ......... ................................ ............................... 68

2-7 Night escape trials ........................ ....... .. ... .. .. .................. 69

2-8 D ay escape trials ....................................................... .................. 70

2-9 Percent mosquitoes recovered from experimental cage and exit trap..............................71

4-1 Orange County water bodies with sentinel sites, TBP and MP ......................................102

4 -2 A erial v iew of T B P .............. ................................................................................... 10 3

4-3 A erial view of M P ........ ......................... .......... .. .... .. ...... .............. 104

5-1 Total number of Culex nigripalpus captured in light traps in Orange County by month ..113

7-1 Anopheles crucians captured by rotator trap at TBP and MP .......................... .........138

7-2 Coquillettidia perturbans captured by rotator trap at TBP and MP .............................. 139

7-3 Culex erraticus captured by rotator trap at TBP and MP .............................................140

7-4 Culex nigripalpus captured by rotator trap at TBP and MP ........................ ............141

7-5 Culex quinquefasciatus captured by rotator trap at TBP....................... ...............142

7-6 Culiseta melanura captured by rotator trap at TBP ........................................ ................143

7-7 Mansonia titillans captured by rotator trap at TBP..........................................144

7-8 All species captured by rotator trap at TBP and MP .............. .... .................145














Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

MOSQUITO AND SENTINEL CHICKEN INTERACTIONS WITH AN
ASSESSMENT OF EXPERIMENTAL CAGE DESIGN AND FLIGHT
ACTIVITY OF MOSQUITOES IN ORANGE COUNTY, FLORIDA: 2005-2006

By

Kevin C. Kobylinski

December 2006

Chair: Cynthia Roxanne Rutledge-Connelly
Major Department: Entomology and Nematology

An experimental cage was created to attract and contain mosquitoes. Mosquitoes are

funneled into the cage through a baffle. Mosquitoes are contained in the cage, in the exit trap, or

they escape through the baffle slit. Trials were conducted in Gainesville, FL from May 26

through July 25, 2005 to determine mosquito escape rates from the experimental cage. Five

colony-raised mosquito species were used: Culex nigripalpus Theobald, Culex quinquefasciatus

Say, Anopheles quadrimaculatus Say, Aedes albopictus (Skuse), and Aedes aegypti (Linnaeus).

Mosquito escape rates from the cage with the baffle slit open and two restrained chickens inside

the cage were Cx. nigripalpus 77.7 6.8%, Cx. quinquefasciatus 47.4 10.9%, An.

quadrimaculatus 65.8 12.1%, Ae. albopictus 58.7 4.6%, and Ae. aegypti 57.0 5.5%.

Chickens display a variety of defensive behaviors in response to host seeking mosquitoes.

From July 26 through August 13, 2005, in Gainesville, Florida, adult chickens were exposed to

known densities of mosquitoes. A restrained chicken and an unrestrained chicken were housed

in two separate experimental cages and exposed to 300 Cx. nigripalpus or 300 Ae. albopictus.









Blood-feeding success ofAe. albopictus varied significantly between the unrestrained chicken

(6.1%) trials and the restrained chicken (22.2%) trials (Wilcoxon Rank Sum test, P < 0.025).

Groups of four chickens were exposed in the experimental cage once a week at two field

sites, Tibet-Butler Preserve and Moss Park, in Orange County, FL from October 12, 2005 to July

5, 2006. All mosquitoes were aspirated from the cage the morning after exposure, collected from

the exit trap, identified to species, pooled by species, and frozen at -800C. Twelve species of

mosquitoes were captured in the experimental cage at Tibet-Butler Preserve and seven species at

Moss Park. Coquillettidaperturbans (Walker), Cx. nigripalpus, and Culex erraticus (Dyar and

Knab) were frequently captured in the experimental cage at both sites. Mansonia titillans

(Walker) was captured frequently at Tibet-Butler Preserve. Mosquitoes were more frequently

captured in the cage than in the exit trap at Tibet-Butler Preserve (T-test, t = 14.86, P < 0.0001)

and Moss Park (Wilcoxon Rank Sum test, Z = 2.80, P = 0.0052). Coquillettidiaperturbans

(Wilcoxon Rank Sum test, Z = 2.39, P = 0.0085), Cx. erraticus (Wilcoxon Rank Sum test,

Z = 1.86, P = 0.0317), and Ma. titillans (Wilcoxon Rank Sum test, Z = 2.59, P = 0.0048) were

more likely to fly upward and become captured in the exit trap when blood-fed.

A rotator trap, baited with incandescent light and C02, was operated on all nights that

experimental chickens were exposed from November 7, 2005 through July 5, 2006 to determine

periods of mosquito flight activity at night. Coquillettidaperturbans (ANOVA, F = 4.74,

P < 0.0001) and Cx. erraticus (Kruskal-Wallis, Chi-square = 25.2, P = 0.0007) exhibited

crepuscular host-seeking patterns. Anopheles crucians Wiedemann (ANOVA, F = 4.14,

P = 0.0003), Cx. nigripalpus (ANOVA, F=13.96, P < 0.0001), Cx. quinquefasciatus (Kruskal-

Wallis, Chi-square = 22.4, P = 0.0022), and Culiseta melanura (Coquillett) (ANOVA, F = 6.55,

P < 0.0001) exhibited nocturnal host-seeking patterns.









CHAPTER 1
INTRODUCTION

Sentinel chickens are maintained at field sites in Orange County for the detection of

arboviruses. Blood samples are taken weekly from sentinel chickens and analyzed for arboviral

antibodies. Exit traps were attached to the county sentinel chicken cages to assess the species

diversity and abundance of mosquitoes that were attracted to the chickens.

An experimental cage, similar in design to the sentinel chicken cages used by Orange

County Mosquito Control Division (OCMCD), was created to draw mosquitoes into the cage and

contain the mosquitoes. Mosquitoes that approach the sentinel chickens are funneled upward

into the cage by a baffle that narrows to an entry slit. The mosquitoes are contained in the cage

or fly upward into the exit trap. Pre-field trials were conducted to determine the percent of

mosquitoes that could be recovered within the experimental cage and the percent of mosquitoes

that could escape from the experimental cage.

Vertebrates display a variety of defensive behaviors in response to host seeking

mosquitoes (Anderson and Brust 1995, Cully et al. 1991, Waage and Nondo 1982, Webber and

Edman 1972). Adult chicken defensive behavior and its influence on mosquito recovery rates

and mosquito blood-feeding success was assessed. Predetermined numbers of mosquitoes were

released into two cages with one restrained in one cage, and one unrestrained chicken in the

other cage. The effect of mosquito density and chicken defensive behavior on mosquito

recovery rates and mosquito blood-feeding success was assessed. Predetermined numbers of

mosquitoes were released into two cages with one unrestrained chicken in one cage, and four

unrestrained chickens in the other cage. The location in the cage that mosquitoes were recovered

from and their blood-fed status was recorded. The effect of chicken defensive behavior on









mosquito blood-feeding success and location of mosquito recovery in the experimental cage was

assessed.

Experimental chickens were placed at two field sites in Orange County from October 2005

to July 2006 for the detection of arbovirus antibodies. Mosquitoes were collected in the

experimental cage in attempt to determine which mosquitoes vectored arboviruses to sentinel

chickens in Orange County, FL. The diversity and abundance of mosquito species attracted to

chickens was documented. The location of mosquito recovery within the experimental cage was

documented to determine how representative the exit trap collections are of the mosquitoes that

approach sentinel chickens. Blood-feeding success was observed to determine how often blood-

fed mosquitoes fly upward into the exit trap.

The host-seeking times of nocturnally active mosquitoes differ among species. Knowledge

of the host-seeking times of nocturnally active mosquitoes is important to determine the most

appropriate time to conduct night-time adulticide spraying. A rotator trap was operated at the

field sites to determine periods of peak host-seeking activity.

Arboviruses in Florida

The diverse habitats found in Florida support a wide range of vertebrate hosts, mosquito

vectors, and arboviruses. Humans find the Florida climate attractive and millions of susceptible,

elderly humans have made Florida their home (Bond et al. 1963). The elderly tend to suffer

greater complications to arboviruses that occur in Florida such as St. Louis encephalitis (family

Flaviviridae, genus Flavivirus, SLEV) and West Nile virus (family Flaviviridae, genus

Flavivirus, WNV) (Bond et al. 1963, Petersen and Marfin 2002). As the population of north-

central Florida (greater Orlando area) continues to grow, human habitation pushes closer to

swamp foci, increasing the potential of human infection with Eastern equine encephalitis virus

(family Togaviridae, genus Alphavirus, EEEV) (Edman et al. 1972b, Morris 1992).









St. Louis Encephalitis Virus

St. Louis encephalitis virus was first documented in Florida from a thirty year-old man in

Miami in 1952 (Sanders et al. 1953). Since then there have been five major epidemics of SLEV

in three distinct regions of Florida (Bond 1969, Nelson et al. 1983, Day and Stark 2000). The

outbreaks of 1959, 1961, and 1962 were centered in the Tampa Bay Area of Florida (Pinellas,

Hillsborough, Manatee, and Sarasota counties) (Bond 1969). There were 68 clinical human

cases of SLEV and five deaths in 1959, 25 cases and seven deaths in 1961, and 231 cases and 43

deaths in 1962 (CDC 2006a, Bond 1969). As with most outbreaks of SLEV, all deaths were

persons over the age of 45 years (Bond 1969). The primary epidemic vector during these

outbreaks and the 1977 and 1990 outbreaks in peninsular Florida, was Culex nigripalpus

Theobald (Dow et al. 1964, Shroyer 1991). In direct response to the outbreak of SLEV in 1962

the Encephalitis Research Center (ERC) of the Florida State Board of Health in Tampa was

established (Bond 1969).

From 1963 to 1976, only two human cases of SLEV were reported in Florida and thus

surveillance and research interests in SLEV declined (Nelson et al. 1983). The outbreak of

SLEV in 1977 lacked an urban focus and covered a broad region (20 counties) of central and

south Florida (Nelson et al. 1983). This outbreak differed from the previous outbreaks in Tampa

Bay because there was a high attack rate of SLEV in males between 15 and 24 years of age, most

likely due to outdoor occupations (Nelson et al. 1983). There were 110 presumptive and

confirmed human cases of SLEV in Florida in 1977 with 11 cases from Orange County (Nelson

et al. 1983). Interest in surveillance of arboviruses was heightened after this outbreak and the

establishment of the Florida Sentinel Chicken Surveillance Program began in 1978 (Day 1989).

Focal outbreaks and sporadic cases of SLEV in humans occurred from 1979 to 1984 (CDC

2006a). In 1990, there was an epidemic of SLEV in humans that spanned 28 counties in central









and south Florida with 226 clinical human cases and 11 deaths. In Orange County there were 28

human cases of SLEV (Day and Stark 2000). With the exception of two sporadic cases in 1969,

(FDOH 2006a) the 1977 and 1990 outbreaks were the only instances of reported human infection

of SLEV in Orange County. Focal and sporadic cases of SLEV in humans occurred from 1991

to 2002 throughout Florida (CDC 2006a).

Human infection with SLEV is typically a relatively mild illness characterized by fever

and headache, followed by complete recovery (Hayes 2000). Persons over the age of 45 may

experience severe central nervous system involvement of meningitis or encephalitis leading to

death (Bond 1969, Hayes 2000). Inapparent (asymptomatic) human SLEV infection rates vary

from 0.2 to 5.2% (Hayes 2000). Human SLEV infections in Florida typically occur between

August and December (Bond 1969, Nelson et al. 1983).

West Nile Virus

West Nile virus was first documented in Florida in 2001 (Blackmore et al. 2003). In 2001

WNV activity was dispersed throughout the entire state with 12 human cases ranging from the

panhandle to the Keys (Blackmore et al. 2003). In 2002 WNV activity was dispersed throughout

the entire state with 35 human cases, two deaths and three additional cases attributed to blood

transfusions and organ transplants. The only human case of WNV infection in Orange County

occurred in 2002 with an onset date of August 12. In 2003 WNV activity was dispersed

throughout the state with a heavy concentration of cases (61 cases in 10 counties) in the

northwest portion of the panhandle. There were 93 human cases of WNV and six deaths in

2003. In 2004 there were 39 human cases of WNV, two deaths and three additional cases that

were suspected to have been acquired out of state. In 2005 there were 21 human WNV cases and

one death. As of July, 2006, no human infections of WNV have been reported in Florida for

2006 (FDOH 2006b).









During the initial two years of WNV activity there were human cases throughout Florida

with no discernable pattern of transmission or centralized outbreak foci. In 2003 there were

human cases throughout Florida but 66% (61/93) of human cases were from the northwest region

of the Florida Panhandle. In 2004 and 2005 there were urban foci in Dade and Broward

Counties (24 of 39 human cases) and in Pinellas County (18 of 21 human cases), respectively

(FDOH 2006b).

In humans, WNV is capable of causing either a febrile illness known as West Nile fever

(Watson et al. 2004b) or a neuroinvasive disease such as meningitis, encephalitis, or acute

flaccid paralysis leading to death (Petersen and Marfin 2002). Symptoms of West Nile fever

include fever, malaise, anorexia, nausea, vomiting, eye pain, headache, myalgia, rash, and

lymphadenopathy (Petersen and Marfin 2002) with symptoms lasting 3 to 6 days or sometimes

longer (Watson et al. 2004b). Symptoms of neuroinvasive WNV disease include fever, muscle

weakness, gastrointestinal symptoms, headache, and changes in mental status (Petersen and

Martin 2002). A seroprevalence study in New York City during 1999 determined that

approximately 20% of WNV infected persons developed febrile illness and less than 1%

developed neuroinvasive disease (Mostashari et al. 2001). Mostashari et al. (2001) estimated

that for every one meningoencephalitis case there were 140 symptomatic and mildly

symptomatic cases. In a large scale screening of blood donations Busch et al. (2006) found the

incidence of WNV neuroinvasive disease to inapparent infection in humans to be 1 in 256.

Eastern Equine Encephalitis Virus

Transmission of Eastern equine encephalitis virus has only occurred in focal or sporadic

outbreaks in Florida (Bigler et al. 1976, Day and Stark 1996b, Morris 1992). Most cases of

EEEV in humans are confined to the panhandle and north-central Florida (Day and Stark 1996b).

The first documented human case of EEEV in Florida was in 1952, but there were unconfirmed









reports of human cases in Florida prior to 1952 (Bigler et al. 1976). To date there have been 68

reported human cases of EEEV in Florida (CDC 2006a). Four human cases of EEEV have

occurred in Orange County with one case in 1973, 1990, 1995, and 2003 (FDOH 2006a).

Eastern equine encephalitis virus is a severely debilitating disease, causing death in

approximately 50 to 90% of symptomatic human victims, with most deaths being adults (Scott

and Weaver 1989, Villari et al. 1995). Symptoms of encephalitic EEEV infection are high fever

(390 to 410C), irritability, restlessness, drowsiness, muscle tremors, neck rigidity, anorexia,

vomiting, diarrhea, headache, cyanosis, convulsions, and coma (Morris 1992). Villari et al.

(1995) found adults that survive encephalitic EEEV infection often have full recovery but

children less than fifteen often suffer lifelong debilitations caused by persistent and severe

neurologic disease. Human cases can occur throughout the year in Florida but generally occur

between May and August (Bigler et al. 1976). Most human cases of EEEV are rural in

distribution and generally associated with wooded areas adjacent to swamps and marshes (Morris

1992).

Western Equine Encephalitis Virus: Highlands J Virus

Western equine encephalitis virus (family Togaviridae, genus Alphavirus, WEEV) has

been isolated from Culiseta melanura (Coquillett), birds, and horses in Florida (Bigler et al.

1976, Henderson et al. 1962, Karabatsos et al. 1988). It was later determined that all isolates of

WEEV in the Eastern United States were Highlands J (HJ) virus, an antigenically distinct virus

that is part of the WEEV complex (Calisher et al. 1980, Karabatsos et al. 1988). Highlands J

virus has no known public or veterinary health importance (Karabatsos et al. 1988). The primary

vector of HJ virus in the eastern United States is Cs. melanura (Karabatsos et al. 1988, Morris

1988). Highlands J virus can complicate avian serosurvey results as it cross reacts with EEEV in

the Hemaggluttination Inhibition (HI) test. It was not until 2004 that HJ was included in the









laboratory testing of EEEV positive sentinel chickens in Florida. Plaque reduction neutralization

tests (PRNT) are now used to differentiate between HJ and EEEV (L. M. Stark, personal

communication).

Culex nigripalpus and Virus Transmission in Florida

St. Louis Encephalitis Virus

Culex nigripalpus is the most abundant Culex (Culex) species in peninsular Florida

(Edman 1974). An outbreak of SLEV in the Tampa Bay Area of Florida in 1962 resulted in 22

of 23 isolations of SLEV from wild Cx. nigripalpus, establishing this species as the primary

epidemic vector of SLEV in south Florida (Dow et al. 1964). Provost (1969) observed that Cx.

nigripalpus activity was directly influenced by humidity with nights of 90% humidity or greater

triggering the greatest amount of flight activity. He hypothesized that rainless periods during the

Cx. nigripalpus breeding season punctuated by intermittent rainfall leads to SLEV transmission

(Provost 1969).

Day and Curtis (1999) defined four phases of SLEV transmission in subtropical Florida:

maintenance, amplification, early season transmission, and late season transmission phases. The

maintenance phase occurs from December through March and SLEV transmission is maintained

at low levels between avian hosts and mosquitoes. The amplification phase occurs from April

through June and coincides with the avian nesting season. Early season transmission of SLEV

occurs from July through September at which time the first human cases of SLEV may be

reported. Late season transmission of SLEV occurs from October through December when more

human cases of SLEV may be reported. St. Louis encephalitis virus transmission to humans

typically ceases by the end of December (Day and Curtis 1999).

Winter and springtime (maintenance and amplification phases) drought restricts Cx.

nigripalpus to wooded hammocks where many avian species nest (Day 2001). This brings Cx.









nigripalpus into contact with adult and nestling birds (Day 2001). Culex nigripalpus had higher

blood feeding success on nestling than adult ciconiiform birds most likely due to the lack of

plumage on nestling birds (Kale et al. 1972). Nestling birds have been shown to circulate a

higher viremia of SLEV for a longer duration than adult birds and display fewer defensive

behaviors making them more likely to infect mosquitoes (Scott et al. 1988, Scott et al. 1990a).

Culex nigripalpus shifts host preference from avian hosts during January through May to

mammalian hosts during June through November or December (Edman 1974, Edman and Taylor

1968). The mammalian feeding preference shift by Cx. nigripalpus coincides with the

occurrence of epidemic activity of SLEV in humans during early and late season transmission

(Day and Curtis 1999, Edman 1974). The Cx. nigripalpus host preference shift from avian to

mammal is directly linked with rainfall (Edman 1974).

Day and Curtis (1989) showed that Cx. nigripalpus blood-feeding and oviposition (Day et

al. 1990) behavior was regulated by heavy rainfall (>30mm). When heavy rainfalls occur every

10 to 14 days, parous Cx. nigripalpus were forced to delay oviposition. This allowed time for

the extrinsic incubation of SLEV virus resulting in infected salivary glands of the mosquito. The

next heavy rainfall triggered the infective Cx. nigripalpus to oviposit and then blood feed,

potentially infecting further hosts (Day et al. 1990). This amplifying cycle continues until

minimum infection rates of SLEV in Cx. nigripalpus are >1:1000 at which point the likelihood

of incidental human infection can reach epidemic levels (Day and Stark 2000). Once the drought

ends, the repeated rainfall raises humidity levels and Cx. nigripalpus disperse from the

hammocks furthering the potential of SLEV transmission to humans (Shaman et al. 2002).

West Nile Virus

Culex nigripalpus has been incriminated in transmission of WNV in Florida and was

suspected of being the primary epidemic vector in 2001 (Rutledge et al. 2003). Culex









nigripalpus was capable of transmitting WNV in the laboratory (Turell et al. 2005). Two WNV

infected pools of Cx. nigripalpus were found in Jefferson County, FL in 2001 (Godsey et al.

2005a). The same mechanisms that make Cx. nigripalpus an efficient vector of SLEV were

thought to make it an efficient vector of WNV (Shaman et al. 2005).

Eastern Equine Encephalitis Virus

Culex nigripalpus was suggested as a possible bridge vector for EEEV in Florida by Nayar

(1982). Eastern equine encephalitis virus was isolated from sixteen pools of Cx. nigripalpus in

the Tampa Bay area from 1962 to 1967 (Taylor et al. 1969) and from 94 pools from 1962 to

1970 (Wellings et al. 1972). Three often pools of Cx. nigripalpus were positive for EEEV

during a focal outbreak of EEEV in emus in Volusia County, Florida in 1994 (Day and Stark

1996a). Eastern equine encephalomyelitis virus has also been isolated from Cx. nigripalpus in

Trinidad (Downs et al. 1959). Sudia and Chamberlain (1964) found Cx. nigripalpus to be a poor

vector of EEEV under laboratory conditions. Culex nigripalpus is often collected from swamp

foci with known EEEV activity in Orange County (OCMCD data). The abundance of Cx.

nigripalpus in swamp habitat, generalist feeding habits (Edman 1974) and rainfall mediated host-

feeding patterns (Day and Curtis 1989) increase its potential as a vector of EEEV in Florida.

Other Potential Vectors of Arboviruses in Florida

Chamberlain (1958) used five criteria for the incrimination of potential vectors of

arboviruses which were simplified and restated by Vaidyanathan et al. (1997): history of

frequent virus isolation; flight range overlapping host habitat; host seeking and blood feeding

coinciding with disease incidence; varied host choice; and laboratory demonstration of vector

competence. Many of the mosquito species in Florida suspected as potential arbovirus vectors

have not fulfilled all five of the requirements to be incriminated as vectors of arboviruses.









Potential Vectors of St. Louis Encephalitis Virus

Culexpipiens quinquefasciatus Say and Culex salinarius Coquillett are the second most

abundant species of Culex (Culex) in Florida (Edman 1974). Both reach peak abundance in late

winter and spring in south Florida (O'Meara and Evans 1983, Provost 1969). Culex

quinquefasciatus is a generalist feeder that frequently feeds on man and domestic animals such

as cats, dogs, and chickens (Edman 1974). Culex quinquefasciatus was a competent vector of

SLEV in the laboratory (Chamberlain et al. 1959). Culex quinquefasciatus is the primary vector

of SLEV in the southeastern United States (Tsai and Mitchell 1989) and in the absence of Cx.

nigripalpus, it is most likely the vector of SLEV in northern Florida. Culex quinquefasciatus

was the suspected vector of the 1980 focal outbreak of SLEV in Fort Walton Beach, Florida

(McCaig et al. 1994). Culex salinarius is a generalist feeder that was less associated with man

and domestic animals than Cx. quinquefasciatus (Edman 1974). Culex salinarius, a highly

competent vector of SLEV in the laboratory, (Chamberlain 1958, Chamberlain et al. 1959) may

be an enzootic vector of SLEV in Florida (Zyzak et al. 2002).

Culex restuans Theobald is a competent vector of SLEV in the laboratory (Chamberlain et

al. 1959). The host preference of Cx. restuans appears to vary by locality but was thought to be

primarily ornithophilic in the United States (Edman 1974, Mitchell et al. 1980). Culex restuans is

most abundant in Florida during late fall, winter and spring which makes it most likely an

enzootic vector of SLEV during the maintenance and amplification phases of SLEV transmission

(Edman 1974). When summertime temperatures drop below 200C Cx. restuans may become

active and aging mosquitoes may vector SLEV (Reiter 1988).

There have been no documented isolations of SLEV from Cx. quinquefasciatus, Cx.

restuans, or Cx. salinarius in Florida. This limits their incrimination as vectors of SLEV in









Florida by not fulfilling the first criteria of vector incrimination (Chambelain 1958): history of

frequent virus isolation.

Potential Vectors of West Nile virus

Of the sixty species of mosquitoes that have tested positive for WNV in

North America, thirty-eight species occur in Florida (CDC 2006b). Isolation of WNV with

various polymerase chain reaction (PCR) techniques only shows that mosquitoes were infected

with WNV (Rutledge et al. 2003). Isolation of WNV from mosquitoes does not establish their

capability of transmitting WNV to vertebrate hosts (Rutledge et al. 2003). Of the thirty-eight

mosquito species that had WNV isolated from them and occur in Florida, sixteen were tested in

the laboratory for vector competence (Turell et al. 2005).

Culex restuans, Cx. quinquefasciatus, and Cx. salinarius were found to be efficient vectors

of WNV in the laboratory (Turell et al. 2005). The varied host preference of Cx.

quinquefasciatus and Cx. salinarius (Edman 1974) make these species potential enzootic or

bridge vectors of WNV. Five WNV infected pools of Cx. quinquefasciatus were found in

Jefferson County, FL in 2001 (Godsey et al. 2005a). Four WNV infected pools of Cx. salinarius

were found in Jefferson County, FL in 2001 (Godsey et al. 2005a). The ornithophilic feeding

behavior of Cx. restuans (Edman 1974) makes this species a potential enzootic vector of WNV.

Aedes albopictus (Skuse) and Ochlerotatus triseriatus (Say) were found to be efficient

vectors of WNV in the laboratory and their mammalophilic host feeding habits make them

potential bridge vectors (Sardelis et al. 2002, Turell et al. 2005). Aedes albopictus and Oc.

triseriatus are very localized mosquitoes (flight range 200 m), so involvement in WNV

transmission may be limited to areas where Culex spp. are present and transmitting WNV to

amplifying hosts (Sardelis et al. 2002, Turell et al. 2005). Aedes vexans (Meigen), Culiseta

inornata (Williston), and Ochlerotatus canadensis canadensis (Theobald) are mammalophagic,









have flight ranges of 2 km or greater, capable of WNV transmission in the laboratory and may

play minor roles as bridge vectors of WNV to humans in Florida (Turell et al. 2005).

Three WNV infected pools of Cs. melanura were found in Jefferson County, FL in 2001

(Godsey et al. 2005a). Culiseta melanura (Coquillett) was a poor vector of WNV in the

laboratory (Turell et al. 2005). Culiseta melanura feeds primarily on birds (Edman et al. 1972b)

and may be an enzootic vector of WNV in Florida.

West Nile virus has only recently been introduced to Florida and there have been very few

studies demonstrating WNV transmission to vertebrates in the field, the exception being Cx.

nigripalpus (Rutledge et al. 2003). Laboratory studies of vector competence and field isolations

of WNV only fulfill part of the requirements for vector incrimination (Chamberlain 1958). More

field studies on WNV transmission in Florida are required to fulfill all requirements of vector

incrimination for the species discussed in this section (Chamberlain 1958).

Potential Vectors of Eastern Equine Encephalitis Virus

Culiseta melanura is the primary enzootic vector of EEEV in Florida (Mitchell et al. 1996,

Morris 1992, Scott and Weaver 1989). Culiseta melanura breeding tends to be restricted to

swamp habitats (Morris 1992). Most human cases of EEEV occur within five miles of swamp

foci (Morris 1992). In Florida, Cs. melanura was found to be ornithophilic, feeding

predominantly on passerine birds (Edman et al. 1972b). Since Cs. melanura is ornithophilic it is

considered an enzootic vector of EEEV and other species of mosquitoes must serve as bridge

vectors to humans and horses (Edman et al. 1972b).

Numerous species of mosquitoes that have been suggested as potential bridge vectors of

EEEV in the United States occur in Florida including Coquillettidiaperturbans (Walker), Ae.

vexans, Ochlerotatus sollicitans (Walkeri), Culex erraticus (Dyar and Knab), Cx. salinarius, Oc.

canadensis, Anopheles quadrimaculatus Say, and Aedes albopictus (Edman et al. 1972b, Scott









and Weaver 1989, Scott et al. 1990b). Bridge vectors of EEEV to humans and horses vary from

location to location (Crans and Schulze 1986). In New Jersey, Oc. sollicitans is the primary

bridge vector of EEEV to humans in coastal areas and Cq. perturbans is the primary epizootic

vector of EEEV to horses at inland sites (Crans and McCuiston 1993, Crans and Schulze 1986).

Eastern equine encephalitis virus was isolated from Cx. salinarius from Tampa Bay,

Florida (Wellings et al. 1972) and other locations in the eastern United States (Scott and Weaver

1989). Cx. salinarius was able to transmit EEEV fourteen days post-infection in the laboratory

(Vaidyanathan et al. 1997). Culex salinarius may be a potential bridge vector because its

winter/spring abundance in Florida means that aging Cx. salinarius would be present during

EEEV transmission periods (Zyzak et al. 2002). The generalist feeding habits of Cx. salinarius

(Edman 1974) make it a potential enzootic and bridge vector of EEEV.

Both Oc. canadensis and An. quadrimaculatus were capable of transmitting EEEV in the

laboratory (Vaidyanathan et al. 1997). Both species are mammalophilic and could potentially

transmit EEEV to horses or humans (Edman 1971, Vaidyanathan et al. 1997).

Neither Ae. vexans nor Oc. sollicitans are commonly found near swamp foci in Florida,

which limits their potential involvement as bridge vectors of EEEV in Florida (Edman 1974).

Aedes vexans was not capable of transmitting EEEV in the laboratory (Vaidyanathan et al. 1997).

Minimum infection rates (MIR) of Cx. erraticus for EEEV in central Alabama were 3.2

per 1000 (Cupp et al. 2003). Positive Cx. erraticus pools were found from mid-June to mid-

September (Cupp et al. 2003), the same time period that EEEV is most active in Florida (Bigler

et al. 1976). An isolation of EEEV from Cx. erraticus in Polk County, Florida was made in 1993

(Mitchell et al. 1996). Chamberlain et al. (1954) found Cx. erraticus to be a competent vector of

EEEV in the laboratory. Robertson et al. (1993) documented Cx. erraticus feeding on humans in









North Carolina. These findings suggest that Cx. erraticus may play a role as an enzootic or

bridge vector of EEEV transmission in Florida.

Aedes albopictus was found naturally infected with EEEV in Polk County, Florida (MIR of

1.5 per 1000) (Mitchell et al. 1992). Aedes albopictus was capable of transmitting EEEV in the

laboratory (Scott et al. 1990b). Junkyards and tire piles in rural and urban areas near swamp foci

may harbor populations ofAe. albopictus (OCMCD data) and may lead to focal amplification of

EEEV (Mitchell et al. 1996).

Coquillettidiaperturbans has been incriminated as a bridge vector of EEEV in the eastern

United States (Morris 1988, Scott and Weaver 1989). Coquillettidiaperturbans was the vector

of EEEV in an epizootic among horses in New Jersey in 1983 (Crans and Schulze 1986).

Wellings (1972) obtained five EEEV isolates from Cq. perturbans in Florida. In Florida, 9% of

blood meals from Cq. perturbans were avian and 91% were of mammalian origin (Edman 1971).

Coquillettidiaperturbans was able to transmit EEEV seven days post-infection in the laboratory

(Vaidyanathan et al. 1997). Coquillettidiaperturbans feeds on humans in Florida (Provost

1969). These findings suggest that Cq. perturbans may be a potential bridge vector of EEEV to

humans and horses in Florida.

Surveillance of Arboviruses

Arboviral surveillance is the systematic collection of data regarding arboviral activity to

predict or recognize epidemics and assess the size and progression of current epidemics (Bowen

and Francy 1980). The objective of arbovirus surveillance is to predict the likelihood of

arbovirus transmission to humans so that public health measures and mosquito control activity

may limit or abort human outbreaks (Eldridge 1987). There are two types of arbovirus

surveillance methods: passive and active. Passive surveillance is the accumulation of

unsolicited information concerning arboviral infection in humans or other vertebrates whereas









active surveillance is the intentional collection of specimens or data. There are four methods for

conducting active arboviral surveillance: detection of human disease or infection, observation of

weather patterns, surveillance of vector populations, and the testing of non-human vertebrates

(Bowen and Francy 1980).

Human Surveillance

Human surveillance for arboviral encephalitis incidence is conducted by public health

officials. A passive human surveillance system requires local physicians and hospital staff to

record and report suspect human cases of arboviral infection to local health departments. In the

past, cases were not being reported in a timely manner or reported case information was not

assimilated properly, allowing epidemics to go unnoticed (Chamberlain 1980). In many regions

where there is little money for mosquito control operations and arboviral surveillance, the only

affordable method for arboviral surveillance is human surveillance (Bowen and Francy 1980,

CDC 2003).

Once public health officials have determined that an arboviral epidemic is imminent or

already in progress the most common measures of control are emergency aerial adulticide

spraying and public service announcements (PSA's). The drawback to human surveillance is

that arboviral amplification is already at high levels in vector and amplifying host populations by

the time incidental human cases are reported. The delayed vector control will likely have little

impact on epidemic transmission if most arboviral transmission has already occurred (Nelson et

al. 1983). Day and Lewis (1992) discussed the difficulties of human surveillance due to the

several week time lag between the date of onset, the first (acute phase) and second (convalescent

phase) blood samples and the diagnostic tests. Nelson et al. (1983) suggested that clinically

suspected SLEV cases should be reported to reduce the reaction time of mosquito control efforts.









Human arboviral case infections in the United States are defined as possible, probable, and

confirmed cases. Possible and probable human case definitions of arboviral disease require the

display of symptoms associated with the disease and positive antibody results from one acute or

convalescent human sera, cerebrospinal fluid (CSF) or other body fluids. Confirmed human case

definition requires the display of symptoms and positive results with a fourfold increase between

acute and convalescent serum, CSF, or other body fluid or the direct isolation of arbovirus or

demonstration of arboviral genomic sequences in human tissue, blood, CSF, or other body fluid

(Petersen and Marfin 2002).

In Florida human diseases are monitored by the Florida Department of Health (FDOH) and

County Health Departments (CHD). Human sera are screened for arboviruses with a

hemagluttination inhibition (HI) assay and all positives are confirmed with an immunoglobulin

M-capture enzyme-linked immunosorbent assay (MAC-ELISA). A plaque reduction

neutralization test (PRNT) is used to differentiate between WNV and SLEV. Cerebrospinal fluid

samples are tested for WNV with MAC-ELISA. All tests are performed free of charge at the

DOH Bureau of Laboratory Services in Tampa or Jacksonville (Blackmore et al. 2003). Human

case reports are entered into Merlin, a real time electronic reporting system for all diseases that

affect humans (FDOH 2005). All probable and confirmed human cases are then entered into

Arbo-NET, a cooperative West Nile virus surveillance database and reported to the Centers for

Disease Control (Marfin et al. 2001). Even with a nationwide human surveillance system the

data generated are problematic because of the variability and inaccuracies in disease reporting.

This makes the estimation of disease incidence and the interpretation of potential risks difficult

(Brownstein et al. 2004).









Weather Patterns

Weather patterns that preceded past arboviral epidemics can be analyzed to formulate

predictions of future arboviral transmission potential. These predictions can provide an early

warning of pending arboviral epidemics so that public health officials and mosquito control

programs can increase surveillance and control efforts to reduce epidemic potential. Weather

patterns are not indicative of arboviral presence or amplification in mosquito vectors or

vertebrate hosts. Field isolations of arbovirus in mosquitoes or vertebrate hosts or isolation of

arboviral antibodies from sentinel and wild animals is required for confirmation of viral activity.

Tracking weather patterns is an easy and low cost method that when used in combination with

human surveillance, is the best alternative for areas lacking a budget for a standing mosquito

control and surveillance operations (Bowen and Francy 1980).

Recently, several models have been created which correlate water table depth and surface

wetness in peninsular Florida with SLEV or WNV transmission to birds and humans (Shaman et

al. 2002, 2003a, 2004a, 2004b, 2005). A topographically based hydrology model for south-

central Florida was created to predict surface level wetness (Shaman et al. 2002). This model

was validated using groundwater well measurements and surface water levels throughout Florida

(Shaman et al. 2003b). Sentinel chicken seroconversion data from 1978 to 2002 in Indian River

County and the topographically based hydrology model were used to assess past epidemic and

non-epidemic levels of SLEV transmission with the expectation that this model will be able to

forecast future SLEV epidemics (Shaman et al. 2004b).

Vector Surveillance

Vector surveillance for arbovirus transmission potential can involve monitoring of

mosquito abundance, parity status of the mosquito population and minimum infection rates of

mosquitoes. Vector surveillance can be costly to a mosquito control district or public health









department which means that not all three factors can always be assessed (Bowen and Francy

1980, Scott et al. 2001).

Monitoring of Mosquito Abundance

There are many techniques available to assess mosquito populations (Service 1976).

Long-term baseline data sets of mosquito abundance for each trap site with consistent sampling

methods demonstrate fluctuations in mosquito abundance over time (Day and Lewis 1991). In

peninsular Florida the primary arboviral vector of concern, Cx. nigripalpus, can be readily

caught by light traps, CDC baited light traps, ground aspirations, avian baited lard can traps (Day

and Lewis 1991), and sentinel chicken coop exit traps (T. P. Breaud and D. A. Shroyer, personal

communication). Indian River County Mosquito Control District (Vero Beach, FL) has

discontinued the use of light traps in favor of aspiration of resting mosquitoes (Day and Lewis

1991). Mosquitoes collected by aspiration can be analyzed to determine mosquito population

dynamics such as abundance, migration, blood feeding, parity, and age (Day and Lewis 1991).

Mosquito abundance has been correlated with arbovirus transmission with mixed success.

In California, Reeves (1968) predicted that if Culex tarsalis Coquillett light trap catches were

kept below ten mosquitoes per night SLEV human cases would not occur. Blackmore et al.

(1962) found no correlation between Cx. tarsalis abundance and SLEV or WEEV transmission

in Colorado. Olson et al. (1979) found that SLEV and WEEV transmission was affected by

varied abundance of Cx. tarsalis in rural and urban California habitats. The 1961 SLEV

epidemic in Tampa Bay, FL was characterized by high transmission rates to humans, low

numbers of complaint calls and low populations of mosquitoes in urban areas (Mulrennan 1969,

Rogers 1969). Recent focal outbreaks of WNV in the eastern United States have been associated

with low densities of Cx. pipiens and Cx. quinquefasciatus (CDC 2003). These disparities in

correlation of mosquito abundance and epidemic transmission render mosquito abundance an









invalid predictor of arboviral transmission and therefore a poor surveillance technique.

Mosquito abundance alone may be a poor predictor of epidemic transmission but when

combined with parity information it can be important for the timing and assessment of mosquito

control efforts during arboviral epidemics (Day 1991).

Parity Analysis

Parity analysis is a technique that is used to estimate the physiological age of an

anautogenous (mosquito that requires a blood meal before the development of eggs) mosquito.

Parity analysis involves dissection and examination of the ovaries for the presence or absence of

tracheolar skeins. Coiled tracheole skeins indicate a nulliparous (has not laid eggs) mosquito.

Extended tracheoles indicate a parous (has laid eggs) mosquito (Crans and McCuiston 1993). A

parous mosquito indicates that it has obtained at least one blood meal and may have lived long

enough to become infective. A period of time, known as the extrinsic incubation period, must

pass after the ingestion of blood before a mosquito can infect another host (Foster and Walker

2002). The process of extrinsic incubation requires the ingestion of arbovirus by a mosquito

vector, multiplication of arbovirus in the midgut cells, escape from the midgut, dissemination

throughout the hemocoel, and infection of the salivary glands (Foster and Walker 2002).

Assessment of parity status in mosquito populations provides an estimate of how many

mosquitoes might be infective (Bowen and Francy 1980).

Since climatological regulation of Cx. nigripalpus oviposition is crucial for SLEV

epidemics (Day et al. 1990) the determination of parity status of Cx. nigripalpus populations

during an epidemic is important for the timing of adult mosquito control operations (Day 1991).

Minimum Infection Rates in Mosquitoes

Virus isolation from mosquitoes is a part of vector incrimination (Chamberlain 1958) and

has been used for the detection of new arboviruses in an area (Scott et al. 2001). Virus isolation









is often conducted during arboviral epidemics but rarely during non-epidemic periods (Bowen

and Francy 1980, Shroyer 1991).

Minimum infection rates (MIR) are determined by the collection, identification, pooling,

and processing of mosquitoes for arbovirus isolation. Reeves et al. (1961) calculated minimum

infection rates using Equation 1-1.

Number of positive pools x 1,000 (Equation 1-1)
Number of mosquitoes tested


The MIR only represents the minimum number of infected mosquitoes, not the minimum

number of infective mosquitoes. Similar to mosquito abundance data, there must be baseline

data of MIRs during non-epidemic periods to understand what epidemic MIR levels represent as

a risk assessment of arboviral infection to humans (Shroyer 1991).

Transmission rates of arboviruses to sentinel vertebrates more accurately reflect the

potential for mosquitoes to transmit arboviruses. To determine transmission rates sentinel hosts

must be placed in the field, determine the number of seroconversions in hosts, and capture as

many mosquitoes that come to feed on the hosts as possible (Reeves et al. 1961). Reeves et al.

(1961) calculated transmission rates using Equation 1-2.

Number of birds infected x 1,000 (Equation 1-2)
Total mosquitoes feeding

Reeves et al. (1961) found both SLEV and WEEV virus infection rates in Cx. tarsalis to be

higher than transmission rates to sentinel chickens. Rutledge et al. (2003) found WNV infection

rates in Cx. nigripalpus to be higher than WNV transmission rates to sentinel chickens. One

factor attributed to higher MIRs than transmission rates is the extrinsic incubation period.

Mosquitoes may be infected with arbovirus which would raise the MIR but sufficient time may

not have passed for the mosquito to be infective which would not increase the transmission rate.









Shroyer (1991) found that during non-epidemic periods the MIRs for SLEV in Indian

River County, FL were very low. It was very difficult to statistically determine confidence

intervals and sampling error was very high because of the low frequencies of SLEV, which

makes direct estimates of human risk from MIRs difficult (Shroyer 1991). While isolation of

arboviruses from mosquito vectors is important for vector incrimination (Chamberlain 1958), it

is of little value as a surveillance tool to predict the threat of human arboviral infection in Florida

(Shroyer 1991).

Non-Human Vertebrate Surveillance

Surveillance of non-human vertebrates involves the testing of blood or tissue specimens

from vertebrates for the presence of arboviruses or antibodies to arboviruses (Bowen and Francy

1980). Surveillance of non-human vertebrates uses wild, domestic and sentinel animals. Wild

animal surveillance requires the collection of live or dead wild vertebrates and obtainment of

serum or tissue for detection of serologic or virologic evidence of arboviruses. Domestic animal

surveillance uses animals associated with humans that are exposed to natural vector populations

(i.e., backyard chicken flocks, cattle). Sentinel animals are those that are purposely exposed to

natural vector populations in a permanent location and routinely tested for antibodies to

arboviruses. Any vertebrate surveillance system used requires years of baseline data during

epidemic and non-epidemic periods to determine antibody and virus isolation thresholds for

prediction and assessment of arboviral activity (Bowen and Francy 1980, Day 1989).

Live, Wild Vertebrate Surveillance

Live, wild vertebrate surveillance requires the active capture of the vertebrate and

collection of serum or secretions that may contain virus or antibodies to the arbovirus of concern.

Virus isolation is sometimes performed in conjunction with antibody detection to establish

potential vertebrate reservoir and amplification hosts (Scott 1988). The use of mobile, wild









animals to monitor arboviral activity provides a greater range of area to be sampled per sampling

effort (Komar 2001, Trainer 1970).

Wild bird surveillance

Wild birds are the primary vertebrates that are tested for serosurveys and virus isolation in

eastern United States as they are the primary reservoir and amplification hosts of EEEV, SLEV,

WNV, and HJ. Wild bird surveillance can target either peridomestic birds associated with

epidemic transmission cycles (Gruwell et al. 2000) or non-peridomestic birds associated with

enzootic transmission cycles (Crans et al. 1994). Peridomestic birds associated with epidemic

transmission cycles are those that have been introduced to the United States such as pigeons

(Columba livia), House Sparrows (Passer domesticus), House Finches (Carpodacus mexicanus),

Mourning Doves (Zenaida macroura), and European Starlings (Sturnis vulgaris) (Allison et al.

2004, Bowen and Francy 1980, Gruwell et al. 2000, Komar et al. 1999, Komar et al. 2001,

Reisen et al. 2004). When virus levels rise sharply in peridomestic birds, there is an imminent

threat to humans (McLean et al. 1983). Bowen and Francy (1980) suggest that the risk of

arboviral transmission of SLEV to humans is "high when the wild bird antibody prevalence in

urban settings is above 15 percent." The level of risk of arbovirus transmission to humans in

relation to antibody prevalence in wild birds varies based on the vector/host/disease relationships

of the region. When a bird population involved in epizootic transmission cycles has high levels

of non-immune birds the likelihood of an epizootic is greater as there are many susceptible birds

present (Day 2001). The strength of wild bird surveillance is that it can accurately sample the

arboviral activity of amplification hosts over a wide area as opposed to sentinel birds which

represent focal arbovirus transmission in a fixed location.

Direct arbovirus isolation is rare but it does provide the best indicator of vertebrate

involvement in transmission cycles (Scott 1988). As direct virus isolation is rare, many









investigators have relied on antibody surveys to determine past arbovirus infection history of

wild birds. Birds with high field seroprevalances only indicate past infections, not a greater

potential for viral amplification. Species of birds with high field seropevalance need to have

laboratory studies of viremia titers and antibody responses to arboviral challenge to determine

their importance as arborviral amplifiers (Scott 1988).

Birds hatched the year that a serosurvey was conducted show recent arbovirus transmission

(Beveroth et al. 2006, Holden et al. 1973). Birds less than one month old with low level

antibody response most likely have maternal antibodies present (CDC 2003). Birds older than

one year may have been infected before the year of the serosurvey. In the eastern United States,

hatching year pigeons and House Sparrows are often sampled for SLEV and WNV virus

isolation (Allison et al. 2004, Bowen and Francy 1980, Komar et al. 2001). Large populations of

nestling pigeons and House Sparrows occur in peridomestic habitat and an increase in arbovirus

isolation rate from these species can predict transmission to humans (Holden et al. 1973).

Previously banded and blood sampled birds that are recaptured provide a precise time frame of

arboviral infection. Unfortunately there are generally low recapture rates during serosurveys

(Day and Stark 1999, Scott et al. 2001).

When wild bird surveillance data are collected in a systematic manner, wild bird

surveillance can provide an early warning of arbovirus transmission to humans (Bowen and

Francy 1980, Holden et al. 1973).

Wild mammal surveillance

Wild mammals, primarily rodents, have been used for serosurveys and virus isolation of

California Encephalitis virus complex (family Bunyaviridae, genus Bunyavirus, CEV),

Everglades virus (family Togaviridae, genus Alphavirus, EVEV), EEEV, and SLEV in Florida

(Bigler and Hoff 1975, Chamberlain et al. 1969, Day et al. 1996, Jennings et al. 1968, Wellings









et al. 1972). Wild mammals are the reservoir hosts for CEV (Bigler and Hoff 1975) and EVEV

(Day et al. 1996), possibly SLEV (Bigler and Hoff 1975, Day et al. 1995, Herbold et al. 1983),

and are incidental, dead end hosts for WNV and EEEV.

Everglades virus is only found in south Florida (Calisher and Karabatsos 1988). The

primary reservoir hosts of EVEV are rodents, especially Hispid Cotton Rats (Sigmodon hispidus)

(Coffey et al. 2004) and Cotton Mice (Peromyscus gossypinus) (Chamberlain et al. 1969).

Rodents most commonly found infected with virus or presence of antibodies to EVEV are the

Cotton Rat, Rattus sp., and Cotton Mice (Bigler and Hoff 1975, Chamberlain et al. 1969, Day et

al. 1996). In addition to rodents, the raccoon (Procyon lotor) was commonly found positive for

antibodies to EVEV in South Florida (Bigler 1971). Clinical EVEV disease in humans is rare in

South Florida and there is no routine surveillance of wild mammals (Sudia et al. 1969).

St. Louis encephalitis virus has been isolated from several wild bat species in the United

States (Allen et al. 1970, Herbold et al. 1983). Antibodies to SLEV virus have been found in

deer (Odocoileus virginianus) (Trainer 1970), cotton mouse (Peromyscus gossypinus), opossums

(Didelphis marsupialis), and raccoons (Bigler 1971). The raccoon specimens were obtained

during rabies surveys and demonstrates the benefits of testing mammals for arboviruses that

were passively obtained from other wild animal collections (Bigler 1971). Day et al. (1995)

found 31% (59 /189), armadillos positive for SLEV antibodies in Florida.

Wild mammal populations have not been associated with WNV amplification in the United

States. Pilipski et al. (2004) found 2% (2/83) bats positive for WNV antibodies in New Jersey

and New York. One percent (2/149) of Mexican free-tailed bats (Tadarida brasiliensis) from

Louisiana were positive for WNV antibodies (Davis et al. 2005). Neither Big brown bats

(Eptesicusfuscus) or the Mexican free-tailed bats are likely amplification hosts of WNV because









of absent or low viremia titers (Davis et al. 2005). In laboratory tests, the Eastern cottontail

rabbit, Sylvilagusfloridanus, developed sufficient WNV viremia (>105 CID5os/mL) to infect

feeding mosquitoes (Tiawsirisup et al. 2005). Cottontail rabbits are found in peridomestic

habitats and could play a role in peridomestic transmission cycles of WNV (Tiawsirisup et al.

2005).

Wild reptile surveillance

Reptiles have been suggested as potential overwintering hosts for EEEV and WEEV

because of their long lasting viremias (Bowen 1977, Gebhardt and Hill 1960, Hayes et al. 1964)

but have not been used in routine surveillance operations in the United States. This is due to the

safety risks associated with the capture (e.g., snakes and alligators) and difficulty of bleeding

reptiles (e.g., turtles) (Sudia et al. 1970). Klenk and Komar (2003) found the Green Iguana

(Iguana iguana), Florida Garter Snake (Thamnophis sirtalis sirtalis), and Red-ear Slider

(Trachymes script elegens) to be unlikely reservoir hosts for WNV because of low tittered

viremia < 103.2 PFU/mL of serum. Young alligators circulate WNV viremia titers as high as

106.2 PFU/mL for a duration of one to two weeks which makes them potential amplifying hosts

(Klenk et al. 2004). The role that wild alligators play in WNV epidemiology has not yet been

determined (Klenk et al. 2004).

Dead, Wild Bird Surveillance

West Nile virus epidemics in the United States have been characterized by high mortality

in American Crows (Corvus brachyrhynchos) which has led to the development of dead bird

surveillance (Eidson 2001, Eidson et al. 2001a, 2001b, 2001c; Johnson et al. 2006, Julian et al.

2002, Komar 2001, Marfin et al. 2001). Dead bird surveillance is a passive system that is

established by state and local health departments for the public to report dead birds (Eidson

2001). Dead bird surveillance systems compile dead bird sightings, and dead bird testing









involves collection and testing dead bird specimens for West Nile virus infection (Eidson 2001).

Results from both dead bird sightings and dead bird testing systems are tabulated and reported by

county health departments to Arbo-NET (Marfin et al. 2001).

Collection of dead or dying bird specimens for testing of WNV can be used to determine

WNV presence in an area. Although this system is considered active it still relies on passive

public reports for the location of specimens (Eidson 2001). By 2006 the Centers for Disease

Control and Prevention had reported 285 species of birds that tested positive for WNV (CDC

2006c). In the northern United States crows have been found to be the most sensitive indicator

of WNV, whereas in the southern United States Blue Jays have been the most sensitive indicator

of WNV (CDC 2003).

Problems with dead bird testing include: delays in test results (Eidson 2001), reliance on

public interest for dead bird reports (Eidson 2001, Julian et al. 2002, Ward et al. 2006), the

location the dead bird was found may not be the area that the bird was infected (CDC 2003),

dead bird surveillance data from different areas with different collection protocols are difficult to

compare regionally (CDC 2003), and the inability to estimate incidence of virus infection in the

wild bird population (Eidson 2001).

In Florida, the Fish and Wildlife Conservation Commission (FWCC) manages a bird

mortality database. This system allows the public, county health departments, or FWCC

personnel to report dead bird sightings. Dead birds are necropsied and tested by reverse

transcriptase-polymerase chain reaction (RT-PCR) for WNV at the DOH Bureau of Laboratory

Services, Tampa (Blackmore et al. 2003).

There have been numerous retrospective studies demonstrating the effectiveness of dead

bird sightings or dead bird testing systems at foreshadowing human WNV incidence in the









northeastern United States (Eidson et al. 2001a, 2001b, 2001c, 2005; Guptill et al. 2003, Julian et

al. 2002, Watson et al. 2004a). Unfortunately dead bird surveillance and dead bird testing have

been implemented in many areas without proper real-time analysis of the data collected. This

leads to systems that merely detect WNV in a given area but do not function as surveillance

because they cannot predict the risk of human infection (Guptill et al. 2003).

Domestic Animal Surveillance

Domestic animal surveillance is the use of domesticated or farmed animals for either

passive or active arbovirus surveillance. Passive animal surveillance relies on local veterinarians

to diagnose infected animals and report cases to animal morbidity reporting systems or public

health officials (Nichols and Bigler 1967). In Florida, specimens or tissue samples are sent to the

Department of Agricultural and Consumer Services (DACS) Division of Animal Industry and

Bureau of Diagnostic Laboratory Activities or the Department of Health (DOH) Laboratory,

Tampa for arbovirus isolation or antibody tests (FDOH 2005).

Active domestic animal surveillance involves the occasional or routine sampling of

domestic animals for arbovirus infection. Occasional antibody serosurveys of domestic animal

populations are considered retrospective and should be used in areas that lack routine arbovirus

surveillance (Nichols and Bigler 1967). Routine antibody serosurveys or virus isolations from

domestic animals provide a known period of infection. Drawbacks to the use of domestic animal

surveillance include a high turnover of animals making repeated sampling from the same

individual difficult (Endy and Nisalak 2002) and occasionally, inadequate knowledge of the

animal's life history (Dickerman and Scherer 1983).

Domestic mammals that have been used for arboviral surveillance include horses (Monath

et al. 1985), cattle (Gard et al. 1988), goats (Peiris et al. 1993), and pigs (Geevarghese et al.

1987).









Passive horse surveillance is most commonly used for monitoring EEEV, WEEV, and

WNV as horses may develop infections of the central nervous system in response to these

viruses but are not considered important reservoir hosts (Scott and Weaver 1989, Trock et al.

2001). Horse cases have preceded human cases in some epidemics of WNV (CDC 2003) and

EEEV (Scott and Weaver 1989). In Florida, before the statewide use of sentinel chickens,

equine morbidity was the best indicator of EEEV activity (Bigler et al. 1976), but Hayes and

Hess (1964) found no correlation of EEEV between horse and human cases in Florida from 1938

through 1961.

In Florida, horse serum is screened by HI assay at the Animal Disease Diagnostic

Laboratory (Kissimmee, FL) for Flavivirus and Alphavirus antibodies. Confirmation tests for

WNV or EEEV by MAC-ELISA and PRNT are performed at the National Veterinary Services

Laboratories in Ames, Iowa (Blackmore et al. 2003). Virus isolation from blood, brain, and

spinal cord tissues can be performed with rabbit kidney and Vero cell cultures with Alphavirus

isolates identified by complement fixation tests (Ostlund et al. 2001).

Many birds, such as emus (Dromaius novaehollandiae) and pheasants (Phasianus

colchicus) have been introduced to North America are farmed for game animals and commodity

markets (Tully et al. 1992). Massive die offs of farm raised birds in the United States caused by

EEEV have been observed in emus (Day and Stark 1996a, Tully et al. 1992), pheasants (Tyzzer

et al. 1938), White Pekin ducklings (Dougherty and Price 1960), and Chukar Partridges

(Alectoris chukar) (Ranck et al. 1965). Many EEEV outbreaks in pheasants are driven by bird-

to-bird transmission (Holden 1955). In Florida case fatality rates due to EEEV in emus have

been reported as high as 14% (Day and Stark 1996a), pheasants up to 45%, and partridges up to









19% (Bigler et al. 1976). Human epidemics of EEEV have been preceded by epizootics among

gamebirds (Scott and Weaver 1989).

Zoological parks should be monitored for avian mortality (CDC 2003). Many bird species

in zoo collections are exotic and may experience high mortality when first introduced to WNV.

In New York in 1999, eight different species of exotic birds had deaths attributable to WNV

(Eidson et al. 2001a, Rappole et al. 2000). Of the 285 bird species found dead and positive for

WNV, 54 species were captive and exotic to North America (CDC 2006c).

Backyard chicken flocks have been used in active surveillance programs, especially in

areas where no other sentinel surveillance is routinely conducted. Flocks used for surveillance

should be no more than 30 birds (Bigler et al. 1976). Large commercial flocks are usually reared

indoors, inaccessible to mosquitoes, and in numbers that are so large the chance of finding

antibodies in birds is remote. Unless chickens are bled on a routine basis, backyard chicken

flocks are only useful in determination of past arboviral activity and do little to determine current

arboviral activity. Chickens less than one year old should be used in serosurveys to detect

arboviral activity in the last year (Nichols and Bigler 1967). Backyard chickens and domestic

geese were found to be ideal sentinels for WNV in New York City during 1999 (Komar et al.

2001).

In the southeast there have been several reports of farmed American alligator (Alligator

mississippiensis) die-offs caused by WNV (Jacobson et al. 2005, Klenk et al. 2004, Miller et al.

2003). On one alligator farm in Georgia more than 10% of the young alligators succumbed to

WNV (Miller et al. 2003). It has been shown that alligators can directly transmit WNV to

tankmates (Klenk et al. 2004). Horsemeat infected with WNV and fed to alligators was shown to

be the cause of an outbreak in Georgia (Miller et al. 2003). With multiple routes of infection for









WNV in alligators it is difficult to use alligator die-offs as an indication of arboviral activity

circulating in local mosquito populations. Outbreaks on alligator farms are cause for public

health concern as direct transmission of WNV to human alligator handlers has been documented

(Klenk et al. 2004).

Sentinel Animal Surveillance

Sentinel animals are vertebrates maintained at a specific location, exposed to natural

populations of mosquitoes, and routinely tested for evidence of arboviral infection. Evidence of

arbovirus infection is determined indirectly by testing blood for a seroconversion (production of

specific antibodies to arboviruses) or by direct virus isolation (Komar and Spielman 1995).

Vertebrate sentinels are critical for establishing proof of virus transmission in the absence of

human and horse arbovirus cases or bird die-offs (Day and Lewis 1991). Caged sentinel animals

provide information of arbovirus transmission at an exact location (Komar and Spielman 1995).

Komar (2001) stated that the ideal sentinel {vertebrate} is uniformly susceptible to infection,

resistant to disease, rapidly develops a detectable immune response, easily maintained, presents

negligible health risks to handlers, does not contribute to local pathogen transmission cycles, and

seroconverts to the target pathogen before the onset of disease outbreaks in the community.

Other factors to consider when selecting a sentinel animal are aggressiveness toward other

animals sharing the cage, its ability to withstand the local environmental conditions, and the

host/vector/disease dynamics for the area (Sudia et al. 1970).

Birds are commonly used as sentinels but may only represent enzootic transmission cycles

between predominantly ornithophilic mosquitoes (i.e., Cx. pipiens, Cs. melanura) and the viruses

they vector: whereas a sentinel mammal would indicate spillover of virus into an epidemic

transmission cycle by incidental feeding on non-target hosts or feeding by mammalophilic bridge

vectors (i.e., Cq. perturbans, Ae. albopictus). Selection of the ideal sentinel vertebrate is









complicated as there is no one vertebrate that works well for detection of all arboviruses in every

geographic region. Sentinel animal use is further complicated by the need to select proper

monitoring sites to represent enzootic or epidemic transmission (Komar 2001).

Sentinel mammals

Mammals are infrequently used as sentinels for arboviruses as small mammals may

experience high levels of mortality (Ventura and Ehrenkranz 1975), larger species are difficult to

maintain, and cross reactivity with other viruses may confound test results (Day et al. 1996).

Rabbits and rodents have been used as sentinels for CEV and EVEV in Florida (Jennings et al.

1968, Ventura and Ehrenkranz 1975). Sentinel mammals can be bled routinely to monitor

seroconversion and mammals that die during routine exposure should be submitted to a

laboratory for virus isolation (Sudia et al. 1970). Ventura and Ehrenkranz (1975) exposed

sentinel hamsters in South Florida for the detection of EVEV. Over 40% of the hamsters

exposed were killed by predators or stolen (Ventura and Ehrenkranz 1975).

Sentinel birds

Birds have been used routinely in arboviral surveillance programs for many decades

(Moore et al. 1993). The birds most often used as avian sentinels are chickens, pigeons,

pheasant, and quail (Bowen and Francy 1980, Morris et al. 1994, Reisen et al. 1992, Williams et

al. 1972). Within the United States bird species vary in their usefulness as sentinel animals from

region to region based on vector/host/disease dynamics. In the United States, sentinel birds have

been used for arboviral surveillance in Alabama, Arizona, Colorado, Delaware, Florida, Iowa,

Louisiana, Maryland, Nebraska, Nevada, New Jersey, North Carolina, Tennessee, Texas, and

Utah (Komar 2001). The various organizations conducting avian sentinel surveillance in a wide

range of geographic locations have used different bird species, varying numbers of sites,









numbers of birds exposed per site, heights and types of exposure devices, various bleeding

regimens and techniques, and laboratory tests.

The number of birds used per site ranges from 1 to 50 and depends on the species and type

of exposure device (Bellamy and Reeves 1952, Rainey et al. 1962). Public health and mosquito

control resources are limited so the number of birds exposed at a site and the number of sites

must be chosen carefully (Bowen and Francy 1980). In densely populated areas fewer sentinel

sites are available (Gergis and Presser 1988). Large numbers of chickens (25) and pigeons (30)

were exposed at ten sites between two counties in Los Angeles, the second largest city in

America (Gergis and Presser 1988). The use of fewer sentinels, 6 to 10, would reduce space

requirements, maintenance and eyesore to the public. By reducing these problems more

sentinels could be dispersed over a wider area and increase the chance of detecting focal

transmission points. The reduction of space requirements, maintenance and eyesore issues may

allow for the deployment of sentinels at public facilities (e.g., public health departments, pump

stations, wastewater treatment plants, electrical substations) or private residences. In urbanized

areas where sentinels cannot be placed wild bird populations should be tested (Gergis and

Presser 1988). Sentinel birds should be placed in the field prior to mosquito activity in northern

states or kept in the field throughout the year in southern states that experience mosquito activity

throughout the year (CDC 2003).

Birds should be bled and tested for previous arboviral infection before placement in the

field. Sentinel birds are bled on weekly (Day and Stark 1996b), bi-weekly (Crans 1986), or

monthly (Reisen et al. 1992) schedules. Blood (0.5 to 4.0cc) is drawn by brachial or jugular

venipuncture (Day 1989, Reisen et al. 1992) or by lancet comb prick (Reisen et al. 2004).









In Florida, during 2005, 33 of 67 counties conducted sentinel chicken surveillance (Stark

and Kazanis 2005). Most programs typically expose six chickens, with a range of four to eight

chickens per site throughout the year. Blood samples are taken via veinipuncture of the jugular

or brachial veins with a syringe and plunger. The blood is dispensed into vacutainers and usually

centrifuged at Mosquito Control for the separation of blood and serum. Serum samples are then

packaged and shipped to the DOH Bureau of Laboratory Services in Tampa, FL.

Hemagglutination Inhibition test is the initial screening test used to determine antibody response

to Alphaviruses or Flaviviruses. Positive HI results for Flaviviruses are identified by IgM-

capture enzyme-linked immunosorbent assay (MAC-ELISA). West Nile virus negative birds are

re-bled and screened by the HI test. Samples positive during the second HI test are confirmed

and identified as SLEV or WNV by the plaque reduction neutralization (PRNT). Alphavirus

positives are confirmed and identified as EEEV or HJ by PRNT. The DOH Bureau of

Laboratory Services in Tampa processes 1200 to 1500 HI screening tests per week. On Fridays

the results of all positive chickens in the state are then faxed to every participating county and

negative/positive results of each chicken are faxed to the counties that submitted the samples (L.

M. Stark, personal communication).

Cherry et al. (2001) reported that fourteen sentinel chicken flocks failed to seroconvert for

WNV before human infection on Staten Island and in New York City. These areas cover

roughly 830 km2 and contain approximately 8.1 million people. With roughly 84 sentinel

chickens exposed and 8.1 million humans, there was a very low chance that the chickens were

bitten by infected mosquitoes before human infection. In New York State, sentinel chicken

surveillance has been ceased in favor of mosquito and dead bird surveillance (Darbro and

Harrington 2006).









Sentinel chickens were found to be relatively insensitive as predictors of human EEEV

cases in Florida (Bigler et al. 1976, Day and Stark 1996b). This may be due to lower feeding

rates on caged sentinels by Cs. melanura, the primary vector of EEEV (Scott and Weaver 1989).

Sentinel chickens in California were tested for antibodies to Cx. tarsalis salivary gland antigens

(Trevejo and Reeves 2005). Chickens seropositive for antibodies to Cx. tarsalis salivary gland

antigens were more likely to seroconvert to SLEV than salivary gland antigen seronegative

chickens (Trevejo and Reeves 2005). Culex tarsalis is the primary vector of SLEV in California

(Mitchell et al. 1980) so it follows that its absence would lead to an absence of SLEV

transmission. If this same technique of antibody detection to salivary gland antigens from Cs.

melanura, then seronegative chickens for antibodies to Cs. melanura salivary gland antigens

could explain a lack of EEEV transmission to sentinel chickens in Florida.

An important consideration when placing sentinels in the field is the height at which

sentinels are exposed. Sentinel chickens maintained at ground level did not predict WNV

transmission before human incidence in New York City (Cherry et al. 2001). Culexpipiens and

Cx. restuans are suspected vectors of WNV in New York (Anderson et al. 2004, Apperson et al.

2004). In the northeast, both Cx. pipiens and Cx. restuans host-seek in the tree canopy and at

ground level (Anderson et al. 2004, Darbro and Harrington 2006, Deegan et al. 2005). Darbro

and Harrington (2006) caught significantly more mosquitoes in the tree canopy than on the

ground with house sparrows and chickens as bait and they recommend the use of ground level

and tree canopy sentinel surveillance. Deegan et al. (2005) exposed sentinel pigeons at canopy

and ground levels and found significantly higher seroconversion rates for WNV at the canopy

level. Sentinel birds exposed at canopy levels are contained in modified lard cans (Bellamy and









Reeves 1952) or the Ehrenberg pigeon trap (Downing and Crans 1977) and typically exposed for

one night.

Interpretation of sentinel bird results can be complicated. Years of baseline data in an area

during epidemic and non-epidemic years are required. When conducting regional surveillance,

mean annual seroconversion rates (MASR), excluding epidemic years should be calculated

separately for each county (Day and Stark 1996b). When the real time ASR is higher than the

MASR there is generally cause for concern. At a regional level the comparison of real time

ASRs to the MASR shows increases and decreases of arbovirus activity between counties and

identifies viral foci that cross county lines (Day et al. 1991).

At the county level, certain sentinel sites will be more likely to have positive sentinels

depending on host/vector/disease dynamics (Day et al. 1991). Certain sites within counties will

more accurately represent enzootic/sylvan transmission (e.g., swamps, wooded areas) and other

sites will represent peridomestic/urban transmission cycles (e.g., backyards, urban parks). Single

serocoversions of sentinels in enzootic/sylvan transmission sites are typically not cause for

public health concern as these sites will experience background transmission quite often.

Williams et al. (1972) found that sites deep inside swamps were likely to have elevated levels of

EEEV transmission earlier in the season than sites at the edge of the swamp and upland sites. An

increase in sentinel seroconversion rates in open, dry sites in south Florida may represent

imminent threat of arbovirus transmission to humans by Cx. nigripalpus (Day and Carlson

1985). As autumn rainfalls occur and humidity levels increase Cx. nigripalpus will exit previous

resting sites (e.g., wooded hammocks) and blood feed in open dry sites (e.g., fields, backyards,

urban areas) increasing the likelihood of SLEV transmission to humans (Day and Carlson 1985,

Day and Edman 1988). In Florida, seroconversion rates of SLEV above 30 percent over a three









week period are cause for concern (Day and Lewis 1992). High seroconversion rates earlier in

the transmission season (May to July) represent greater threat to human cases than equally high

seroconverison rates later in the season (August to December) (Day and Lewis 1992).

There are many positives to using chickens as arbovirus sentinels. Chickens are

commercially available and relatively inexpensive -$5 to 7 (Komar and Spielman 1995).

Several strains of chickens are capable of enduring extreme outdoor conditions (O'Bryan and

Jefferson 1991). Chickens are highly attractive to Culex spp. mosquitoes. In the most extreme

case in Orange County, FL an estimated 2.3 million mosquitoes, mostly Culex, were recovered

from one exit trap collection (OCMCD data). Sentinel chickens are inactive after sunset, during

the period of Culex spp. blood feeding activity (Day and Edman 1984). Adult chickens survive

EEEV (Byrne and Robbins 1961), WEEV, SLEV (LaMotte et al. 1967), and WNV (Senne et al.

2000) infections. Chickens develop detectable antibodies to EEEV within 4 days (Olson et al.

1991), WEEV within 10 days, SLEV within 14 days (Reisen et al. 1994), and WNV within eight

days (Senne et al. 2000). Chickens run short duration, low-level viremias to WEEV, SLEV

(Reisen et al. 1994), and WNV (Langevin et al. 2001) with little threat of passing the virus to

other mosquitoes. During the 1990 SLEV outbreak in Florida sentinel chicken seroconversion

rates above the MASR were found in several counties before human infection (Day and Stark

1996b).

There are disadvantages to using sentinel chickens. Chickens are larger than most other

sentinel birds and therefore more expensive to feed and require more cage space (Komar and

Spielman 1995). During epidemics when many sentinel chickens are seroconverting they need

to be rapidly replaced it may be difficult to find poultry distributors with the particular breed

(White leghorns in Florida) that is desired (O'Bryan and Jefferson 1991). Chickens may become









aggressive toward cage mates leading to mortality (OCMCD data). Delayed results because of

antibody development and lab confirmation are another major drawback to sentinel chicken

surveillance or any avian surveillance (O'Bryan and Jefferson 1991).

It takes 8 to 21 days for HI antibodies to SLEV to develop in chickens (O'Bryan and

Jefferson 1991). It would not be until day 9 (minimum) or day 15 (maximum) post infection that

the antibody positive blood sample would be drawn. Initial report of suspected positive would

not occur until day 16 to 22 post infection and confirmation would follow 23 to 29 days post-

infection (DPI) (O'Bryan and Jefferson 1991).

In the 1990 SLEV outbreak in Indian River County, suspected positives were treated as

confirmed so a more rapid mosquito control response could be mounted (O'Bryan and Jefferson

1991). Even with this system, the actual transmission event occurred two to three weeks prior.

For WNV, it takes a minimum of seven days for HI antibody development (Senne et al. 2000), so

suspect cases would be reported within 15 to 21 DPI and confirmed 22 to 28 DPI. For EEEV it

takes a minimum of 4 days for HI antibody development (Olson et al. 1991), so suspect cases

would be reported within 8 to 14 DPI. However, confirmation with PRNT takes 1 to 2 weeks so

it would be 15 to 28 DPI before a confirmed EEEV case was reported. It is important to wait for

confirmation of EEEV tests because Highlands J virus is not of public health importance and

therefore control measures for HJ virus are not necessary.

As with other avian species (Sooter et al. 1954) adult hens may transovarialy transmit

maternal HI antibodies to SLEV virus (Bond et al. 1965) and NT antibodies to SLEV and WEEV

viruses to their offspring (Reeves et al. 1954). Maternal HI antibody levels tend to be lower than

HI levels found after mosquito-borne arboviral infection (Bond et al. 1965). Chickens less than

15 days old may have detectable maternal antibodies (Day 2001), which would produce









seropositive chickens that were not infected during field exposure. Often times chickens used as

bait in lard can traps are less than one month old (Rutledge et al. 2003). Chickens should be bled

and tested before field exposure (Day 1989). Chickens less than nine days of age should not be

used for the surveillance of EEEV as they exhibit high levels of mortality to EEEV infection

(Byrne and Robbins 1961). Chickens should be housed in a mosquito proof enclosure before

placement in the field (Day 1989). During the Florida SLEV outbreak of 1990, the Indian River

Mosquito Control District had numerous sentinel chickens seroconvert before placement in the

field because of SLEV infection at the IRMCD because chickens were not held in mosquito

proof enclosures (O'Bryan and Jefferson 1991).









CHAPTER 2
CAGE ESCAPE TRIALS

Introduction

Sentinel chickens can provide important information about arbovirus activity in an area,

but do not provide information about which mosquito species are transmitting arboviruses.

Rainey et al. (1962) implemented a mosquito trap designed to catch mosquitoes as they approach

the chickens, before mosquitoes entered the cage. A recent addition to some sentinel cage

designs in Florida is an exit trap (T. P. Breaud and D. A. Shroyer, personal communication).

These exit traps differ from the traps discussed in Rainey et al. (1962) because they capture

mosquitoes after they enter the cages, attempt to obtain a blood meal, and try to exit the cage.

Exit traps are placed on top (OCMCD and Volusia County Mosquito Control District) or on the

sides of cages (Manatee County and Indian River County Mosquito Control Districts). While

exit traps provide evidence of the mosquito species attracted to sentinel chickens, they may

disproportionately represent the abundance of mosquito species that enter the cages and attempt

to feed. This means that there is a potential to miss certain mosquito species that enter sentinel

cages and are able to transmit arboviruses, thereby underestimating their importance in arbovirus

transmission cycles.

There is no universal blueprint for the design of the sentinel chicken cages for mosquito

control districts. This has led to the creation of many different types of sentinel cages. Sentinel

cages have been designed to hold two chickens (King 1983) to thirty chickens (Rainey et al.

1962). Most sentinel chicken cages used in Florida hold six chickens (T. P. Breaud and D. A.

Shroyer, personal communication).

An experimental cage was designed in January, 2005 and built at OCMCD for this study.

The experimental cage was designed to capture as many host seeking mosquitoes that come to









feed on the chickens as possible. Mosquitoes were caught in the exit trap or contained within the

sentinel cage where the chickens are held. The experimental sentinel chicken cage design is

similar to that used by OCMCD with a few modifications (Figures 2-1 and 2-2). Modifications

to the original OCMCD sentinel cage include: the addition of a mesh baffle to funnel

mosquitoes into the cage, plastic sheets to prevent chicken feces from contacting the mesh baffle,

entry portals to allow personnel access to inside of the cage for aspiration, and a reduced inner

cage to allow room for personnel to aspirate the cage and as a refuge for mosquitoes.

It is important to determine the limitations of any mosquito trap used in surveillance. Prior

to field trials, the experimental cage was tested for its efficiency at drawing mosquitoes into the

cage and its ability to contain mosquitoes once inside the cage. To determine the efficiency of

the experimental cage the baffle angle was altered through a range of angles from 0-60. This

was done to determine the baffle angle that maximized the number of mosquitoes and mosquito

species caught in the cage. The mark-release-recapture technique was used with five species of

colonized mosquitoes with CO2 and chickens as the attractant. The mark-release-recapture

technique is the release of a known number of mosquitoes that have been marked by fluorescent

powders (Pylam Products Co., Tempe, AZ) and measuring the rates of mosquito recovery.

Materials and Methods

Laboratory colonies of Ae. albopictus, Aedes aegypti (Linnaeus), An. quadrimaculatus

Say, Cx. quinquefasciatus, and Cx. nigripalpus (USDA-CMAVE strains) were used to determine

whether or not they were able to escape from the experimental cage, and the likelihood that each

species was attracted to and entered the cage. Diurnally active species, Ae. aegypti and Ae.

albopictus were released at 9:30 am and aspirated from the cage at 5:00 pm. Nocturnally active

species, An. quadrimaculatus, Cx. quinquefasciatus, and Cx. nigripalpus were released at sunset

and aspirated from the cage at 8:00 am the next morning.









All trials were conducted within the confines of an outdoor mosquito cage (9.1 m x 18.3 m

x 4.9 m gabled to 5.5 m) at the United States Department of Agriculture, Center for Medical

Agricultural and Veterinary Entomology, Gainesville, FL (USDA-CMAVE). The experimental

cage was placed in the center of an outdoor mosquito cage. All release trials were conducted

within the confines of the outdoor mosquito cage. A Hobo data logger (Onset, Bourne, MA),

was placed on top of the inner cage to monitor changes in temperature, humidity, and light. A

Vantage Pro 2 Weather Envoy (Davis, Hayward, CA), was placed next to experimental cage to

monitor wind speed and direction. Rainfall data were obtained from the University of Florida

Physics department (UFDP 2005).

Mosquito Rearing

Aedes aegypti and An. quadrimaculatus eggs were provided weekly by the USDA. Aedes

albopictus, Cx. nigripalpus, and Cx. quinquefasciatus eggs were collected from lab-reared

specimens. Five days after Cx. nigripalpus blood feeding, 5 to 7 day old hay infusion was

provided for as an oviposition source for gravid adult Cx. nigripalpus females. Four days after

Cx. quinquefasciatus blood feeding, distilled well water was provided as an oviposition source

for gravid adult Cx. quinquefasciatus females. Four to five egg rafts from either species were

immediately placed into plastic containers with distilled well water. After larvae hatched they

were transferred to larval pans with distilled well water. Eggs from adult Ae. albopictus were

collected on moistened brown paper, dried and stored for approximately one week in a cool, dry

Ziploc (Johnson, Racine, WI) bag.

Larvae were reared in plastic pans 50.8 cm x 38.1 cm x 7.6 cm and fed according to the

feeding schedule in Table 2-1. Adults were held in 41 cm x 41 cm x 46 cm cages and were

provided a constant source of 10% sugar water solution. All mosquitoes were reared at 28 +

1.0C and 90 4% RH on a 14 : 10, light: dark cycle. Adult females were provided the









opportunity to blood feed on restrained chickens obtained by the USDA-CMAVE. All chickens

used for colony feeding were cared for according protocols approved by the University

Institutional Animal Care and Use Committee (#D469) and were housed at the USDA-CMAVE.

Marking

Mosquitoes were aspirated from emergence cages with a hand-held aspirator (Hausherr's

Machine Works, Toms River, N.J.). Except for An. quadrimaculatus, mosquitoes were

temporarily immobilized with CO2 at a rate of 500 mL/min. The high knock-down threshold of

colony-reared Anopheles quadrimaculatus to CO2, required them to be chilled for 20 minutes in

a refrigerator and then placed on a chill table maintained at 4 oC.

The number of individuals of each mosquito species required for each trial were counted

placed into cardboard cups with screen lids. The inside of the cups were lightly dusted with

either red, white, blue, yellow, green, or orange fluorescent marking powder. Each trial and cage

were assigned specific colors for that day or night. Mosquitoes used in release trials were six to

eighteen days post emergence. A water moistened cotton ball was placed on top of the screen to

keep the humidity high. Mosquitoes were given at least half an hour to recover from the cold or

CO2 knockdown treatments before being released in trials. At the completion of every trial all

mosquitoes were aspirated from the experimental cage. Any mosquitoes that were dead inside

the cardboard cup were noted as mortality caused by the marking process and excluded from

statistical analysis. Recovered mosquitoes were observed under a black light (Adams Apple

Distributing LP, Glenview, IL) to determine the trial in which they were released.

Chickens

White leghorn chickens, 19 to 25 weeks old, were used as the attractant in the experimental

cages. Protocols for the care and use of the chickens were approved by the UF IACUC, #D996.

The chickens were housed behind the Entomology and Nematology Department building on









campus, in Gainesville, FL. Chickens housed in this location were exposed to natural

populations of mosquitoes prior to experimental exposure.

Experimental Cage Design

An outer cage measuring 1.7 m x 1.3 m x 1.2 m (H x W x D) was constructed with 2" x 4"

and 4" x 4" treated pine lumber and 2.5 cm x 5.1 cm hardware cloth (Figure 2-3). An inner cage

measuring 0.7 m x 1 m x 1 m was constructed out of 2" x 2" treated pine lumber and 1.3 cm x

2.5 cm hardware cloth (Figure 2-4). Chickens were housed in the inner cage. The purpose of the

inner cage was to create a double barrier of hardware cloth to prevent predators from entering

cage, to allow easy removal of the chickens, provide enough room to maneuver an aspirator, and

to provide a refuge for mosquitoes to rest without threat of consumption by chickens.

A board measuring 2" x 2" on which chickens roosted during the night, extended

diagonally across the inner cage (Figure 2-4). The inner cage was supported by two 2" x 4"

boards spaced 1.9 cm apart, which extended horizontally across the cage to form the baffle entry

slit (Figure 2-3). The inner cage rested on two 5 mm thick plastic sheets. The plastic sheets

were cut to the shape of the outer cage and the baffle slit. The plastic sheets kept chicken feces

from falling on the mesh baffle and prevented mosquitoes from flying underneath the inner cage.

Two pieces of 0.5" x 2" treated pine lumber were attached to the bottom of the inner cage to

prevent chickens from contacting their feces (Figure 2-4). The inner cage was coated in water

sealant to prevent feces and urine from soaking into the wood. A one-gallon water container and

feeder were placed in the cage when chickens were present.

The outer cage was sealed with 12-count mesh screening. Male Velcro was stapled to

the outer cage. Twelve-count mesh screening was attached to the cage by female Velcro. The

mesh screening was removed when the cage was not used.









Portals (20 cm x 23 cm) on three sides of the outer cage functioned as doors that allowed

personnel to reach inside the cage. The portals were constructed with 1.3 cm x 2.5 cm hardware

cloth and wire ring clamps. The portals were used to release mosquitoes or to aspirate

mosquitoes. The portals were sealed with 12-count mesh screening attached to the cage with

Velcro.

An exit trap was constructed in the center of the sheet metal roof (Figure 2-5 A.). A five

gallon water container was used to construct the exit trap. The water container was cut in half

and the top half was permanently attached to a circular, sliding piece of sheet metal. The bottom

half of the water container was inverted and rested on the top half of the permanently mounted

water container (Figure 2-5 B.). Mosquitoes flying upward into the exit trap were funneled

through the neck, out the spout, and were potentially contained within the removable portion of

the exit trap. Tubular Stockinette (Bioquip, Rancho Dominguez, CA) was attached to the

bottom half of the water container and functioned as a seal to keep mosquitoes in the

containment portion of the exit trap.

Aspirator

An aspirator was designed to remove mosquitoes from the experimental cage (Figure 2-6).

A modified Bioquip DC Insect Vac 12 volt aspirator (Rancho Dominguez, CA) provided the

suction power. A containment jar was connected to the end of the aspirator. Two meters of 10

mm diameter braided vinyl tubing extended from the top of the containment jar. The braided

vinyl tubing was the only part of the aspirator that entered the cage. Mosquitoes were sucked

through the braided vinyl tubing and held inside the containment jar. Mosquitoes were not

damaged by the aspirator. Fluorescent marking powders and important morphological characters

used for species identification were not removed or destroyed during the aspiration process.









Escape Rates

The experimental cage was completely sealed as it would be used in the field with only the

baffle slit open (Experiments 2 and 3) or closed (Experiment 1). The cardboard cups that

contained one-hundred marked mosquitoes were placed inside the sealed, experimental outer

cage, on top of the inner cage. The cardboard cups were opened rapidly and the outer cage portal

and mesh were sealed before mosquitoes could escape the cage. Three escape experiments were

conducted: 1) no attractant with the baffle slit sealed, 2) no attractant with the baffle slit open,

and 3) baited with two restrained chickens with the baffle slit open. Chickens were individually

restrained inside mesh laundry bags that were cinched tight and allowed no movement of the

chickens. The mesh bag holding the chickens were suspended in the center of the inner cage.

Three night time and four day time non-baited, closed slit trials were conducted. Five night time

and seven daytime non-baited, open slit trials were conducted. Eight night time and day time

chicken baited, open slit trials were conducted. A Mosquito Magnet Pro (American

Biophysics, North Kingstown, RI) was operated three meters from the experimental cage. The

purpose of the Mosquito Magnet was to attract mosquitoes out of the experimental cage.

Any trials with greater than 15% marking mortality for a species were not included in the

means analysis. The mean percent of mosquitoes recovered for each species was calculated.

The percent of mosquitoes recovered from the exit trap and cage were calculated for the non-

baited open slit and chicken baited open slit trials. The Wilcoxon Rank Sum test was used to

compare the percent of mosquitoes captured in the exit trap and cage for the non-baited, open slit

and the chicken baited, open slit experiments. The Wilcoxon Rank Sum test (P = 0.025) was

used to compare the percent of blood-fed mosquitoes recovered in the cage and the exit trap.

This was done to determine if blood-fed mosquitoes were more likely to be captured in the exit

trap or the cage.









Entry Rates

To determine the best baffle angle for trapping the greatest number of mosquitoes, mark-

release-recapture experiments were conducted using six different baffle angles (0, 100, 200, 36,

48, and 560). The temporary baffle angles were created by attaching fiberglass mesh panels to

the baffle slit by Velcro and to the cage legs by thumb push pins. Sewing pins were used to

connect the mesh panels. The cage was baited with CO2 (flow rate of 500 mL/min) for six night

and day trials. The CO2 tank was placed outside the experimental cage and tubing that extended

into the middle of the inner cage released the CO2. A different baffle angle (0, 100, 200, 36,

48, and 560) was used for each trial where CO2 was used as bait. There were six night and day

trials with three restrained chickens used as bait and either 0 or 36 baffle angles. Two-hundred

and fifty female mosquitoes of each species were released at their respective times of activity.

Mosquitoes were released outside of the experimental cage. The number of mosquitoes that

were contained in the experimental cage and the exit trap at the end of each trial were recorded.

Cardboard cups that contained marked mosquitoes of each species were placed against the

back wall of the outdoor mosquito cage. The experimental cage was then baited with either CO2

or chickens. After the CO2 or the chickens were set in place the mosquitoes were released.

Mosquitoes were aspirated from the cage at the termination of each trial.

Results

Escape Trial Results

Mean recovery rates from the no-bait, slit closed trials were Cx. nigripalpus 89.4 2.4%,

Cx. quinquefasciatus 84.9%, An. quadrimaculatus 91.1% (Figure 2-7), Ae. albopictus 74.1 +

4.8%, and Ae. aegypti 86.6 2.8% (Figure 2-8).

Mean recovery rates from the no-bait, slit open trials were Cx. nigripalpus 74.0 + 8.4%,

Cx. quinquefasciatus 55.3 + 6.4%, An. quadrimaculatus 80.0 11.5% (Figure 2-7), Ae.









albopictus 65.6 5.8%, and Ae. aegypti 71.1 5.6% (Figure 2-8). Of the mosquitoes recovered,

the mean percent recovered from the exit trap were Cx. nigripalpus 43.6 13.3%, Cx.

quinquefasciatus 25.2 11.8%, An. quadrimaculatus 33.7 6.1%, Ae. albopictus 4.4 1.4%,

and Ae. aegypti 11.4 + 3.5%. Cage aspiration collections were significantly greater than cage

aspiration collections for Cx. quinquefasciatus, An. quadrimaculatus, Ae. albopictus, and Ae.

aegypti (Wilcoxon Rank Sum test, P < 0.025).

Mean recovery rates from the chicken baited, slit open trials were Cx. nigripalpus 77.7 +

6.8%, Cx. quinquefasciatus 47.4 10.9%, An. quadrimaculatus 65.8 12.1% (Figure 2-7), Ae.

albopictus 58.7 4.6%, and Ae. aegypti 57.0 5.5% (Figure 2-8). Of the mosquitoes recovered

the mean percent recovered from the exit trap were Cx. nigripalpus 28.8 9.2%, Cx.

quinquefasciatus 23.0 11.7%, An. quadrimaculatus 34.2 12.3%, Ae. albopictus 4.2 0.7%,

and Ae. aegypti 3.2 1.0% (Figure 2-9). Cage aspiration collections were significantly greater

than exit trap collections for Cx. nigripalpus, Cx. quinquefasciatus, Ae. albopictus, and Ae.

aegypti (Wilcoxon Rank Sum test, P < 0.025).

The mean percent of blood-fed mosquitoes caught in the cage compared to exit trap for the

chicken baited, open slit trials were Cx. nigripalpus cage 51.0 16.7% and exit trap 37.2

10.9%, Cx. quinquefasciatus cage 35.3 12.8% and exit trap 0 + 0%, An. quadrimaculatus cage

57.8 11.0% and exit trap 7.3 5.7%, Ae. albopictus cage 42.8 8.6% and exit trap 12.5 +

12.5%, and Ae. aegypti cage 42.6 8.7% and exit trap 0 + 0%. The percent of blood-fed

mosquitoes were significantly higher in the cage than exit trap for Cx. quinquefasciatus, An.

quadrimaculatus, Ae. albopictus, and Ae. aegypti (Wilcoxon Rank Sum test, P < 0.025).

Cage Entry Results

Mean percent recovery rates for the 0 baffle trials were Cx. nigripalpus 0.27 0.27%, Cx.

quinquefasciatus 0.14 0.14%, An. quadrimaculatus 1.61 0.83%, Ae. albopictus 0.4 0.0%,









and Ae. aegypti 0.4 0.23%. Mean percent recovery rates for the 36 baffle trials were Cx.

nigripalpus 0.14 0.14%, Cx. quinquefasciatus 0.14 0.14%, An. quadrimaculatus 1.34 +

1.34%, Ae. albopictus 0.13 0.13%, and Ae. aegypti 1.07 0.13%. No mosquitoes were

recovered from the exit trap during the cage entry trials.

Discussion

Trials with mosquito marking mortality greater than 15% were associated with older

females (>16 days post emergence) and with lowered blood-feeding success in the chicken

baited, open slit trials. This behavioral change and increased marking mortality suggested that

older (>16 days post emergence), colony-raised mosquitoes may not behave the same as younger

mosquitoes (<16 days post emergence). Therefore, trials with mosquito marking mortality

greater than 15% were not included in data analysis.

The Mosquito Magnet rarely caught mosquitoes that were released during trials. The

Mosquito Magnet did capture mosquitoes after the trial had ended as determined by analyzing

mosquitoes for fluorescent marking powders. The Mosquito Magnet data was not presented

because of the low capture rate of escaped mosquitoes during release trials.

The no-bait, closed slit trials represented the likelihood that mosquitoes would be captured

in the cage and recovered during the aspiration process. Mean recovery rates for Cx.

nigripalpus, Cx. quinquefasciatus, An. quadrimaculatus, and Ae. aegypti ranged from 84.9% to

91.1% (Figures 2-7 and 2-8). Aedes albopictus mean recovery rate was 74.1 4.8% (Figure 2-

8). Aedes albopictus may be more adept at escaping from the cage or avoiding aspiration than

the other four species. Colony raised Ae. albopictus appeared to be smaller than field caught

specimens (personal observation). Numerous Ae. albopictus individuals were caught in the

fiberglass mesh screen. It appeared as if the Ae. albopictus may have been trying to twist their









way through the mesh. If colony-raised Ae. albopictus can escape through the mesh, then this

might explain the lowered recovery rates for this species.

Mean recovery rates from the non-baited, open slit trials were less than non-baited, closed

slit trials (Ae. albopictus 8.5%, An. quadrimaculatus 11.5%, Cx. nigripalpus 15.4%, Ae. aegypti

15.5%, and Cx. quinquefasciatus 29.6%) (Figures 2-7 and 2-8). This shows that mosquitoes may

escape through the baffle slit. Culex quinquefasciatus appeared to be able to escape the cage

through the baffle slit. When there was no bait, lab-raised Cx. quinquefasciatus, An.

quadrimaculatus, Ae. albopictus, and Ae. aegypti were caught in the cage more often than the

exit trap (Wilcoxon Rank Sum test, P < 0.025). Aedes albopictus was observed to fly freely

between the exit trap and cage which shows that the exit trap collections may not contain all

mosquitoes that originally entered the exit trap.

Mean recovery rates from the chicken-baited, open slit trials were less than non-baited,

open slit trials for Ae. albopictus 6.8%, Cx. quinquefasciatus 7.9%, An. quadrimaculatus 13.8%,

and Ae. aegypti 14.1% (Figures 2-7 and 2-8). The two restrained chickens were not capable of

ingesting mosquitoes, which would have reduced recovery rates. Some mosquito species may be

more likely to escape the cage after blood-feeding. The percent mosquitoes blood-fed and the

percent captured in the cage compared to the exit trap were significantly higher for Cx.

quinquefasciatus, Ae. albopictus, and Ae. aegypti. These three colony raised species avoided

entry into the exit trap when blood-fed. No blood-fed Cx. quinquefasciatus or Ae. aegypti were

recovered from the exit trap. The percent of mosquitoes captured in the exit trap shows that Ae.

albopictus and Ae. aegypti are infrequently captured in the exit trap (Figure 2-9).

Culex nigripalpus had the highest percent of blood-fed mosquitoes in the exit trap (37.2 +

10.9%) and was the only species not to have significantly (Wilcoxon Rank Sum test, P > 0.025)









more blood-fed mosquitoes caught in the cage than the exit trap. These findings show that

blood-fed Cx. nigripalpus were more likely to be collected in the exit trap than the other four

species.

These results should be interpreted with caution because colony-raised mosquitoes may

behave differently from wild mosquitoes. These results should only be used to show the

potential range of differences that may exist among wild mosquito species with respect to their

ability to enter, fly within, and exit the cage.

Recovery rates of mosquitoes released in CO2 and chicken baited cage entry trials were

poor. Low recovery rates were contributed to many factors. It was very dry and hot during the

time periods that mosquitoes were released, with maximum daytime temperatures reaching

40.5C on many days. There was only one evening with rainfall (>20 mm). Increased

temperatures combined with reduced humidity levels may have reduced mosquito flight activity.

Within the outdoor mosquito cage there were numerous species of lizards and frogs that may

have consumed mosquitoes before they entered the experimental cage. Glue boards were used to

reduce lizard populations within the outdoor mosquito cage. Directly adjacent to the outdoor

mosquito cage were several trees that harbored populations of wild birds. Released mosquitoes

may have been more attracted to wild birds than the CO2 or chickens inside the experimental

cage. Time constraints only allowed a single day and night trial to be conducted for CO2 baited

cage entry trials which eliminated statistical analyses of these data.

Cage entry trials provided little insight to which baffle angle would maximize mosquito

catches in the field. An angle of 33 was chosen as the permanent baffle angle for the

experimental cage. This angle was chosen because it allowed a gap of 0.4 m from underneath

the front and rear edges of the baffle to the ground (Figures 2.2 and 2.4). The gap allowed









mosquitoes flying at ground level to 0.4m above ground level to approach the cage from all four

sides. The permanent baffle was constructed with 2" x 4" and 2" x 2" treated pine lumber

(Figure 2-4). Male Velcro was stapled to the baffle. Twelve-count mesh screen was attached

to the baffle by female Velcro.

Conclusions

The experimental cage was capable of containing released mosquitoes. Mosquitoes were

capable of exiting the cage through the baffle slit. Certain species such as Cx. quinquefasciatus,

Ae. aegypti, and Ae. albopictus were more adept at escaping from the experimental cage than Cx.

nigripalpus and An. quadrimaculatus. Culex nigripalpus, Cx. quinquefasciatus, and An.

quadrimaculatus were frequently recovered from the exit trap, whereas, Ae. aegypti and Ae.

albopictus were infrequently recovered from the exit trap.

Orange County Mosquito Control Division relies only on exit trap collections to assess

mosquito species diversity and abundance of mosquitoes attracted to sentinel chickens. The

experimental cages were used in the field to compare the number of mosquitoes successfully

collected in the exit trap to the number of mosquitoes collected in the cage (Chapter 6).









Table 2-1. Larval feeding regimen


Species
Cx. quinquefasciatus
Cx. nigripalpus
Ae. aegypti
Ae. albopictus
An. quadrimaculatus*


Slurry Ingredients
1%BLP and 1%BY
1%BLP and 1%BY
3% BLP and 2% BY
3% BLP and 2% BY
1% BLP, 1%BY, 1%HC
1% BLP, 1%BY, 1%HC
1% BLP, 1%BY, 1%HC**


BLP = bovine liver powder, BY
** powder


brewers yeast, HC


Amount Fed
20mL
20mL
50mL
50mL
50, 25mL
50mL
dusted


Feeding Schedule
daily
daily
every other day
every other day
first and third day
fifth and seventh day
fourth and sixth day


hog chow, fed slurry and powder,


Figure 2-1. Original sentinel chicken cage used by OCMCD



































Figure 2-2. Experimental sentinel chicken cage used for research


















12m












I -------------------------- -----------------------
.-


Figure 2-3. Line drawing of experimental cage A) Front view B) Side view


~ ------------------------------ --------------------- -------


13m


I~
,,
I~


II


~















Ior I







0
-j


A
Figure 2-4. Line drawing of experimental inner cage A) Front view B) Side view


























A B
Figure 2-5. Sentinel chicken cage exit trap A) Sliding sheet metal with permanently mounted portion of exit trap B) Removable exit
trap collection device













































figure z-o. AssemDlea aspirator


,...-~; R.















80


60


40


20


T
T


ST


T


0
0 -- -------- -- --------^--
Cx. nigripalpus Cx. quinquefasciatus An.
quadrimaculatus

O no bait, closed slit U no bait, open slit U chicken, open slit

Figure 2-7. Night escape trials percent mosquitoes recovered with no bait, closed slit; no bait, open slit; and chicken bait, open slit







































Ae. albopictus

O no bait, slit closed


* no bait, slit open


Ae. aegypti

* chicken, slit open


Figure 2-8. Day escape trials percent mosquitoes recovered with no bait, closed slit; no bait, open slit; and chicken bait, open slit
















80

0

S60
o ocage
U exit
o 40
E



g 20



0
Cx. nigripalpus Cx. quinquefasciatus An. quadrimaculatus Ae. albopictus Ae. aegypti

Figure 2-9. Percent mosquitoes recovered from experimental cage and exit trap with chicken bait and baffle slit open









CHAPTER 3
SENTINEL CHICKEN DEFENSIVE BEHAVIOR

Introduction

Studies conducted by Edman and others concerning the defensive behaviors of

Ciconiiformes against mosquitoes and the influence of these behaviors on Cx. nigripalpus blood

feeding success (Edman and Kale 1971, Edman et al. 1972a, Kale et al. 1972, Maxwell and Kale

1977, Webber and Edman 1972). Edman and Kale (1971), Edman et al. (1972a), and Kale et al.

(1972) showed the blood feeding success by Cx. nigripalpus to be influenced by host defensive

behaviors, host age, host species and mosquito density. Ciconiiform birds (i.e., Little Blue

Heron, Cattle Egret, Snowy Egret, and White Ibis) actively display host defensive behaviors after

sunset (Edman and Kale 1971), whereas the domestic white leghorn chickens used in this

experiment roost at sunset and do not display defensive behaviors until awakening at sunrise

(Appleby et al. 2004). These differing patterns in host defensive behavior may influence the

blood feeding success of Cx. nigripalpus.

Edman et al. (1974) used unrestrained chickens of various ages to look at the blood feeding

success of wild Cx. nigripalpus. Three experiments exposed two-week old chicks, eight-week

old chicks, and adult chickens to 200 host-seeking Cx. nigripalpus. Culex nigripalpus recovery

rates and blood feeding success increased with older chickens. Recovery rates and percent of

blood-fed Cx. nigripalpus were: two-week old chicks 79% and 24%, eight-week old chickens

84% and 59%, and adult chickens 90% and 78%. Chickens six to eight weeks old were found to

be moderately tolerant to Cx. nigripalpus when compared to other species of birds and mammals

(Edman et al. 1974). Sentinel chickens used for arboviral surveillance in Orange County, FL are

a minimum of 12 weeks old when first placed in the field (OCMCD data). In Orange County

sentinel chickens are kept in the field for approximately one year. Therefore, it is important to









evaluate mosquito interactions with chickens at an age representative of sentinel chickens used

for arboviral surveillance in Orange County, FL. For the research presented here, chicken

defensive behaviors are defined as any behavior that causes mortality to mosquitoes or interferes

with mosquito blood feeding success.

The present study was designed to assess the effects of adult chickens (25 to 28 weeks old)

defensive behaviors on the recovery rates and blood feeding success of Cx. nigripalpus and

Aedes albopictus (Skuse) during their normal flight and activity periods. Exit traps were added

to sentinel chicken cages in Orange County, FL in 2003. These exit traps capture mosquitoes as

they fly upward from the cage. Exit traps provide an assessment of the mosquito species that

approach sentinel chickens. In the field, exit traps can only collect mosquitoes that enter the

cage, survive chicken defensive behaviors, and enter the exit trap. Exit trap collections may not

accurately quantify the mosquito species attracted to sentinel chickens if chicken defensive

behavior adversely impacts the survival of host-seeking mosquitoes. The effects of adult sentinel

chicken defensive behavior on mosquito recovery rates and the location of recovery within the

cage were assessed. This assessment will provide insight into how sentinel chicken defensive

behaviors might influence mosquito exit trap collections in the field.

Materials and Methods

Mosquito Rearing

Too few wild Cx. nigripalpus were captured in Orange County to produce enough F

progeny for the experiments and time constraints limited the creation of an entirely new colony.

Therefore, a cross mating of colony and wild Cx. nigripalpus were used for the present study.

Wild, virgin male Cx. nigripalpus and colony, virgin females were cross mated, as this was

proven the most successful mating combination by Haeger and O'Meara (1970). Wild, blood-

fed Cx. nigripalpus females were obtained from Orange County on May 21, 2006 by using white









leghorn chicken-baited lardcan traps (Bellamy and Reeves 1952) and sentinel chicken cage exit

traps. Chickens used as bait for field collected Cx. nigripalpus were obtained by OCMCD.

Mosquitoes were aspirated from traps, placed in cardboard cups, provided a 10% sugar water

solution, and transported to Gainesville. Once in the laboratory, the field-collected mosquitoes

were immobilized with chloroform and placed on a chill table maintained at 40C. Chilled

mosquitoes were identified to species and all mosquitoes that were not blood-fed Cx. nigripalpus

females were discarded.

Blood-fed Cx. nigripalpus were placed into a 61 cm x 46 cm x 61 cm cage. All adults

were provided cotton soaked in 10% sugar water solution. Mosquitoes were maintained at 26.5

2.5C and 90 10% RH in a room exposed to the natural outdoor light cycle. Larvae were fed

a 20 mL slurry of 1% bovine liver powder and 1% brewers yeast once per day. Adults were

provided blood meals from restrained chickens of various breeds (IACUC #D469).

Colony and wild Cx. nigripalpus were offered restrained chickens to blood feed from on

the same day. Five days after blood feeding, 5 to 7 day old, hay-infusion in black plastic cups

was provided for wild and colony adult, gravid female oviposition. This blood feeding and

oviposition schedule synchronized the emergence of adult colony and wild Cx. nigripalpus. Ten

hours after emergence, adult mosquitoes were placed in a walk-in cooler (Eskimo Panels Inc.,

Miami, FL) at 40C to temporarily immobilize them. Immobilized mosquitoes were aspirated

from the cage, transported to the chill table, and sorted by sex. Wild Cx. nigripalpus males were

placed into a 61 cm x 46 cm x 61 cm cage and wild females were discarded. This was done three

times every 10 hours with each cohort. The same method was used to obtain colony females and

discard colony males. By removing recently emerged Cx. nigripalpus from the emergence cage

every ten hours it was assured that males would not be able to inseminate colony females since









there is a 12 to 24 hour period required for male genitals to rotate into the proper position for

copulation. Colony females were placed into the same 61 cm x 46 cm x 61 cm cage with the

wild males. A 40-watt incandescent bulb on a dimmer switch was used to create an artificial

sunset in the room that would extend dusk, the primary period of mating activity, one hour past

sunset. The Fl progeny of this wild x colony cross mated and produced viable offspring. The

F2 progeny were used for the experiments reported here.

Wild Ae. albopictus and other mosquitoes were aspirated as they landed on a human host

at a field site in Orange County between the hours of 11:00 am and 1:00 pm on May 22, 2005.

Mosquitoes were transported to Gainesville in cardboard cups with 10% sugar water solution

soaked cotton balls. Mosquitoes were temporarily immobilized using 500 mL/min of CO2

lightly blown over them. Aedes albopictus were placed into a 41 cm x 41 cm x 46 cm cage and

all other mosquito species were discarded. Adult Ae. albopictus were offered restrained chickens

of various breeds for blood meals (IACUC #D469). Eggs were collected on moistened brown

paper, dried, and stored for approximately one week in a cool, dry Ziploc bag. Adults were

raised on 10% sugar water solution. Mosquitoes were maintained at 28 1C and 90 4% RH

on a 14: 10 L: D cycle. Larvae were fed a 50 mL slurry of 3% bovine liver powder and 2%

brewers yeast provided every other day. The F2 progeny were used for the experiments reported

here.

Chickens used for routine colony feeding were of various breeds and were obtained by the

USDA-CMAVE. All chickens used for colony feeding were housed at the USDA-CMAVE.

Mark-Release-Recapture Studies

Mosquitoes were aspirated from emergence cages with a hand-held aspirator (Hausherr's

Machine Works, Toms River, N.J.). They were temporarily immobilized with CO2 at a rate of

500 mL/min. The exact number of females required for each experiment were counted and









placed into cardboard cups with screen lids. The inside of the cups were lightly dusted with red,

white, blue, yellow, green, or orange fluorescent powder (Pylam Products Co., Tempe, AZ).

Each trial and cage was assigned a specific color for that day or night. A water-moistened cotton

ball was placed on top of the screen to keep the humidity high. Mosquitoes were given at least

30 minutes to recover from the CO2 before being released in trials. The cup was placed on top of

the inner experimental cage, the lid rapidly removed, and the cage portal and mesh sealed. At

the completion of the trial all mosquitoes were aspirated from the cage. Any mosquitoes that

were dead inside the cup were noted and excluded from statistical analysis. Mosquitoes were

observed under black light (Adams Apple Distributing LP, Glenview, IL) to determine which

trial they were released. Mosquito blood-fed status was recorded. Partial blood meals were

counted as blood-fed. The location where each mosquito was recovered in the cage at the end of

each trial was recorded. All statistical analyses were performed using Wilcoxon Rank Sum test

(Ott and Longnecker 2001).

Chickens

Chickens used for behavior experiments were 25 to 28 week old white leghorns obtained

by Orange County Mosquito Control (OCMCD). All chickens used for experiments were cared

for under #D996 University of Florida IACUC protocols and were housed on the University of

Florida main campus behind the Entomology and Nematology building in Gainesville, FL.

Chickens, housed behind the Entomology and Nematology building, were exposed to natural

populations of mosquitoes prior to experimental exposure.

Host Defensive Behavior

Chicken behavior was monitored during three different time periods (sunset, sunrise, and

11:00 am). Chickens were observed at distances of 26.5 m or 12.2 m, dependent on light levels

and visibility, with 10 x 42 Eagle Optics Ranger Platinum Class binoculars. Some behaviors









were also observed at close range (i.e., eye blink and the physical consumption of mosquitoes)

while mosquitoes were aspirated from the sentinel cages. Host defensive behaviors were

observed immediately after the release of Cx. nigripalpus until all the chickens roosted; at

sunrise for 30 min with Cx. nigripalpus; and for 30 min immediately after release ofAe.

albopictus. Chicken behavior was observed for several minutes prior to release of mosquitoes to

determine routine maintenance behaviors. Every type of maintenance and defensive behavior

displayed was recorded. Defensive behaviors observed are presented in Table 3-3. Descriptive

comparisons were made among the types of behaviors displayed in response to the various

species of mosquitoes, densities of mosquitoes and densities of chickens (Table 3-4).

Effect of Host Defensive Behavior

Host defensive behavior trials were conducted in outdoor cages at the U.S.D.A. -

C.M.A.V.E. facility in Gainesville, FL between July 26 and August 13, 2005. One chicken was

placed in each sentinel cage 1 to 2 h before mosquitoes were released. Three-hundred 6 to 14-

day-old adult Cx. nigripalpus females were released into each cage at sunset. This experiment

was replicated four times. Mosquitoes were recovered from the cages at 8:00 am the following

morning.

Three-hundred 7 to 12-day-old adult Ae. albopictus females were released into each cage at

11:00 am. This experiment was replicated four times. Aedes albopictus females were recovered

by aspiration from the cage at 5:00 pm at the end of each trial.

For trials with each mosquito species, one cage contained an unrestrained chicken that was

allowed total freedom of movement. The second cage contained a restrained chicken that was fit

snugly into a nylon mesh laundry bag that restricted any head, wing, or leg movement. Three

individual chickens were used for these experiments. Chickens were rotated between the

restrained and unrestrained treatments and between cages. Cages containing restrained and









unrestrained chickens were rotated between each trial to negate differences in the ability of each

cage to contain mosquitoes. All statistical analyses were performed using Wilcoxon Rank Sum

test (Ott and Longnecker 2001).

To evaluate the importance of mosquito and vertebrate host density on mosquito blood

feeding success one chicken was placed into one of the cages and four chickens were placed into

the second cage one or two hours before the mosquitoes were released. An equal predetermined

(100, 500, or 700) number of 6 to 15-day-old adult Cx. nigripalpus females were released into

both cages at sunset. Released mosquitoes were aspirated from both cages at 8:00 am the

following morning. Four trials were conducted with each density of mosquitoes. Five individual

chickens were used for these experiments. Chickens were rotated so that no chicken was used

more than once by itself in the single chicken cage at each mosquito density. Cages containing

one and four chickens were rotated between each trial to negate differences in the cages ability to

contain mosquitoes. All statistical analyses were performed using Wilcoxon Rank Sum test (Ott

and Longnecker 2001).

Results

Chicken defensive behaviors had no significant effect on the percent of Cx. nigripalpus

females recovered (Wilcoxon Rank Sum test, P > 0.05), their blood feeding success (Wilcoxon

Rank Sum test, P > 0.025), or the location in the cage from which the mosquitoes were recovered

(Wilcoxon Rank Sum test, P > 0.025).

Chicken defensive behaviors were found to significantly lower the blood feeding success

ofAe. albopictus females (Wilcoxon Rank Sum test P < 0.025) (Table 3-1); however, these

behaviors had no significant effect on the percent ofAe. albopictus recovered (Wilcoxon Rank

Sum test, P > 0.05) or the location in the cage from which the mosquitoes were recovered

(Wilcoxon Rank Sum test, P > 0.025) (Table 3-1).









No significant differences were found between the one and four chicken trials at any

mosquito density (100, 500, or 700) for the percent of Cx. nigripalpus recovered (Wilcoxon

Rank Sum test, P > 0.05), the percent that were blood-fed (Wilcoxon Rank Sum test, P > 0.025);

or the location in the cage from which the mosquitoes were recovered (Wilcoxon Rank Sum test,

P > 0.025). Mosquito blood feeding success in trials with one chicken were significantly

different between the 100 and 700 mosquito densities and between the 100 and 500 mosquito

densities (Wilcoxon Rank Sum test, P < 0.025).

An inverse correlation was observed as Cx. nigripalpus density increased, blood feeding

success decreased when one chicken was present (Table 3-2). When one chicken was present the

percent of Cx. nigripalpus recovered from the exit trap differed significantly between the 300

and 500 densities and between the 500 and 700 densities (Wilcoxon Rank Sum test, P < 0.025)

with the 500 density being significantly less than the 300 and 700 densities (Table 3-2).

Discussion

Anderson and Brust (1996) exposed pairs of Japanese Quail (Coturnixjaponica) to Cx.

nigripalpus or Ae. aegypti and video tapped Quail activity before and after the release of

mosquitoes. They concluded that anti-mosquito behaviors were exaggerations of pre-existing

behavioral patterns (Anderson and Brust 1996). This increase of normal body maintenance

behaviors in response to mosquito (Cx. nigripalpus) biting pressure has also been observed with

ciconiiform birds (Webber and Edman 1972). In the current experiments, it appeared that

chickens used routine maintenance behaviors as defensive behaviors in response to mosquito

biting pressure. Through direct observation it was shown that all of the maintenance behaviors

displayed in the absence of mosquitoes were displayed more often in the presence of mosquitoes.

The preening and head scratching behavior observed in the present study may have been in

response to mosquito or louse biting pressure. Menopon gallinae (Linnaeus) (Phthiraptera:









Philopteridae), the shaft louse, was occasionally found on the chickens used for these

experiments. Estimates of louse load were not made in the study but symptoms of heavy

infestation (i.e., feather loss and lameness) were never observed. Whether preening and head

scratching behaviors were in response to mosquitoes or lice could not be determined. Preening

and head scratching, however, occurred more often in the presence of mosquitoes than in their

absence. The only defensive behaviors not observed in the absence of mosquitoes were the wing

shrug and head tuck. Wing shrugging behavior was only displayed when four chickens were

present at the 700 mosquito density, suggesting that it may only be reserved for periods of highly

elevated agitation.

Several times in the commentary section of Table 3-4 defensive behavior was described as

occurring often in the presence of a particular mosquito species. This means that while the

behavior was displayed in the presence of both mosquito species, it was displayed much more

frequently in the presence of the mentioned species and therefore is attributed to that species.

Aedes albopictus appeared to attempt to blood feed around the face and head regions due to the

many defensive behaviors displayed by the chickens such as the head rub, head shake, head jerk,

head tuck, pecking at the air, and eye blink. Culex nigripalpus appeared to attempt to feed

around the legs and feet because the primary responses elicited from chickens in the presence of

this mosquito species were the foot kick/stamp and pecking at the ground and legs. Webber and

Edman (1972) observed that Cx. nigripalpus primarily feed around the leg region of ciconiiform

birds.

Chicken defensive behaviors had a negative influence on the blood feeding success

(Wilcoxon rank sum P < 0.025) and caused high mortality ofAe. albopictus (Table 3-1). Similar

results were found by Waage and Nondo (1982) in restraint trials with rabbits and Ae. aegypti.









Klowden and Lea (1979) found unrestrained rabbits significantly lowered the percent blood

feeding success of wild, host seeking mosquitoes when compared to restrained rabbits. All three

of these experiments were conducted during the photophase period of mosquito host seeking

activity and coincided with diurnally active time periods of their hosts. In the experiments

reported here, unrestrained chickens displayed defensive behaviors in response to mosquito

biting pressure throughout the entire duration of the trials with Ae. albopictus. It could not be

determined whether the constant chicken activity was due to the time of day or the host feeding

behaviors ofAe. albopictus. The increased activity of chickens during the daytime, the reduced

(80.0 4.1% restrained > 40.1 + 12.5% unrestrained) recovery rate ofAe. albopictus and

significantly (Wilcoxon Rank Sum test, P < 0.025) lowered blood feeding on unrestrained

chickens indicate that unrestrained sentinel chickens in the field may reduce the number of

daytime biting mosquitoes that enter the cage and reduce the likelihood of them obtaining a

blood meal.

Culex nigripalpus trials with restrained and unrestrained chickens showed a slight

reduction in recovery rates of mosquitoes and no differences in blood feeding success (Table

3-1). These results differ from the restraint trials conducted by Edman and Kale (1971) with

nocturnally active ciconiiform birds and Cx. nigripalpus. Unrestrained chickens displayed very

few defensive behaviors when Cx. nigripalpus were released at sunset because chickens roost at

sunset and wake at sunrise (Appleby et al. 2004). However, unrestrained chickens displayed

more defensive behaviors shortly after sunrise than at sunset. Day and Edman (1984) found that

mosquito blood feeding success occurs more often during host seeking periods that overlap with

periods of inactivity by hosts. Since both the restrained and the unrestrained chicken were









inactive throughout the night, when Cx. nigripalpus feeds, there were no differences in blood

feeding success between the two treatments.

Trials in which one chicken was exposed to increasing densities of Cx. nigripalpus showed

a significant negative correlation (Wilcoxon Rank Sum test, P < 0.025) between the number of

mosquitoes released and the percent that successfully obtained blood meals (Table 3-2). This

negative correlation between increasing density and blood meal success was counterintuitive. If

chickens were inactive during the primary host seeking period of Cx. nigripalpus then there

should have been no significant difference in blood feeding success at the various densities. At

the 500 and 700 Cx. nigripalpus densities little defensive behavior activity was displayed by

chickens at sunset, but at sunrise chickens were very active. The bulk of chicken morning

activities involved pecking at their legs and the ground. It was observed at close range that

pecking at the ground was done in an attempt to consume mosquitoes. After Cx. nigripalpus had

engorged it would usually travel a short distance to the nearest resting spot. Dozens of blood-fed

Cx. nigripalpus were found resting on the inner cage where they were accessible to chickens. If

chickens peck and consume primarily blood fed mosquitoes resting near them, then the chickens

would skew the percent of blood fed mosquitoes that were recovered. The heightened activity of

chickens and pecking behavior in the early morning make this the most likely time period during

which Cx. nigripalpus were consumed. In field settings, this means that mosquitoes should be

aspirated shortly after sunrise to reduce the likelihood of mosquitoes being consumed by the

sentinel chickens.

No significant trend was found between blood meal success or recovery rates and mosquito

density when four chickens were exposed to varying densities of Cx. nigripalpus. Roughly 80%

percent of all Cx. nigripalpus that were released were recovered. Although there were no









significant differences found when comparing the recovery rates and blood feeding success

between one and four chickens, with four chickens there was an increase in the percent of blood-

fed mosquitoes recovered at the 500 (45.7%) and 700 (56.7%) mosquito density trials compared

with trials containing one chicken at the 500 (35.8%) and 700 (35.7%) mosquito densities.

Chickens in groups of four appeared more concerned with the presence of other chickens than

with the presence of mosquitoes. Fewer defensive behaviors toward mosquitoes were observed

in cages containing four chickens than cages containing one chicken. The four chickens would

peck at each other to assert dominance. Aggression toward other chickens likely distracted

chickens from the presence of mosquitoes which allowed more mosquitoes to obtain blood meals

and avoid consumption by chickens. In field settings, four or more chickens should be exposed

in the same cage to increase the survival and number of day feeding mosquitoes that obtain blood

meals.

Exit trap collections were not significantly different between restrained and unrestrained

chicken trials for Cx. nigripalpus orAe. albopictus (Table 3-1). There were no discernable

patterns of exit trap recovery rates at different densities of Cx. nigripalpus (Table 3-2). It does

not appear that chicken defensive behaviors influence the percent of mosquitoes captured in the

exit trap.

An important point to note is the lower percent ofAe. albopictus (8.1%) recovered in the

exit trap during restrained chicken trials compared to the percent of Cx. nigripalpus (41.3%)

recovered during similar trials. This difference in the percent of mosquitoes recovered between

Ae. albopictus and Cx. nigripalpus during a time when chickens could not display defensive

behaviors, demonstrates that Ae. albopictus is not as likely to be recovered in the exit trap as Cx.

nigripalpus. Exit trap collections from sentinel chicken cages in the field may not accurately









represent all mosquito species or the abundance of those species that come to feed on the sentinel

chickens. The absence, or low abundance of certain mosquito species in exit traps collections

from sentinel chicken cages does not exclude these species from being potential enzootic vectors

of viral disease in avian populations.

Conclusions

Chicken defensive behaviors are maintenance behaviors displayed at greater frequencies in

the presence of mosquitoes. Chickens may display different defensive behaviors depending

upon the preference of the mosquito blood feeding site. Sentinel chicken defensive behaviors

can influence mosquito recovery rates and the blood feeding success of recovered mosquitoes.

Diurnally active mosquitoes were more likely than nocturnal mosquitoes to be consumed and

have reduced blood feeding success. Increasing densities of mosquitoes feeding on chickens can

influence the percent of blood-fed mosquitoes recovered when one chicken was present. At least

four chickens should be exposed at sentinel sites to reduce the number of mosquitoes that are

consumed and increase the number of mosquitoes that obtain blood. Exit trap collections did not

appear to be influenced by chicken defensive behaviors. Exit trap collections may not accurately

sample all of the mosquito species that are attracted to sentinel chicken surveillance cages.









Table 3-1. Effect of host defensive behavior on recovery, blood feeding success, and location recovered*
Percent mosquitoes recovered Percent mosquitoes engorged Percent mosquitoes recovered from exit trap
Species Restrained Unrestrained Restrained Unrestrained Restrained Unrestrained
Cx. nigripalpus 90.2 3.1 77.7 11.0 50.2 7.2 52.7 8.6 41.3 10.3 56.5 3.8
Ae. albopictus 80.0 4.1 40.1 + 12.5 22.2 3.4a 6.1 + 1.9b 8.1 + 1.4 18.9 6.7
* Means and standard errors of recovered mosquitoes as a percent of total mosquitoes released, engorgement and exit trap as a
percent of mosquitoes recovered.

Table 3-2. Effect of Cx. nigripalpus density on recovery, blood feeding success, and location recovered when one or four chickens
were present*
Percent mosquitoes recovered Percent mosquitoes engorged Percent mosquitoes recovered from exit trap
Density One chicken Four chickens One chicken Four chickens One chicken Four chickens
100 74.2 + 12.2 80.1 + 3.6 58.9 + 4.2a 58.2 + 5.8 54.2 + 6.9a,b 43 10.9
300a 77.7 + 11.0 -- 52.7 8.6a,b -- 56.5 3.8a --
500 86.3 + 1.7 76.2 + 6.9 35.8 + 4.1b 45.7 + 8.8 37.8 5.2b 45.2 + 7.2
700 82.6 + 3.3 85.6 + 1.9 35.7 + 6.2b 56.7 + 8.5 67.1 + 3.6a 63.0 + 2.5
* Means and standard errors of recovered mosquitoes as a percent of total mosquitoes released, engorgement and exit trap as
percent of mosquitoes recovered.
a Data from unrestrained trials presented earlier.









Table 3-3. Behaviors
Behavior
Eye blink
Head jerk
Peck

Head shake
Head scratch
Foot kick/stamp
Wing flap
Hind feather ruffle
Preening
Head rub
Head tuck

Wing shrug


displayed by chickens in response to mosquitoes
Description
An eye was closed and opened rapidly
Head was moved forward and backward or side to side in a rapid jerking motion
The head was moved forward and backward in a pecking manner
The behavior was used to peck at the air, ground or legs
The head was vigorously shaken side to side
The foot was used to scratch the neck and head region
The foot was quickly raised into the air and lowered
The wings were spread out and retracted in a manner similar to flapping
The rear portion of the chicken shook, ruffling the feathers outward away from the body
Chicken used its beak to scratch amongst feathered regions
Chicken rubbed its head against vigorously against its body
Chicken buried its head beneath its wing for extended for several minutes without preening
Chicken lowered its body close to the ground and shrugged its wings upward and kept wings pressed
against its body









Table 3-4. Defensive behaviors displayed by chickens in the presence of mosquitoes with commentary


Behavior
Eye blink
Head jerk
Peck

Head shake

Head scratch

Foot kick/stamp
Wing flap

Hind feather
ruffle
Preening

Head rub

Head tuck


Wing shrug


Comments
This behavior occurred when Ae. albopictus flew near the chicken's face.
A head jerk was often performed in response to Ae. albopictus presence.
Pecking at the air was often performed in response to Ae. albopictus presence.
Pecking at the ground or legs was often performed in response to Cx. nigripalpus.
This was most often performed in the presence ofAe. albopictus. It was probably done in attempt to
dislodge a feeding mosquito from the face, head or neck region.
This behavior may have been performed in response to mosquitoes or to the shaft louse, Menopon
gallinae(L.). This behavior was observed more frequently in the presence ofAe. albopictus.
This behavior was frequently displayed at densities of 500 and 700 Cx. nigripalpus.
This behavior was usually observed when a chicken repositions itself before going to roost and may
not have been in response to mosquito attack.

This behavior was performed in response to Ae. albopictus and Cx. nigripalpus.
This may have been a response to mosquitoes or lice. Preening was performed over all feathered
regions of the body and in the presence of both mosquito species.
This behavior was displayed in response to Ae. albopictus in attempt to dislodge mosquitoes from the face
and head regions.
This behavior was displayed only in presence ofAe. albopictus and was likely done to prevent mosquitoes
from feeding near face and head regions.
This behavior was only observed at the 700 density of Cx. nigripalpus with four chickens present.









CHAPTER 4
DESCRIPTION OF FIELD SITES: ARBOVIRUSES/VECTORS/HOSTS

Introduction

Orange County Mosquito Control Division (OCMCD) conducts arbovirus surveillance

with sentinel chickens year round. There are twelve sentinel chicken sites maintained by

OCMCD in Orange County. Two of the twelve sentinel chicken sites were chosen for

experimental sentinel chicken surveillance and mosquito collection for the present study. The

two sites were Tibet-Butler Preserve (TBP) and Moss Park (MP).

Three factors led to the decision to conduct field research at TBP and MP. The first factor

was frequent seroconversions of sentinel chickens from October through June. Field studies

were conducted from October 12, 2005 to July 4, 2006. Historically, several EEEV and WNV

sentinel chicken seroconversions have occurred during October through June in Orange County.

The second factor for site selection was higher mosquito species diversity as determined by

OCMCD sentinel chicken exit trap collections from 2003 and 2004. Mosquito species diversity

ranged from fourteen (Winter Park and Clarcona-Horseman Park) to twenty-nine (Fort

Christmas) mosquito species in sentinel chicken exit trap collections (OCMCD data). The third

factor for site selection was lower mosquito abundance as determined by sentinel chicken exit

trap collections from 2003 and 2004. Sentinel chicken exit traps at some locations in Orange

County have collected 60,000 (Blanchard Park) Cx. nigripalpus in one night (September 21,

2004; OCMCD data). Lower mosquito abundance in exit traps would reduce the cost of testing

mosquitoes for arboviruses for this project.

Tibet-Butler Preserve and MP are managed by Orange County Parks and Recreation

(OCPR). Both parks are located outside of the Orlando metropolitan area. As the human

population continues to expand and spread away from Orlando, the parks have become









surrounded by residential areas. While both parks allow the preservation of unique habitats and

provide excellent opportunities for human recreation, they are also reservoirs for large numbers

of birds and mosquitoes which may circulate arboviruses that affect humans. The close

association of humans and wildlife at both parks calls for in-depth study to understand the

ecology of arbovirus transmission at these sites. Understanding arbovirus transmission cycles at

TBP and MP would be beneficial for OCMCD and the Orange County Public Health Department

as they attempt to reduce the risk of arbovirus transmission to humans.

Orange County is located along the Atlantic Flyway, a major bird migration route along

the east coast of the United States (Lord and Calisher 1970). The hot and moist environment of

Orange County provides suitable habitat for migrant birds, transient (birds migrating further

south or north that use Florida as a temporary stopover site) birds, and resident birds. Some

birds, such as the Red-winged Blackbird and Wood Stork, have populations that make north-

central Florida their permanent area of residence, while other populations of the same species use

this area to overwinter (Favorite 1960).

Site Descriptions

Tibet-Butler Preserve

The 438 acres known as the TBP is owned by the South Florida Water Management

District. Tibet-Butler Preserve is located in the southwest corner of Orange County along the

Tibet-Butler chain of lakes (Figure 4-1). Public activity within the park is limited to daytime

use, which reduces the potential for arbovirus transmission to humans by nocturnally active

mosquitoes inside park boundaries. Immediately to the south of the park are active, drip-

irrigated orange groves, and undeveloped land (Figure 4-2). Adjacent to the southwest corner of

TBP is residential housing. Adjacent to the east and west of TBP, orange groves have been









converted to high-end housing communities. The preserve abuts Lake Tibet to the northeast.

Houses are located along the banks of Lake Tibet.

Many habitats are present in TBP including bayhead and cypress swamps, freshwater

marsh, scrub, and pine flatwoods. The bayhead swamp is dominated by loblolly (Gordonia

lasianthus), red (Persea borbonia), and sweet bay (Magnolia virginiana) trees and comprises the

northwestern portion of the park. The freshwater marsh is dominated by sawgrass (Cladium

jamaicense) and waterlily (Nymphaea odorata). Numerous aquatic plant species, such as Water

Lettuce (Psitia stratiotes), Water Hyacinth (Eichhornia crassipes), Maidencane (Panicum

hemitomon), Cattail (Typha spp.), Pickerelweed (Pontederia cordata), and several sedge species

were found in the marsh and lake margins of TBP. These aquatic plant species create ideal larval

habitat for Cq. perturbans, Ma. titillans, and Ma. dyari (Morris et al. 1990, Lounibos and Escher

1985, Slaff and Haefner 1985, Callahan and Morris 1987).

To the south of Lake Tibet is a cypress dome where large sections of trees were knocked

down by the hurricanes of 2004. Upturned cypress (Cupressus spp.) tree roots create ideal

breeding habitat for mosquitoes, especially Cs. melanura (Foster and Walker 2002). To the

southwest of Lake Tibet is an undamaged cypress dome. In the eastern half of the park, there is

a mixed wetland forest with pine (Pinus spp.) flatwoods, bayhead and cypress swamp

communities. Various pine, bay, and cypress trees create the forest canopy depending on the

ground elevation. Various areas of the east side of the park flood and drain throughout the year

which provides temporary and semi-permanent pools of water, ideal habitat for Aedes,

Psorophora, Culex, and Anopheles larvae (Foster and Walker 2002, personal observation,

OCMCD data). The east side of the park is adjacent to residential neighborhoods along the









easternmost edge and Aedes albopictus larvae have been found in containers inside this area

(personal observation).

The southwest portion of the park is xeric, elevated two to three feet above the swamp,

composed of scrub and pine flatwood communities, and an understory dominated by saw

palmetto (Serenoa repens). The OCMCD sentinel chickens and those used in this investigation,

were located in this section of the park near the visitor center. The chickens were held in areas

cleared as fire breaks. In the southwestern corner of the park, there was a small section of oak

hammock dominated by Turkey (Quercus laevis), Live (Quercus viginiana), and Chapman's

(Quercus chapmanii) oak. Mosquito larvae were rarely found on this side of the park (OCMCD

data). Aedes albopictus have been found in containers near the visitor center (personal

observation).

Moss Park

Moss Park is located in the southeast portion of Orange County between Lakes Hart and

Mary Jane which are part of the St. John's Water Management District (Figure 4-1). The

surrounding banks of both Lakes Hart and Mary Jane have residential housing. Moss Park is a

peninsula of land with Lake Hart to the west, Lake Mary Jane to the east, and Pickerelweed

marsh establishing the southern boundary of the park. Moss Park is predominantly mature

Laurel (Quercus hemisphaerica) and Live Oak hammocks, and Pine trees with the understory

managed for human recreational activities. In the southern half of the park camping is allowed

and many visitors take advantage of this throughout the year. To the west of Moss Park and

south of Lake Hart is Split Oak Preserve. Split Oak Preserve consists of pine flatwoods and oak

hammocks interspersed with small ponds. Moss Park and Split Oak Preserve together span 3351

acres (Figure 4-3). There were two locations at MP that experimental sentinel chickens were

housed. The original experimental cage location was toward the back half of the park near the









pickerel reed marsh (Figure 4-3). After five trap nights, the cage had to be removed by request

of the park manager. It was relocated in a fenced off area near the county sentinel chickens.

At times during the wet season, Moss Park will flood in various regions along the marsh

and lake edges. This periodic flooding creates temporary and semi-permanent pools, ideal

habitat for Culex, Anopheles, Aedes, and Psorophora larvae (Foster and Walker 2002, OCMCD

data). For the rest of the year larval habitats are restricted to the marsh and shallow edges of

Lakes Hart and Mary Jane (OCMCD data, personal observation). Lakes Hart and Mary Jane

contain many aquatic plant species including Cattail, Pickerelweed, Arrow arum (Peltandra

virginica), Maidencane, Water Hyacinth, and numerous sedge species. These plants create ideal

larval habitat for Cq. perturbans, Ma. titillans, and Ma. dyari (Morris et al. 1990, Lounibos and

Escher 1985, Slaff and Haefner 1985, Callahan and Morris 1987). Anopheles spp. and Cx.

erraticus larvae have been found along the edges of Lake Mary Jane. Several areas of the park

are designated as repositories for human refuse which provides larval habitats for Ae. albopictus

and other container inhabiting mosquitoes.

Arboviruses

Tibet-Butler Preserve

Sentinel chicken surveillance began at TBP in 1995. Eastern equine encephalitis virus

transmission at TBP was the highest of the twelve sentinel chicken surveillance sites operated by

OCMCD. From 1995 through 2004, the mean annual seroconversion rate (MASR) of sentinel

chickens for EEEV at TBP was 33.6% (44/131). Mean annual seroconversion rates for EEEV at

the other eleven sites ranged from 0 to 16.5% (43/261) (Table 4-1). In 2002, four often

seroconversions to EEEV occurred between January and March at TBP.

St. Louis encephalitis virus transmission at TBP was the seventh highest of the twelve

sentinel chicken surveillance sites operated by OCMCD. From 1995 through 2004, the MASR









of sentinel chickens for SLEV at TBP was 5.3% (7/131). All SLEV transmission to sentinel

chickens occurred in 1996 and 1997. Mean annual seroconversion rates for SLEV at the other

eleven sites ranged from 0 to 17.4% (8/46) (Table 4-2).

West Nile virus transmission at TBP is the sixth highest of the twelve sentinel chicken

surveillance sites operated by OCMCD. From 2001 through 2004, the MASR for WNV was

30.8% (20/65). However, WNV was not detected by sentinel chickens in Orange County until

2001, so this data only represents four years of surveillance. Mean annual seroconversion rates

for WNV at the other eleven sites ranged from 7.7% (2/26) to 40% (16/40) (Table 4-3). In 2002

and 2003, four of twenty-one and four of twenty WNV seroconversions occurred between

October and December at TBP.

Moss Park

Sentinel chicken surveillance began at MP in 1991. Eastern equine encephalitis virus

transmission at MP was the eighth highest of the twelve sentinel chicken surveillance sites

operated by OCMCD. From 1991 through 2004 the MASR of sentinel chickens for EEEV was

4.5% (5/111). Mean annual seroconversion rates for EEEV at the other eleven sites ranged from

0% to 33.6% (44/131) (Table 4-1).

St. Louis encephalitis virus transmission at MP was the sixth highest of all twelve sentinel

chicken surveillance sites operated by OCMCD. From 1991 through 2004 the MASR of sentinel

chickens for SLEV was 7.2% (8/111) with all seroconversions occurring in 1997. Mean annual

seroconversion rates for SLEV at the other eleven sites ranged from 0 to 17.4% (8/46) (Table 4-

2). The only SLEV seroconversion between 2001 and 2006 in Orange County occurred at MP in

December of 2001.

West Nile virus transmission at MP was the second highest of all twelve sentinel chicken

surveillance sites operated by OCMCD. From 2001 through 2004 the MASR of sentinel









chickens for WNV was 35.9% (14/39). The high level of WNV activity is a threat to people

camping at MP. The MASR for MP only represents the first four years of surveillance data

which may account for the higher transmission rate. Mean ASRs for WNV at the other eleven

sites ranged from 7.7% (2/26) to 40.0% (16/40) (Table 4-3). In 2002 and 2003, five of twenty-

one and two of twenty WNV seroconversions occurred between October and December at MP.

Vectors of Arboviruses

Tibet-Butler Preserve had twenty different species of mosquitoes captured in exit traps

from July 2003 through December 2004. Moss Park had nineteen different species of

mosquitoes captured in exit traps from July 2003 through December 2004.

Culiseta melanura, the primary epizootic vector of EEEV (Scott and Weaver 1989), is

frequently captured in light traps at TBP. Culiseta melanura is infrequently captured in light

traps at MP. Frequently captured, potential secondary epizootic or bridge vectors of EEEV at

TBP and MP include Ae. albopictus (Scott et al. 1990b, Mitchell et al. 1992), Cq. perturbans

(Scott and Weaver 1989), Cx. erraticus (Cupp et al. 2003), An. quadrimaculatus (Vaidyanathan

et al. 1997), and Cx. nigripalpus (Nayar 1982).

Frequently captured vectors of SLEV at MP and TBP include the primary epidemic vector

of SLEV in peninsular Florida, Cx. nigripalpus (Dow et al. 1964) and the potential, secondary

vector of SLEV in the Florida panhandle, Cx. quinquefasciatus (Tsai and Mitchell 1989, McCaig

et al. 1994).

Frequently captured, potential vectors of WNV at TBP include Cx. nigripalpus (Rutledge

et al. 2003, Turell et al. 2005), Cx. quinquefasciatus (Turell et al. 2005), Cs. melanura (Sardelis

et al. 2002, Turell et al. 2005), and Ae. albopictus (Sardelis et al. 2002, Turell et al. 2005,).

Frequently captured, potential vectors of WNV at MP include Cx. nigripalpus (Rutledge et al.









2003, Turell et al. 2005), Cx. quinquefasciatus (Turell et al. 2005), and Ae. albopictus (Turell et

al. 2005, Sardelis et al. 2002).

Avian Hosts of Arboviruses

An amplification host is a vertebrate that circulates an arbovirus titer of sufficient

magnitude and duration to infect blood-feeding vectors (Scott 1988). Minimum infection

threshold virus titers required to infect mosquito vectors of EEEV (Cs. melanura) were >103

PFU/mL of blood (Komar et al. 1999), and for SLEV and WNV were >105 PFU/mL of blood

(Sardelis et al. 2001).

Almost 100 bird species have been identified in Tibet-Butler Preserve and Moss Park

(OCPR data). Of these birds only a few are critical for arbovirus amplification.

Several potential amplifying hosts for EEEV in TBP and MP are residents in Florida.

Resident birds are important amplification hosts for arboviruses (Crans et al. 1994). The

Northern Cardinal (Cardinalis cardinalis), Common Grackle (Quiscalus quiscula), and Red-

winged Blackbird (Agelaiusphoeniceus) were proven to be efficient hosts of EEEV in the

laboratory (Komar et al. 1999). All three species are common at TBP and MP and were

seropositive for EEEV in Florida in 1958 and 1960-1961 (Favorite 1960, Henderson et al. 1962).

Eastern equine encephalitis virus seropositive rates for the three bird species in Florida were:

Northern Cardinal 34% (10/29), Common Grackle 12% (17/143), and Red-winged Blackbird

19% (3/16) (Favorite 1960, Henderson et al. 1962). Virus isolations of EEEV, from species that

are common residents at TBP and MP include the Common Grackle, Tufted Titmouse

(Baeolophus bicolor), Northern Mockingbird (Mimuspolyglottos), and Carolina Wren

(Thiiy,,th1u, ,n ludovicianus) (Favorite 1960, Henderson et al. 1962, Lord and Calisher 1970).

Several serosurveys performed outside of Florida have isolated EEEV from the Common

yellowthroat (Ge,,11)ypi\ trichas) (Howard et al. 2004), Northern cardinal, and White-eyed vireo









(Vireo griseus) (Lord and Calisher 1970, Stamm et al. 1962, Stamm and Newman 1963). The

Common yellowthroat and Northern cardinal are common residents at TBP and MP. The White-

eyed vireo is a common resident at MP.

Overwintering birds may contribute to EEEV enzootic transmission during amplification

and maintenance phases. Winter migrants, such as the American Robin (Turdus migratorius),

may introduce EEEV from northern breeding sites (Stamm and Newman 1963, Lord and

Calisher 1970). The American Robin and Swamp Sparrow (Melospiza georgiana) were shown

to be efficient amplifying hosts of EEEV in the laboratory (Komar et al. 1999). The American

Robin is common at TBP. The American Robin and Swamp Sparrow are common at MP. The

American Robin, American Goldfinch (Carduelis tristis), House Wren (Troglodytes aedon), and

Ruby-crowned Kinglet (Regulus calendula) had EEEV isolates from serosurveys conducted

outside of Florida (Howard et al. 2004, McLean et al. 1985, Stamm et al. 1962). The American

Goldfinch and House Wren are common overwintering residents at TBP. The House Wren and

Ruby-crowned Kinglet are common overwintering residents at MP. There are resident and

overwintering populations of the Eastern Towhee (Pipilo erythrophthalmus) and Gray Catbird

(Dumetella carolinensis) at TBP and MP. Both species are common at TBP and MP. There

have been several virus isolations of EEEV from the Eastern Towhee and Gray Catbird (Favorite

1960, Howard et al. 2004, Lord and Calisher 1970, Stamm and Newman 1963). The Eastern

Towhee population was 18% (4/22) seropositive for EEEV in Orlando (Favorite 1960).

Four common, resident species of TBP and MP, the Blue Jay (Cyanocitta cristata),

Northern Cardinal, Common Grackle, and Mourning Dove (Zenaida macroura) were frequently

found seropositive for SLEV in 1990 in Florida (Day and Stark 1999). St. Louis encephalitis

virus rates for the four common, resident birds were: Blue Jay 20% (9/46), Northern Cardinal









6% (1/18), Common Grackle 7% (9/123), and Mourning Dove 56% (75/135) (Day and Stark

1999). The Blue Jay has been implicated in outbreaks of SLEV (McLean and Bowen 1980).

The Common Grackle and Mourning Dove each had two isolations of SLEV in Florida in 1990

(Day and Stark 1999). One SLEV isolate was obtained from the Northern Cardinal in Florida in

1989 (Day and Stark 1999). No migrant species at TBP or MP have been implicated in SLEV

epidemics or had SLEV isolated from them.

Numerous resident species in Florida at TBP are potential amplifying hosts for WNV. The

American Crow (Corvus brachyrhynchos), Common Grackle, Blue Jay, Fish Crow (Corvus

ossifragus), Great Homed Owl (Bubo virginianus), and Red-tailed hawk (Buteojamaicensis)

were found to be efficient amplifying hosts of WNV in the laboratory (Komar et al. 2003,

Nemeth et al. 2006). The American Crow, Common Grackle, Blue Jay, Fish Crow, Great

Horned Owl, and Red-tailed Hawk are common residents at TBP. The American Crow,

Common Grackle, Blue Jay, and Fish Crow are common residents at MP. The role of the

American Crow and Blue Jay in WNV transmission may be limited due to the high levels of

mortality these species encounter with WNV infection (Lord et al. 2006). Fifteen percent (2/13)

of Common Grackles were seropositive for WNV in Jefferson County, FL in 2001 (Godsey et al.

2005a). The Red-winged Blackbird has resident and overwintering populations in Florida and is

common at TBP and MP. The Red-winged Blackbird was proven an efficient amplifying host of

WNV in the laboratory (Komar et al. 2003).

Birds that overwinter in Florida may potentially introduce WNV acquired at their northern

breeding grounds (Rappole et al. 2000). The American Robin and Mallard (Anasplatyrhynchos)

were efficient amplification hosts of WNV in the laboratory (Komar et al. 2003). The American

Robin and Mallard are common overwintering birds at TBP. The American Robin is a common









overwintering bird at MP. Seven percent (2/27) of Mallards were seropositive for WNV in

Jefferson County, FL in 2001 (Godsey et al. 2005a).

Discussion

Young nestling House sparrows display fewer defensive behaviors and have elevated

viremias for longer durations than adult counterparts (Scott et al. 1990a). Young nestling birds

would be more likely to be fed upon by mosquitoes due to their reduced host defensive behaviors

compared to their adult counterparts (Scott et al. 1990a, Kale et al. 1972). If young birds remain

viremic at higher magnitudes and for longer duration then the presence of susceptible, young

nestlings may facilitate epidemics. In both TBP and MP amplification transmission is most

likely facilitated by viremic hatching year resident birds.

The highest level of EEEV transmission to sentinel chickens in Orange County was

observed at TBP (MASR 33.6%, 44/131). The bayhead and cypress swamps present at TBP

provide ideal breeding habitat for Cs. melanura which is frequently found at TBP (Foster and

Walker 2002). In contrast, the eighth highest level of EEEV transmission to sentinel chickens in

Orange County was observed at MP (MASR 4.5%, 5/111). There are no bayhead or cypress

swamps at MP and Culiseta melanura is infrequently captured (OCMCD data). Culex

nigripalpus and Cq. perturbans are often caught in sentinel chicken exit traps (OCMCD data)

and may be the vectors of EEEV to sentinel chickens at MP in the absence of Cs. melanura. The

lowered transmission rate of EEEV at MP demonstrates the importance of Cs. melanura

presence for the transmission of EEEV.

The majority of SLEV activity at both TBP and MP occurred in 1997. Neither site had

sentinel chickens present at them during the 1990 SLEV epidemic. In 1997 monthly sentinel

chicken seroconversion rates were above MASRs throughout Florida but only nine human cases









were reported and none from Orange County (Day and Stark 1999). Both MP and TBP provide

suitable habitat for Cx. nigripalpus and numerous potential amplifying hosts of SLEV including

the Blue Jay, Common Grackle, Northern Cardinal, and Mourning Dove. With numerous

potential amplifying hosts of SLEV and the primary epizootic and epidemic vector of SLEV (Cx.

nigripalpus) present, both of these sites should be considered potential foci for SLEV

amplification during SLEV epidemics.

West Nile virus transmission to sentinel chickens at MP was the second highest of the

twelve sentinel sites in Orange County (MASR 35.9%, 14/39). The first detection of WNV in

Florida was in 2001, so information concerning transmission data, potential amplifying hosts,

and potential vectors is limited (Blackmore et al. 2003). The primary implication of amplifying

hosts for WNV transmission have been laboratory studies (Komar et al. 2003, Nemeth et al.

2006, Reisen et al. 2005). Serosurveys of potential amplifying hosts (Godsey et al. 2005a) and

field incrimination of potential vectors of WNV in Florida (Rutledge et al. 2003) have been

limited. Future field studies should elucidate which habitat characteristics, vectors and

amplifying hosts create ideal foci for WNV transmission in Florida.










Table 4-1. Seroconversions for EEEV at sites operated by OCMCD from 1978-2004

'o

a -f e j
e *S
-tL cci ^ 2> u ^
-0 C -e

rS .2 0 c ^^o S
H S H H s~fi m CI> _
Cij j j -
CL~ j F9 F


Years of
operation EEEV
EEEV


10 14 27 8


Seroconversions 44 5 43 0


7 4 13 11 5

7 0 9 14 0


3 27 8

4 24 2


# chickens used
EEEV MASR
Rank


131 111 261 72 69 33 95 96 46 34 240 52
33.6 4.5 16.5 0 10.1 0 9.5 14.6 0 11.8 10 3.8


1 8 2 10 5 10 7


3 10 4 6 9


Table 4-2. Seroconversions for SLEV at sites operated by OCMCD from 1978-2004






o0
H C H H CA


Years of
operation SLEV
SLEV
Seroconversions


10 14 27 8

7 8 37 8


7 4 13 11 5


3 27 8


2 1 1 9 8 0 19 0


# chickens used 131 111 261 72 69 33 95 96 46 34 240 52
SLEVMASR 5.3 7.2 14.2 11.1 2.9 3 1.1 9.4 17.4 0 7.9 0


7 6 2 3 9 8 10 4


Rank


1 11 5 11




Full Text

PAGE 1

MOSQUITO AND SENTINEL CHICKEN INTERACTIONS WITH AN ASSESSMENT OF EXPERIMENTAL CAGE DESIGN AND FLIGHT ACTIVITY OF MOSQUITOES IN OR ANGE COUNTY, FLORIDA: 2005-2006 By KEVIN C. KOBYLINSKI A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2006

PAGE 2

Copyright 2006 by Kevin C. Kobylinski

PAGE 3

3 ACKNOWLEDGMENTS I thank Robyn Raban for her countless effort s, suggestions, and he lp. I thank Sean McCann, Erin Vrzal, Aissa Doumboya, Melissa Doyle, Aaron Lloyd, Joyce Urban, Haze Brown, and David Hoel for their support, help, and sugge stions concerning rearing and experiments in Gainesville. I thank Gregg Ross, Amador Rodr iguez, Columbus Holland, Terry Hughes, Sue Durand, Armond Cross, and the Orange County Mosquito Control Division Staff for their support, help, and suggestions co ncerning field research. I tha nk Dr. Don Shroyer, Dr. Jerry Hogsette, Dr. Phil Koehler, Frank Wessels, Les lie Rios, and the Urban Entomology Laboratory for their thoughtful suggestions and support. I thank Dr. Cynthia Lord, Dr. Ramon Littell, Robyn Raban, and Alejandro Arevalo for their help with statistical analys is. I thank Dr. Sandra Allan and Dr. Dan Kline for their support, help, suggestions, and use of laboratory supplies and space for work done in Gainesville. I thank Dr. Cynthia Rutledge-Connelly, Dr. Thomas Breaud, Dr. Jon Day, Dr. Sandra Allan, and Dr. Da n Kline for all their support, help, and suggestions during the experiments and writing process. This res earch was funded in part by the National Institutes of Health Grant AI042164, M odeling and Empirical Studies of Arboviruses in Florida. This research was funded in pa rt by the Mosquito Control Division, Health and Family Services Department, Orange County Florida.

PAGE 4

4 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................3 TABLE OF CONTENTS.............................................................................................................. ...4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES................................................................................................................ .........9 ABSTRACT....................................................................................................................... ............10 CHAPTER 1 INTRODUCTION..................................................................................................................12 Arboviruses in Florida......................................................................................................... ...13 St. Louis Encephalitis Virus............................................................................................14 West Nile Virus...............................................................................................................15 Eastern Equine Encephalitis Virus..................................................................................16 Western Equine Encephalitis Vi rus: Highlands J Virus.................................................17 Culex nigripalpus and Virus Transmission in Florida............................................................18 St. Louis Encephalitis Virus............................................................................................18 West Nile Virus...............................................................................................................19 Eastern Equine Encephalitis Virus..................................................................................20 Other Potential Vectors of Arboviruses in Florida.................................................................20 Potential Vectors of St. Louis Encephalitis Virus...........................................................21 Potential Vectors of West Nile virus...............................................................................22 Potential Vectors of Eastern Equine Encephalitis Virus.................................................23 Surveillance of Arboviruses...................................................................................................25 Human Surveillance............................................................................................................. ...26 Weather Patterns............................................................................................................... ......28 Vector Surveillance............................................................................................................ ....28 Monitoring of Mosquito Abundance...............................................................................29 Parity Analysis................................................................................................................ .30 Minimum Infection Rates in Mosquitoes........................................................................30 Non-Human Vertebrate Surveillance.....................................................................................32 Live, Wild Vertebrate Surveillance.................................................................................32 Wild bird surveillance..............................................................................................33 Wild mammal surveillance.......................................................................................34 Wild reptile surveillance..........................................................................................36 Dead, Wild Bird Surveillance..........................................................................................36 Domestic Animal Surveillance........................................................................................38

PAGE 5

5 Sentinel Animal Surveillance..........................................................................................41 Sentinel mammals....................................................................................................42 Sentinel birds............................................................................................................42 2 CAGE ESCAPE TRIALS.......................................................................................................50 Introduction................................................................................................................... ..........50 Materials and Methods.......................................................................................................... .51 Mosquito Rearing............................................................................................................52 Marking........................................................................................................................ ...53 Chickens....................................................................................................................... ...53 Experimental Cage Design..............................................................................................54 Aspirator...................................................................................................................... ....55 Escape Rates................................................................................................................... .56 Entry Rates.................................................................................................................... ..57 Results........................................................................................................................ .............57 Escape Trial Results........................................................................................................57 Cage Entry Results..........................................................................................................58 Discussion..................................................................................................................... ..........59 Conclusions.................................................................................................................... .........62 3 SENTINEL CHICKEN DEFENSIVE BEHAVIOR..............................................................72 Introduction................................................................................................................... ..........72 Materials and Methods.......................................................................................................... .73 Mosquito Rearing............................................................................................................73 Mark-Release-Recapture Studies....................................................................................75 Chickens....................................................................................................................... ...76 Host Defensive Behavior.................................................................................................76 Effect of Host Defensive Behavior..................................................................................77 Results........................................................................................................................ .............78 Discussion..................................................................................................................... ..........79 Conclusions.................................................................................................................... .........84 4 DESCRIPTION OF FIELD SITES: ARBOVIRUSES/VECTORS/HOSTS........................88 Introduction................................................................................................................... ..........88 Site Descriptions.............................................................................................................. .......89 Tibet-Butler Preserve.......................................................................................................89 Moss Park...................................................................................................................... ..91 Arboviruses.................................................................................................................... .........92 Tibet-Butler Preserve.......................................................................................................92 Moss Park...................................................................................................................... ..93 Vectors of Arboviruses......................................................................................................... ..94 Avian Hosts of Arboviruses....................................................................................................95 Discussion..................................................................................................................... ..........98

PAGE 6

6 5 EXPERIMENTAL ARBOVIRAL SURVEI LLANCE IN ORANGE COUNTY, FL.........105 Introduction................................................................................................................... ........105 Materials and Methods.........................................................................................................105 Sentinel Chickens..........................................................................................................105 Mosquitoes....................................................................................................................107 Results........................................................................................................................ ...........108 Discussion..................................................................................................................... ........108 Conclusion..................................................................................................................... .......110 6 CAGE ANALYSIS...............................................................................................................114 Introduction................................................................................................................... ........114 Materials and Methods.........................................................................................................115 Results........................................................................................................................ ...........116 Tibet-Butler Preserve.....................................................................................................116 Moss Park......................................................................................................................117 Discussion..................................................................................................................... ........118 Conclusions.................................................................................................................... .......121 7 ROTATOR STUDIES..........................................................................................................128 Introduction................................................................................................................... ........128 Materials and Methods.........................................................................................................129 Results........................................................................................................................ ...........130 Tibet-Butler Preserve.....................................................................................................130 Moss Park......................................................................................................................131 Discussion..................................................................................................................... ........132 Conclusions.................................................................................................................... .......134 8 CONCLUSIONS..................................................................................................................146 REFERENCES CITED............................................................................................................... .148 BIOGRAPHICAL SKETCH.......................................................................................................166

PAGE 7

7 LIST OF TABLES Table page 2-1 Larval feeding regimen..................................................................................................... ...63 3-1 Effect of host defensiv e behavior on recovery, blood feeding success, and location recovered...................................................................................................................... ........85 3-2 Effect of Cx. nigripalpus density on recovery, blood feeding success, and location recovered when one or four chickens were present.............................................................85 3-3 Behaviors displayed by chickens in response to mosquitoes...............................................86 3-4 Defensive behaviors displayed by chickens in the presence of mosquitoes with commentary..................................................................................................................... .....87 4-1 Seroconversions for EEEV at s ites operated by OCMCD from 1978-2004......................100 4-2 Seroconversions for SLEV at si tes operated by OCMCD from 1978-2004......................100 4-3 Seroconversions for WNV at si tes operated by OCMCD from 1978-2004.......................101 5-1 Rain (mm).................................................................................................................. ........111 5-2 Relative humidity (%)...................................................................................................... ..111 5-3 Temperature (C)........................................................................................................... .....111 5-4 Wind speed (km/h).......................................................................................................... ...112 5-5 Barometric pressure (mm-Hg)...........................................................................................112 6-1 Mosquitoes captured in experimental cage at TBP............................................................122 6-2 Other Diptera captured in experimental cage at TBP........................................................122 6-3 Mosquitoes captured in experimental cage at MP..............................................................123 6-4 Other Diptera captured in experimental cage at MP..........................................................123 6-5 Mean number of mosquitoes caught in the exit trap and the cage at TBP.........................124 6-6 Percent of blood-fed mosquitoes caught in the exit trap and the cage at TBP...................124 6-7 Mean number of blood-fed and non bloodfed mosquitoes by location at TBP................125 6-8 Mean number of mosquitoes caught in the exit trap and the cage at MP..........................126

PAGE 8

8 6-9 Percent of blood-fed mosquitoes caugh t in the exit trap and cage at MP..........................126 6-10 Mean number of blood-fed and non bloodfed mosquitoes by location at MP..................127 7-1 Rotator tim e schedule...................................................................................................... ...135 7-2 Tibet-Butler Preserve rotator catch totals 11/7/05 to 7/4/06..............................................136 7-3 Moss Park rotator catch totals 11/8/05 to 7/5/06................................................................137

PAGE 9

9 LIST OF FIGURES Figure page 2-1 Original sentinel chicken cage used by OCMCD................................................................63 2-2 Experimental sentinel ch icken cage used for research.........................................................64 2-3 Line drawing of experimental cage......................................................................................65 2-4 Line drawing of e xperimental inner cage.............................................................................66 2-5 Sentinel chicken cage exit trap............................................................................................ .67 2-6 Assembled aspirator........................................................................................................ .....68 2-7 Night escape trials........................................................................................................ ........69 2-8 Day escape trials.......................................................................................................... .........70 2-9 Percent mosquitoes recovered from experimental cage and exit trap..................................71 4-1 Orange County water bodies w ith sentinel sites, TBP and MP.........................................102 4-2 Aerial view of TBP......................................................................................................... ...103 4-3 Aerial view of MP.......................................................................................................... ....104 5-1 Total number of Culex nigripalpus captured in light traps in Orange County by month..113 7-1 Anopheles crucians captured by rotator trap at TBP and MP............................................138 7-2 Coquillettidia perturbans captured by rotator trap at TBP and MP...................................139 7-3 Culex erraticus captured by rotator trap at TBP and MP...................................................140 7-4 Culex nigripalpus captured by rotator trap at TBP and MP...............................................141 7-5 Culex quinquefasciatus captured by rotator trap at TBP....................................................142 7-6 Culiseta melanura captured by rotator trap at TBP...........................................................143 7-7 Mansonia titillans captured by rotator trap at TBP............................................................144 7-8 All species captured by ro tator trap at TBP and MP..........................................................145

PAGE 10

10 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science MOSQUITO AND SENTINEL CHICKEN INTERACTIONS WITH AN ASSESSMENT OF EXPERIMENTAL CAGE DESIGN AND FLIGHT ACTIVITY OF MOSQUITOES IN OR ANGE COUNTY, FLORIDA: 2005-2006 By Kevin C. Kobylinski December 2006 Chair: Cynthia Roxanne Rutledge-Connelly Major Department: Entomology and Nematology An experimental cage was created to attract and contain mosquitoes. Mosquitoes are funneled into the cage through a baffle. Mosquitoes are contained in the cage, in the exit trap, or they escape through the baffle slit. Trials we re conducted in Gainesville, FL from May 26 through July 25, 2005 to determine mosquito escape rates from the experimental cage. Five colony-raised mosquito species were used: Culex nigripalpus Theobald, Culex quinquefasciatus Say, Anopheles quadrimaculatus Say, Aedes albopictus (Skuse), and Aedes aegypti (Linnaeus). Mosquito escape rates from the cage with the baffl e slit open and two restra ined chickens inside the cage were Cx. nigripalpus 77.7 6.8%, Cx. quinquefasciatus 47.4 10.9%, An. quadrimaculatus 65.8 12.1%, Ae. albopictus 58.7 4.6%, and Ae. aegypti 57.0 5.5%. Chickens display a variety of defensive behavi ors in response to host seeking mosquitoes. From July 26 through August 13, 2005, in Gainesvill e, Florida, adult chic kens were exposed to known densities of mosquitoes. A restrained chic ken and an unrestrained chicken were housed in two separate experimental cages and exposed to 300 Cx. nigripalpus or 300 Ae. albopictus

PAGE 11

11 Blood-feeding success of Ae. albopictus varied significantly between the unrestrained chicken (6.1%) trials and the restrained chic ken (22.2%) trials (Wilcoxon Rank Sum test, P < 0.025) Groups of four chickens were exposed in the experimental cage once a week at two field sites, Tibet-Butler Preserve and Moss Park, in Orange County, FL from October 12, 2005 to July 5, 2006. All mosquitoes were aspirated from the cage the morning after exposure, collected from the exit trap, identified to species, pooled by spec ies, and frozen at -80 C. Twelve species of mosquitoes were captured in the experimental cage at Tibet-Butler Preserve and seven species at Moss Park. Coquillettida perturbans (Walker), Cx. nigripalpus and Culex erraticus (Dyar and Knab) were frequently captured in the experimental cage at both sites. Mansonia titillans (Walker) was captured frequently at Tibet-Butler Preserve. Mos quitoes were more frequently captured in the cage than in the exit trap at Ti bet-Butler Preserve (T-test, t = 14.86, P < 0.0001) and Moss Park (Wilcoxon Rank Sum test, Z = 2.80, P = 0.0052). Coquillettidia perturbans (Wilcoxon Rank Sum test, Z = 2.39, P = 0.0085), Cx. erraticus (Wilcoxon Rank Sum test, Z = 1.86, P = 0.0317), and Ma. titillans (Wilcoxon Rank Sum test, Z = 2.59, P = 0.0048) were more likely to fly upward and become cap tured in the exit trap when blood-fed. A rotator trap, baited with incandescent light and CO2, was operated on a ll nights that experimental chickens were exposed from November 7, 2005 through July 5, 2006 to determine periods of mosquito flight activity at night. Coquillettida perturbans (ANOVA, F = 4.74, P < 0.0001) and Cx. erraticus (Kruskal-Wallis, Chi-square = 25.2, P = 0.0007) exhibited crepuscular host-seeking patterns. Anopheles crucians Wiedemann (ANOVA, F = 4.14, P = 0.0003), Cx. nigripalpus (ANOVA, F=13.96, P < 0.0001), Cx. quinquefasciatus (KruskalWallis, Chi-square = 22.4, P = 0.0022), and Culiseta melanura (Coquillett) (ANOVA, F = 6.55, P < 0.0001) exhibited nocturna l host-seeking patterns.

PAGE 12

12 CHAPTER 1 INTRODUCTION Sentinel chickens are maintained at field sites in Orange County for the detection of arboviruses. Blood samples are taken weekly from sentinel chickens and analyzed for arboviral antibodies. Exit traps were attached to the coun ty sentinel chicken cages to assess the species diversity and abundance of mosquitoes th at were attracted to the chickens. An experimental cage, similar in design to the sentinel chicken cages used by Orange County Mosquito Control Division (OCMCD), was created to draw mosquitoes into the cage and contain the mosquitoes. Mosquitoes that appr oach the sentinel chic kens are funneled upward into the cage by a baffle that narrows to an entry slit. The mosquitoes are contained in the cage or fly upward into the exit trap. Pre-field trials were conducte d to determine the percent of mosquitoes that could be recove red within the experimental cage and the percent of mosquitoes that could escape from the experimental cage. Vertebrates display a variety of defensiv e behaviors in response to host seeking mosquitoes (Anderson and Brust 1995, Cully et al. 1991, Waage and Nondo 1982, Webber and Edman 1972). Adult chicken defensive behavior and its influence on mosquito recovery rates and mosquito blood-feeding success was assessed. Predetermined numbers of mosquitoes were released into two cages with one restrained in one cage, and one unrestrained chicken in the other cage. The effect of mosquito density and chicken defensive behavior on mosquito recovery rates and mosquito blood-feeding su ccess was assessed. Predetermined numbers of mosquitoes were released into two cages with one unrestrained chicken in one cage, and four unrestrained chickens in the othe r cage. The location in the cage that mosquitoes were recovered from and their blood-fed status was recorded. The effect of chicken defensive behavior on

PAGE 13

13 mosquito blood-feeding success and location of mosquito recovery in the experimental cage was assessed. Experimental chickens were placed at two fi eld sites in Orange County from October 2005 to July 2006 for the detection of arbovirus an tibodies. Mosquitoes were collected in the experimental cage in attempt to determine which mosquitoes vectored arboviruses to sentinel chickens in Orange County, FL. The diversity and abundance of mosquito species attracted to chickens was documented. The location of mosqu ito recovery within the experimental cage was documented to determine how repres entative the exit trap collections are of the mosquitoes that approach sentinel chickens. Blood-feeding succ ess was observed to determine how often bloodfed mosquitoes fly upward into the exit trap. The host-seeking times of nocturnally active mosquitoes differ among species. Knowledge of the host-seeking times of nocturnally active mo squitoes is important to determine the most appropriate time to conduct night-time adulticide spraying. A rotator trap was operated at the field sites to determine periods of peak host-s eeking activity. Arboviruses in Florida The diverse habitats found in Florida support a wide range of vertebrate hosts, mosquito vectors, and arboviruses. Humans find the Florida climate attrac tive and millions of susceptible, elderly humans have made Florida their home (B ond et al. 1963). The elderly tend to suffer greater complications to arboviruses that occur in Florida such as St. Louis encephalitis (family Flaviviridae genus Flavivirus SLEV) and West Nile virus (family Flaviviridae genus Flavivirus WNV) (Bond et al. 1963, Petersen and Ma rfin 2002). As the population of northcentral Florida (greater Orlando area) continues to grow, human habitation pushes closer to swamp foci, increasing the potenti al of human infection with Ea stern equine encephalitis virus (family Togaviridae genus Alphavirus EEEV) (Edman et al. 1972b, Morris 1992).

PAGE 14

14 St. Louis Encephalitis Virus St. Louis encephalitis virus was first documented in Florida from a thirty year-old man in Miami in 1952 (Sanders et al. 1953). Since then there have been five major epidemics of SLEV in three distinct regions of Florida (Bond 1969, Nelson et al. 1983, Day and Stark 2000). The outbreaks of 1959, 1961, and 1962 were centered in th e Tampa Bay Area of Florida (Pinellas, Hillsborough, Manatee, and Sarasota counties) (Bond 1969). There were 68 clinical human cases of SLEV and five deaths in 1959, 25 cases and seven deaths in 1961, and 231 cases and 43 deaths in 1962 (CDC 2006a, Bond 1969). As with most outbreaks of SLEV, all deaths were persons over the age of 45 years (Bond 1969). The primary epidemic vector during these outbreaks and the 1977 and 1990 outbreaks in peninsular Florida, was Culex nigripalpus Theobald (Dow et al. 1964, Shroyer 1991). In di rect response to the ou tbreak of SLEV in 1962 the Encephalitis Research Center (ERC) of the Florida State Board of Health in Tampa was established (Bond 1969). From 1963 to 1976, only two human cases of S LEV were reported in Florida and thus surveillance and research interests in SLEV declined (Nel son et al. 1983). The outbreak of SLEV in 1977 lacked an urban focus and covere d a broad region (20 counties) of central and south Florida (Nelson et al. 1983). This outbreak differed from the previous outbreaks in Tampa Bay because there was a high attack rate of SLEV in males between 15 and 24 years of age, most likely due to outdoor occupations (Nelson et al. 1983). Th ere were 110 presumptive and confirmed human cases of SLEV in Florida in 1977 with 11 cases from Orange County (Nelson et al. 1983). Interest in surv eillance of arboviruses was heighten ed after this outbreak and the establishment of the Florida Sentinel Chicke n Surveillance Program began in 1978 (Day 1989). Focal outbreaks and sporadic cases of SLEV in humans occurred from 1979 to 1984 (CDC 2006a). In 1990, there was an epid emic of SLEV in humans that spanned 28 counties in central

PAGE 15

15 and south Florida with 226 clinical human cases and 11 deaths. In Orange County there were 28 human cases of SLEV (Day and Stark 2000). W ith the exception of two sporadic cases in 1969, (FDOH 2006a) the 1977 and 1990 outbreaks were the on ly instances of reported human infection of SLEV in Orange County. Focal and sporadic cases of SLEV in humans occurred from 1991 to 2002 throughout Florida (CDC 2006a). Human infection with SLEV is typically a re latively mild illness characterized by fever and headache, followed by complete recovery (Hayes 2000). Persons over the age of 45 may experience severe central nervous system involve ment of meningitis or encephalitis leading to death (Bond 1969, Hayes 2000). Inapparent (asymp tomatic) human SLEV infection rates vary from 0.2 to 5.2% (Hayes 2000). Human SLEV inf ections in Florida t ypically occur between August and December (Bond 1969, Nelson et al. 1983). West Nile Virus West Nile virus was first documented in Fl orida in 2001 (Blackmore et al. 2003). In 2001 WNV activity was dispersed throughout the entire state with 12 human cases ranging from the panhandle to the Keys (Blackmore et al. 2003). In 2002 WNV activity was dispersed throughout the entire state with 35 human cases, two deaths and three additional cases attributed to blood transfusions and organ transplants. The only human case of WNV infection in Orange County occurred in 2002 with an onset date of August 12. In 2003 WNV activity was dispersed throughout the state with a hea vy concentration of cases (61 cases in 10 counties) in the northwest portion of the panhandl e. There were 93 human cases of WNV and six deaths in 2003. In 2004 there were 39 human cases of WNV, two deaths and three additional cases that were suspected to have been acquired out of st ate. In 2005 there were 21 human WNV cases and one death. As of July, 2006, no human infections of WNV have been reported in Florida for 2006 (FDOH 2006b).

PAGE 16

16 During the initial two years of WNV activity there were human cases throughout Florida with no discernable pattern of transmission or centralized outbreak foci. In 2003 there were human cases throughout Florida bu t 66% (61/93) of human cases were from the northwest region of the Florida Panhandle. In 2004 and 2005 ther e were urban foci in Dade and Broward Counties (24 of 39 human cases) and in Pinellas County (18 of 21 human cases), respectively (FDOH 2006b). In humans, WNV is capable of causing either a febrile illness known as West Nile fever (Watson et al. 2004b) or a neuroinvasive diseas e such as meningitis, encephalitis, or acute flaccid paralysis leading to deat h (Petersen and Marfin 2002). Sy mptoms of West Nile fever include fever, malaise, anorexia, nausea, vom iting, eye pain, headache, myalgia, rash, and lymphadenopathy (Petersen and Marfin 2002) with symptoms lasting 3 to 6 days or sometimes longer (Watson et al. 2004b). Symptoms of neur oinvasive WNV disease in clude fever, muscle weakness, gastrointestinal symptoms, headache, and changes in mental status (Petersen and Marfin 2002). A seroprevalence study in New York City during 1999 determined that approximately 20% of WNV infected persons developed febrile illness and less than 1% developed neuroinvasive disease (Mostashari et al. 2001). Mostas hari et al. (2001) estimated that for every one meningoencephalitis cas e there were 140 symptomatic and mildly symptomatic cases. In a large scale screening of blood donations Busch et al. (2006) found the incidence of WNV neuroinvasive disease to in apparent infection in humans to be 1 in 256. Eastern Equine Encephalitis Virus Transmission of Eastern equine encephalitis virus has only occurred in focal or sporadic outbreaks in Florida (Bigler et al. 1976, Day and Stark 1996b, Mo rris 1992). Most cases of EEEV in humans are confined to the panhandle a nd north-central Florida (Day and Stark 1996b). The first documented human case of EEEV in Fl orida was in 1952, but th ere were unconfirmed

PAGE 17

17 reports of human cases in Florida prior to 1952 (Big ler et al. 1976). To date there have been 68 reported human cases of EEEV in Florida (CDC 2006a). Four human cases of EEEV have occurred in Orange County with one case in 1973, 1990, 1995, and 2003 (FDOH 2006a). Eastern equine encephalitis vi rus is a severely debilitati ng disease, causing death in approximately 50 to 90% of symptomatic human vi ctims, with most deat hs being adults (Scott and Weaver 1989, Villari et al. 1995). Symptoms of encephalitic EEEV in fection are high fever (39 to 41C), irritability, restlessness, drowsine ss, muscle tremors, neck rigidity, anorexia, vomiting, diarrhea, headache, cyanosis, convulsion s, and coma (Morris 1992). Villari et al. (1995) found adults that survive encephalitic EE EV infection often ha ve full recovery but children less than fifteen ofte n suffer lifelong debilitations cau sed by persistent and severe neurologic disease. Human cases can occur th roughout the year in Florida but generally occur between May and August (Bigler et al. 1976). Most human cases of EEEV are rural in distribution and generally associ ated with wooded areas adjacent to swamps and marshes (Morris 1992). Western Equine Encephalitis Virus: Highlands J Virus Western equine encephalitis virus (family Togaviridae genus Alphavirus WEEV) has been isolated from Culiseta melanura (Coquillett), birds, and horses in Florida (Bigler et al. 1976, Henderson et al. 1962, Karabatsos et al. 1988). It was later dete rmined that all isolates of WEEV in the Eastern United States were Highlands J (HJ) virus, an antigenically distinct virus that is part of the WEEV complex (Calisher et al. 1980, Karabatsos et al. 1988). Highlands J virus has no known public or veterinary health im portance (Karabatsos et al. 1988). The primary vector of HJ virus in the eastern United States is Cs. melanura (Karabatsos et al. 1988, Morris 1988). Highlands J virus can complicate avian serosu rvey results as it cros s reacts with EEEV in the Hemaggluttination Inhibition (HI) test. It was not until 2004 that HJ was included in the

PAGE 18

18 laboratory testing of EEEV positive sentinel chickens in Florida. Plaque reduction neutralization tests (PRNT) are now used to differentiate between HJ and EEEV (L. M. Stark, personal communication). Culex nigripalpus and Virus Transmission in Florida St. Louis Encephalitis Virus Culex nigripalpus is the most abundant Culex ( Culex ) species in peninsular Florida (Edman 1974). An outbreak of SLEV in the Tampa Bay Area of Florida in 1962 resulted in 22 of 23 isolations of SLEV from wild Cx. nigripalpus establishing this species as the primary epidemic vector of SLEV in south Florida (D ow et al. 1964). Provos t (1969) observed that Cx. nigripalpus activity was directly influenced by humid ity with nights of 90% humidity or greater triggering the greatest amount of flight activity. He hypothesized that rainless periods during the Cx. nigripalpus breeding season punctuated by intermittent rainfall leads to SLEV transmission (Provost 1969). Day and Curtis (1999) defined four phases of SLEV transmission in subtropical Florida: maintenance, amplification, early season transmi ssion, and late season transmission phases. The maintenance phase occurs from December throug h March and SLEV transmission is maintained at low levels between avian hosts and mosquitoes The amplification phase occurs from April through June and coincides with the avian nes ting season. Early season transmission of SLEV occurs from July through September at which time the first human cas es of SLEV may be reported. Late season transmission of SLEV occu rs from October through December when more human cases of SLEV may be re ported. St. Louis encephalitis virus transmission to humans typically ceases by the end of December (Day and Curtis 1999). Winter and springtime (maintenance and amplification phases) drought restricts Cx. nigripalpus to wooded hammocks where many avian spec ies nest (Day 2001). This brings Cx.

PAGE 19

19 nigripalpus into contact with adult a nd nestling birds (Day 2001). Culex nigripalpus had higher blood feeding success on nestling than adult cicon iiform birds most likely due to the lack of plumage on nestling birds (Kale et al. 1972). Ne stling birds have been shown to circulate a higher viremia of SLEV for a longer duration than adult bird s and display fewer defensive behaviors making them more likely to infect mo squitoes (Scott et al. 1988, Scott et al. 1990a). Culex nigripalpus shifts host preference from avian hosts during January through May to mammalian hosts during June through November or December (Edman 1974, Edman and Taylor 1968). The mammalian feeding preference shift by Cx. nigripalpus coincides with the occurrence of epidemic activity of SLEV in humans during earl y and late season transmission (Day and Curtis 1999, Edman 1974). The Cx. nigripalpus host preference sh ift from avian to mammal is directly linked with rainfall (Edman 1974). Day and Curtis (1989) showed that Cx. nigripalpus blood-feeding and oviposition (Day et al. 1990) behavior was regulated by heavy rainfa ll (>30mm). When heavy rainfalls occur every 10 to 14 days, parous Cx. nigripalpus were forced to delay ovipos ition. This allowed time for the extrinsic incubation of SLEV virus resulting in infected salivary glands of the mosquito. The next heavy rainfall triggered the infective Cx. nigripalpus to oviposit and then blood feed, potentially infecting further hosts (Day et al 1990). This amplifying cycle continues until minimum infection rates of SLEV in Cx. nigripalpus are >1:1000 at whic h point the likelihood of incidental human infection can reach epidemic levels (Day and Stark 2000). Once the drought ends, the repeated rainfall raises humidity levels and Cx. nigripalpus disperse from the hammocks furthering the potential of SLEV tr ansmission to humans (Shaman et al. 2002). West Nile Virus Culex nigripalpus has been incriminated in transm ission of WNV in Florida and was suspected of being the primary epidemic vector in 2001 (Rutledge et al. 2003). Culex

PAGE 20

20 nigripalpus was capable of transmitting WNV in the la boratory (Turell et al. 2005). Two WNV infected pools of Cx. nigripalpus were found in Jefferson Count y, FL in 2001 (Godsey et al. 2005a). The same mechanisms that make Cx. nigripalpus an efficient vector of SLEV were thought to make it an efficient v ector of WNV (Shaman et al. 2005). Eastern Equine Encephalitis Virus Culex nigripalpus was suggested as a possible bridge vector for EEEV in Florida by Nayar (1982). Eastern equine en cephalitis virus was isolat ed from sixteen pools of Cx. nigripalpus in the Tampa Bay area from 1962 to 1967 (Taylor et al. 1969) and from 94 pools from 1962 to 1970 (Wellings et al. 1972). Three of ten pools of Cx. nigripalpus were positive for EEEV during a focal outbreak of EEEV in emus in Vo lusia County, Florida in 1994 (Day and Stark 1996a). Eastern equine encephalomyelitis virus has also been isolated from Cx. nigripalpus in Trinidad (Downs et al. 1959). Sudia and Chamberlain (1964) found Cx. nigripalpus to be a poor vector of EEEV under laboratory conditions. Culex nigripalpus is often collected from swamp foci with known EEEV activity in Orange County (OCM CD data). The abundance of Cx. nigripalpus in swamp habitat, generalis t feeding habits (Edman 1974) and rainfall mediated hostfeeding patterns (Day and Curtis 1989) increase its pot ential as a vector of EEEV in Florida. Other Potential Vectors of Arboviruses in Florida Chamberlain (1958) used five criteria for the incrimination of potential vectors of arboviruses which were simplified and restated by Vaidyanathan et al. (1997): history of frequent virus isolation; fli ght range overlapping host habita t; host seeking and blood feeding coinciding with disease incidence; varied host ch oice; and laboratory demonstration of vector competence. Many of the mosquito species in Fl orida suspected as potential arbovirus vectors have not fulfilled all five of the requirements to be incriminated as vectors of arboviruses.

PAGE 21

21 Potential Vectors of St. Louis Encephalitis Virus Culex pipiens quinquefasciatus Say and Culex salinarius Coquillett are the second most abundant species of Culex ( Culex ) in Florida (Edman 1974). Both reach peak abundance in late winter and spring in south Florida (OMeara and Evans 1983, Provost 1969). Culex quinquefasciatus is a generalist feeder that frequently feeds on man and domestic animals such as cats, dogs, and chickens (Edman 1974). Culex quinquefasciatus was a competent vector of SLEV in the laboratory (Chamberlain et al. 1959). Culex quinquefasciatus is the primary vector of SLEV in the southeastern United States (Tsai and Mitchell 1989) and in the absence of Cx. nigripalpus it is most likely the vector of SLEV in northern Florida. Culex quinquefasciatus was the suspected vector of the 1980 focal outbr eak of SLEV in Fort Walton Beach, Florida (McCaig et al. 1994). Culex salinarius is a generalist feeder that was less associated with man and domestic animals than Cx. quinquefasciatus (Edman 1974). Culex salinarius a highly competent vector of SLEV in the laborator y, (Chamberlain 1958, Chambe rlain et al. 1959) may be an enzootic vector of SLEV in Florida (Zyzak et al. 2002). Culex restuans Theobald is a competent vector of S LEV in the laboratory (Chamberlain et al. 1959). The host preference of Cx. restuans appears to vary by locality but was thought to be primarily ornithophilic in the United St ates (Edman 1974, Mitc hell et al. 1980). Culex restuans is most abundant in Florida during late fall, winter and spring which makes it most likely an enzootic vector of SLEV during the maintenance and amplification phases of SLEV transmission (Edman 1974). When summertim e temperatures drop below 20C Cx. restuans may become active and aging mosquitoes may vector SLEV (Reiter 1988). There have been no documented isolations of SLEV from Cx. quinquefasciatus Cx. restuans or Cx. salinarius in Florida. This limits their incrimination as vectors of SLEV in

PAGE 22

22 Florida by not fulfilling the first criteria of vect or incrimination (Chambelain 1958): history of frequent virus isolation. Potential Vectors of West Nile virus Of the sixty species of mosquitoes that have tested positive for WNV in North America, thirty-eight sp ecies occur in Florida (CDC 2006b). Isolation of WNV with various polymerase chain reaction (PCR) techniqu es only shows that mosquitoes were infected with WNV (Rutledge et al. 2003). Isolation of WNV from mosquitoes do es not establish their capability of transmitting WNV to vertebrate hosts (Rutledge et al. 2003). Of the thirty-eight mosquito species that had WNV isolated from them and occur in Florida, sixteen were tested in the laboratory for vector co mpetence (Turell et al. 2005). Culex restuans Cx. quinquefasciatus and Cx. salinarius were found to be efficient vectors of WNV in the laboratory (T urell et al. 2005). The vari ed host preference of Cx. quinquefasciatus and Cx. salinarius (Edman 1974) make these species potential enzootic or bridge vectors of WNV. Fi ve WNV infected pools of Cx. quinquefasciatus were found in Jefferson County, FL in 2001 (Godsey et al. 2005a). Four WNV infected pools of Cx. salinarius were found in Jefferson County, FL in 2001 (Godsey et al. 2005a). The ornithophilic feeding behavior of Cx. restuans (Edman 1974) makes this species a potential enzootic vector of WNV. Aedes albopictus (Skuse) and Ochlerotatus triseriatus (Say) were found to be efficient vectors of WNV in the laboratory and their mammalophilic host feeding habits make them potential bridge vectors (Sardelis et al. 2002, Turell et al. 2005). Aedes albopictus and Oc. triseriatus are very localized mosquitoes (flight range 200 m), so involvement in WNV transmission may be limited to areas where Culex spp. are present and transmitting WNV to amplifying hosts (Sardelis et al. 2002, Turell et al. 2005). Aedes vexans (Meigen), Culiseta inornata (Williston), and Ochlerotatus canadensis canadensis (Theobald) are mammalophagic,

PAGE 23

23 have flight ranges of 2 km or greater, capable of WNV transmission in the laboratory and may play minor roles as bridge vectors of WNV to humans in Florida (Turell et al. 2005). Three WNV infected pools of Cs. melanura were found in Jefferson County, FL in 2001 (Godsey et al. 2005a). Culiseta melanura (Coquillett) was a poor ve ctor of WNV in the laboratory (Turell et al. 2005). Culiseta melanura feeds primarily on birds (Edman et al. 1972b) and may be an enzootic vector of WNV in Florida. West Nile virus has only recently been introduced to Florida and there have been very few studies demonstrating WNV tran smission to vertebrates in the field, the exception being Cx. nigripalpus (Rutledge et al. 2003). Labor atory studies of vector comp etence and field isolations of WNV only fulfill part of the requirements for v ector incrimination (Chamberlain 1958). More field studies on WNV transmission in Florida are required to fulfill all requirements of vector incrimination for the species discussed in this section (C hamberlain 1958). Potential Vectors of Eastern Equine Encephalitis Virus Culiseta melanura is the primary enzootic vector of EEEV in Florida (M itchell et al. 1996, Morris 1992, Scott and Weaver 1989). Culiseta melanura breeding tends to be restricted to swamp habitats (Morris 1992). Most human cases of EEEV occur within five miles of swamp foci (Morris 1992). In Florida, Cs. melanura was found to be ornithophilic, feeding predominantly on passerine birds (Edman et al. 1972b). Since Cs. melanura is ornithophilic it is considered an enzootic vector of EEEV and othe r species of mosquitoes must serve as bridge vectors to humans and horses (Edman et al. 1972b). Numerous species of mosquitoes that have b een suggested as potenti al bridge vectors of EEEV in the United States occur in Florida including Coquillettidia perturbans (Walker), Ae. vexans Ochlerotatus sollicitans (Walkeri), Culex erraticus (Dyar and Knab), Cx. salinarius Oc. canadensis Anopheles quadrimaculatus Say, and Aedes albopictus (Edman et al. 1972b, Scott

PAGE 24

24 and Weaver 1989, Scott et al. 1990b). Bridge vectors of EEEV to humans and horses vary from location to location (Crans a nd Schulze 1986). In New Jersey, Oc. sollicitans is the primary bridge vector of EEEV to humans in coastal areas and Cq. perturbans is the primary epizootic vector of EEEV to horses at inland sites (C rans and McCuiston 1993, Crans and Schulze 1986). Eastern equine encephalitis virus was isolated from Cx. salinarius from Tampa Bay, Florida (Wellings et al. 1972) and other locations in the eastern United States (Scott and Weaver 1989). Cx. salinarius was able to transmit EEEV fourteen days post-infection in the laboratory (Vaidyanathan et al. 1997). Culex salinarius may be a potential bri dge vector because its winter/spring abundance in Florida means that aging Cx. salinarius would be present during EEEV transmission periods (Zyzak et al. 2002 ). The generalist feeding habits of Cx. salinarius (Edman 1974) make it a potential enz ootic and bridge vector of EEEV. Both Oc. canadensis and An. quadrimaculatus were capable of transmitting EEEV in the laboratory (Vaidyanathan et al. 1997). Both species are mammalophilic and could potentially transmit EEEV to horses or humans (Edman 1971, Vaidyanathan et al. 1997). Neither Ae. vexans nor Oc. sollicitans are commonly found near swamp foci in Florida, which limits their potential invo lvement as bridge vectors of EEEV in Florida (Edman 1974). Aedes vexans was not capable of transmitting EEEV in th e laboratory (Vaidyanathan et al. 1997). Minimum infection rates (MIR) of Cx. erraticus for EEEV in central Alabama were 3.2 per 1000 (Cupp et al. 2003). Positive Cx. erraticus pools were found from mid-June to midSeptember (Cupp et al. 2003), the same time peri od that EEEV is most active in Florida (Bigler et al. 1976). An isolation of EEEV from Cx. erraticus in Polk County, Florida was made in 1993 (Mitchell et al. 1996). Cham berlain et al. (1954) found Cx. erraticus to be a competent vector of EEEV in the laboratory. Robert son et al. (1993) documented Cx. erraticus feeding on humans in

PAGE 25

25 North Carolina. These findings suggest that Cx. erraticus may play a role as an enzootic or bridge vector of EEEV transmission in Florida. Aedes albopictus was found naturally infected with EEEV in Polk County, Florida (MIR of 1.5 per 1000) (Mitchell et al. 1992). Aedes albopictus was capable of transmitting EEEV in the laboratory (Scott et al. 1990b). J unkyards and tire piles in rural and urban areas near swamp foci may harbor populations of Ae. albopictus (OCMCD data) and may lead to focal amplification of EEEV (Mitchell et al. 1996). Coquillettidia perturbans has been incriminated as a brid ge vector of EEEV in the eastern United States (Morris 1988, Scott and Weaver 1989). Coquillettidia perturbans was the vector of EEEV in an epizootic among horses in New Jersey in 1983 (Crans and Schulze 1986). Wellings (1972) obtained fi ve EEEV isolates from Cq. perturbans in Florida. In Florida, 9% of blood meals from Cq. perturbans were avian and 91% were of mammalian origin (Edman 1971). Coquillettidia perturbans was able to transmit EEEV seven da ys post-infection in the laboratory (Vaidyanathan et al. 1997). Coquillettidia perturbans feeds on humans in Florida (Provost 1969). These findings suggest that Cq. perturbans may be a potential bridge vector of EEEV to humans and horses in Florida. Surveillance of Arboviruses Arboviral surveillance is the syst ematic collection of data re garding arboviral activity to predict or recognize epid emics and assess the size and progres sion of current epidemics (Bowen and Francy 1980). The objective of arbovirus su rveillance is to pred ict the likelihood of arbovirus transmission to humans so that public health measures and mo squito control activity may limit or abort human outbreaks (Eldridge 1987). There are two types of arbovirus surveillance methods: passive and active. Passive surveillance is the accumulation of unsolicited information concerning arboviral infec tion in humans or other vertebrates whereas

PAGE 26

26 active surveillance is the intentional collection of specimens or data. There are four methods for conducting active arboviral surveillan ce: detection of human diseas e or infection, observation of weather patterns, surveillance of vector populations, and the tes ting of non-human vertebrates (Bowen and Francy 1980). Human Surveillance Human surveillance for arboviral encephalitis incidence is conducte d by public health officials. A passive human surveillance system requires local physicians and hospital staff to record and report suspect human cases of arboviral in fection to local health departments. In the past, cases were not being reported in a timely manner or reported case information was not assimilated properly, allowing epidemics to go unnoticed (Chamberlain 1980). In many regions where there is little money for mosquito cont rol operations and arbovira l surveillance, the only affordable method for arboviral surveillance is human surveillance (Bowen and Francy 1980, CDC 2003). Once public health officials have determined that an arboviral epidemic is imminent or already in progress the most common measures of control are emergenc y aerial adulticide spraying and public service announcements (PSAs). The drawback to human surveillance is that arboviral amplification is al ready at high levels in vector and amplifying host populations by the time incidental human cases are reported. The delayed vector control will likely have little impact on epidemic transmission if most arbovira l transmission has already occurred (Nelson et al. 1983). Day and Lewis (1992) discussed the difficulties of human surveillance due to the several week time lag between the date of onset, the first (acute phase) and second (convalescent phase) blood samples and the diagnostic tests. Nelson et al. (1983) sugg ested that clinically suspected SLEV cases should be reported to redu ce the reaction time of mos quito control efforts.

PAGE 27

27 Human arboviral case infections in the United States are defined as possible, probable, and confirmed cases. Possible and probable human case definitions of arbovi ral disease require the display of symptoms associated with the disease and positive antibody results from one acute or convalescent human sera, cerebrospinal fluid (C SF) or other body fluids. Confirmed human case definition requires the display of symptoms and po sitive results with a fourfold increase between acute and convalescent serum, CSF, or other body fluid or the dire ct isolation of arbovirus or demonstration of arboviral genomic sequences in human tissue, blood, CSF, or other body fluid (Petersen and Marfin 2002). In Florida human diseases are monitored by th e Florida Department of Health (FDOH) and County Health Departments (CHD). Human se ra are screened for arboviruses with a hemagluttination inhibition (HI) assay and all po sitives are confirmed with an immunoglobulin M-capture enzyme-linked immunosorbent as say (MAC-ELISA). A plaque reduction neutralization test (PRNT) is used to differentiate between WNV and SLEV. Cerebrospinal fluid samples are tested for WNV with MAC-ELISA. A ll tests are performed free of charge at the DOH Bureau of Laboratory Services in Tampa or Jacksonville (Blackmore et al. 2003). Human case reports are entered into Merlin, a real time el ectronic reporting system for all diseases that affect humans (FDOH 2005). All probable and c onfirmed human cases are then entered into Arbo-NET, a cooperative West Nile virus surveillance database a nd reported to the Centers for Disease Control (Marfin et al. 2001). Even with a nationwide human surveillance system the data generated are problematic because of the va riability and inaccuracies in disease reporting. This makes the estimation of disease incidence an d the interpretation of potential risks difficult (Brownstein et al. 2004).

PAGE 28

28 Weather Patterns Weather patterns that preceded past arbovira l epidemics can be analyzed to formulate predictions of future arboviral transmission potential. These pr edictions can provide an early warning of pending arboviral epidemics so that public health officials and mosquito control programs can increase surveillanc e and control efforts to reduce epidemic potential. Weather patterns are not indicativ e of arboviral presence or amplification in mosquito vectors or vertebrate hosts. Field isolations of arbovirus in mosquitoes or vertebrate hosts or isolation of arboviral antibodies from sentinel and wild animals is required for confirma tion of viral activity. Tracking weather patterns is an easy and low co st method that when used in combination with human surveillance, is the best alternative fo r areas lacking a budget for a standing mosquito control and surveillance operati ons (Bowen and Francy 1980). Recently, several models have been created wh ich correlate water table depth and surface wetness in peninsular Florida with SLEV or WN V transmission to birds and humans (Shaman et al. 2002, 2003a, 2004a, 2004b, 2005). A topographical ly based hydrology model for southcentral Florida was created to predict surface le vel wetness (Shaman et al. 2002). This model was validated using groundwater well measuremen ts and surface water le vels throughout Florida (Shaman et al. 2003b). Sentinel chicken seroconve rsion data from 1978 to 2002 in Indian River County and the topographically based hydrology model were used to assess past epidemic and non-epidemic levels of SLEV transmission with th e expectation that this model will be able to forecast future SLEV epidemics (Shaman et al. 2004b). Vector Surveillance Vector surveillance for arbovi rus transmission potential can involve monitoring of mosquito abundance, parity status of the mos quito population and minimum infection rates of mosquitoes. Vector surveillance can be costly to a mosquito control dist rict or public health

PAGE 29

29 department which means that not all three factor s can always be assessed (Bowen and Francy 1980, Scott et al. 2001). Monitoring of Mosquito Abundance There are many techniques available to a ssess mosquito populations (Service 1976). Long-term baseline data sets of mosquito abunda nce for each trap site with consistent sampling methods demonstrate fluctuations in mosquito abundance over time (Day and Lewis 1991). In peninsular Florida the primary arboviral vector of concern, Cx. nigripalpus can be readily caught by light traps, CDC baited light traps, gr ound aspirations, avian baited lard can traps (Day and Lewis 1991), and sentinel chicken coop exit trap s (T. P. Breaud and D. A. Shroyer, personal communication). Indian River County Mosqui to Control District (Vero Beach, FL) has discontinued the use of light traps in favor of aspiration of resting mosquitoes (Day and Lewis 1991). Mosquitoes collected by aspiration can be analyzed to determine mosquito population dynamics such as abundance, migration, blood fe eding, parity, and age (Day and Lewis 1991). Mosquito abundance has been correlated with arbovirus transmission with mixed success. In California, Reeves (1 968) predicted that if Culex tarsalis Coquillett light trap catches were kept below ten mosquitoes per night SLEV huma n cases would not occur. Blackmore et al. (1962) found no correlation between Cx. tarsalis abundance and SLEV or WEEV transmission in Colorado. Olson et al. (1979) found that SLEV and WEEV transmission was affected by varied abundance of Cx. tarsalis in rural and urban Califor nia habitats. The 1961 SLEV epidemic in Tampa Bay, FL was characterized by high transmission rates to humans, low numbers of complaint calls and low populations of mosquitoes in urban areas (Mulrennan 1969, Rogers 1969). Recent focal outbreaks of WNV in th e eastern United States have been associated with low densities of Cx. pipiens and Cx. quinquefasciatus (CDC 2003). These disparities in correlation of mosquito abundance and epidemic transmission render mosquito abundance an

PAGE 30

30 invalid predictor of arbovira l transmission and therefore a poor surveillance technique. Mosquito abundance alone may be a poor pred ictor of epidemic transmission but when combined with parity information it can be impor tant for the timing and assessment of mosquito control efforts during arbovi ral epidemics (Day 1991). Parity Analysis Parity analysis is a technique that is used to estimate the physiological age of an anautogenous (mosquito that requires a blood meal before the development of eggs) mosquito. Parity analysis involves dissection and examinati on of the ovaries for the presence or absence of tracheolar skeins. Coiled tracheole skeins indica te a nulliparous (has not laid eggs) mosquito. Extended tracheoles indicate a parous (has laid eggs) mosquito (Crans and McCuiston 1993). A parous mosquito indicates that it has obtained at least one blo od meal and may have lived long enough to become infective. A period of time, known as the extrinsic incubation period, must pass after the ingestion of blood before a mosquito can infect another host (Foster and Walker 2002). The process of extrinsic incubation requi res the ingestion of arbovirus by a mosquito vector, multiplication of arbovi rus in the midgut cells, escape from the midgut, dissemination throughout the hemocoel, and infection of the sa livary glands (Foster and Walker 2002). Assessment of parity status in mosquito popul ations provides an es timate of how many mosquitoes might be infec tive (Bowen and Francy 1980). Since climatological regulation of Cx. nigripalpus oviposition is crucial for SLEV epidemics (Day et al. 1990) the de termination of parity status of Cx. nigripalpus populations during an epidemic is important for the timing of adult mosquito control operations (Day 1991). Minimum Infection Rates in Mosquitoes Virus isolation from mosquitoes is a part of vector incrimination (Chamberlain 1958) and has been used for the detection of new arboviruses in an area (Scott et al 2001). Virus isolation

PAGE 31

31 is often conducted during arboviral epidemics but rarely during non-epidemic periods (Bowen and Francy 1980, Shroyer 1991). Minimum infection rates (MIR) are determin ed by the collection, identifica tion, pooling, and processing of mosquitoes for arbovirus isol ation. Reeves et al. (1961) calculated minimum infection rates using Equation 1-1. Number of positive pools x 1,000 (Equation 1-1) Number of mosquitoes tested The MIR only represents the minimum number of infected mosquitoes, not the minimum number of infective mosquitoes. Similar to mosquito abundance data, there must be baseline data of MIRs during non-epidemic periods to unde rstand what epidemic MIR levels represent as a risk assessment of arboviral infection to humans (Shroyer 1991). Transmission rates of arboviruses to sentinel vertebrates more accurately reflect the potential for mosquitoes to transmit arboviruses. To determine transmissi on rates sentinel hosts must be placed in the field, determine the numbe r of seroconversions in hosts, and capture as many mosquitoes that come to f eed on the hosts as possible (Reeves et al. 1961). Reeves et al. (1961) calculated transmissi on rates using Equation 1-2. Number of birds infected x 1,000 (Equation 1-2) Total mosquitoes feeding Reeves et al. (1961) found both SLEV and WEEV virus infection rates in Cx. tarsalis to be higher than transmission rates to sentinel chic kens. Rutledge et al. (2003) found WNV infection rates in Cx. nigripalpus to be higher than WNV transmission rates to sentinel chickens. One factor attributed to higher MI Rs than transmission rates is the extrinsic incubation period. Mosquitoes may be infected with arbovirus which would raise the MIR but sufficient time may not have passed for the mosquito to be infective which would not increase the transmission rate.

PAGE 32

32 Shroyer (1991) found that during non-epidemic periods the MIRs for SLEV in Indian River County, FL were very low. It was very difficult to statistically determine confidence intervals and sampling error was very high becau se of the low frequencies of SLEV, which makes direct estimates of human risk from MI Rs difficult (Shroyer 1991). While isolation of arboviruses from mosquito vectors is important for vector incrimination (Chamberlain 1958), it is of little value as a su rveillance tool to predict the threat of human arbovi ral infection in Florida (Shroyer 1991). Non-Human Vertebrate Surveillance Surveillance of non-human vertebrates involve s the testing of blood or tissue specimens from vertebrates for the presence of arboviruses or antibodies to arboviruses (Bowen and Francy 1980). Surveillance of non-human ve rtebrates uses wild, domestic and sentinel animals. Wild animal surveillance requires the collection of liv e or dead wild vertebrates and obtainment of serum or tissue for detection of serologic or viro logic evidence of arboviruses. Domestic animal surveillance uses animals associated with humans that are exposed to natu ral vector populations (i.e., backyard chicken flocks, cattle). Sentinel animals are those that are purposely exposed to natural vector populations in a permanent location and routinely tested for antibodies to arboviruses. Any vertebrate su rveillance system used requires years of baseline data during epidemic and non-epidemic periods to determin e antibody and virus isolation thresholds for prediction and assessment of arboviral ac tivity (Bowen and Francy 1980, Day 1989). Live, Wild Vertebrate Surveillance Live, wild vertebrate surveillance requires the active capture of the vertebrate and collection of serum or secretions that may contain virus or antibodies to the arbovirus of concern. Virus isolation is sometimes performed in c onjunction with antibody de tection to establish potential vertebrate reservoir and amplification hosts (Scott 1 988). The use of mobile, wild

PAGE 33

33 animals to monitor arboviral activity provides a gr eater range of area to be sampled per sampling effort (Komar 2001, Trainer 1970). Wild bird surveillance Wild birds are the primary verteb rates that are tested for serosurveys and virus isolation in eastern United States as they are the primary reservoir and am plification hosts of EEEV, SLEV, WNV, and HJ. Wild bird surveillance can target either peridomestic birds associated with epidemic transmission cycles (Gruwell et al. 200 0) or non-peridomestic birds associated with enzootic transmission cycles (Crans et al. 1994). Peridomestic birds associated with epidemic transmission cycles are those that have been introduced to the United States such as pigeons ( Columba livia ), House Sparrows ( Passer domesticus ), House Finches ( Carpodacus mexicanus ), Mourning Doves ( Zenaida macroura ), and European Starlings ( Sturnis vulgaris ) (Allison et al. 2004, Bowen and Francy 1980, Gruwell et al. 2000, Komar et al. 1999, Komar et al. 2001, Reisen et al. 2004). When virus levels rise sharply in peridomes tic birds, there is an imminent threat to humans (McLean et al. 1983). Bowe n and Francy (1980) sugge st that the risk of arboviral transmission of SLEV to humans is h igh when the wild bird antibody prevalence in urban settings is above 15 percent. The level of risk of arbovirus transmission to humans in relation to antibody prevalence in wild birds varies based on the vector/host/disease relationships of the region. When a bird popul ation involved in epizootic tran smission cycles has high levels of non-immune birds the lik elihood of an epizootic is greater as there ar e many susceptible birds present (Day 2001). The strength of wild bird surveilla nce is that it can accurately sample the arboviral activity of amplifica tion hosts over a wide area as o pposed to sentinel birds which represent focal arbovirus tran smission in a fixed location. Direct arbovirus isolation is rare but it does provide the be st indicator of vertebrate involvement in transmission cycles (Scott 1988). As direct virus isolation is rare, many

PAGE 34

34 investigators have relied on antibody surveys to de termine past arbovirus infection history of wild birds. Birds with high field seroprevalan ces only indicate past infections, not a greater potential for viral amplification. Species of bi rds with high field seropevalance need to have laboratory studies of viremia titers and antibody responses to arboviral challenge to determine their importance as arborviral amplifiers (Scott 1988). Birds hatched the year that a serosurvey was conducted show recent arbovirus transmission (Beveroth et al. 2006, Holden et al. 1973). Birds less than on e month old with low level antibody response most likely have maternal anti bodies present (CDC 2003). Birds older than one year may have been infected before the year of the serosurvey. In the eastern United States, hatching year pigeons and House Sparrows are often sampled for SLEV and WNV virus isolation (Allison et al. 2004, Bowen and Francy 1980, Komar et al. 2001). Large populations of nestling pigeons and House Sparrows occur in pe ridomestic habitat and an increase in arbovirus isolation rate from these species can predict transmission to humans (Holden et al. 1973). Previously banded and blood sampled birds that are recaptured provide a precise time frame of arboviral infection. Unfortunately there are generally low reca pture rates during serosurveys (Day and Stark 1999, Scott et al. 2001). When wild bird surveillance data are coll ected in a systematic manner, wild bird surveillance can provide an ear ly warning of arbovirus transmission to humans (Bowen and Francy 1980, Holden et al. 1973). Wild mammal surveillance Wild mammals, primarily rodents, have been used for serosurveys and virus isolation of California Encephalitis virus complex (family Bunyaviridae genus Bunyavirus CEV), Everglades virus (family Togaviridae genus Alphavirus EVEV), EEEV, and SLEV in Florida (Bigler and Hoff 1975, Chamberlain et al. 1969, Day et al. 1996, Jennings et al. 1968, Wellings

PAGE 35

35 et al. 1972). Wild mammals are the reservoir hosts for CEV (Bigler a nd Hoff 1975) and EVEV (Day et al. 1996), possibly SLEV (Bigler a nd Hoff 1975, Day et al. 1995, Herbold et al. 1983), and are incidental, dead end hosts for WNV and EEEV. Everglades virus is only found in south Fl orida (Calisher and Karabatsos 1988). The primary reservoir hosts of EVEV are rode nts, especially Hispid Cotton Rats ( Sigmodon hispidus ) (Coffey et al. 2004) and Cotton Mice ( Peromyscus gossypinus ) (Chamberlain et al. 1969). Rodents most commonly found infected with viru s or presence of antibodies to EVEV are the Cotton Rat, Rattus sp., and Cotton Mice (Bigle r and Hoff 1975, Chamberl ain et al. 1969, Day et al. 1996). In addition to rodents, the raccoon ( Procyon lotor ) was commonly found positive for antibodies to EVEV in South Flor ida (Bigler 1971). Clinical EVEV disease in humans is rare in South Florida and there is no routine survei llance of wild mammal s (Sudia et al. 1969). St. Louis encephalitis virus has been isolated from several wild bat species in the United States (Allen et al. 1970, Herbol d et al. 1983). Antibodies to S LEV virus have been found in deer ( Odocoileus virginianus ) (Trainer 1970), cotton mouse ( Peromyscus gossypinus ), opossums ( Didelphis marsupialis ), and raccoons (Bigler 1971). The raccoon specimens were obtained during rabies surveys and demonstrates the bene fits of testing mammals for arboviruses that were passively obtained from other wild animal collections (Bigler 1971). Day et al. (1995) found 31% (59 /189), armadillos positive for SLEV antibodies in Florida. Wild mammal populations have not been associ ated with WNV amplification in the United States. Pilipski et al. (2004) found 2% (2/83) bats positive for WNV antibodies in New Jersey and New York. One percent (2/149) of Mexican free-tailed bats ( Tadarida brasiliensis ) from Louisiana were positive for WNV antibodies (Dav is et al. 2005). Ne ither Big brown bats ( Eptesicus fuscus ) or the Mexican free-taile d bats are likely amplifi cation hosts of WNV because

PAGE 36

36 of absent or low viremia titers (Davis et al. 2005). In laboratory tests, the Eastern cottontail rabbit, Sylvilagus floridanus developed sufficient WNV viremia (>105 CID50s/mL) to infect feeding mosquitoes (Tiawsirisup et al. 2005). Cottontail rabbits are found in peridomestic habitats and could play a role in peridomestic transmission cycl es of WNV (Tiawsirisup et al. 2005). Wild reptile surveillance Reptiles have been suggested as poten tial overwintering hosts for EEEV and WEEV because of their long lasting viremias (Bowen 1977, Gebhardt and Hill 1960, Hayes et al. 1964) but have not been used in routine surveillance oper ations in the United States. This is due to the safety risks associated with th e capture (e.g., snakes and alligat ors) and difficulty of bleeding reptiles (e.g., turtles) (Sudia et al. 1970). Kle nk and Komar (2003) found the Green Iguana ( Iguana iguana ), Florida Garter Snake ( Thamnophis sirtalis sirtalis ), and Red-ear Slider ( Trachymes scripta elegens ) to be unlikely reservoir hosts for WNV because of low tittered viremia < 103.2 PFU/mL of serum. Young alligators circ ulate WNV viremia titers as high as 106.2 PFU/mL for a duration of one to two weeks which makes them potential amplifying hosts (Klenk et al. 2004). The role that wild alligators play in WN V epidemiology has not yet been determined (Klenk et al. 2004). Dead, Wild Bird Surveillance West Nile virus epidemics in the United Stat es have been characterized by high mortality in American Crows ( Corvus brachyrhynchos ) which has led to the de velopment of dead bird surveillance (Eidson 2001, Eids on et al. 2001a, 2001b, 2001c; Johnson et al. 2006, Julian et al. 2002, Komar 2001, Marfin et al. 2001). Dead bird surveillance is a passive system that is established by state and local h ealth departments for the public to report dead birds (Eidson 2001). Dead bird surveillance systems compile dead bird sightings, and dead bird testing

PAGE 37

37 involves collection and testing dead bird specimens for West Nile virus infection (Eidson 2001). Results from both dead bird sightings and dead bi rd testing systems are tabulated and reported by county health departments to Arbo-NET (Marfin et al. 2001). Collection of dead or dying bird specimens fo r testing of WNV can be used to determine WNV presence in an area. Although this system is considered active it still relies on passive public reports for the location of specimens (Eidson 2001). By 2006 the Centers for Disease Control and Prevention had reported 285 species of birds that tested positive for WNV (CDC 2006c). In the northern United States crows have been found to be the most sensitive indicator of WNV, whereas in the southern United States Blue Jays have b een the most sensitive indicator of WNV (CDC 2003). Problems with dead bird testing include: dela ys in test results (E idson 2001), reliance on public interest for dead bird reports (Eidson 2001, Julian et al. 2002, Ward et al. 2006), the location the dead bird was found may not be th e area that the bird was infected (CDC 2003), dead bird surveillance data from different areas with different co llection protocols are difficult to compare regionally (CDC 2003), and the inability to estimate incidence of virus infection in the wild bird population (Eidson 2001). In Florida, the Fish and Wildlife Conser vation Commission (FWCC) manages a bird mortality database. This system allows th e public, county health departments, or FWCC personnel to report dead bird si ghtings. Dead birds are necr opsied and test ed by reverse transcriptase-polymerase chain reaction (RT-PC R) for WNV at the DOH Bureau of Laboratory Services, Tampa (Blackmore et al. 2003). There have been numerous retrospective studies demonstrating the effectiveness of dead bird sightings or dead bird testing systems at foreshadowing human WNV incidence in the

PAGE 38

38 northeastern United States (Eidson et al. 2001a 2001b, 2001c, 2005; Guptill et al. 2003, Julian et al. 2002, Watson et al. 2004a). Unfortunately dead bird surveillance and dead bird testing have been implemented in many areas wi thout proper real-time analysis of the data co llected. This leads to systems that merely detect WNV in a given area but do not f unction as surveillance because they cannot predict the risk of human infection (Guptill et al. 2003). Domestic Animal Surveillance Domestic animal surveillance is the use of domesticated or farmed animals for either passive or active arbovirus surveillance. Passive animal surveillance relies on local veterinarians to diagnose infected animals and report cases to animal morbidity reporting systems or public health officials (Nichols and Bigl er 1967). In Florida, specimens or tissue samples are sent to the Department of Agricultural and Consumer Servic es (DACS) Division of Animal Industry and Bureau of Diagnostic Laboratory Activities or the Department of Health (DOH) Laboratory, Tampa for arbovirus isolation or antibody tests (FDOH 2005). Active domestic animal surveillance involve s the occasional or routine sampling of domestic animals for arbovirus infection. Occasi onal antibody serosurveys of domestic animal populations are considered retrospe ctive and should be used in ar eas that lack routine arbovirus surveillance (Nichols and Bigler 1967). Routine antibody serosu rveys or virus isolations from domestic animals provide a known period of infecti on. Drawbacks to the use of domestic animal surveillance include a high tur nover of animals making repeated sampling from the same individual difficult (Endy and Nisalak 2002) a nd occasionally, inadequate knowledge of the animals life history (Dic kerman and Scherer 1983). Domestic mammals that have been used for arboviral surveillance include horses (Monath et al. 1985), cattle (Gard et al. 1988), goats (Peiris et al. 1993), and pigs (Geevarghese et al. 1987).

PAGE 39

39 Passive horse surveillance is most co mmonly used for monitoring EEEV, WEEV, and WNV as horses may develop infections of the central nervous system in response to these viruses but are not considered important reser voir hosts (Scott and Weaver 1989, Trock et al. 2001). Horse cases have preceded human cases in some epidemics of WNV (CDC 2003) and EEEV (Scott and Weaver 1989). In Florida, befo re the statewide use of sentinel chickens, equine morbidity was the best indicator of EEEV activity (Bigler et al 1976), but Hayes and Hess (1964) found no correlation of EEEV between horse and human cases in Florida from 1938 through 1961. In Florida, horse serum is screened by HI assay at the Animal Disease Diagnostic Laboratory (Kissimmee, FL) for Flavivirus and Alphavirus antibodies. Confirmation tests for WNV or EEEV by MAC-ELISA and PRNT are perf ormed at the National Veterinary Services Laboratories in Ames, Iowa (Blackmore et al 2003). Virus isolati on from blood, brain, and spinal cord tissues can be performed with rabb it kidney and Vero cell cultures with Alphavirus isolates identified by complement fixation tests (Ostlund et al. 2001). Many birds, such as emus ( Dromaius novaehollandiae ) and pheasants ( Phasianus colchicus ) have been introduced to North America ar e farmed for game animals and commodity markets (Tully et al. 1992). Massive die offs of farm raised birds in the United States caused by EEEV have been observed in emus (Day and St ark 1996a, Tully et al. 1992), pheasants (Tyzzer et al. 1938), White Pekin duc klings (Dougherty and Price 196 0), and Chukar Partridges ( Alectoris chukar ) (Ranck et al. 1965). Many EEEV outbr eaks in pheasants are driven by birdto-bird transmission (Holden 1955). In Florida case fatality ra tes due to EEEV in emus have been reported as high as 14% (Day and Stark 199 6a), pheasants up to 45%, and partridges up to

PAGE 40

40 19% (Bigler et al. 1976). Human epidemics of EEEV have been preceded by epizootics among gamebirds (Scott and Weaver 1989). Zoological parks should be monitored for avia n mortality (CDC 2003). Many bird species in zoo collections are exotic and may experience high mortality when first introduced to WNV. In New York in 1999, eight different species of exotic birds had deaths attributable to WNV (Eidson et al. 2001a, Rappole et al. 2000). Of the 285 bird species found dead and positive for WNV, 54 species were captive and e xotic to North America (CDC 2006c). Backyard chicken flocks have been used in active surveillance programs, especially in areas where no other sentinel surveillance is rou tinely conducted. Flocks used for surveillance should be no more than 30 birds (Bigler et al. 1976). Large commercial fl ocks are usually reared indoors, inaccessible to mosquitoes, and in num bers that are so large the chance of finding antibodies in birds is remote. Unless chickens are bled on a routine ba sis, backyard chicken flocks are only useful in determination of past arboviral activity and do little to determine current arboviral activity. Chickens less than one year old should be used in serosurveys to detect arboviral activity in the last ye ar (Nichols and Bigler 1967). B ackyard chickens and domestic geese were found to be ideal sentinels for WN V in New York City during 1999 (Komar et al. 2001). In the southeast there have been several reports of farmed American alligator ( Alligator mississippiensis ) die-offs caused by WNV (Jacobson et al 2005, Klenk et al. 2004, Miller et al. 2003). On one alligator farm in Georgia more than 10% of the young alligators succumbed to WNV (Miller et al. 2003). It has been shown that alligators can directly transmit WNV to tankmates (Klenk et al. 2004). Horsemeat infected with WNV and fed to alligators was shown to be the cause of an outbreak in Georgia (Miller et al. 2003). With multiple routes of infection for

PAGE 41

41 WNV in alligators it is difficult to use alligato r die-offs as an indication of arboviral activity circulating in local mosquito populations. Outb reaks on alligator farms are cause for public health concern as direct transmission of WNV to human alligator handlers has been documented (Klenk et al. 2004). Sentinel Animal Surveillance Sentinel animals are vertebrates maintained at a specific locati on, exposed to natural populations of mosquitoes, and routinely tested fo r evidence of arboviral in fection. Evidence of arbovirus infection is determined indirectly by testing blood fo r a seroconversion (production of specific antibodies to arboviruses) or by direct virus isolation (Komar and Spielman 1995). Vertebrate sentinels are critical for establishing proof of virus transmission in the absence of human and horse arbovirus cases or bird die-offs (Day and Lewi s 1991). Caged sentinel animals provide information of arbovirus transmission at an exact locati on (Komar and Spielman 1995). Komar (2001) stated that the ideal sentinel {vertebrate} is unifo rmly susceptible to infection, resistant to disease, rapidly de velops a detectable immune res ponse, easily maintained, presents negligible health risks to handler s, does not contribute to local pa thogen transmission cycles, and seroconverts to the target pathogen before the onset of disease outbreaks in the community. Other factors to consider when selecting a se ntinel animal are aggressiveness toward other animals sharing the cage, its ability to withst and the local environmental conditions, and the host/vector/disease dynamics for the area (Sudia et al. 1970). Birds are commonly used as sentinels but may only represent enzootic transmission cycles between predominantly ornithophilic mosquitoes (i.e., Cx. pipiens Cs. melanura ) and the viruses they vector: whereas a sentinel mammal would indicate spillover of vi rus into an epidemic transmission cycle by incidental feeding on non-ta rget hosts or feeding by mammalophilic bridge vectors (i.e., Cq. perturbans Ae. albopictus ). Selection of the ideal sentinel vertebrate is

PAGE 42

42 complicated as there is no one ve rtebrate that works well for dete ction of all arbovi ruses in every geographic region. Sentinel animal use is furt her complicated by the n eed to select proper monitoring sites to represen t enzootic or epidemic transmission (Komar 2001). Sentinel mammals Mammals are infrequently used as sentin els for arboviruses as small mammals may experience high levels of mortality (Ventura an d Ehrenkranz 1975), larger species are difficult to maintain, and cross reactivity wi th other viruses may confound te st results (Day et al. 1996). Rabbits and rodents have been used as sentinel s for CEV and EVEV in Florida (Jennings et al. 1968, Ventura and Ehrenkranz 1975). Sentinel ma mmals can be bled routinely to monitor seroconversion and mammals that die during routine exposure should be submitted to a laboratory for virus isolation (Sudia et al. 1970). Ventura and Ehrenkranz (1975) exposed sentinel hamsters in South Florida for the de tection of EVEV. Over 40% of the hamsters exposed were killed by predators or stolen (Ventura and Ehrenkranz 1975). Sentinel birds Birds have been used routinely in arbovira l surveillance programs for many decades (Moore et al. 1993). The birds most often used as avian sentinels are chickens, pigeons, pheasant, and quail (Bowen and Francy 1980, Morri s et al. 1994, Reisen et al. 1992, Williams et al. 1972). Within the United States bird species va ry in their usefulness as sentinel animals from region to region based on vector/host/disease dynamic s. In the United States, sentinel birds have been used for arboviral surveillance in Alabam a, Arizona, Colorado, Delaware, Florida, Iowa, Louisiana, Maryland, Nebraska, Nevada, New Je rsey, North Carolina, Tennessee, Texas, and Utah (Komar 2001). The various organizations c onducting avian sentinel su rveillance in a wide range of geographic locations have used differe nt bird species, varying numbers of sites,

PAGE 43

43 numbers of birds exposed per site, heights a nd types of exposure de vices, various bleeding regimens and techniques, and laboratory tests. The number of birds used per site ranges from 1 to 50 and depends on the species and type of exposure device (Bellamy and Reeves 1952, Rainey et al. 1962). Public health and mosquito control resources are limited so the number of bi rds exposed at a site and the number of sites must be chosen carefully (Bowen and Francy 19 80). In densely populated areas fewer sentinel sites are available (Gergis and Pr esser 1988). Large numbers of ch ickens (25) and pigeons (30) were exposed at ten sites between two counties in Los Angeles, the second largest city in America (Gergis and Presser 1988). The use of fewer sentinels, 6 to 10, would reduce space requirements, maintenance and eyesore to th e public. By reducing these problems more sentinels could be dispersed over a wider area and increase the chance of detecting focal transmission points. The reduction of space requi rements, maintenance and eyesore issues may allow for the deployment of sentinels at public facilities (e.g., public heal th departments, pump stations, wastewater treatment plan ts, electrical substations) or pr ivate residences. In urbanized areas where sentinels cannot be placed wild bi rd populations should be tested (Gergis and Presser 1988). Sentinel birds should be placed in the field prior to mosquito activity in northern states or kept in the field thr oughout the year in southern states that experience mosquito activity throughout the year (CDC 2003). Birds should be bled and tested for previous arboviral infection before placement in the field. Sentinel birds are bled on weekly (Day and Stark 1996b), bi-week ly (Crans 1986), or monthly (Reisen et al. 1992) schedules. Blood (0 .5 to 4.0cc) is drawn by brachial or jugular venipuncture (Day 1989, Reisen et al. 1992) or by lancet comb prick (Reisen et al. 2004).

PAGE 44

44 In Florida, during 2005, 33 of 67 counties conduc ted sentinel chicken surveillance (Stark and Kazanis 2005). Most programs typically expose six chickens, with a range of four to eight chickens per site throughout the year. Blood samples are taken vi a veinipuncture of the jugular or brachial veins with a syringe and plunger. Th e blood is dispensed into vacutainers and usually centrifuged at Mosquito Control for the separatio n of blood and serum. Serum samples are then packaged and shipped to the DOH Bureau of Laboratory Services in Tampa, FL. Hemagglutination Inhibition test is the initial screening test us ed to determine antibody response to Alphaviruses or Flaviviruses. Positive HI results for Flaviviruses are identified by IgMcapture enzyme-linked immunosorbent assay (MAC-ELI SA). West Nile virus negative birds are re-bled and screened by the HI test. Samples positive during the second HI test are confirmed and identified as SLEV or WNV by the plaque reduction neutralization (PRNT). Alphavirus positives are confirmed and identified as EEEV or HJ by PRNT. The DOH Bureau of Laboratory Services in Tampa pr ocesses 1200 to 1500 HI screening tests per week. On Fridays the results of all positive chickens in the state are then faxed to every participating county and negative/positive results of each chicken are faxed to the counties that submitted the samples (L. M. Stark, personal communication). Cherry et al. (2001) reported th at fourteen sentinel chicken fl ocks failed to seroconvert for WNV before human infection on Staten Island and in New York City. These areas cover roughly 830 km2 and contain approximately 8.1 million people. With roughly 84 sentinel chickens exposed and 8.1 million humans, there was a very low chance that the chickens were bitten by infected mosquitoes before human inf ection. In New York State, sentinel chicken surveillance has been ceased in favor of mos quito and dead bird su rveillance (Darbro and Harrington 2006).

PAGE 45

45 Sentinel chickens were found to be relative ly insensitive as pred ictors of human EEEV cases in Florida (Bigler et al. 1976, Day and Stark 1996b). This may be due to lower feeding rates on caged sentinels by Cs. melanura the primary vector of EEEV (Scott and Weaver 1989). Sentinel chickens in California were tested for antibodies to Cx. tarsalis salivary gland antigens (Trevejo and Reeves 2005). Chicke ns seropositive for antibodies to Cx. tarsalis salivary gland antigens were more likely to seroconvert to S LEV than salivary gland antigen seronegative chickens (Trevejo and Reeves 2005). Culex tarsalis is the primary vector of SLEV in California (Mitchell et al. 1980) so it follo ws that its absence would l ead to an absence of SLEV transmission. If this same technique of an tibody detection to salivar y gland antigens from Cs. melanura then seronegative chickens for antibodies to Cs. melanura salivary gland antigens could explain a lack of EEEV transmissi on to sentinel chickens in Florida. An important consideration when placing sentin els in the field is the height at which sentinels are exposed. Sentinel chickens ma intained at ground level did not predict WNV transmission before human incidence in New York City (Cherry et al. 2001). Culex pipiens and Cx. restuans are suspected vectors of WNV in New York (Anderson et al. 2004, Apperson et al. 2004). In the northeast, both Cx. pipiens and Cx. restuans host-seek in the tree canopy and at ground level (Anderson et al. 2004, Darbro and Harrington 2006, Deegan et al. 2005). Darbro and Harrington (2006) caught significantly more mosquitoes in the tree canopy than on the ground with house sparrows and chickens as bait and they recommend the use of ground level and tree canopy sentinel surveilla nce. Deegan et al. (2005) ex posed sentinel pigeons at canopy and ground levels and found significantly highe r seroconversion rates for WNV at the canopy level. Sentinel birds exposed at canopy levels are contained in modified lard cans (Bellamy and

PAGE 46

46 Reeves 1952) or the Ehrenberg pigeon trap (Dow ning and Crans 1977) and typically exposed for one night. Interpretation of sentinel bird results can be complicated. Ye ars of baseline data in an area during epidemic and non-epidemic years are requ ired. When conducting regional surveillance, mean annual seroconversion rates (MASR), excl uding epidemic years should be calculated separately for each county (Day and Stark 1996b). When the real time ASR is higher than the MASR there is generally cause for concern. At a regional level the comparison of real time ASRs to the MASR shows increases and decrea ses of arbovirus activity between counties and identifies viral foci that cro ss county lines (Day et al. 1991). At the county level, certain sentinel sites wi ll be more likely to have positive sentinels depending on host/vector/disease dynamics (Day et al. 1991). Certain si tes within counties will more accurately represent enzootic/sylvan tran smission (e.g., swamps, wooded areas) and other sites will represent peridomestic/urban transmissi on cycles (e.g., backyards, urban parks). Single serocoversions of sentinels in enzootic/sylva n transmission sites are typically not cause for public health concern as these sites will e xperience background transmission quite often. Williams et al. (1972) found that sites deep inside sw amps were likely to have elevated levels of EEEV transmission earlier in the season than sites at the edge of the swamp and upland sites. An increase in sentinel seroconversion rates in open, dry sites in sout h Florida may represent imminent threat of arbovirus transmission to humans by Cx. nigripalpus (Day and Carlson 1985). As autumn rainfalls occur and humidity levels increase Cx. nigripalpus will exit previous resting sites (e.g., wooded hammocks) and blood feed in open dry sites (e.g., fields, backyards, urban areas) increasing the like lihood of SLEV transmission to humans (Day and Carlson 1985, Day and Edman 1988). In Florida, seroconversion rates of SLEV above 30 percent over a three

PAGE 47

47 week period are cause for concern (Day and Le wis 1992). High seroconver sion rates earlier in the transmission season (May to July) represent gr eater threat to human cases than equally high seroconverison rates later in the season (August to December) (Day and Lewis 1992). There are many positives to using chickens as arbovirus sentinels. Chickens are commercially available and relatively inexpensive ~$5 to 7 (Komar and Spielman 1995). Several strains of chickens are capable of e nduring extreme outdoor cond itions (OBryan and Jefferson 1991). Chickens are highly attractive to Culex spp. mosquitoes. In the most extreme case in Orange County, FL an estimated 2.3 million mosquitoes, mostly Culex were recovered from one exit trap collection (OCMCD data). Sen tinel chickens are inacti ve after sunset, during the period of Culex spp. blood feeding activity (Day and Edman 1984). Adult chickens survive EEEV (Byrne and Robbins 1961), WEEV, SLEV (LaM otte et al. 1967), and WNV (Senne et al. 2000) infections. Chickens develop detectable antibodies to EEEV within 4 days (Olson et al. 1991), WEEV within 10 days, SLEV within 14 days (Reisen et al. 1994), and WNV within eight days (Senne et al. 2000). Chickens run shor t duration, low-level viremias to WEEV, SLEV (Reisen et al. 1994), and WNV (Lange vin et al. 2001) with little threat of passing the virus to other mosquitoes. During the 1990 SLEV outbreak in Florida sentinel chicken seroconversion rates above the MASR were found in several counties before hu man infection (Day and Stark 1996b). There are disadvantages to using sentinel chic kens. Chickens are larger than most other sentinel birds and therefore more expensive to feed and require more cage space (Komar and Spielman 1995). During epidemics when many sen tinel chickens are sero converting they need to be rapidly replaced it may be difficult to find poultry distri butors with the particular breed (White leghorns in Florida) that is desired (O Bryan and Jefferson 1991). Chickens may become

PAGE 48

48 aggressive toward cage mates leading to mortalit y (OCMCD data). Dela yed results because of antibody development and lab confirmation are a nother major drawback to sentinel chicken surveillance or any avian survei llance (OBryan and Jefferson 1991). It takes 8 to 21 days for HI antibodies to SLEV to develop in chickens (OBryan and Jefferson 1991). It would not be until day 9 (minimum) or day 15 (maximum) post infection that the antibody positive blood sample would be draw n. Initial report of suspected positive would not occur until day 16 to 22 post infection and confirmation would follow 23 to 29 days postinfection (DPI) (OBryan and Jefferson 1991). In the 1990 SLEV outbreak in Indian River C ounty, suspected positives were treated as confirmed so a more rapid mosquito control re sponse could be mounted (OBryan and Jefferson 1991). Even with this system, the actual transmi ssion event occurred two to three weeks prior. For WNV, it takes a minimum of seven days fo r HI antibody development (Senne et al. 2000), so suspect cases would be reported within 15 to 21 DPI and confirmed 22 to 28 DPI. For EEEV it takes a minimum of 4 days for HI antibody development (Olson et al. 1991), so suspect cases would be reported within 8 to 14 DPI. However, confirmation with PRNT takes 1 to 2 weeks so it would be 15 to 28 DPI before a confirmed EEEV case was reported. It is important to wait for confirmation of EEEV tests becaus e Highlands J virus is not of public health importance and therefore control measures for HJ virus are not necessary. As with other avian species (Sooter et al 1954) adult hens may transovarialy transmit maternal HI antibodies to SLEV virus (Bond et al. 1965) and NT antibodies to SLEV and WEEV viruses to their offspring (Reeves et al. 1954). Maternal HI anti body levels tend to be lower than HI levels found after mosquito-bor ne arboviral infection (Bond et al. 1965). Chickens less than 15 days old may have detectable maternal antibodies (Day 2001), which would produce

PAGE 49

49 seropositive chickens that were not infected during field exposure. Often times chickens used as bait in lard can traps are less th an one month old (Rutledge et al 2003). Chickens should be bled and tested before field exposure (Day 1989). Chic kens less than nine days of age should not be used for the surveillance of EEEV as they exhi bit high levels of mortality to EEEV infection (Byrne and Robbins 1961). Chickens should be housed in a mosquito proof enclosure before placement in the field (Day 1989). During the Fl orida SLEV outbreak of 1990, the Indian River Mosquito Control District had numerous sentinel chickens seroconvert before placement in the field because of SLEV infection at the IRMCD because chickens were not held in mosquito proof enclosures (OBryan and Jefferson 1991).

PAGE 50

CHAPTER 2 CAGE ESCAPE TRIALS Introduction Sentinel chickens can provide important informati on about arbovirus acti vity in an area, but do not provide information about which mo squito species are transmitting arboviruses. Rainey et al. (1962) implemented a mosquito trap designed to catch mosquitoes as they approach the chickens, before mosquitoes entered the cag e. A recent addition to some sentinel cage designs in Florida is an exit trap (T. P. Brea ud and D. A. Shroyer, personal communication). These exit traps differ from the traps discussed in Rainey et al. (1962) because they capture mosquitoes after they enter the cages, attempt to obtain a blood meal, and try to exit the cage. Exit traps are placed on top (OCMCD and Volusia County Mosquito Control District) or on the sides of cages (Manatee County and Indian Rive r County Mosquito Control Districts). While exit traps provide evidence of the mosquito speci es attracted to sentinel chickens, they may disproportionately represen t the abundance of mosquito species that enter the cages and attempt to feed. This means that there is a potential to miss certain mosquito species that enter sentinel cages and are able to transmit arboviruses, thereby underestimati ng their importance in arbovirus transmission cycles. There is no universal blueprint for the design of the sentinel chicken cages for mosquito control districts. This has led to the creation of many different t ypes of sentinel cages. Sentinel cages have been designed to hold two chickens (King 1983) to thirty chickens (Rainey et al. 1962). Most sentinel chicken cages used in Flor ida hold six chickens (T P. Breaud and D. A. Shroyer, personal communication). An experimental cage was designed in Janua ry, 2005 and built at OCMCD for this study. The experimental cage was designed to capture as many host seeking mosquitoes that come to

PAGE 51

51 feed on the chickens as possible. Mosquitoes were caught in the exit trap or contained within the sentinel cage where the chickens are held. Th e experimental sentinel chicken cage design is similar to that used by OCMCD with a few modifi cations (Figures 2-1 and 2-2). Modifications to the original OCMCD sentin el cage include: the additi on of a mesh baffle to funnel mosquitoes into the cage, plastic sheets to preven t chicken feces from contacting the mesh baffle, entry portals to allow personnel access to inside of the cage fo r aspiration, and a reduced inner cage to allow room for personne l to aspirate the cage and as a refuge for mosquitoes. It is important to determine the limitations of any mosquito trap used in surveillance. Prior to field trials, the experimental cage was tested fo r its efficiency at draw ing mosquitoes into the cage and its ability to contain mosquitoes once insi de the cage. To determine the efficiency of the experimental cage the baffle a ngle was altered through a range of angles from 0-60. This was done to determine the baffle angle that maxi mized the number of mo squitoes and mosquito species caught in the cage. The mark-release-recapture technique was used with five species of colonized mosquitoes with CO2 and chickens as the attractant The mark-release-recapture technique is the release of a known number of mos quitoes that have been marked by fluorescent powders (Pylam Products Co., Tempe, AZ) and measuring the rates of mosquito recovery. Materials and Methods Laboratory colonies of Ae. albopictus Aedes aegypti (Linnaeus), An. quadrimaculatus Say, Cx. quinquefasciatus and Cx. nigripalpus (USDA-CMAVE strains) we re used to determine whether or not they were able to escape from th e experimental cage, and the likelihood that each species was attracted to and entered the cage. Diurnally active species, Ae. aegypti and Ae. albopictus were released at 9:30 am and aspirated fr om the cage at 5:00 pm. Nocturnally active species, An. quadrimaculatus Cx. quinquefasciatus and Cx. nigripalpus were released at sunset and aspirated from the cage at 8:00 am the next morning.

PAGE 52

52 All trials were conducted with in the confines of an outdoor mosquito cage (9.1 m x 18.3 m x 4.9 m gabled to 5.5 m) at the United States Department of Agriculture, Center for Medical Agricultural and Veterinary Entomology, Gaines ville, FL (USDA-CMAVE). The experimental cage was placed in the center of an outdoor mosquito cage. Al l release trials were conducted within the confines of the outdoor mosquito cage A Hobo data logger (Onset, Bourne, MA), was placed on top of the inner cage to monitor ch anges in temperature, humidity, and light. A Vantage Pro 2 Weather Envoy (Davis, Hayward, CA ), was placed next to experimental cage to monitor wind speed and direction. Rainfall data were obtained from the University of Florida Physics department (UFDP 2005). Mosquito Rearing Aedes aegypti and An. quadrimaculatus eggs were provided weekly by the USDA. Aedes albopictus Cx. nigripalpus and Cx. quinquefasciatus eggs were collected from lab-reared specimens. Five days after Cx. nigripalpus blood feeding, 5 to 7 day old hay infusion was provided for as an oviposition source for gravid adult Cx. nigripalpus females. Four days after Cx. quinquefasciatus blood feeding, distilled well water was provided as an oviposition source for gravid adult Cx. quinquefasciatus females. Four to five egg rafts from either species were immediately placed into plastic containers with di stilled well water. After larvae hatched they were transferred to larval pans with di stilled well water. Eggs from adult Ae. albopictus were collected on moistened brown paper, dried and stored for approxima tely one week in a cool, dry Ziploc (Johnson, Racine, WI) bag. Larvae were reared in plastic pans 50.8 cm x 38.1 cm x 7.6 cm and fed according to the feeding schedule in Table 2-1. Adults were he ld in 41 cm x 41 cm x 46 cm cages and were provided a constant source of 10% sugar water so lution. All mosquitoes were reared at 28 1.0C and 90 4% RH on a 14 : 10, light : dark cycle. Adult females were provided the

PAGE 53

53 opportunity to blood feed on restrained chicke ns obtained by the USDACMAVE. All chickens used for colony feeding were cared for according protocols approved by the University Institutional Animal Care a nd Use Committee (#D469) and were housed at the USDA-CMAVE. Marking Mosquitoes were aspirated from emergence ca ges with a hand-held aspirator (Hausherrs Machine Works, Toms River, N.J.). Except for An. quadrimaculatus mosquitoes were temporarily immobilized with CO2 at a rate of 500 mL/min. Th e high knock-down threshold of colony-reared Anopheles quadrimaculatus to CO2, required them to be chilled for 20 minutes in a refrigerator and then placed on a chill table maintained at 4 C. The number of individuals of each mosquito sp ecies required for each trial were counted placed into cardboard cups with screen lids. Th e inside of the cups were lightly dusted with either red, white, blue, yellow, green, or orange fluorescent marking powder. Each trial and cage were assigned specific colors for that day or night. Mosquitoes used in release trials were six to eighteen days post emergence. A water moistened cotton ball was placed on top of the screen to keep the humidity high. Mosquitoes were given at least half an hour to recover from the cold or CO2 knockdown treatments before being released in tr ials. At the completion of every trial all mosquitoes were aspirated from the experimental cage. Any mosquitoes that were dead inside the cardboard cup were noted as mortality cau sed by the marking process and excluded from statistical analysis. Recovered mosquitoes we re observed under a black light (Adams Apple Distributing LP, Glenview, IL) to determine the trial in which they were released. Chickens White leghorn chickens, 19 to 25 weeks old, were used as the attractant in the experimental cages. Protocols for the care and use of the ch ickens were approved by the UF IACUC, #D996. The chickens were housed behind the Entomo logy and Nematology Department building on

PAGE 54

54 campus, in Gainesville, FL. Chickens housed in this location were exposed to natural populations of mosquitoes prio r to experimental exposure. Experimental Cage Design An outer cage measuring 1.7 m x 1.3 m x 1.2 m (H x W x D) wa s constructed with 2 x 4 and 4 x 4 treated pine lumber and 2.5 cm x 5.1 cm hardware cloth (Figure 2-3). An inner cage measuring 0.7 m x 1 m x 1 m was c onstructed out of 2 x 2 treat ed pine lumber and 1.3 cm x 2.5 cm hardware cloth (Figure 2-4) Chickens were housed in the inner cage. The purpose of the inner cage was to create a double barrier of hardware cloth to prevent predators from entering cage, to allow easy removal of the chickens, pr ovide enough room to maneuver an aspirator, and to provide a refuge for mosquitoes to rest without threat of consumption by chickens. A board measuring 2 x 2 on which chic kens roosted during the night, extended diagonally across the inner cage (Figure 2-4). The inner cage was supported by two 2 x 4 boards spaced 1.9 cm apart, which extended horizont ally across the cage to form the baffle entry slit (Figure 2-3). The inner cage rested on two 5 mm thick plastic sheets. The plastic sheets were cut to the shape of the outer cage and the ba ffle slit. The plastic sheets kept chicken feces from falling on the mesh baffle and prevented mos quitoes from flying underneath the inner cage. Two pieces of 0.5 x 2 treated pine lumber were attached to the bottom of the inner cage to prevent chickens from contacting their feces (Figure 2-4). The inner cage was coated in water sealant to prevent feces and urine from soaking into the wood. A one-gallon water container and feeder were placed in the cage when chickens were present. The outer cage was sealed with 12-count mesh screening. Male Velcro was stapled to the outer cage. Twelve-count mesh screening was attached to the cage by female Velcro. The mesh screening was removed when the cage was not used.

PAGE 55

55 Portals (20 cm x 23 cm) on three sides of th e outer cage functioned as doors that allowed personnel to reach inside the cage. The portals were constructe d with 1.3 cm x 2.5 cm hardware cloth and wire ring clamps. The portals were used to release mosquitoes or to aspirate mosquitoes. The portals were sealed with 12-co unt mesh screening attached to the cage with Velcro. An exit trap was constructed in the center of the sheet metal r oof (Figure 2-5 A.). A five gallon water container was used to construct the exit trap. The water container was cut in half and the top half was permanently attached to a ci rcular, sliding piece of sheet metal. The bottom half of the water container was inverted and rested on the top ha lf of the permanently mounted water container (Figure 2-5 B.). Mosquitoes flying upward into the exit trap were funneled through the neck, out the spout, a nd were potentially contained w ithin the removable portion of the exit trap. Tubular Stockinette (Bioqui p, Rancho Dominguez, CA) was attached to the bottom half of the water contai ner and functioned as a seal to keep mosquitoes in the containment portion of the exit trap. Aspirator An aspirator was designed to remove mosquitoes from the experimental cage (Figure 2-6). A modified Bioquip DC Insect Vac 12 volt as pirator (Rancho Dominguez, CA) provided the suction power. A containment jar was connected to the end of the aspirator. Two meters of 10 mm diameter braided vinyl tubing extended from the top of the containment jar. The braided vinyl tubing was the only part of the aspirator that entered the cage. Mosquitoes were sucked through the braided vinyl tubing a nd held inside the containment jar. Mosquitoes were not damaged by the aspirator. Fluorescent marking powders and important morphological characters used for species identification were not remove d or destroyed during th e aspiration process.

PAGE 56

56 Escape Rates The experimental cage was completely sealed as it would be used in the field with only the baffle slit open (Experiments 2 and 3) or closed (Experiment 1). The cardboard cups that contained one-hundred marked mosquitoes were placed inside the sealed, experimental outer cage, on top of the inner cage. The cardboard cu ps were opened rapidly and the outer cage portal and mesh were sealed before mosquitoes could escape the cage. Three escape experiments were conducted: 1) no attractant with the baffle slit sealed, 2) no attr actant with the baffle slit open, and 3) baited with two restrained chickens with the baffle slit open. Chickens were individually restrained inside mesh laundry bags that were cinched tight and allowed no movement of the chickens. The mesh bag holding the chickens were suspended in the center of the inner cage. Three night time and four day time non-baited, closed slit trials we re conducted. Five night time and seven daytime non-baited, open slit trials were conducted. Eight night time and day time chicken baited, open slit trials were conduc ted. A Mosquito Magnet Pro (American Biophysics, North Kingstown, RI) was operated thr ee meters from the experimental cage. The purpose of the Mosquito Magnet was to attract mosquitoes out of the experimental cage. Any trials with greater than 15% marking mort ality for a species were not included in the means analysis. The mean percent of mosquitoes recovered for each species was calculated. The percent of mosquitoes recovered from the exit trap and cage were calculated for the nonbaited open slit and chicken baited open slit trials. The Wilcoxon Rank Sum test was used to compare the percent of mosquitoes captured in the exit trap and cage for the non-baited, open slit and the chicken baited, open slit experi ments. The Wilcoxon Rank Sum test ( P = 0.025) was used to compare the percent of blood-fed mosquitoes recovered in the cage and the exit trap. This was done to determine if blood-fed mosquitoes were more likely to be captured in the exit trap or the cage.

PAGE 57

57 Entry Rates To determine the best baffle angle for trappi ng the greatest number of mosquitoes, markrelease-recapture experiments were conducted using si x different baffle angles (0, 10, 20, 36, 48, and 56). The temporary baffl e angles were created by attach ing fiberglass mesh panels to the baffle slit by Velcro and to the cage legs by thumb push pins. Sewing pins were used to connect the mesh panels. The cage was baited with CO2 (flow rate of 500 mL/min) for six night and day trials. The CO2 tank was placed outside the experime ntal cage and tubing that extended into the middle of the inner cage released the CO2. A different baffle angle (0, 10, 20, 36, 48, and 56) was used for each trial where CO2 was used as bait. There were six night and day trials with three restrained chickens used as ba it and either 0 or 36 baffle a ngles. Two-hundred and fifty female mosquitoes of each species were released at their respective times of activity. Mosquitoes were released outside of the experi mental cage. The number of mosquitoes that were contained in the experimental cage and the ex it trap at the end of each trial were recorded. Cardboard cups that contained marked mosqu itoes of each species were placed against the back wall of the outdoor mosquito cage. The expe rimental cage was then baited with either CO2 or chickens. After the CO2 or the chickens were set in place the mosquitoes were released. Mosquitoes were aspirated from the cage at the termination of each trial. Results Escape Trial Results Mean recovery rates from the nobait, slit closed trials were Cx. nigripalpus 89.4 2.4%, Cx. quinquefasciatus 84.9%, An. quadrimaculatus 91.1% (Figure 2-7), Ae. albopictus 74.1 4.8%, and Ae. aegypti 86.6 2.8% (Figure 2-8). Mean recovery rates from the no-bait, slit open trials were Cx. nigripalpus 74.0 8.4%, Cx. quinquefasciatus 55.3 6.4%, An. quadrimaculatus 80.0 11.5% (Figure 2-7), Ae.

PAGE 58

58 albopictus 65.6 5.8%, and Ae. aegypti 71.1 5.6% (Figure 2-8). Of the mosquitoes recovered, the mean percent recovered from the exit trap were Cx. nigripalpus 43.6 13.3%, Cx. quinquefasciatus 25.2 11.8%, An. quadrimaculatus 33.7 6.1%, Ae. albopictus 4.4 1.4%, and Ae. aegypti 11.4 3.5%. Cage aspiration collections were significantly greater than cage aspiration collections for Cx. quinquefasciatus An. quadrimaculatus Ae. albopictus and Ae. aegypti (Wilcoxon Rank Sum test, P < 0.025). Mean recovery rates from the ch icken baited, slit open trials were Cx. nigripalpus 77.7 6.8%, Cx. quinquefasciatus 47.4 10.9%, An. quadrimaculatus 65.8 12.1% (Figure 2-7), Ae. albopictus 58.7 4.6%, and Ae. aegypti 57.0 5.5% (Figure 2-8). Of the mosquitoes recovered the mean percent recovered from the exit trap were Cx. nigripalpus 28.8 9.2%, Cx. quinquefasciatus 23.0 11.7%, An. quadrimaculatus 34.2 12.3%, Ae. albopictus 4.2 0.7%, and Ae. aegypti 3.2 1.0% (Figure 2-9). Cage aspiration collections were significantly greater than exit trap collections for Cx. nigripalpus Cx. quinquefasciatus Ae. albopictus and Ae. aegypti (Wilcoxon Rank Sum test, P < 0.025). The mean percent of blood-fed mosquitoes caugh t in the cage compared to exit trap for the chicken baited, open slit trials were Cx. nigripalpus cage 51.0 16.7% an d exit trap 37.2 10.9%, Cx. quinquefasciatus cage 35.3 12.8% and exit trap 0 0%, An. quadrimaculatus cage 57.8 11.0% and exit trap 7.3 5.7%, Ae. albopictus cage 42.8 8.6% and exit trap 12.5 12.5%, and Ae. aegypti cage 42.6 8.7% and exit trap 0 0%. The percent of blood-fed mosquitoes were significantly higher in the cage than exit trap for Cx. quinquefasciatus An. quadrimaculatus Ae. albopictus and Ae. aegypti (Wilcoxon Rank Sum test, P < 0.025). Cage Entry Results Mean percent recovery rates fo r the 0 baffle trials were Cx. nigripalpus 0.27 0.27%, Cx. quinquefasciatus 0.14 0.14%, An. quadrimaculatus 1.61 0.83%, Ae. albopictus 0.4 0.0%,

PAGE 59

59 and Ae. aegypti 0.4 0.23%. Mean percen t recovery rates for the 36 baffle trials were Cx. nigripalpus 0.14 0.14%, Cx. quinquefasciatus 0.14 0.14%, An. quadrimaculatus 1.34 1.34%, Ae. albopictus 0.13 0.13%, and Ae. aegypti 1.07 0.13%. No mosquitoes were recovered from the exit trap during the cage entry trials. Discussion Trials with mosquito marking mortality grea ter than 15% were associated with older females (>16 days post emergence) and with lowered blood-feeding su ccess in the chicken baited, open slit trials. This behavioral change and increased marking mortality suggested that older (>16 days post emergence), colony-raised mosquitoes may not behave the same as younger mosquitoes (<16 days post emergence). Therefore, trials with mosquito marking mortality greater than 15% were not included in data analysis. The Mosquito Magnet rarely ca ught mosquitoes that were released during trials. The Mosquito Magnet did capture mosquitoes after the trial had ended as determined by analyzing mosquitoes for fluorescent marking powders. The Mosquito Magnet data was not presented because of the low capture rate of escap ed mosquitoes during release trials. The no-bait, closed slit trials represented th e likelihood that mosquito es would be captured in the cage and recovered dur ing the aspiration process. Mean recovery rates for Cx. nigripalpus Cx. quinquefasciatus An. quadrimaculatus and Ae. aegypti ranged from 84.9% to 91.1% (Figures 2-7 and 2-8). Aedes albopictus mean recovery rate was 74.1 4.8% (Figure 28). Aedes albopictus may be more adept at escaping from the cage or avoidi ng aspiration than the other four species. Colony raised Ae. albopictus appeared to be smaller than field caught specimens (personal observation). Numerous Ae. albopictus individuals were caught in the fiberglass mesh screen. It appeared as if the Ae. albopictus may have been trying to twist their

PAGE 60

60 way through the mesh. If colony-raised Ae. albopictus can escape through the mesh, then this might explain the lowered rec overy rates for this species. Mean recovery rates from the non-baited, open s lit trials were less than non-baited, closed slit trials ( Ae. albopictus 8.5%, An. quadrimaculatus 11.5%, Cx. nigripalpus 15.4%, Ae. aegypti 15.5%, and Cx. quinquefasciatus 29.6%) (Figures 2-7 and 2-8). This shows that mosquitoes may escape through the baffle slit. Culex quinquefasciatus appeared to be able to escape the cage through the baffle slit. When there was no bait, lab-raised Cx. quinquefasciatus An. quadrimaculatus Ae. albopictus and Ae. aegypti were caught in the cage more often than the exit trap (Wilcoxon Rank Sum test, P < 0.025). Aedes albopictus was observed to fly freely between the exit trap and cage wh ich shows that the exit trap co llections may not contain all mosquitoes that originally entered the exit trap. Mean recovery rates from the chicken-baited, open slit trials were less than non-baited, open slit trials for Ae. albopictus 6.8%, Cx. quinquefasciatus 7.9%, An. quadrimaculatus 13.8%, and Ae. aegypti 14.1% (Figures 2-7 and 2-8). The two re strained chickens were not capable of ingesting mosquitoes, which would have reduced r ecovery rates. Some mosquito species may be more likely to escape the cage after blood-feedin g. The percent mosquitoes blood-fed and the percent captured in the cage compared to the exit trap were significantly higher for Cx. quinquefasciatus Ae. albopictus and Ae. aegypti These three colony raised species avoided entry into the exit trap wh en blood-fed. No blood-fed Cx. quinquefasciatus or Ae. aegypti were recovered from the exit trap. The percent of mo squitoes captured in the exit trap shows that Ae. albopictus and Ae. aegypti are infrequently captured in the exit trap (Figure 2-9). Culex nigripalpus had the highest percent of blood-fed mosquitoes in the exit trap (37.2 10.9%) and was the only species not to have significantly (Wilcoxon Rank Sum test, P > 0.025)

PAGE 61

61 more blood-fed mosquitoes caught in the cage th an the exit trap. These findings show that blood-fed Cx. nigripalpus were more likely to be collected in the exit trap than the other four species. These results should be interpreted with caution because colony-raised mosquitoes may behave differently from wild mosquitoes. Th ese results should only be used to show the potential range of differences th at may exist among wild mosquito species with respect to their ability to enter, fly w ithin, and exit the cage. Recovery rates of mosquitoes released in CO2 and chicken baited cage entry trials were poor. Low recovery rates were c ontributed to many factors. It was very dry and hot during the time periods that mosquitoes were released, with maximum daytime temperatures reaching 40.5C on many days. There was only one evening with rainfall (>20 mm). Increased temperatures combined with reduced humidity leve ls may have reduced mosquito flight activity. Within the outdoor mosquito cage there were num erous species of lizards and frogs that may have consumed mosquitoes before they entered th e experimental cage. Gl ue boards were used to reduce lizard populations within the outdoor mosqu ito cage. Directly adjacent to the outdoor mosquito cage were several trees that harbored p opulations of wild birds. Released mosquitoes may have been more attracted to wild birds than the CO2 or chickens inside the experimental cage. Time constraints only allowed a singl e day and night trial to be conducted for CO2 baited cage entry trials which eliminated st atistical analyses of these data. Cage entry trials provided little insight to which baffle angle w ould maximize mosquito catches in the field. An angle of 33 was chosen as the permanent baffle angle for the experimental cage. This angle was chosen b ecause it allowed a gap of 0.4 m from underneath the front and rear edges of the baffle to the ground (Figures 2.2 and 2.4). The gap allowed

PAGE 62

62 mosquitoes flying at ground level to 0.4m above ground level to appr oach the cage from all four sides. The permanent baffle was constructed with 2 x 4 and 2 x 2 treated pine lumber (Figure 2-4). Male Velcro was stapled to the baffle. Twelve-c ount mesh screen was attached to the baffle by female Velcro. Conclusions The experimental cage was capable of containi ng released mosquitoes. Mosquitoes were capable of exiting the cage through the ba ffle slit. Certain species such as Cx. quinquefasciatus Ae. aegypti and Ae. albopictus were more adept at escaping fr om the experimental cage than Cx. nigripalpus and An. quadrimaculatus Culex nigripalpus Cx. quinquefasciatus and An. quadrimaculatus were frequently recovered fr om the exit trap, whereas, Ae. aegypti and Ae. albopictus were infrequently rec overed from the exit trap. Orange County Mosquito Control Division relies only on exit trap co llections to assess mosquito species diversity and abundance of mos quitoes attracted to se ntinel chickens. The experimental cages were used in the field to compare the number of mosquitoes successfully collected in the exit trap to the number of mosquitoes collect ed in the cage (Chapter 6).

PAGE 63

63 Table 2-1. Larval feeding regimen Species Slurry Ingredients Amount Fed Feeding Schedule Cx. quinquefasciatus 1%BLP and 1%BY 20mL daily Cx. nigripalpus 1%BLP and 1%BY 20mL daily Ae. aegypti 3% BLP and 2% BY 50mL every other day Ae. albopictus 3% BLP and 2% BY 50mL every other day An. quadrimaculatus* 1% BLP, 1%BY, 1%HC 50, 25mL first and third day 1% BLP, 1%BY, 1%HC 50 mL fifth and seventh day 1% BLP, 1%BY, 1%HC** dusted fourth and sixth day BLP = bovine liver powder, BY = brewers yeas t, HC = hog chow, fed slurry and powder, ** powder Figure 2-1. Original sentinel chicken cage used by OCMCD

PAGE 64

64 Figure 2-2. Experimental sentinel chicken cage used for research

PAGE 65

65 A B Figure 2-3. Line drawing of experimental cage A) Fr ont view B) Side view

PAGE 66

66 A B Figure 2-4. Line drawing of e xperimental inner cage A) Front view B) Side view

PAGE 67

67 A B Figure 2-5. Sentinel chicken cage exit trap A) Sliding sheet metal with permanently mounted portion of exit trap B) Removabl e exit trap collection device

PAGE 68

68 Figure 2-6. Assembled aspirator

PAGE 69

69 Percent mosquitoes recovered no bait, closed slit no bait, open slit chicken, open slit100 80 60 40 20Cx. nigripalpus Cx. quinquefasciatus An. q uadrimaculatus0 Figure 2-7. Night escape trials percent mosquitoes recovered with no bait, closed slit; no bait, open slit; and chicken bait open slit

PAGE 70

70 Percent mosquitoes recovered no bait, slit closed no bait, slit open chicken, slit open Ae. albopictus Ae. aegypti100 80 60 40 20 0 Figure 2-8. Day escape trials percent mo squitoes recovered with no bait, closed slit; no bait, open slit; and chicken bait, open slit

PAGE 71

71 0 20 40 60 80 100 Cx. nigripalpusCx. quinquefasciatusAn. quadrimaculatusAe. albopictusAe. aegyptiPercent mosquitoes recovered cage exit Figure 2-9. Percent mosquitoes recovere d from experimental cage and exit trap with chicken bait and baffle slit open

PAGE 72

72 CHAPTER 3 SENTINEL CHICKEN DEFENSIVE BEHAVIOR Introduction Studies conducted by Edman and others concerning the defensive behaviors of Ciconiiformes against mosquitoes and the influence of these behaviors on Cx. nigripalpus blood feeding success (Edman and Kale 1971, Edman et al. 1972a, Kale et al. 1972, Maxwell and Kale 1977, Webber and Edman 1972). Edman and Kale ( 1971), Edman et al. (1972a), and Kale et al. (1972) showed the blood feeding success by Cx. nigripalpus to be influenced by host defensive behaviors, host age, host species and mosquito density. Ciconiiform bi rds (i.e., Little Blue Heron, Cattle Egret, Snowy Egret, and White Ibis ) actively display host defensive behaviors after sunset (Edman and Kale 1971), whereas the dom estic white leghorn chickens used in this experiment roost at sunset a nd do not display defensive behavi ors until awakening at sunrise (Appleby et al. 2004). These differing patterns in host defensive behavior may influence the blood feeding success of Cx. nigripalpus Edman et al. (1974) used unrestr ained chickens of various ages to look at the blood feeding success of wild Cx. nigripalpus. Three experiments exposed tw o-week old chicks, eight-week old chicks, and adult chic kens to 200 host-seeking Cx. nigripalpus Culex nigripalpus recovery rates and blood feeding success in creased with older chickens. Recovery rates and percent of blood-fed Cx. nigripalpus were: two-week old chicks 79% and 24%, eight-week old chickens 84% and 59%, and adult chickens 90% and 78%. Chickens six to eight weeks old were found to be moderately tolerant to Cx. nigripalpus when compared to other species of birds and mammals (Edman et al. 1974). Sentinel chickens used fo r arboviral surveillance in Orange County, FL are a minimum of 12 weeks old when first placed in the field (OCMCD data ). In Orange County sentinel chickens are kept in th e field for approximately one year. Therefore, it is important to

PAGE 73

73 evaluate mosquito interactions wi th chickens at an age representative of sentinel chickens used for arboviral surveillance in Or ange County, FL. For the rese arch presented here, chicken defensive behaviors are defined as any behavior that caus es mortality to mosquitoes or interferes with mosquito blood feeding success. The present study was designed to assess the effects of adult ch ickens (25 to 28 weeks old) defensive behaviors on the recovery rates and blood feeding success of Cx. nigripalpus and Aedes albopictus (Skuse) during their normal flight and activity periods. Exit traps were added to sentinel chicken cages in Or ange County, FL in 2003. These exit traps capture mosquitoes as they fly upward from the cage. Exit traps provid e an assessment of the mosquito species that approach sentinel chickens. In the field, exit traps can only collect mosquitoes that enter the cage, survive chicken defensive behaviors, and ente r the exit trap. Exit trap collections may not accurately quantify the mosquito species attracted to sentinel chickens if chicken defensive behavior adversely impacts the surv ival of host-seeking mosquitoes. The effects of adult sentinel chicken defensive behavior on mosquito recovery rates and the location of recovery within the cage were assessed. This assessment will provide insight into how sentinel chicken defensive behaviors might influence mosquito exit trap collections in the field. Materials and Methods Mosquito Rearing Too few wild Cx. nigripalpus were captured in Orange County to produce enough F1 progeny for the experiments and time constraints lim ited the creation of an entirely new colony. Therefore, a cross mating of colony and wild Cx. nigripalpus were used for the present study. Wild, virgin male Cx. nigripalpus and colony, virgin females were cross mated, as this was proven the most successful ma ting combination by Haeger a nd OMeara (1970). Wild, bloodfed Cx. nigripalpus females were obtained from Ora nge County on May 21, 2006 by using white

PAGE 74

74 leghorn chicken-baited lardcan traps (Bellamy a nd Reeves 1952) and sentinel chicken cage exit traps. Chickens used as bait for field collected Cx. nigripalpus were obtained by OCMCD. Mosquitoes were aspirated from traps, placed in cardboard cups, provi ded a 10% sugar water solution, and transported to Gaines ville. Once in the laboratory, the field-collected mosquitoes were immobilized with chloroform and placed on a chill table maintained at 4C. Chilled mosquitoes were identified to species and all mosquitoes that were not blood-fed Cx. nigripalpus females were discarded. Blood-fed Cx. nigripalpus were placed into a 61 cm x 46 cm x 61 cm cage. All adults were provided cotton soaked in 10% sugar water solution. Mosquitoes were maintained at 26.5 2.5C and 90 10% RH in a room exposed to the natural outdoor light cycle. Larvae were fed a 20 mL slurry of 1% bovine liver powder and 1% brewers yeast once per day. Adults were provided blood meals from re strained chickens of va rious breeds (IACUC #D469). Colony and wild Cx. nigripalpus were offered restrained chic kens to blood feed from on the same day. Five days after blood feeding, 5 to 7 day old, hay-infusion in black plastic cups was provided for wild and colony adult, grav id female oviposition. This blood feeding and oviposition schedule synchronized the em ergence of adult colony and wild Cx. nigripalpus Ten hours after emergence, adult mosquitoes were pla ced in a walk-in cooler (Eskimo Panels Inc., Miami, FL) at 4C to temporarily immobilize th em. Immobilized mosquitoes were aspirated from the cage, transported to the chill table, and sorted by sex. Wild Cx. nigripalpus males were placed into a 61 cm x 46 cm x 61 cm cage and wild females were discarded. This was done three times every 10 hours with each cohort. The same method was used to obtain colony females and discard colony males. By removing recently emerged Cx. nigripalpus from the emergence cage every ten hours it was assured that males would no t be able to inseminate colony females since

PAGE 75

75 there is a 12 to 24 hour period required for male genitals to rotate into the proper position for copulation. Colony females were placed into the same 61 cm x 46 cm x 61 cm cage with the wild males. A 40-watt incandescent bulb on a d immer switch was used to create an artificial sunset in the room that would extend dusk, the primary period of mating activity, one hour past sunset. The F1 progeny of this wild x colony cross mated and produced viable offspring. The F2 progeny were used for the experiments reported here. Wild Ae. albopictus and other mosquitoes were aspirate d as they landed on a human host at a field site in Orange C ounty between the hours of 11:00 am and 1:00 pm on May 22, 2005. Mosquitoes were transported to Gainesville in cardboard cups with 10% sugar water solution soaked cotton balls. Mosquitoes were tem porarily immobilized using 500 mL/min of CO2 lightly blown over them. Aedes albopictus were placed into a 41 cm x 41 cm x 46 cm cage and all other mosquito species were discarded. Adult Ae. albopictus were offered restrained chickens of various breeds for blood meal s (IACUC #D469). Eggs were collected on moistened brown paper, dried, and stored for approximately one week in a cool, dry Ziploc bag. Adults were raised on 10% sugar water solution. Mosquitoes were maintained at 28 1C and 90 4% RH on a 14: 10 L: D cycle. Larvae were fed a 50 mL slurry of 3% bovine liver powder and 2% brewers yeast provided every othe r day. The F2 progeny were used for the experiments reported here. Chickens used for routine colony feeding were of various breeds and were obtained by the USDA-CMAVE. All chickens used for col ony feeding were housed at the USDA-CMAVE. Mark-Release-Recapture Studies Mosquitoes were aspirated from emergence ca ges with a hand-held aspirator (Hausherrs Machine Works, Toms River, N.J.). Th ey were temporarily immobilized with CO2 at a rate of 500 mL/min. The exact number of females re quired for each experiment were counted and

PAGE 76

76 placed into cardboard cups with screen lids. The inside of the cups were lightly dusted with red, white, blue, yellow, green, or orange fluorescent powder (Pylam Products Co., Tempe, AZ). Each trial and cage was assigned a specific color fo r that day or night. A water-moistened cotton ball was placed on top of the screen to keep the humidity high. Mosquitoes were given at least 30 minutes to recover from the CO2 before being released in trials The cup was placed on top of the inner experimental cage, the lid rapidly re moved, and the cage portal and mesh sealed. At the completion of the trial all mosquitoes were aspirated from the cage. Any mosquitoes that were dead inside the cup were noted and excluded from statistical analysis. Mosquitoes were observed under black light (Adams Apple Distri buting LP, Glenview, IL) to determine which trial they were released. Mosquito blood-fed status was recorded. Partial blood meals were counted as blood-fed. The location where each mosqu ito was recovered in the cage at the end of each trial was recorded. All st atistical analyses were perfor med using Wilcoxon Rank Sum test (Ott and Longnecker 2001). Chickens Chickens used for behavior experiments we re 25 to 28 week old white leghorns obtained by Orange County Mosquito Contro l (OCMCD). All chickens used for experiments were cared for under #D996 University of Florida IACUC protocols and were housed on the University of Florida main campus behind the Entomology and Nematology building in Gainesville, FL. Chickens, housed behind the Entomology and Nematology building, were exposed to natural populations of mosquitoes prio r to experimental exposure. Host Defensive Behavior Chicken behavior was monitored during three di fferent time periods (sunset, sunrise, and 11:00 am). Chickens were observed at distance s of 26.5 m or 12.2 m, dependent on light levels and visibility, with 10 x 42 Eagle Optics Ranger Platinum Class binoculars. Some behaviors

PAGE 77

77 were also observed at close range (i.e., eye blink and the physical cons umption of mosquitoes) while mosquitoes were aspirated from the sen tinel cages. Host defensive behaviors were observed immediately after the release of Cx. nigripalpus until all the chickens roosted; at sunrise for 30 min with Cx. nigripalpus ; and for 30 min immediat ely after release of Ae. albopictus Chicken behavior was observed for several minutes prior to release of mosquitoes to determine routine maintenance behaviors. Ever y type of maintenance and defensive behavior displayed was recorded. Defensiv e behaviors observed are presen ted in Table 3-3. Descriptive comparisons were made among the types of beha viors displayed in response to the various species of mosquitoes, densities of mosquito es and densities of chickens (Table 3-4). Effect of Host Defensive Behavior Host defensive behavior tr ials were conducted in outdoor cages at the U.S.D.A. C.M.A.V.E. facility in Gainesville, FL be tween July 26 and August 13, 2005. One chicken was placed in each sentinel cage 1 to 2 h before mosquitoes were released. Three-hundred 6 to 14day-old adult Cx. nigripalpus females were released into each cage at sunset. This experiment was replicated four times. Mosquitoes were re covered from the cages at 8:00 am the following morning. Three-hundred 7 to 12-day-old adult Ae. albopictus females were released into each cage at 11:00 am. This experiment was replicated four times. Aedes albopictus females were recovered by aspiration from the cage at 5:00 pm at the end of each trial. For trials with each mosquito species, one cag e contained an unrestrained chicken that was allowed total freedom of movement. The second cag e contained a restrained chicken that was fit snugly into a nylon mesh laundry bag that restri cted any head, wing, or leg movement. Three individual chickens were used for these expe riments. Chickens were rotated between the restrained and unrestrained treatments and be tween cages. Cages containing restrained and

PAGE 78

78 unrestrained chickens were rotated between each tria l to negate differences in the ability of each cage to contain mosquitoes. All statistical analyses were performed using Wilcoxon Rank Sum test (Ott and Longnecker 2001). To evaluate the importance of mosquito a nd vertebrate host density on mosquito blood feeding success one chicken was placed into one of the cages and four chickens were placed into the second cage one or two hours before the mosqu itoes were released. An equal predetermined (100, 500, or 700) number of 6 to 15-day-old adult Cx. nigripalpus females were released into both cages at sunset. Released mosquitoes we re aspirated from both cages at 8:00 am the following morning. Four trials were conducted with each density of mosquito es. Five individual chickens were used for these experiments. Chic kens were rotated so that no chicken was used more than once by itself in the single chicken ca ge at each mosquito density. Cages containing one and four chickens were rotated between each tria l to negate differences in the cages ability to contain mosquitoes. All statistical analyses were performed using W ilcoxon Rank Sum test (Ott and Longnecker 2001). Results Chicken defensive behaviors had no significant effect on the percent of Cx. nigripalpus females recovered (Wilcoxon Rank Sum test, P > 0.05), their blood feeding success (Wilcoxon Rank Sum test, P > 0.025), or the location in the cage from which the mosquitoes were recovered (Wilcoxon Rank Sum test, P > 0.025). Chicken defensive behaviors were found to si gnificantly lower the blood feeding success of Ae. albopictus females (Wilcoxon Rank Sum test P < 0.025) (Table 3-1); however, these behaviors had no significant effect on the percent of Ae. albopictus recovered (Wilcoxon Rank Sum test, P > 0.05) or the location in the cage from which the mosquitoes were recovered (Wilcoxon Rank Sum test, P > 0.025) (Table 3-1).

PAGE 79

79 No significant differences were found between the one and four chicken trials at any mosquito density (100, 500, or 700) for the percent of Cx. nigripalpus recovered (Wilcoxon Rank Sum test, P > 0.05), the percent that were blood-fed (Wilcoxon Rank Sum test, P > 0.025); or the location in the cage from which the mo squitoes were recovered (Wilcoxon Rank Sum test, P > 0.025). Mosquito blood feeding success in tr ials with one chicken were significantly different between the 100 and 700 mosquito de nsities and between the 100 and 500 mosquito densities (Wilcoxon Rank Sum test, P < 0.025). An inverse correlation was observed as Cx. nigripalpus density increased, blood feeding success decreased when one chicken was present (T able 3-2). When one chicken was present the percent of Cx. nigripalpus recovered from the exit trap differed significantly between the 300 and 500 densities and between the 500 and 700 densities (Wilcoxon Rank Sum test, P < 0.025) with the 500 density being significantly less than the 300 and 700 densities (Table 3-2). Discussion Anderson and Brust (1996) exposed pairs of Japanese Quail ( Coturnix japonica ) to Cx. nigripalpus or Ae. aegypti and video tapped Quail activity be fore and after the release of mosquitoes. They concluded that anti-mosquito behaviors were exaggerations of pre-existing behavioral patterns (Anderson and Brust 1996) This increase of normal body maintenance behaviors in response to mosquito ( Cx. nigripalpus ) biting pressure has also been observed with ciconiiform birds (Webber and Edman 1972). In the current experiments, it appeared that chickens used routine maintenance behaviors as defensive behaviors in response to mosquito biting pressure. Through direct observation it wa s shown that all of the maintenance behaviors displayed in the absence of mosqu itoes were displayed more often in the presence of mosquitoes. The preening and head scratching behavior obser ved in the present study may have been in response to mosquito or louse biting pressure. Menopon gallinae (Linnaeus) (Phthiraptera:

PAGE 80

80 Philopteridae), the shaft louse, was occasio nally found on the chickens used for these experiments. Estimates of louse load were not made in the study but symptoms of heavy infestation (i.e., feather loss and lameness) we re never observed. Whet her preening and head scratching behaviors were in response to mosqu itoes or lice could not be determined. Preening and head scratching, however, occurred more often in the presence of mos quitoes than in their absence. The only defensive behaviors not observe d in the absence of mosquitoes were the wing shrug and head tuck. Wing shrugging behavior was only displayed when four chickens were present at the 700 mosquito densit y, suggesting that it may only be reserved for periods of highly elevated agitation. Several times in the commentary section of Ta ble 3-4 defensive behavi or was described as occurring often in the presence of a particular mosquito species. This means that while the behavior was displayed in the presence of both mosquito species, it was displayed much more frequently in the presence of the mentioned species and therefore is attributed to that species. Aedes albopictus appeared to attempt to blood feed ar ound the face and head regions due to the many defensive behaviors displayed by the chickens su ch as the head rub, head shake, head jerk, head tuck, pecking at the air, and eye blink. Culex nigripalpus appeared to attempt to feed around the legs and feet because the primary respons es elicited from chickens in the presence of this mosquito species were the foot kick/stamp and pecking at the ground and legs. Webber and Edman (1972) observed that Cx. nigripalpus primarily feed around the leg region of ciconiiform birds. Chicken defensive behaviors had a negative influence on the blood feeding success (Wilcoxon rank sum P < 0.025) and caused high mortality of Ae. albopictus (Table 3-1). Similar results were found by Waage and Nondo (1982) in restraint trials with rabbits and Ae. aegypti

PAGE 81

81 Klowden and Lea (1979) found unrestrained rabb its significantly lowered the percent blood feeding success of wild, host seek ing mosquitoes when compared to restrained rabbits. All three of these experiments were conducted during th e photophase period of mosquito host seeking activity and coincided with diurnally active time periods of their hosts. In the experiments reported here, unrestrained chicke ns displayed defensive behavi ors in response to mosquito biting pressure throughout the enti re duration of the trials with Ae. albopictus It could not be determined whether the constant chicken activity wa s due to the time of day or the host feeding behaviors of Ae. albopictus The increased activity of chickens during the daytime, the reduced (80.0 4.1% restrained > 40.1 12.5% unr estrained) recovery rate of Ae. albopictus and significantly (Wilcoxon Rank Sum test, P < 0.025) lowered blood feeding on unrestrained chickens indicate that unrestrained sentinel chickens in the field may reduce the number of daytime biting mosquitoes that enter the cage and reduce the likelihood of them obtaining a blood meal. Culex nigripalpus trials with restrained and unres trained chickens showed a slight reduction in recovery rates of mosquitoes and no differences in blood feeding success (Table 3-1). These results differ from the restraint trials conducted by Edma n and Kale (1971) with nocturnally active ciconiiform birds and Cx. nigripalpus Unrestrained chic kens displayed very few defensive behaviors when Cx. nigripalpus were released at sunset because chickens roost at sunset and wake at sunrise (Appleby et al. 2004 ). However, unrestrained chickens displayed more defensive behaviors shortly after sunrise th an at sunset. Day and Edman (1984) found that mosquito blood feeding success occurs more often during host seeking periods that overlap with periods of inactivity by hosts. Since both the restrained and the unrestrained chicken were

PAGE 82

82 inactive throughout the night, when Cx. nigripalpus feeds, there were no differences in blood feeding success between the two treatments. Trials in which one chicken was e xposed to increasing densities of Cx. nigripalpus showed a significant negative correla tion (Wilcoxon Rank Sum test, P < 0.025) between the number of mosquitoes released and the pe rcent that successfully obtained blood meals (Table 3-2). This negative correlation between incr easing density and blood meal suc cess was counterintuitive. If chickens were inactive during the primary host seeking period of Cx. nigripalpus then there should have been no significant difference in bloo d feeding success at the various densities. At the 500 and 700 Cx. nigripalpus densities little defensive be havior activity was displayed by chickens at sunset, but at sunr ise chickens were very active. The bulk of chicken morning activities involved pecki ng at their legs and the ground. It was observed at close range that pecking at the ground was done in an attempt to consume mosquitoes. After Cx. nigripalpus had engorged it would usually travel a short distance to the nearest resting spot. Dozens of blood-fed Cx. nigripalpus were found resting on the inner cage where they were accessible to chickens. If chickens peck and consume primarily blood fed mos quitoes resting near them, then the chickens would skew the percent of blood fe d mosquitoes that were recovere d. The heighten ed activity of chickens and pecking behavior in the early morn ing make this the most likely time period during which Cx. nigripalpus were consumed. In field settings, this means that mosquitoes should be aspirated shortly after sunrise to reduce the lik elihood of mosquitoes being consumed by the sentinel chickens. No significant trend was found between blood meal success or recovery rates and mosquito density when four chickens were exposed to varying densities of Cx. nigripalpus Roughly 80% percent of all Cx. nigripalpus that were released were r ecovered. Although there were no

PAGE 83

83 significant differences found when comparing th e recovery rates and blood feeding success between one and four chickens, with four chicke ns there was an increase in the percent of bloodfed mosquitoes recovered at the 500 (45.7%) and 700 (56.7%) mosquito density trials compared with trials containing one chic ken at the 500 (35.8%) and 700 (35.7%) mosquito densities. Chickens in groups of four appe ared more concerned with the pr esence of other chickens than with the presence of mosquitoes. Fewer defens ive behaviors toward mosquitoes were observed in cages containing four chickens than cages co ntaining one chicken. Th e four chickens would peck at each other to assert dominance. Aggr ession toward other chickens likely distracted chickens from the presence of mosquitoes whic h allowed more mosquitoes to obtain blood meals and avoid consumption by chickens. In field set tings, four or more chickens should be exposed in the same cage to increase the survival and num ber of day feeding mosquitoes that obtain blood meals. Exit trap collections were not significantly different between restra ined and unrestrained chicken trials for Cx. nigripalpus or Ae. albopictus (Table 3-1). There were no discernable patterns of exit trap recovery ra tes at different densities of Cx. nigripalpus (Table 3-2). It does not appear that chicken defensiv e behaviors influence the percent of mosquitoes captured in the exit trap. An important point to note is the lower percent of Ae. albopictus (8.1%) recovered in the exit trap during restrained chicken trials compared to the percent of Cx. nigripalpus (41.3%) recovered during similar trials. This difference in the percent of mosqu itoes recovered between Ae. albopictus and Cx. nigripalpus during a time when chickens could not display defensive behaviors, demonstrates that Ae. albopictus is not as likely to be r ecovered in the exit trap as Cx. nigripalpus Exit trap collections from sentinel chic ken cages in the field may not accurately

PAGE 84

84 represent all mosquito species or the abundance of those species that come to feed on the sentinel chickens. The absence, or low abundance of cer tain mosquito species in exit traps collections from sentinel chicken cages does not exclude thes e species from being potential enzootic vectors of viral disease in avian populations. Conclusions Chicken defensive behaviors are maintenance be haviors displayed at greater frequencies in the presence of mosquitoes. Chickens may di splay different defensiv e behaviors depending upon the preference of the mosquito blood feeding site. Sentinel chicken defensive behaviors can influence mosquito recovery rates and the blood feeding success of recovered mosquitoes. Diurnally active mosquitoes were more likely th an nocturnal mosquitoes to be consumed and have reduced blood feeding success. Increasing de nsities of mosquitoes feeding on chickens can influence the percent of blood-fed mosquitoes recovered when one chicken was present. At least four chickens should be exposed at sentinel si tes to reduce the number of mosquitoes that are consumed and increase the number of mosquitoes th at obtain blood. Exit tr ap collections did not appear to be influenced by chicken defensive beha viors. Exit trap collections may not accurately sample all of the mosquito species that are attracted to sentinel chicken surveillance cages.

PAGE 85

85 Table 3-1. Effect of host defe nsive behavior on recovery, blood feed ing success, and location recovered* Percent mosquitoes recovered Percent mosquitoes e ngorged Percent mosquitoes recovered from exit trap Species Restrained Unrestrained Restrained Unrestrained Restrained Unrestrained Cx. nigripalpus 90.2 3.1 77.7 11.0 50.2 7.2 52.7 8.6 41.3 10.3 56.5 3.8 Ae. albopictus 80.0 4.1 40.1 12.5 22.2 3.4a 6.1 1.9b 8.1 1.4 18.9 6.7 Means and standard errors of recovered mosquitoes as a percen t of total mosquitoes released, engorgement and exit trap as a percent of mo squitoes recovered. Table 3-2. Effect of Cx. nigripalpus density on recovery, blood feeding success, and location recovered when one or four chickens were present* Percent mosquitoes recovered Percent mosquitoes e ngorged Percent mosquitoes recovered from exit trap Density One chicken Four chickens One chicke n Four chickens One chicken Four chickens 100 74.2 12.2 80.1 3.6 58.9 4.2a 58.2 5.8 54.2 6.9a,b 43 10.9 300a 77.7 11.0 -52.7 8.6a,b -56.5 3.8a -500 86.3 1.7 76.2 6.9 35.8 4.1b 45.7 8.8 37.8 5.2b 45.2 7.2 700 82.6 3.3 85.6 1.9 35.7 6.2b 56.7 8.5 67.1 3.6a 63.0 2.5 Means and standard errors of recovered mosquitoes as a percen t of total mosquitoes released, engorgement and exit trap as percent of mosquitoes recovered. a Data from unrestrained tr ials presented earlier.

PAGE 86

86 Table 3-3. Behaviors displayed by ch ickens in response to mosquitoes Behavior Description Eye blink An eye was closed and opened rapidly Head jerk Head was moved forward and backward or side to side in a rapid jerking motion Peck The head was moved forward and backward in a pecking manner The behavior was used to p eck at the air, ground or legs Head shake The head was vigorously shaken side to side Head scratch The foot was used to scratch the neck and head region Foot kick/stamp The foot was quickly raised into the air and lowered Wing flap The wings were spread out and retracted in a manner similar to flapping Hind feather ruffle The rear portion of the chicken s hook, ruffling the feathers outward away from the body Preening Chicken used its beak to scratch amongst feathered regions Head rub Chicken rubbed its head against vigorously against its body Head tuck Chicken buried its head beneath its wing for extended for seve ral minutes without preening Wing shrug Chicken lowered its body close to the ground and shrugged its wings upward and kept wings pressed against its body

PAGE 87

87 Table 3-4. Defensive behaviors di splayed by chickens in the presence of mosquitoes with commentary Behavior Comments Eye blink This behavior occurred when Ae. albopictus flew near the chickens face. Head jerk A head jerk was often performed in response to Ae. albopictus presence. Peck Pecking at the air was often performed in response to Ae. albopictus presence. Pecking at the ground or legs was often performed in response to Cx. nigripalpus Head shake This was most often performed in the presence of Ae. albopictus It was probably done in attempt to dislodge a feeding mosquito from the face, head or neck region. Head scratch This behavior may have been performed in response to mosquitoes or to the shaft louse, Menopon gallinae (L.). This behavior was observed mo re frequently in the presence of Ae. albopictus. Foot kick/stamp This behavior was freque ntly displayed at densities of 500 and 700 Cx. nigripalpus. Wing flap This behavior was usually observed when a ch icken repositions itself before going to roost and may not have been in response to mosquito attack. Hind feather ruffle This behavior was performed in response to Ae. albopictus and Cx. nigripalpus Preening This may have been a response to mosquitoes or lice. Preening was performed over all feathered regions of the body and in the presence of both mosquito species. Head rub This behavior was displayed in response to Ae. albopictus in attempt to dislodge mosquitoes from the face and head regions. Head tuck This behavior was displayed only in presence of Ae. albopictus and was likely done to prevent mosquitoes from feeding near face and head regions. Wing shrug This behavior was only observed at the 700 density of Cx. nigripalpus with four chickens present.

PAGE 88

88 CHAPTER 4 DESCRIPTION OF FIELD SITES: ARBOVIRUSES/VECTORS/HOSTS Introduction Orange County Mosquito Control Division (OCMCD) conducts arbov irus surveillance with sentinel chickens year round. There are twelve sentinel chicke n sites maintained by OCMCD in Orange County. Two of the twelve sentinel chicken sites were chosen for experimental sentinel chicken surveillance and mosquito collec tion for the pres ent study. The two sites were Tibet-Butler Pres erve (TBP) and Moss Park (MP). Three factors led to the decisi on to conduct field research at TBP and MP. The first factor was frequent seroconversions of sentinel chickens from Oct ober through June. Field studies were conducted from October 12, 2005 to July 4, 2006. Historically, several EEEV and WNV sentinel chicken seroconversions have occurred during October through June in Orange County. The second factor for site selection was higher mosquito species diversity as determined by OCMCD sentinel chicken exit trap collections from 2003 and 2004. Mosquito species diversity ranged from fourteen (Winter Park and Clar cona-Horseman Park) to twenty-nine (Fort Christmas) mosquito species in sentinel chicken exit trap collections (OCM CD data). The third factor for site selection was lower mosquito ab undance as determined by sentinel chicken exit trap collections from 2003 and 2004. Sentinel ch icken exit traps at some locations in Orange County have collected 60,000 (Blanchard Park) Cx. nigripalpus in one night (September 21, 2004; OCMCD data). Lower mosquito abundance in exit traps would reduce the cost of testing mosquitoes for arboviruses for this project. Tibet-Butler Preserve and MP are mana ged by Orange County Parks and Recreation (OCPR). Both parks are located outside of the Orlando metropolitan area. As the human population continues to expand and spread aw ay from Orlando, the parks have become

PAGE 89

89 surrounded by residential areas. While both parks allow the preservation of unique habitats and provide excellent opportunities for human recreati on, they are also reserv oirs for large numbers of birds and mosquitoes which may circulate arboviruses that affect humans. The close association of humans and wildlife at both pa rks calls for in-depth study to understand the ecology of arbovirus transmission at these sites. Understanding ar bovirus transmission cycles at TBP and MP would be beneficial for OCMCD and the Orange County Public Health Department as they attempt to reduce the risk of arbovirus transmission to humans. Orange County is located along the Atlantic Flyway, a major bird migration route along the east coast of the United States (Lord and Ca lisher 1970). The hot a nd moist environment of Orange County provides suitable habitat for migran t birds, transient (birds migrating further south or north that use Florida as a temporary stopover site) bird s, and resident birds. Some birds, such as the Red-winged Blackbird and Wood Stork, have populations that make northcentral Florida their permanent area of residence, while other populations of the same species use this area to overwinte r (Favorite 1960). Site Descriptions Tibet-Butler Preserve The 438 acres known as the TBP is owned by the South Florida Water Management District. Tibet-Butler Preserve is located in the southwest co rner of Orange County along the Tibet-Butler chain of lakes (Figure 4-1). Public activity within the park is limited to daytime use, which reduces the potential for arbovirus transmission to humans by nocturnally active mosquitoes inside park boundaries. Immediatel y to the south of the park are active, dripirrigated orange groves, and undeveloped land (Figur e 4-2). Adjacent to the southwest corner of TBP is residential housing. Adjacent to the east and west of TBP, orange groves have been

PAGE 90

90 converted to high-end housing communities. The preserve abuts Lake Tibet to the northeast. Houses are located along the banks of Lake Tibet. Many habitats are present in TBP including bayhead and cypress swamps, freshwater marsh, scrub, and pine flatwoods. The ba yhead swamp is dominated by loblolly ( Gordonia lasianthus ), red ( Persea borbonia ), and sweet bay ( Magnolia virginiana ) trees and comprises the northwestern portion of the park. The fres hwater marsh is dominated by sawgrass ( Cladium jamaicense ) and waterlily ( Nymphaea odorata ). Numerous aquatic plan t species, such as Water Lettuce ( Psitia stratiotes ), Water Hyacinth ( Eichhornia crassipes ), Maidencane ( Panicum hemitomon ), Cattail ( Typha spp.), Pickerelweed ( Pontederia cordata ), and several sedge species were found in the marsh and lake margins of TBP. These aquatic plant speci es create ideal larval habitat for Cq. perturbans Ma. titillans and Ma. dyari (Morris et al. 1990, Lounibos and Escher 1985, Slaff and Haefner 1985, Callahan and Morris 1987). To the south of Lake Tibet is a cypress dom e where large sections of trees were knocked down by the hurricanes of 2004. Upturned cypress ( Cupressus spp.) tree roots create ideal breeding habitat for mosquitoes, especially Cs. melanura (Foster and Walker 2002). To the southwest of Lake Tibet is an undamaged cypress dom e. In the eastern half of the park, there is a mixed wetland forest with pine ( Pinus spp.) flatwoods, bayhead and cypress swamp communities. Various pine, bay, and cypress tr ees create the forest canopy depending on the ground elevation. Various areas of the east side of the park flood and drain throughout the year which provides temporary and semi-permanent pools of water, ideal habitat for Aedes Psorophora Culex and Anopheles larvae (Foster and Walker 2002, personal observation, OCMCD data). The east side of the park is adjacent to residentia l neighborhoods along the

PAGE 91

91 easternmost edge and Aedes albopictus larvae have been found in c ontainers inside this area (personal observation). The southwest portion of the park is xeric, el evated two to three feet above the swamp, composed of scrub and pine flatwood communities, and an understory dominated by saw palmetto ( Serenoa repens ). The OCMCD sentinel chickens a nd those used in th is investigation, were located in this section of the park near the visitor center. The chicke ns were held in areas cleared as fire breaks. In the southwestern corn er of the park, there wa s a small section of oak hammock dominated by Turkey ( Quercus laevis ), Live ( Quercus viginiana ), and Chapmans ( Quercus chapmanii ) oak. Mosquito larvae were rarely found on this side of the park (OCMCD data). Aedes albopictus have been found in containers ne ar the visitor center (personal observation). Moss Park Moss Park is located in the southeast porti on of Orange County be tween Lakes Hart and Mary Jane which are part of the St. Johns Wa ter Management District (Figure 4-1). The surrounding banks of both Lakes Hart and Mary Jane have reside ntial housing. Moss Park is a peninsula of land with Lake Hart to the west, Lake Mary Jane to the east, and Pickerelweed marsh establishing the southern boundary of the park. Moss Park is predominantly mature Laurel ( Quercus hemisphaerica ) and Live Oak hammocks, and Pi ne trees with the understory managed for human recreational activities. In th e southern half of the park camping is allowed and many visitors take advantage of this through out the year. To the west of Moss Park and south of Lake Hart is Split Oak Preserve. Split Oak Preserve consists of pine flatwoods and oak hammocks interspersed with small ponds. Moss Pa rk and Split Oak Preserve together span 3351 acres (Figure 4-3). There were two locations at MP that experimental sentinel chickens were housed. The original experimental cage location wa s toward the back half of the park near the

PAGE 92

92 pickerel reed marsh (Figure 4-3). After five tr ap nights, the cage had to be removed by request of the park manager. It was relocated in a fe nced off area near the county sentinel chickens. At times during the wet season, Moss Park will flood in various regions along the marsh and lake edges. This periodic flooding create s temporary and semi-permanent pools, ideal habitat for Culex Anopheles Aedes and Psorophora larvae (Foster and Walker 2002, OCMCD data). For the rest of the year larval habitats are restricted to the marsh and shallow edges of Lakes Hart and Mary Jane (OCMCD data, person al observation). Lakes Hart and Mary Jane contain many aquatic plant species includi ng Cattail, Pickerelweed, Arrow arum ( Peltandra virginica ), Maidencane, Water Hyacinth, and numerous sedge species. These plants create ideal larval habitat for Cq. perturbans Ma. titillans and Ma. dyari (Morris et al. 1990, Lounibos and Escher 1985, Slaff and Haefner 1985, Callahan and Morris 1987). Anopheles spp. and Cx. erraticus larvae have been found along the edges of Lake Mary Jane. Several areas of the park are designated as repositories for human re fuse which provides la rval habitats for Ae. albopictus and other container inhabiting mosquitoes. Arboviruses Tibet-Butler Preserve Sentinel chicken surveillance began at TBP in 1995. Eastern equine encephalitis virus transmission at TBP was the highest of the twelve sentinel chicken survei llance sites operated by OCMCD. From 1995 through 2004, the mean annual seroconversion rate (MASR) of sentinel chickens for EEEV at TBP was 33.6% (44/131). Mean annual seroconversion rates for EEEV at the other eleven sites ranged from 0 to 16.5% (43/261) (Table 4-1). In 2002, four of ten seroconversions to EEEV occurred be tween January and March at TBP. St. Louis encephalitis virus tr ansmission at TBP was the seve nth highest of the twelve sentinel chicken surv eillance sites operated by OCMCD. From 1995 through 2004, the MASR

PAGE 93

93 of sentinel chickens for SLEV at TBP was 5.3% (7/131). All SLEV transmission to sentinel chickens occurred in 1996 and 1997. Mean annual seroconversion rates fo r SLEV at the other eleven sites ranged from 0 to 17.4% (8/46) (Table 4-2). West Nile virus transmission at TBP is the si xth highest of the twelve sentinel chicken surveillance sites operated by OCMCD. Fr om 2001 through 2004, the MASR for WNV was 30.8% (20/65). However, WNV wa s not detected by sentinel chic kens in Orange County until 2001, so this data only represents four years of surveillance. Mean a nnual seroconversion rates for WNV at the other eleven sites ranged from 7. 7% (2/26) to 40% (16/40) (Table 4-3). In 2002 and 2003, four of twenty-one and four of tw enty WNV seroconversions occurred between October and December at TBP. Moss Park Sentinel chicken surveillance began at MP in 1991. Eastern equi ne encephalitis virus transmission at MP was the eighth highest of the twelve sentinel chicken surveillance sites operated by OCMCD. From 1991 through 2004 the MA SR of sentinel chickens for EEEV was 4.5% (5/111). Mean annual sero conversion rates for EEEV at the other eleven sites ranged from 0% to 33.6% (44/131) (Table 4-1). St. Louis encephalitis virus tran smission at MP was the sixth hi ghest of all twelve sentinel chicken surveillance sites operated by OCMCD. From 1991 through 2004 the MASR of sentinel chickens for SLEV was 7.2% (8/111) with all seroconversions occurring in 1997. Mean annual seroconversion rates for SLEV at the other eleven sites ranged from 0 to 17.4% (8/46) (Table 42). The only SLEV seroconversion between 2001 and 2006 in Orange County occurred at MP in December of 2001. West Nile virus transmission at MP was the s econd highest of all twel ve sentinel chicken surveillance sites operated by OCMCD. Fr om 2001 through 2004 the MASR of sentinel

PAGE 94

94 chickens for WNV was 35.9% (14/39 ). The high level of WNV ac tivity is a threat to people camping at MP. The MASR for MP only represents the first four years of surveillance data which may account for the higher transmission rate Mean ASRs for WNV at the other eleven sites ranged from 7.7% (2/26) to 40.0% (16/40) (Table 4-3). In 2002 and 2003, five of twentyone and two of twenty WNV seroconversions o ccurred between October and December at MP. Vectors of Arboviruses Tibet-Butler Preserve had twenty different sp ecies of mosquitoes captured in exit traps from July 2003 through December 2004. Moss Pa rk had nineteen different species of mosquitoes captured in exit traps from July 2003 through December 2004. Culiseta melanura the primary epizootic vector of EEEV (Scott and Weaver 1989), is frequently captured in light traps at TBP. Culiseta melanura is infrequently captured in light traps at MP. Frequently capture d, potential secondary epizootic or bridge vectors of EEEV at TBP and MP include Ae. albopictus (Scott et al. 1990b, Mitchell et al. 1992), Cq. perturbans (Scott and Weaver 1989), Cx. erraticus (Cupp et al. 2003), An. quadrimaculatus (Vaidyanathan et al. 1997), and Cx. nigripalpus (Nayar 1982) Frequently captured vectors of SLEV at MP and TBP include the primary epidemic vector of SLEV in peninsular Florida, Cx. nigripalpus (Dow et al. 1964) and the potential, secondary vector of SLEV in the Florida panhandle, Cx. quinquefasciatus (Tsai and Mitchell 1989, McCaig et al. 1994). Frequently captured, potential vectors of WNV at TBP include Cx. nigripalpus (Rutledge et al. 2003, Turell et al. 2005), Cx. quinquefasciatus (Turell et al. 2005), Cs melanura (Sardelis et al. 2002, Turell et al. 2005), and Ae. albopictus (Sardelis et al. 2002, Turell et al. 2005,). Frequently captured, potential vectors of WNV at MP include Cx. nigripalpus (Rutledge et al.

PAGE 95

95 2003, Turell et al. 2005), Cx. quinquefasciatus (Turell et al. 2005), and Ae. albopictus (Turell et al. 2005, Sardelis et al. 2002). Avian Hosts of Arboviruses An amplification host is a vertebrate that circulates an arbovirus titer of sufficient magnitude and duration to infect blood-feed ing vectors (Scott 1988) Minimum infection threshold virus titers required to in fect mosquito vectors of EEEV ( Cs. melanura ) were 103 PFU/mL of blood (Komar et al. 199 9), and for SLEV and WNV were 105 PFU/mL of blood (Sardelis et al. 2001). Almost 100 bird species have been identifie d in Tibet-Butler Pres erve and Moss Park (OCPR data). Of these birds only a few ar e critical for arbovirus amplification. Several potential amplifying hosts for EEEV in TBP and MP are residents in Florida. Resident birds are important amplification hos ts for arboviruses (Crans et al. 1994). The Northern Cardinal ( Cardinalis cardinalis ), Common Grackle ( Quiscalus quiscula ), and Redwinged Blackbird ( Agelaius phoeniceus ) were proven to be effi cient hosts of EEEV in the laboratory (Komar et al. 1999). All three sp ecies are common at TBP and MP and were seropositive for EEEV in Flor ida in 1958 and 1960-1961 (Favorit e 1960, Henderson et al. 1962). Eastern equine encephalitis virus seropositive rate s for the three bird species in Florida were: Northern Cardinal 34% (10/29), Common Grack le 12% (17/143), and Red-winged Blackbird 19% (3/16) (Favorite 1960, Henderson et al. 1962). Virus isolations of EEEV, from species that are common residents at TBP and MP in clude the Common Grackle, Tufted Titmouse ( Baeolophus bicolor ), Northern Mockingbird ( Mimus polyglottos ), and Carolina Wren ( Thryothorus ludovicianus ) (Favorite 1960, Henderson et al 1962, Lord and Calisher 1970). Several serosurveys performed outside of Fl orida have isolated EEEV from the Common yellowthroat ( Geothlypis trichas ) (Howard et al. 2004), Northern cardinal, and White-eyed vireo

PAGE 96

96 ( Vireo griseus ) (Lord and Calisher 1970, Stamm et al 1962, Stamm and Newman 1963). The Common yellowthroat and Northern cardinal are common residents at TBP and MP. The Whiteeyed vireo is a common resident at MP. Overwintering birds may contribute to EEEV enzootic transmission during amplification and maintenance phases. Winter migr ants, such as the American Robin ( Turdus migratorius), may introduce EEEV from northern breeding sites (Stamm and Newman 1963, Lord and Calisher 1970). The American Robin and Swamp Sparrow ( Melospiza georgiana ) were shown to be efficient amplifying hosts of EEEV in th e laboratory (Komar et al. 1999). The American Robin is common at TBP. The American Robi n and Swamp Sparrow are common at MP. The American Robin, American Goldfinch ( Carduelis tristis ), House Wren ( Troglodytes aedon ), and Ruby-crowned Kinglet ( Regulus calendula ) had EEEV isolates from serosurveys conducted outside of Florida (Howard et al. 2004, McLean et al. 1985, Stamm et al. 1962). The American Goldfinch and House Wren are common overwinteri ng residents at TBP. The House Wren and Ruby-crowned Kinglet are common overwintering residents at MP There are resident and overwintering populations of the Eastern Towhee ( Pipilo erythrophthalmus ) and Gray Catbird ( Dumetella carolinensis ) at TBP and MP. Both species are common at TBP and MP. There have been several virus isolations of EEEV from the Eastern Towhee and Gray Catbird (Favorite 1960, Howard et al. 2004, Lord and Calisher 197 0, Stamm and Newman 1963). The Eastern Towhee population was 18% (4/22) seropositiv e for EEEV in Orlando (Favorite 1960). Four common, resident species of TBP and MP, the Blue Jay ( Cyanocitta cristata ), Northern Cardinal, Common Grackle, and Mourning Dove ( Zenaida macroura ) were frequently found seropositive for SLEV in 1990 in Florida (D ay and Stark 1999). St. Louis encephalitis virus rates for the four common, resident birds we re: Blue Jay 20% (9/46), Northern Cardinal

PAGE 97

97 6% (1/18), Common Grackle 7% (9/123), and Mourning Dove 56% (75/135) (Day and Stark 1999). The Blue Jay has been implicated in outbreaks of SLEV (McLean and Bowen 1980). The Common Grackle and Mourning Dove each had two isolations of SLEV in Florida in 1990 (Day and Stark 1999). One SLEV isolate was obtaine d from the Northern Cardinal in Florida in 1989 (Day and Stark 1999). No migrant species at TBP or MP have been implicated in SLEV epidemics or had SLEV isolated from them. Numerous resident species in Florida at TBP are potential amplifying hosts for WNV. The American Crow ( Corvus brachyrhynchos ), Common Grackle, Blue Jay, Fish Crow ( Corvus ossifragus ), Great Horned Owl ( Bubo virginianus ), and Red-tailed hawk ( Buteo jamaicensis ) were found to be efficient amplifying hosts of WNV in the laborator y (Komar et al. 2003, Nemeth et al. 2006). The American Crow, Common Grackle, Blue Ja y, Fish Crow, Great Horned Owl, and Red-tailed Hawk are common residents at TBP. The American Crow, Common Grackle, Blue Jay, and Fi sh Crow are common residents at MP. The role of the American Crow and Blue Jay in WNV transmissi on may be limited due to the high levels of mortality these species encounter with WNV infec tion (Lord et al. 2006). Fifteen percent (2/13) of Common Grackles were seropo sitive for WNV in Jefferson County, FL in 2001 (Godsey et al. 2005a). The Red-winged Blackbird has resident and overwintering populations in Florida and is common at TBP and MP. The Red-winged Blackbird was proven an efficient amplifying host of WNV in the laboratory (Komar et al. 2003). Birds that overwinter in Flor ida may potentially introduce WNV acquired at their northern breeding grounds (Rappole et al. 2000). The American Robin and Mallard ( Anas platyrhynchos ) were efficient amplification hosts of WNV in th e laboratory (Komar et al. 2003). The American Robin and Mallard are common over wintering birds at TBP. Th e American Robin is a common

PAGE 98

98 overwintering bird at MP. Seven percent (2/ 27) of Mallards were seropositive for WNV in Jefferson County, FL in 2001 (Godsey et al. 2005a). Discussion Young nestling House sparrows display fewer defensive behaviors and have elevated viremias for longer durations than adult counterparts (Scott et al. 1990a). Young nestling birds would be more likely to be fed upon by mosquitoes due to their reduced host defensive behaviors compared to their adult counterpa rts (Scott et al. 1990a, Kale et al. 1972). If young birds remain viremic at higher magnitudes and for longer durat ion then the presence of susceptible, young nestlings may facilitate epidemics. In both TBP and MP amplification transmission is most likely facilitated by viremic hatching year resident birds. The highest level of EEEV transmission to sentinel chickens in Orange County was observed at TBP (MASR 33.6%, 44/ 131). The bayhead and cypre ss swamps present at TBP provide ideal breeding habitat for Cs. melanura which is frequently found at TBP (Foster and Walker 2002). In contrast, the ei ghth highest level of EEEV transmi ssion to sentinel chickens in Orange County was observed at MP (MASR 4.5%, 5/111). There are no bayhead or cypress swamps at MP and Culiseta melanura is infrequently captured (OCMCD data). Culex nigripalpus and Cq. perturbans are often caught in sentinel chicken exit traps (OCMCD data) and may be the vectors of EEEV to sentin el chickens at MP in the absence of Cs. melanura The lowered transmission rate of EEEV at MP demonstrates the importance of Cs. melanura presence for the transmission of EEEV. The majority of SLEV activity at both TBP and MP occurred in 1997. Neither site had sentinel chickens present at them during the 1990 SLEV epidemic. In 1997 monthly sentinel chicken seroconversion rates were above MASR s throughout Florida but only nine human cases

PAGE 99

99 were reported and none from Orange County (Day and Stark 1999). Both MP and TBP provide suitable habitat for Cx. nigripalpus and numerous potential amp lifying hosts of SLEV including the Blue Jay, Common Grackle, Northern Ca rdinal, and Mourning Dove. With numerous potential amplifying hosts of SLEV and the prim ary epizootic and epidemic vector of SLEV ( Cx. nigripalpus ) present, both of these sites should be considered potential foci for SLEV amplification during SLEV epidemics. West Nile virus transmission to sentinel ch ickens at MP was the second highest of the twelve sentinel sites in Ora nge County (MASR 35.9%, 14/39). Th e first detection of WNV in Florida was in 2001, so information concerning transmission data, potential amplifying hosts, and potential vectors is limited (Blackmore et al. 2003). The primary implication of amplifying hosts for WNV transmission have been laborator y studies (Komar et al. 2003, Nemeth et al. 2006, Reisen et al. 2005). Serosurveys of potenti al amplifying hosts (Godsey et al. 2005a) and field incrimination of potential vectors of WNV in Florida (Ru tledge et al. 2003) have been limited. Future field studies should elucidate which habitat characte ristics, vectors and amplifying hosts create ideal foci fo r WNV transmission in Florida.

PAGE 100

100 Table 4-1. Seroconversions for EEEV at sites operated by OCMCD from 1978-2004 Tibet-Butler Moss Park Turkey Lake Trimble Park Kelly Park Cl arcon a Horseman Bear Creek Blanchard Park Winter Park Landfill Fort Christmas Mosquito Control Years of operation EEEV 10 14 27 8 7 4 13 11 5 3 27 8 EEEV Seroconversions 44 5 43 0 7 0 9 14 0 4 24 2 # chickens used 131 111 261 72 69 33 95 96 46 34 240 52 EEEV MASR 33.6 4.5 16.50 10.10 9.5 14.60 11.8 10 3.8 Rank 1 8 2 10 5 10 7 3 10 4 6 9 Table 4-2. Seroconversions for SLEV at sites operated by OCMCD from 1978-2004 Tibet-Butler Moss Park Turkey Lake Trimble Park Kelly Park Cl arcon a Horseman Bear Creek Blanchard Park Winter Park Landfill Fort Christmas Mosquito Control Years of operation SLEV 10 14 27 8 7 4 13 11 5 3 27 8 SLEV Seroconversions 7 8 37 8 2 1 1 9 8 0 19 0 # chickens used 131 111 261 72 69 33 95 96 46 34 240 52 SLEV MASR 5.3 7.2 14.211.12.9 3 1.1 9.4 17.4 0 7.9 0 Rank 7 6 2 3 9 8 10 4 1 11 5 11

PAGE 101

101 Table 4-3. Seroconversions for WNV at sites operated by OCMCD from 1978-2004 Tibet-Butler Moss Park Turkey Lake Trimble Park Kelly Park Clarcon a Horseman Bear Creek Blanchard Park Winter Park Landfill Fort Christmas Mosquito Control Years of operation WNV 4 4 4 4 4 4 4 3 3 3 4 4 WNV Seroconversions 20 14 4 16 17 8 7 5 8 11 17 2 # chickens 65 39 29 40 49 33 35 28 26 34 50 26 WNV MASR 30.8 35.9 13.840 34.724.220 17.930.8 32.4 34 7.7 Rank 6 2 11 1 3 8 9 10 6 5 4 12

PAGE 102

102 Figure 4-1. Orange County water bodies with sentinel sites, TBP and MP

PAGE 103

103 Figure 4-2. Aerial view of TBP with experimental se ntinel chicken cage location

PAGE 104

104 Figure 4-3. Aerial view of MP with the original a nd permanent experimental sen tinel chicken cage locations

PAGE 105

105 CHAPTER 5 EXPERIMENTAL ARBOVIRAL SURVEILLANCE IN ORANGE COUNTY, FL Introduction Mosquito species that occur at Tibet Butler and Moss Park (i.e., Cx. nigripalpus Cq. perturbans and Ae. albopictus ) have been incriminated or sugg ested as potential epizootic and bridge vectors of arboviruses to humans and horses. Crans and Schulze (1986) stated that bridge vectors of EEEV vary from region to region. It is important for OCMCD to know the potential bridge and enzootic vectors of arboviruses in the varied habitats of Orange County. To attempt to determine potential vectors of arboviruses (EEEV, SLEV, WNV, and HJ) in Orange County, experimental sent inel chickens were exposed to mosquitoes and other biting Diptera at TBP and MP. Diptera other than mosquitoes, such as Stomoxys calcitrans and Culicoides sp. can mechanically transmit EEEV (C hamberlain 1958, Chamberlain and Sudia 1961). Tabanids and Culicoides spp. have been captured in OC MCD sentinel chicken exit traps (OCMCD data). All Diptera of medical impor tance (Mullen and Durden 2002) captured from the experimental cage were iden tified. If an experimental chicken seroconverted for EEEV, SLEV, WNV, or HJ then all Dipt era of medical importance that we re captured the night of that chickens exposure were tested for arboviruses. Materials and Methods Sentinel Chickens Chickens were obtained by OCMCD from Hill andale Farms on October 4, 2005. Chickens were bled upon arrival to the OCMCD, held in a mosquito-proof cage, and bled again just prior to field exposure to ensure that they were se ronegative. The chickens were cared for according to University of Florida IACUC protocol #D 996 and housed at OCMCD in two mosquito-proof cages.

PAGE 106

106 Mosquito-proof cage dimens ions measured 1.2 m x 2.4 m x 1.7 m (W x L x H). The section that housed the chickens was 0.8 m from the ground. The mosquito-proof cages were constructed out of wood 2 x 4s, wood 4 x 4s, had an aluminum roof, and were enclosed with 1.25 cm x 2.5 cm hardware cloth. A second layer of hardware cloth was attached 15 cm below the bottom of the cage to prevent predators from reaching the chickens. Both mosquito-proof cages were sealed in 12-count fiberglass mesh th at was attached to the cages by Velcro. All spaces between the aluminum roof and the cage were sealed with silicone sealant. In each mosquito-proof cage, four groups of four chicke ns were separated by ha rdware cloth dividers that could be removed. Each section of cages had one 11.3 L water holder and a feeder. The mesh could be removed when the cages needed to be cleaned or chickens needed to be removed. Sixteen chickens were placed in the field at TBP and thirty-two chickens at MP weekly from October 12, 2005 to July 4, 2006. All chicke ns were pre-bled the Monday prior to field placement. Chickens were placed at the site s between 10 am to 2 pm and were collected between sunrise and 8 am the next morning. Mo squitoes were removed from the cage with an aspirator. Generally chickens were placed at TBP on Mondays and at Moss Park on Tuesdays. If the predicted high temperature was not above 10C then chickens were not placed in the field on the Monday/Tuesday schedule, but later in the week when te mperatures were warmer. The alternate schedule decision was made after numerous trials on nights when the temperature did not reach 10C and no mosquitoes entered the cage. A Watchdog Weather Station (Spectrum, Plainfield, IL) was used to measure rainfall, re lative humidity, daylight, temperature, wind speed, and barometric pressure at the sentinel sites. Abiotic data were compiled per hour, by month at each site and monthly means were generated from data collected only on nights of experimental

PAGE 107

107 chicken exposure. Monthly means at both sites were compared by the Wilcoxon Rank Sum test (WRSt) to determine significant differences. After field exposure, chickens were transporte d to OCMCD and held in the mosquito-proof cages for 14 to 20 days before being bled. Chickens were exposed multiple times throughout the study period with at least one m onth between field exposures. Chic kens were bled again one to two weeks later just prior to placing them in the field again. Approximately 1.5 to 2.5 cc of blood was drawn from the wing vein of each exposed chicken with a 22 gauge Kendall Monoject syringe (Tyco Healthcare group, Mansfield, MA). The blood wa s transferred to 6 mL BD Vacutainer Blood Collection Tubes (BD Vacutainer Franklin Lakes, NJ). The Vacutainer tubes were centrifuged at ~3400 rpm for 10 m to separate the bl ood and serum. Blood and serum samples were shipped via courier to the Department of Health Bureau of Laboratories in Tampa. Hemagglutination Inhibition (HI) test was the initial screening test used to determine antibody response to Alphaviruses or Flaviviruses Positive HI samples were tested by IgMcapture enzyme-linked immunosorbent assay (MACELISA) to determine the specific virus. West Nile virus negative birds were re-bled a nd screened by the HI test Samples positive during the second HI test were confirmed and identifi ed as SLEV or WNV by the plaque reduction neutralization (PRNT). Alphaviru s positives were confirmed and identified as EEEV or HJ by PRNT. Mosquitoes Mosquitoes and other Diptera were collected be fore the removal of the chickens from the experimental cage, transported to the OCMCD, killed by freezi ng, and identified to species on a chill table (Bioquip, Rancho Dominguez, CA). Mo squito specimens were sorted into pools of 300 or less and stored in a -80C Ultra low freezer (So-Low Environmental Equipment Company, Cincinnati, OH ). Othe r Diptera were identified to the species level when possible

PAGE 108

108 and pooled. In the event of seroconversion in an y of the experimental chickens, the mosquitoes associated with the positive chicken were to be taken from the freezer and have the legs removed for subsequent testing. The legs would be st ored in pools of 300 or less by species. The mosquito leg pools would be shipped to the Depa rtment of Health Bur eau of Laboratories in Tampa for RT-PCR testing. Results Forty-eight chickens were pl aced in the field for a nine month period from October 12, 2005 to July 5, 2006. At TBP 16 chickens were placed in the field on 34 nights for 136 surveillance nights. At MP a to tal of 32 chickens were placed in the field on 34 nights for 131 surveillance nights. There we re no seroconversions to SLEV WNV, EEEV, or HJ in the experimental chickens. Since there were no seroconversions in experimental chickens, no mosquitoes, or any other Dipt era were submitted for testing. Weather data are presented in tables 5.1 through 5.5. January 2006 was not included because of equipment failure. Wind speed was si gnificantly greater at TBP when compared to MP (Wilcoxon Rank Sum test, Z = 2.80, P = 0.0051). No other abiotic factor was significantly different between the two parks: Rainfall (WRSt, Z = 0.11, P = 0.9161), relative humidity (WRSt, Z = -1.42, P = 0.1563), temperature (WRSt, Z = -0.05, P = 0.9581), and barometric pressure (WRSt, Z = -0.15, P = 0.8748). Discussion In June 2005, before the field experiments, four sentinel chic kens from the OCMCD surveillance program were seropo sitive for EEEV at TBP (OCMCD data). At Moss Park two of the OCMCD sentinel chickens were confir med positive for EEEV, one on May 2, 2005 and one on July 5, 2005. The last two EEEV seropositive chickens in Orange County were sampled on July 25, 2005 from the Landfill and Blanchard Park sites. Between October of 2005 and June of

PAGE 109

109 2006, there were no positive seroconversions for SLEV, WNV, EEEV, or HJ in the sentinel chickens used by Orange County as part of their surveillance program (OCMCD data). Orange County experienced severe drought for several months after the hurricanes (FDACS 2006). The drought contributed to redu ced mosquito abundance in Orange County. The number of mosquitoes in regular collections fell below yearly averages, especially for Cx. nigripalpus in Orange County (Figure 5-1, OCMCD da ta). Orange County Mosquito Control Division captures mosquitoes by stationary light trap every day throughout the year. Traps are collected on Monday, Wednesday, and Friday of ever y week and all mosquito es are identified to species. Culex nigripalpus populations during the study period of October 2005 through July of 2006 (solid black line Figure 5-1) were below mo nthly averages compared to the previous two and a half years. Day and Stark (2000) st ated that during extended droughts mosquito populations are low and SLEV transmission is reduced. Abiotic factors at both parks may have reduced Cx. nigripalpus flight activity. Provost (1969) found the flight activity of Cx. nigripalpus was greatly diminished when the temperature drops below 20C. Nightly temperatures were below 20C on most nights from October 2005 through March 2006 (Table 5-3). Provost ( 1973) found mosquito flight activity, and Cx. nigripalpus flight activity (Provost 1969) increases with relative humidity above the 90% level to reach a peak during rain. Relative humidity reached above 90% usually after 1:00 am at TBP and 12:00 pm at MP. Rainfall was limited at bot h sites with only two nights at MP exceeding >30 mm (12/7/05 and 6/28/06). Flight activity of Cx. nigripalpus and other mosquito species may have been limited due to relative humidity not reaching levels greater than 90% and limited rainfall.

PAGE 110

110 Conclusion Reduced populations of mosquitoes, especially Cx. nigripalpus and abiotic factors that suppress mosquito flight activity may have led to the absence of arbovi rus transmission to experimental and sentinel chickens in Orange County from October 2005 to June 2006.

PAGE 111

111 Table 5-1. Rain (mm) TBP MP OCT 2005 0.12 0.53 NOV 2005 0.14 0.01 DEC 2005 0.12 0.84 FEB 2006 0.01 0 MAR 2006 0 0 APR 2006 0.06 0.01 MAY 2006 0.12 0.02 JUN 2006 0.08 0.44 mean 0.08 0.23 Table 5-2. Relative humidity (%) TBP MP OCT 2005 79.11 86.33 NOV 2005 86.37 84.49 DEC 2005 84.71 90.48 FEB 2006 82.60 87.00 MAR 2006 73.76 74.77 APR 2006 74.99 79.47 MAY 2006 70.46 80.81 JUN 2006 80.73 84.20 mean 79.09 83.44 Table 5-3. Temperature (C) TBP MP OCT 2005 16.77 17.54 NOV 2005 17.05 15.36 DEC 2005 12.96 13.55 FEB 2006 12.03 14.38 MAR 2006 15.11 16.56 APR 2006 19.83 18.41 MAY 2006 20.51 20.19 JUN 2006 22.67 22.59 mean 17.12 17.32

PAGE 112

112 Table 5-4. Wind speed (km/h) TBP MP OCT 2005 0.29 0.23 NOV 2005 0.49 0.11 DEC 2005 0.34 0 FEB 2006 0.01 0 MAR 2006 0.26 0.08 APR 2006 0.18 0 MAY 2006 0.21 0.01 JUN 2006 0.45 0 mean 0.28 0.05 Table 5-5. Barometric pressure (mm-Hg) TBP MP OCT 2005 758.0 758.5 NOV 2005 758.0 759.3 DEC 2005 776.9 769.4 FEB 2006 778.7 784.1 MAR 2006 784.2 784.3 APR 2006 784.8 784.6 MAY 2006 774.2 766.9 JUN 2006 778.2 778.9 mean 774.13 773.26

PAGE 113

113 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 22000 JanFebMarAprMayJunJulAugSepOctNovDecMonthTotal number of Cx. nigripalpus captured 2003 2004 2005 Project (10/05 7/06) Figure 5-1. Total number of Culex nigripalpus captured in light traps in Orange County by month (January 2003 to July 2006). The solid black line represents light trap collections that coinci de with the months that experi mental chickens were used (October 12, 2005 to July 5, 2006).

PAGE 114

114 CHAPTER 6 CAGE ANALYSIS Introduction In 2003, sentinel chicken cages used by the OC MCD were modified by adding an exit trap to collect mosquitoes. Exit traps on sentinel ch icken cages are the only means of assessment that OCMCD has to determine the mosquito species th at enter the cages and potentially blood feed on the sentinel chickens. Exit trap collections may not represent all mosquito species that approach or feed on sentinel chickens. Certai n species of mosquitoes may be more likely to enter the exit trap, which biases the exit trap collection results. Culex nigripalpus was the second most frequent species recovered from the exit trap in the chicken-baited, open slit escape trials discussed earlier (Figure 2-9). Culex nigripalpus is the most frequently captured mosquito in exit trap collections in Orange and Indian River Counties (T. P. Breaud and D. A. Shroyer, personal communication). It is im portant to know which mosquito species are likely to enter or avoid entry into the exit trap. There have been no field trials assessing whether the exit trap mosquito catch accurately represents the actual mosquito species composition and abundance of those species that enter th e sentinel cages and attempt to feed on the chickens. It is possible that a mosquito may probe a sentinel chicken without taking a blood meal. Arboviruses may be transferred from the mosquito to the chicken during the initial probe even when no blood meal is taken (Hurlbut 1966). In the experiments perfor med in Orange County, FL and exit trap collections m onitored by OCMCD, mosquito blood-feeding success is the only gauge for assessing whether a mosquito has probed a sentinel chicken. The experimental cage designed for this study was used to determine the ratio of mosquitoes caught within the cage compared with t hose caught in the exit trap. This ratio will be used to estimate the likelihood of a mosquito entering the exit trap after it approaches the

PAGE 115

115 chickens. This information is important to OCMCD to assess how representative the county sentinel chicken exit trap collections are of the mo squitoes that enter the sentinel cage, approach the chickens, and attempt to feed on them. There are anecdotal observations indicating that blood-fed Diptera fly upward (D. L. Kline, personal communication). The comp arison of the percent of bloodfed mosquitoes captured in the cage and the exit trap will assess wh ether or not blood-fed Diptera fly upward. Materials and Methods Mosquitoes were collected from sentinel chic ken cages the morning af ter sentinel chicken exposure, 34 times at both TBP and MP (Chapter 5). Mosquitoes were aspirated from the interior of the cage. The exit trap was removed from the cage. Mosquitoes were killed by freezing, identified to species on a chill table (B ioquip, Rancho Dominguez, CA), sorted into pools, and stored in a -80C Ultra-low freezer (So-Low Environmental Equipment Company, Cincinnati, OH). Data were tabulated by site location, species caught, whethe r the mosquito was caught in the exit trap or the cage, and blood-feeding status. All data were analyzed by PROC GLM in SAS (SAS Institute 2002). Only species that had at least 49 mosquitoes collected at TBP or MP were analyzed. The likelihood that mosquitoes enter the exit trap was assessed by individual species with the Students T-test. Data were transformed by ln (x+1) to normalize the data. Normality was tested by Levenes test for homogeneity of variance. If original or transformed data were not normal, then the Wilcoxon Rank Sum test (WRSt) was used. Ratios of the mean number of mosquitoes caught in the cage and in the exit trap were determined. The Students T-test was used to assess the difference in the percent of blood-fed mosquitoes by species, between the cage and exit trap. Data were transformed by ARCSIN

PAGE 116

116 square root. Normality was tested by Levenes te st for homogeneity of variance. If original or transformed data were not normal then th e Wilcoxon Rank Sum test was used. When mosquitoes were absent from either the cage or the exit trap on a trap night, the data for that location were not included in analysis. The mean number of blood-fed and non blood-fed mosquitoes captured in the cage and the exit trap were reported. Original and transfor med ln (x+1) data were tested by ANOVA for differences in location and blood-fed status betw een the cage and the exit trap. Means found to be significant ( P = 0.05) by ANOVA, were separated by Fishers LSD test. Normality was tested by Levenes test for homogeneity of variance. If original or transformed data were not normal then the Kruskal-Wallis test was used to determine differences in the number of bloodfed mosquitoes between the cage and the exit trap. For terms found to be significant ( P = 0.05) by Kruskal-Wallis test, the means were se parated by Dunns Multiple Comparison test. Results Tibet-Butler Preserve Species of mosquitoes that were frequently ca ptured in the experimental cage at TBP were Cq. perturbans (57.6%), Ma. titillans (18.1%), Cx. nigripalpus (11.6%), Cx. quinquefasciatus (4.5%), Cs. melanura (3.2%), and Cx. erraticus (2.5%) (Table 6-1). Th ese six species represent 97.5% of the 3164 mosquitoes caught in the experimental cage at TB P. Other Diptera of medical importance caught in the experimental cage at TBP include Tabanus sp., Silvius sp., and Culicoides hinmani (Table 6-2). At TBP, the number of mosquitoes captured in the cage was significantly greater than the number caught in the exit trap for the following species: Cx. quinquefasciatus (WRSt, Z = 2.03, P = 0.0466), Cx. nigripalpus (WRSt, Z = 3.72, P = 0.0002), Cx. erraticus (WRSt, Z = 1.99, P =

PAGE 117

117 0.0467), Ma. titillans (WRSt, Z = 3.44, P = 0.0006), and all species combined (T-test, t = 14.86, P < 0.0001) (Tables 6-5). At TBP, the percent of blood-fed mosquitoes captured in the exit trap was significantly greater than the cage for the following species: Cq. perturbans (WRSt, Z = 2.39, P = 0.0085), Cx. erraticus (WRSt, Z = 1.86, P = 0.0317), and Ma. titillans (WRSt, Z = 2.59, P = 0.0048) (Tables 6-6). At TBP, the percent of all blood-fed species (T-test, t = 12.4, P < 0.0001) captured in the cage was significantly greater than the exit trap (Table 6-6). Mosquito species with significantly different location and blood-fed status at TBP were Cq. perturbans (ANOVA, F = 11.94, P < 0.0001), Cx. quinquefasciatus (ANOVA, F = 3.77, P = 0.0142), Cx. nigipalpus (Kruskal-Wallis, Chi-square = 40.14, P < 0.0001), Ma. titillans (ANOVA, F = 4.47, P = 0.0060), and all species (ANOVA, F = 13.86, P < 0.0001) (Table 6-7). Moss Park Species of mosquitoes that were frequently ca ptured in the experime ntal cage at MP were Cq. perturbans (44.4%), Cx. erraticus (34.2%), and Cx. nigripalpus (16.1%) (Table 6-3). These three species represent 94.7% of the 304 mosquitoes caught in the experimental cage at MP. Other than mosquitoes, no Diptera of medical im portance were caught in the experimental cage at MP (Table 6-4). At MP, the number of mosquitoes captured in the cage was significantly greater than the number caught in the exit trap for Cx. erraticus (WRSt, Z = 3.20, P = 0.0014) and all species combined (WRSt, Z = 2.80, P = 0.0052) (Table 6-8). At MP, the percent of blood-fed Cx erraticus (WRSt, Z = 2.24, P = 0.0126) and all bloodfed species combined (WRSt, Z = 2.59, P = 0.0048) captured in the exit trap was significantly greater than the cage (Table 6-9).

PAGE 118

118 Mosquito species with significantly different location and blood-fed status at MP were Cq. perturbans (Kruskal-Wallis, Chi-square = 15.90, P = 0.0012), Cx. erraticus (Kruskal-Wallis, Chi-square = 25.35, P < 0.0001), Cx. nigripalpus (Kruskal-Wallis, Chi-square = 28.16, P < 0.0001), and all species (Krusk al-Wallis, Chi-square = 30.81, P < 0.0001) (Table 6-10). Discussion All species captured in the experimental ca ge during this study had been previously captured in OCMCD sentinel chicken exit traps at TBP and MP (OCMCD data). There were no mosquito species that were captured in the exit trap more frequently than the cage at either site. Ratios of mosquitoes caught in the cage compared to the exit trap at TB P ranged from 1.4 : 1 (all species) to 34.3 : 1 ( Cs. melanura ). Ratios of mosquitoes caught in the cage compared to the exit trap at MP ranged from 1.9 : 1 ( Cq. perturbans ) to 8.3 : 1 ( Cx. nigripalpus ). This means that county sentinel chicken exit trap collections represent only a fr action of the mosquitoes that come to feed on the sentinel chickens. At TBP the exit trap comprised 27% (868/3164) of all the mosquitoes collected. At MP the exit trap comp rised 22% (67/300) of the all the mosquitoes collected. Culiseta melanura was infrequently captured in OCMC D exit traps (OCMCD data). The ratio of Cs. melanura captured in the experimental cag e and exit trap (34.3 : 1) at TBP demonstrates that Cs. melanura is not likely to enter the exit tr ap after approaching the sentinel chickens (Table 6-5). Density levels of Cs. melanura have been shown to directly influence enzootic EEEV transmission (Williams et al. 1972). Culex nigripalpus and Cq. perturbans were two of the most frequently captured species in OCMCD sentinel chicken exit coop traps (OCMCD data). Both species are potential bridge vectors of EEEV in Florida (Nayar 1982, Scott and Weaver 1989). Culiseta melanura Cx. nigripalpus and Cq. perturbans have different larval habitats and require different larval control measures. Adult Cs. melanura Cx.

PAGE 119

119 nigripalpus and Cq. perturbans are active at different times of the night (Chapter 7). If Cs. melanura population levels are not adequately assessed by exit coop traps when EEEV seroconversions occur, then it is difficult to sp eculate their involvement in EEEV transmission in Orange County and the proper cont rol strategies may not need to be implemented to limit EEEV transmission. Anopheles crucians Wiedemann was the only species that was infrequently captured from the experimental cage and exit trap collections but was frequently captured in OCMCD sentinel chicken exit traps. Anopheles crucians may have been deterred from flying upward into the cage by the baffle or by the white color of the bottom of the plastic sheets. Culicoides hinmani was captured on two occasions at TBP. It was likely that Culicoides spp. frequently approached and fed on experimental sentinel chickens bu t were not captured in the experimental cage because they are small enough to escape through the mesh screen on the cage. Every time that Culicoides hinmani was collected, it was from the cage. Culicoides spp. can mechanically transmit EEEV (Chamberlain and Sudia 1961). Mechanical transmission of arboviruses (i.e., EEEV) by Culicoides spp. is dependent upon ac tive arbovirus infection produced by biological vectors (Chamberlain and Sudia 1961). Culicoides spp. likely plays only a minimal, supplementary role, if any at all in EEEV, WNV, or SLEV transmission in Florida. Tabanus sp. and Silvius sp. were collected at TBP in the experimental cage and exit trap and the OCMCD exit traps (OCMCD data). No members of the Tabanidae family have been demonstrated to be capable of EEEV, SLEV, or WNV transmission. Othe r Diptera, such as Fannia sp. and Hylemya sp., captured at MP and TBP (Table s 6-2 and 6-4) were most likely attracted to the chicken feces, do not blood-fee d, and play no role in arbovirus transmission.

PAGE 120

120 Several factors may influence the entry of mosq uitoes into the exit trap. Abiotic factors may influence mosquito flight activity once inside the cage. Moonlight may attract mosquitoes upward into the exit trap. Cloud cover can dire ctly affect the amount of moonlight that can illuminate the cage and exit trap. The reflective plas tic material that the exit trap is constructed from may attract mosquitoes to fly up ward into the exit trap, similar to Stomoxys calcitrans attraction to Alsynite traps (Broce 1988). Wind speed inside the cage was reduced by the mesh screen and probably had limited effect on mosquito flight activity. Humidity and temperature may also influence flight activity inside the trap as well. Chicken defe nsive behaviors did not significantly influence exit trap co llections of nocturnal mosquitoes (Chapter 3). It has been anecdotally suggested that blood-f ed Diptera tend to fly upward a nd this may influence exit trap collections. To assess whether or not blood-fed Diptera tend to fly upward, the percent of blood-fed mosquitoes caught inside the cage were compared to the percent of blood -fed mosquitoes caught inside the exit trap. When all sp ecies of mosquitoes were compar ed, the percent blood-fed in the exit trap was significantly less than the percen t of blood-fed mosquitoes in the cage at TBP (Table 6-6). The percent of blood-fed mosquito es caught in the exit tr ap was significantly greater than the percen t of blood-fed mosquitoes caught in the cage for Cq. perturbans Cx. erraticus and Ma. titillans at TBP, Cx. erraticus and all species combined at MP (Table 6-6). It appears that Cq. perturbans Cx. erraticus and Ma. titillans are more likely to fly upward after obtaining a blood meal and become cap tured in the exit trap. The statement that blood-fed Diptera fly upward may be correct fo r certain species but not for all species of mosquitoes.

PAGE 121

121 Mosquitoes captured by the experimental cage we re not tested for arboviruses. Had they been tested, MIRs could be calculated for multiple species of mosquitoes. This would establish MIRs in mosquitoes during a period of no ar bovirus transmission to sentinel chickens. Minimum infection rates have been shown to be higher than transmission rates (Reeves et al. 1961, Rutledge et al. 2003) and this may hold true for mosquitoes captured by the experimental cage. Conclusions All species of mosquitoes captured in the experimental cage have been captured in OCMCD sentinel chicken exit traps in the past at both sites. At TBP the exit trap comprised 27% (868/3164) of all the mosquitoes collected. At MP the exit trap comprised 22% (68/304) of the all the mosquitoes collected. With the exception of Cs. melanura the exit trap provides an accurate assessment of the mosquito species that frequently approach sentinel chickens at TBP and MP. Coquillettidia perturbans Cx. erraticus and Ma. titillans were the most likely mosquitoes to fly upward after obtaining a blood meal and then enter the exit trap.

PAGE 122

122 Table 6-1. Mosquitoes capture d in experimental cage at TBP Species total percent of composition mean per night Ae. albopictus 3 0.09 0.09 An. crucians 32 1.01 0.94 An. quadrimaculatus 1 0.03 0.03 Cq. perturbans 1823 57.62 53.62 Cx. erraticus 79 2.50 2.32 Cx. nigripalpus 367 11.60 10.79 Cx. quinquefasciatus 141 4.46 4.15 Cx. salinarius 5 0.16 0.15 Cs. melanura 102 3.22 3.00 Ma. dyari 36 1.14 1.06 Ma. titillans 571 18.05 16.79 Oc. infirmatus 4 0.13 0.12 total 3164 Table 6-2. Other Diptera capture d in experimental cage at TBP Species Hylemya sp. Fannia sp. Ophyra aenescens Musca domestica Muscina sp. Muscina assimilis Sarcophaga sp. Silvius sp. Piophilidae Ophyra leucostoma Culicoides hinmani Tabanus sp. Muscina stabulans total total 103 43 15111474 13613202 % composition 50.99 21.29 7.435.450.501.983.471.98 0.501.492.970.501.49 mean per night 3.03 1.26 0.440.320.030.120.210.12 0.030.090.180.030.09

PAGE 123

123 Table 6-3. Mosquitoes capture d in experimental cage at MP Species total % of comp mean per night An. crucians 5 1.64 0.15 Cq. perturbans 135 44.41 3.97 Cx. erraticus 104 34.21 3.06 Cx. nigripalpus 49 16.12 1.44 Cx. quinquefasciatus 7 2.30 0.21 Ma. titillans 3 0.99 0.09 Ps. columbiae 1 0.33 0.03 toal 304 Table 6-4. Other Diptera capture d in experimental cage at MP Species Hylemya sp. Fannia sp. Sarcophaga sp. Ophyra leucostoma total total 15 5 6329 % composition 51.72 17.24 20.6910.34 mean per night 0.44 0.15 0.180.09

PAGE 124

124 Table 6-5. Mean number of mosquitoes ca ught in the exit trap and the cage at TBP Species Exit trap Cage df Z/t P Cq. perturbans 20.00 4.41 36.97 7.44 1 Z = 1.45 0.148 Cx. quinquefasciatus 1.28 0.64a 3.13 1.18b 1 Z = 2.03 0.0466 Cx. nigripalpus 2.50 1.74a 8.94 3.90b 1 Z = 3.72 0.0002 Cx. erraticus 0.46 0.29a 2.00 0.86b 1 Z = 1.99 0.0467 Cs. melanura 0.09 0.09 3.09 2.22 1 Z = 1.39 0.0807 Ma. titillans 2.56 2.03a 15.28 11.84b 1 Z = 3.44 0.0006 All species 27.13 5.90a 71.75 18.36b 62 t = 14.86 < 0.0001 Table 6-6. Percent of blood-fed mosquitoes caught in the exit trap and the cage at TBP Species Exit trap Cage df Z/t P Cq. perturbans 89.95 3.34a 71.57 6.21b 1 Z = 2.39 0.0085 Cx. quinquefasciatus 71.39 11.21 64.88 9.28 1 Z = 0.52 0.301 Cx. nigripalpus 90.94 5.31 88.74 3.39 1 Z = 0.77 0.2217 Cx. erraticus 78.89 12.96a 36.57 12.28b 1 Z = 1.86 0.0317 Cs. melanura 100 0 82.8 6.52 ---Ma. titillans 91.36 6.30a 79.19 6.12b 1 Z = 2.59 0.0048 All species 22.16 3.08a 48.61 3.46b 62 t = 12.4 < 0.0001

PAGE 125

125 Table 6-7. Mean number of blood-fed and non blood-fed mosquitoes by location at TBP Exit trap Cage Species non blood-fed blood-fed non bl ood-fed blood-fed df F/Chi sq P Cq. perturbans 2.63 1.19a 17.38 3.90c 9.72 2.76b 27.25 5.76c 99 F = 11.94 < 0.0001 Cx. quinquefasciatus 0.72 0.51a 0.56 0.18a 0.84 0.30a 2.28 0.96b 75 F = 3.77 0.0142 Cx. nigripalpus 0.68 0.48a 1.88 1.27a 1.84 0.78a 7.09 3.19a 3 Chi sq = 37.42 < 0.0001 Cx. erraticus 0.94 0.55 0.28 0.14 1.06 0.46 0.94 0.55 47 F = 2.80 0.0512 Cs. melanura 0 0 0.09 0.09 0.53 0.44 2.56 1.87 3 Chi sq = 5.96 0.1134 Ma. titillans 1.94 1.62a 0.63 0.42a 10.88 9.32b 4.41 2.56b 83 F = 4.47 0.006 All species 6.19 2.58a 20.94 4.38b 25.97 11.12b 45.78 10.70c 123 F = 13.86 < 0.0001

PAGE 126

126 Table 6-8. Mean number of mosquitoes ca ught in the exit trap and the cage at MP Species Exit trap Cage df Z P Cq. perturbans 1.62 0.66 3.03 1.18 1 0.71 0.4797 Cx. nigripalpus 0.17 0.07 1.41 0.53 1 1.81 0.07 Cx. erraticus 0.45 0.35a 3.14 0.95b 1 3.2 0.0014 All species 2.31 0.70a 8.03 1.92b 1 2.8 0.0052 Table 6-9. Percent of blood-fed mosquitoes caught in the exit trap and cage at MP Species Exit trap Cage df Z P Cq. perturbans 95.20 3.22 85.91 9.16 1 0.61 0.2713 Cx. nigripalpus 100 0 100 0 ---Cx. erraticus 100 0a 65.31 9.66b 1 2.24 0.0126 All species 91.36 6.30a 79.19 6.12b 1 2.59 0.0048

PAGE 127

127 Table 6-10. Mean number of blood-fed a nd non blood-fed mosquitoes by location at MP Exit trap Cage Species non blood-fed blood-fed non blood-fed blood-fed df Chi sq P Cq. perturbans 0.10 0.08a 1.52 0.63a,b 0.14 0.07a 2.90 1.16b 3 15.9 0.0012 Cx. nigripalpus 0 0a 0.17 0.17a 0 0a 1.41 0.53b 3 28.16 < 0.0001 Cx. erraticus 0 0a 0.45 0.35a 0.97 0.34a,b 2.17 0.67b 3 25.35 < 0.0001 All species 0.17 0.10a 2.14 0.67a 1.24 0.34a 6.79 1.73b 3 30.81 < 0.0001

PAGE 128

128 CHAPTER 7 ROTATOR STUDIES Introduction Nighttime ULV adulticide application is an important component of mosquito control in Orange County, FL. An adult mosquito is a ffected by ULV insecticide when the droplet impinges on the mosquito (Mount 1970). Adult mosqu itoes in flight at th e time that adulticides are applied are more likely to contact the ins ecticide (Wright and Knight 1966). To maximize the likelihood that adulticides reach the desired mosquito target, it is necessary to know what times of the night mosquito species are most active (Carroll and Bourg 1977). Many methods have been used in the past to determine periods of mo squito flight activity including truck traps (Carroll and Bourg 1977), timer-equipped light traps (Gladney and Turner 1970), suction traps, monitoring of human biti ng activity (Wright and Knight 1966), and timed collections from resting boxes (Nasci and Edma n 1981). Truck traps and suction traps provide unbiased estimates of mosquito flight act ivity (Bidlingmayer 1966, Carroll and Bourg 1977, Wright and Knight 1966). Timer-e quipped light traps and aspiration of mosquitoes attracted to human hosts provides biased information of flight activity as it pertains to host-seeking (Gladney and Turner 1970, Wright and Knight 1966). Re sting boxes provide biased estimates of nocturnally active mosquitoes that enter resting areas at dawn and exit at dusk (Nasci and Edman 1981). Rotator traps (timer-equipped light trap) a llow multiple collection periods at known times throughout the night. A rota tor trap baited with CO2 and incandescent light will attract primarily host-seeking mosquitoes but will also collect freshly emerged, gravid, and blood-fed mosquitoes. A rotator trap baited with CO2 and incandescent light was used at TBP and MP to determine mosquito host seeking times. This in formation is important for OCMCD to determine when to conduct adulticide spraying at TBP and MP.

PAGE 129

129 Materials and Methods A Collection Bottle Rotator (referred to as a rotator trap for the remainder of the Thesis), model 1512 (John W. Hock, Gainesville, FL) was operated at TBP and MP between November 7, 2005 and July 4, 2006. The rotator trap was oper ated at the same time that the experimental chickens were in the experimental cage. Carbondioxide and incandescent light were used as the attractants. The CO2 was supplied from a 20 lb cylinder (A irgas, Orlando, FL) and regulated by a #1 regulator (Clarke Mosquito C ontrol, Roselle, IL) to provide a flow rate of 500 mL/min. The CDC miniature light trap (model 512 with an air actuated gate system, John W. Hock, Gainesville, FL) attached to th e rotator trap, was modified w ith a 12 V solenoid (Lee Products Ltd., England) so that when the light and fan were turned off CO2 was not released. A 12 V motorcycle battery (Power-Sonic, San Diego, CA) wa s used to power the rotator trap. Collection bottles were removed the morning following opera tion. The mosquitoes we re killed by freezing and then identified to species. Counts were recorded and th e mosquitoes were discarded. Eight timed periods were used for each coll ection night. Between rotations to the next time period, there was a 15 min break. The 15 min br eak allowed mosquito co llection periods to be statistically comparable. Time periods for on/off operation were determ ined as follows: the first collection began 30 min before sunset and the eighth collection ended 30 min after sunrise (Nautical Almanac Office, United States Naval Observatory). Co llection periods 1 to 5 and 8 lasted 75 minutes. The length of the 6th and 7th collection periods varied to coincide with daylength changes that occu r throughout the year. To compensate for these changes, the sixth and seventh collection periods were expanded or reduced to fill the remaining time between th e end of the fifth and beginning of the eighth collection periods (light gray hi ghlighted columns in Table 7-1) Table 7-1 shows all weekly calculated time periods that rotator traps were used. An example of two different nights of

PAGE 130

130 rotator collection periods are hi ghlighted dark gray (rows in Table 7-1) to show the time differences on different collec tion nights. Collection time periods were changed weekly according to the Monday sunrise/sunset times. The variable length of the si xth and seventh collection peri ods make comparison of data between the other six collection pe riods invalid unless transformations of the mosquito collection data were made. The mosquito collection data from periods 6 and 7 were multiplied by the formula 75 min/x min, where x equals the number of minutes that coll ection periods 6 and 7 lasted. Since the value of x changed every week the multiplier value was recalculated each week. The modified data from periods 6 a nd 7 were used for all statistical analyses. Rotator collection data were tabulated by speci es for all collection nights at each park. Any species that comprised at least 2% of the to tal collection counts at ea ch site were analyzed for variation in collection periods throughout the night. All data were analyzed by PROC GLM in SAS (SAS Institute 2002). Original and nor malized (ln (x+1)) data were tested by ANOVA for differences in mean collection periods. Significant ( P = 0.05) periods by ANOVA were separated by Fishers LSD test. Normality was tested by Levenes test for homogeneity of variance. Data that could not be normalized were tested by Kruskal-Wallis test for differences between collection periods. Significant ( P = 0.05) periods by Kruskal-Wallis test were separated by Dunns Multiple Comparison test. Whenever thre e or more sequential collections resulted in no mosquitoes of a particular sp ecies, the data for that mosquito species were not included in statistical analysis. Results Tibet-Butler Preserve Species that comprised greater than 2% of the total mosquitoes caught at TBP were An. crucians (49.47%), Cx. nigripalpus (23.51%), Cq. perturbans (10.87%), Cs. melanura (7.68%),

PAGE 131

131 Cx. quinquefasciatus (2.89%), Cx. erraticus (2.66%), and Ma. titillans (2.00%). These seven species comprised 99.08% (14736/14873) of the total number of mos quitoes caught in the rotator trap at TBP (Table 7-2). Anopheles crucians exhibited significant periods of el evated host-seeking activity from one hour after sunset to the si xth collection period (sometime between 3:00 am and 4:20 am) (ANOVA, F = 4.14, P = 0.0003) (Figure 7-1). Significant pe riods of elevated host-seeking activity for Cq. perturbans occurred one hour after sunset to three and three-quarter hours after sunset (ANOVA, F = 4.74, P < 0.0001) (Figure 7-2). Culex erraticus host-seeking was significantly elevated from half an hour before sunset to three and three-quarter hours after sunset (Kruskall-Wallis, Chi square = 25.2, P = 0.0007) (Figure 7-3). Culex nigripalpus had significant periods of elevated host-seeking ac tivity from one hour afte r sunset to one hour before sunrise (ANOVA, F=13.96, P < 0.0001) (Figure 7-4). Culex quinquefasciatus had significant periods of elevated hos t-seeking activity from one hour after sunset to half an hour after sunrise (Kruskall-Wallis, Chi square = 22.4, P = 0.0022) (Figure 7-5). Culiseta melanura host-seeking activity was significantly elevated from one hour after sunset to one hour before sunrise (ANOVA, F = 6.55, P < 0.0001) (Figure 7-6). Mansonia titillans did not exhibit significant periods of elevated host-seeking activity (ANOVA, F = 1.48, P = 0.1785) (Figure 7-7). All species combined had significan t periods of elevated host-seeking activity from one hour to six and three-quarters hours after sunset (ANOVA, F = 6.58, P < 0.0001) (Figure 7-8). Moss Park Species that comprised greater than 2% of the total mosquitoes caught at MP were An. crucians (41.96%), Cq. perturbans (35.31%), Cx. erraticus (12.17%), and Cx. nigripalpus (6.81%). The four species comprised 96.25% (23 33/2424) of the total mos quitoes caught in the

PAGE 132

132 rotator trap at MP (Table 7-3) The low numbers of mosquitoes captured at MP prevented the determination of significant pe riods of host-seeking activity for most mosquito species. Neither An. crucians (ANOVA, F = 0.22, P = 0.9813) (Figure 7-1) nor Cq. perturbans (ANOVA, F = 1.72, P = 0.1082) (Figure 7-2) exhibited si gnificant periods of host-seeking activity at MP. There were significant peri ods of elevated host-seeking activity for Cx. erraticus from 30 min before sunset to six and three-quarter h after sunset and agai n at sunrise (KruskalWallis, Chi square = 24.8, P = 0.0008) (Figure 7-3). There were significant periods (KruskalWallis, Chi square = 16.40, P = 0.0220) of elevated host-seeking activity for Cx. nigripalpus ; however, this was not detected by Dunns Multiple Comparison test (Fi gure 7-4). All species combined had significant periods of elevated ho st-seeking activity from sunset to three and three-quarters hours afte r sunset (ANOVA, F = 3.54, P = 0.0013) (Figure 7-5). Discussion Coquillettidia perturbans exhibited crepuscular, early evening elevated host-seeking activity during periods 2 to 3 at TBP. Gladney and Turner (1970) found Cq. perturbans hostseeking activity to be greatest in the early evening, from sunset to approximately three h after sunset. Coquillettidia perturbans frequently feeds on humans and of ten is the cause of mosquito complaint calls in Orange C ounty (Provost 1969, OCMCD data). Culiseta melanura displayed a nocturnal feeding pa ttern with elevated host-seeking activity between periods 2 to 7. Nasci and Ed man (1981) found a similar pattern of nocturnal host-seeking activity with a slight peak of feeding activity one h after sunset. Culex nigripalpus displayed a nocturnal f eeding pattern with elevat ed host-seeking activity between periods 2 to 7 at TBP. Provost (1969) found Cx. nigripalpus host-seeking to be nocturnal with slightly elevated periods of host-seeking activity one h after sunset.

PAGE 133

133 Seventy-five percent of the Ma. titillans at Tibet-Butler Preserve were caught on two nights (11/28/05 and 12/15/05) which skewed th e data and made it impossible to determine significant collection periods. Mansonia titillans frequently feeds on humans and often is the cause of mosquito complaint calls in Orange County (Provost 1969, OCMCD data). Gladney and Turner (1970) found Cx. erraticus host-seeking activity to be greatest in the early evening, from sunset to appr oximately three h after sunset. Culex erraticus was crepuscular with significant peri ods of elevated host-seeking activ ity between periods 1 to 3 at TBP (Figure 7-3). Culex erraticus had significant periods of el evated host-seeking activity between periods 1 to 5 and at period 8 at MP There were no significant differences in temperature, humidity, rainfall, and barometric pressure between TBP and MP (Chapter 5). Differences in wind speed between TBP and MP would not have affected Cx. erraticus flight activity (Chapter 5). The differe nces in host seeking activity of Cx. erraticus between these two sites was attributed to low capture rates of Cx. erraticus at MP. All species combined at TBP displayed cre puscular to mid-eveni ng host-seeking activity with peak host-seeking at periods 2 to 4. All sp ecies combined at MP displayed crepuscular to early-evening host-seeking activity with peak host-seeking at periods 1 to 3. Orange County Mosquito Control Division typica lly begins nightly adulticide spraying at sunset and continues spraying for approximately si x hours. The timing of adulticide application should coincide with the primary pe riod of host-seeking activ ity of the target species. On several occasions two or more species may be the target of adulticide operations. If all targeted species exhibit nocturnal host-seeking pa tterns then adulticide should be applied any time one h after sunset. If the targeted species have mixed host-seeking patterns then adulticide should be applied at a time that is most likely to reach as many species as possible. To maximize the

PAGE 134

134 likelihood of affecting the most mo squito species at TBP, adulticide should be sprayed between one h after sunset to five and a quarter h afte r sunset. To maximize th e likelihood of affecting the most mosquito species at MP, adulticide shou ld be applied from sunset to three and threequarter h after sunset. These data only apply to eight months of collections from November 2005 through June 2006. Further rotator trap data should be coll ected throughout the year for multiple years as species diversity and species abundance may change at different times of the year and differ between years. Conclusions Culiseta melanura Cx. nigripalpus Cx. quinquefasciatus and An. crucians displayed nocturnal host-seeking patterns at TBP. Coquillettidia perturbans Cx. erraticus and all species combined displayed crepuscular host-seeking activ ity at TBP. All species combined displayed crepuscular host-seeking activity at MP. To affect the greatest number of mosquitoes, adulticides should be applied at times when the target mosquito species are host seeking.

PAGE 135

135Table 7-1. Rotator time schedule Sunset Sunrise 1 OFF 2 OFF 3 OFF 4 OFF 5 OFF 6 OFF 7 OFF 8 OFF 11/7/05 5:36 6:42 5:06 6:21 6:36 7:51 8:06 9:21 9:36 10:51 11:06 12:21 12:36 3:02 3:17 5:42 5:57 7:12 11/21/05 5:30 6:52 5:00 6:15 6:30 7:45 8:00 9:15 9:30 10:45 11:00 12:15 12:30 3:04 3:17 5:52 6:07 7:22 11/28/05 5:28 6:58 4:58 6:13 6:28 7:43 7:58 9:13 9:28 10:43 10:58 12:13 12:28 3:06 3:21 5:58 6:13 7:28 12/5/05 5:28 7:03 4:58 6:13 6:28 7:43 7:58 9:13 9:28 10:43 10:58 12:13 12:28 3:08 3:23 6:03 6:18 7:33 12/12/05 5:30 7:08 5:00 6:15 6:30 7:45 8:00 9:15 9:30 10:45 11:00 12:15 12:30 3:12 3:27 6:08 6:23 7:38 12/19/05 5:32 7:13 5:02 6:17 6:32 7:47 8:02 9:17 9:32 10:47 11:02 12:17 12:32 3:15 3:30 6:13 6:28 7:43 1/9/06 5:46 7:19 5:16 6:31 6:46 8:01 8:16 9:31 9:46 11:01 11:16 12:31 12:46 3:25 3:40 6:19 6:34 7:49 1/16/06 5:51 7:19 5:21 6:36 6:51 8:06 8:21 9:36 9:51 11:06 11:21 12:36 12:51 3:28 3:43 6:19 6:34 7:49 1/23/06 5:57 7:17 5:27 6:42 6:57 8:12 8:27 9:42 9:57 11:12 11:27 12:42 12:57 3:30 3:45 6:17 6:32 7:47 1/30/06 6:03 7:15 5:33 6:48 7:03 8:18 8:33 9:48 10:03 11:18 11:33 12:48 1:03 3:32 3:47 6:15 6:30 7:45 2/6/06 6:09 7:11 5:39 6:54 7:09 8:24 8:39 9:54 10:09 11:24 11:39 12:54 1:09 3:33 3:48 6:11 6:26 7:41 2/13/06 6:14 7:05 5:44 6:59 7:14 8:29 8:44 9:59 10:14 11:29 11:44 12:59 1:14 3:32 3:47 6:05 6:20 7:35 2/20/06 6:19 6:59 5:49 7:04 7:19 8:34 8:49 10:04 10:19 11:34 11:49 1:04 1:19 3:32 3:47 5:59 6:14 7:29 2/27/06 6:24 6:53 5:54 7:09 7:24 8:39 8:54 10:09 10:24 11:39 11:54 1:09 1:24 3:32 3:47 5:54 6:08 7:23 3/6/06 6:28 6:45 5:58 7:13 7:28 8:43 8:58 10:13 10:28 11:43 11:58 1:13 1:28 3:29 3:44 5:45 6:00 7:15 3/13/06 6:33 6:38 6:03 7:18 7:33 8:48 9:03 10:18 10:33 11:48 12:03 1:18 1:33 3:28 3:43 5:38 5:53 7:08 3/20/06 6:37 6:30 6:07 7:22 7:37 8:52 9:07 10:22 10:37 11:52 12:07 1:22 1:37 3:26 3:41 5:30 5:45 7:00 3/27/06 6:40 6:22 6:10 7:25 7:40 8:55 9:10 10:25 10:40 11:55 12:10 1:25 1:40 3:24 3:39 5:22 5:37 6:52 4/3/06 7:44 7:13 7:14 8:29 8:44 9:59 10:14 11:29 11:44 12:59 1:14 2:29 2:44 4:21 4:36 6:13 6:28 7:43 4/10/06 7:48 7:06 7:18 8:33 8:48 10:03 10:18 11:33 11:48 1:03 1:18 2:33 2:48 4:20 4:35 6:06 6:21 7:36 4/17/06 7:52 6:58 7:22 8:37 8:52 10:07 10:22 11:37 11:52 1:07 1:22 2:37 2:52 4:18 4:33 5:58 6:13 7:28 4/24/06 7:56 6:51 7:26 8:41 8:56 10:11 10:26 11:41 11:56 1:11 1:26 2:41 2:56 4:16 4:31 5:51 6:06 7:21 5/1/06 8:00 6:45 7:30 8:45 9:00 10:15 10:30 11:45 12:00 1:15 1:30 2:45 3:00 4:15 4:30 5:45 6:00 7:15 5/8/06 8:05 6:39 7:35 8:50 9:05 10:20 10:35 11:50 12:05 1:20 1:35 2:50 3:05 4:15 4:30 5:39 5:54 7:09 5/15/06 8:09 6:35 7:39 8:54 9:09 10:24 10:39 11:54 12:09 1:24 1:39 2:54 3:09 4:15 4:30 5:35 5:50 7:05 5/22/06 8:13 6:31 7:43 8:58 9:13 10:28 10:43 11:58 12:13 1:28 1:43 2:58 3:13 4:15 4:30 5:31 5:46 7:01 5/29/06 8:17 6:29 7:47 9:02 9:17 10:32 10:47 12:02 12:17 1:32 1:47 3:02 3:17 4:16 4:31 5:29 5:44 6:59 6/5/06 8:20 6:27 7:50 9:05 9:20 10:35 10:50 12:05 12:20 1:35 1:50 3:05 3:20 4:16 4:31 5:27 5:42 6:57 6/12/06 8:23 6:27 7:53 9:08 9:23 10:38 10:53 12:08 12:23 1:38 1:53 3:08 3:23 4:18 4:33 5:27 5:42 6:57 6/19/06 8:25 6:28 7:55 9:10 9:25 10:40 10:55 12:10 12:25 1:40 1:55 3:10 3:25 4:19 4:34 5:28 5:43 6:58 6/26/06 8:27 6:29 7:57 9:12 9:27 10:42 10:57 12:12 12:27 1:42 1:57 3:12 3:27 4:18 4:33 5:29 5:44 6:59 7/3/06 8:27 6:32 7:57 9:12 9:27 10:42 10:57 12:12 12:27 1:42 1:57 3:12 3:27 4:22 4:37 5:32 5:47 7:02

PAGE 136

136 Table 7-2. Tibet-Butler Preserve rotator catch totals 11/7/05 to 7/4/06 Species Total % of catch Ae. vexans 1 0.01 An. atropos 7 0.05 An. crucians 7357 49.47 An. quadrimaculatus 22 0.15 Cq. perturbans 1616 10.87 Cx. erraticus 396 2.66 Cx. nigripalpus 3496 23.51 Cx. quinquefasciatus 430 2.89 Cx. restuans 1 0.01 Cx. salinarius 23 0.15 Cs. melanura 1142 7.68 Ma. dyari 38 0.26 Ma. titillans 298 2.00 Oc. atlanticus 30 0.20 Oc. infirmatus 14 0.09 Ps. columbiae 2 0.01 Total mosquitoes 14873

PAGE 137

137 Table 7-3. Moss Park rotator catch totals 11/8/05 to 7/5/06 Species Total % of catch An. atropos 2 0.08 An. crucians 1017 41.96 An. quadrimaculatus 6 0.25 Cq. perturbans 856 35.31 Cx. erraticus 295 12.17 Cx. nigripalpus 165 6.81 Cx. quinquefasciatus 16 0.66 Cs. melanura 41 1.69 Ma. dyari 1 0.04 Ma. titillans 2 0.08 Oc. atlanticus 9 0.37 Oc. infirmatus 10 0.41 Oc. triseriatus 1 0.04 Ps. columbiae 1 0.04 Ps. ferox 1 0.04 Total mosquitoes 2424

PAGE 138

138 0.0 10.0 20.0 30.0 40.0 50.0 60.0 12345678Rotator collection periodMean An. crucians captured TBP MPa,c d c,d c,d a,d a,d a,e b,e Figure 7-1. Anopheles crucians captured by rotator trap at TBP and MP: TBP df = 231, F = 4.14, P = 0.0003: MP df = 199, F = 0.22, P = 0.9813

PAGE 139

139 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 12345678Rotator collection periodMean Cq. perturbans captured TBP MPa,d b,c b,c c,d c,d,e d,e a,e a Figure 7-2. Coquillettidia perturbans captured by rotator trap at TBP and MP: TBP df = 183, F = 4.74, P < 0.0001: MP df = 167, F = 1.72, P = 0.1082

PAGE 140

140 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 12345678Rotator collection periodMean Cx. erraticus captured TBP MPa,b a a,b b b b b b x x, y y y x x,y x, y x, y Figure 7-3. Culex erraticus captured by rotator trap at TBP and MP: TBP df = 7, Chi square = 25.2, P = 0.0007: MP df = 7, Chi square = 24.8, P = 0.0008

PAGE 141

141 0.0 5.0 10.0 15.0 20.0 25.0 30.0 12345678Rotator collection periodMean Cx. nigripalpus captured TBP MPa a b b b b b b a a a a a a a a Figure 7-4. Culex nigripalpus captured by rotator trap at TBP and MP: TBP df = 231, F = 13.96, P < 0.0001: MP df = 7, Chi square = 16.4, P = 0.0220

PAGE 142

142 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 12345678Rotator collection periodMean Cx. quinquefasciatus capturedb a a a,b a,b a,ba,b a,b Figure 7-5. Culex quinquefasciatus captured by rotator trap at TBP: TBP df = 7, Chi square = 22.4, P = 0.0022

PAGE 143

143 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 12345678Rotator collection periodMean Cs. melanura captureda a b b bb b b Figure 7-6. Culiseta melanura captured by rotator trap at TBP: TBP df = 231, F = 6.55, P < 0.0001

PAGE 144

144 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 12345678Rotator collection periodMean Ma. titillans captured Figure 7-7. Mansonia titillans captured by rotator trap at TBP: TBP df = 167, F = 1.48, P = 0.1785

PAGE 145

145 0.0 20.0 40.0 60.0 80.0 100.0 120.0 12345678Rotator collection periodMean mosquitoes captured TBP MPa x x c c,e x,z a,c c,d y,z y ,z y,z y ,z y b a,d,e a,e Figure 7-8. All species captur ed by rotator trap at TBP a nd MP: TBP df = 231, F = 6.58, P < 0.0001: MP df = 199, F = 3.54, P = 0.0013

PAGE 146

146 CHAPTER 8 CONCLUSIONS During trials conducted in Gainesville, FL the experimental cage used for arbovirus surveillance wa s found capable of containing mosquitoes. The low recovery rate (47.4 10.9%) of Cx. quinquefasciatus from escape rate trials show that mosquitoes were capable of exiting th e cage through the baffle slit The recovery rates of certain species such as Cx. quinquefasciatus (47.4 10.9%), Ae. aegypti (57.0 5.5%), and Ae. albopictus (58.7 4.6%) show that they were more adept at escaping from the experimental cage. The higher recovery rates of Culex nigripalpus (28.8 9.2%), Cx. quinquefasciatus (23.0 11.7%), and An. quadrimaculatus (34.2 12.3%) from the exit trap compared to Aedes aegypti (3.2 1.0%) and Ae. albopictus (4.2 0.7%) show that some species are captured in the exit trap more frequently than others. Blood-fed Cx. nigripalpus (37.2 10.9%) were freque ntly recovered from the exit trap. Blood-fed An. quadrimaculatus (7.3 5.7%) and Ae. albopictus (12.5 12.5%) were infrequently recovered from the exit trap. Blood-fed Cx. quinquefasciatus and Ae. aegypti were never recovered from the exit trap. Mosquito species, such as An. quadrimaculatus Ae. albopictus Cx. quinquefasciatus and Ae. aegypti were less likely to be captured in the exit trap after they had blood-fed. Chicken defensive behaviors did not reduce the percent recovery or blood-feeding success of nocturnally active Cx. nigripalpus Chicken defensive behaviors significantly reduced the blood-feeding success and redu ced the percent of diurnally active Ae. albopictus recovered from 80% in the restrain ed chicken trials to 40.1% in the unrestrained chicken trials In field setti ngs, sentinel chickens will likely reduce the

PAGE 147

147 blood-feeding success and recovery of diurnal mosquitoes. In field settings, sentinel chickens will not influence blood-feeding su ccess and recovery of nocturnal mosquitoes. There were no arboviruses detected by expe rimental sentinel chickens on 34 nights of exposure from October 12, 2005 through July 4, 2006 at either Tibet-Butler Preserve or Moss Park. No mosquitoes were tested for arboviruses. If mosquitoes were tested for arboviruses then minimum infection rates of mosquitoes could be determined for a known period of time when there was no arbovi rus transmission to sentinel chickens. There were no mosquitoes recovered from th e experimental cage that had not been recovered previously (2003 to 2004) from OCMCD exit traps. Mosquito species frequently captured in the experi mental cage at both sites were Cq. perturbans Cx. nigripalpus and Cx. erraticus Mansonia titillans was captured frequently at TibetButler Preserve. Mosquitoes were more freque ntly captured in the cage than the exit trap at Tibet-Butler Preser ve (T-test, t = 14.86, P < 0.0001) and Moss Park (Wilcoxon Rank Sum test, Z = 2.80, P = 0.0052). Culiseta melanura was infrequently captured in OCMCD exit traps. Culiseta melanura was captured in the experimental cage on four occasions with a cage to exit trap ratio of 34.3: 1. Culiseta melanura is not likely to enter the exit trap after approaching sentinel chickens. Anopheles crucians was frequently captured in OCMCD exit trap collections but was infrequently captured in the experimental cage. Some species were more likely to fly upward and become captured in the exit trap when blood-fed, including Cq. perturbans (Wilcoxon Rank Sum test, Z = 2.39, P = 0.0085), Cx. erraticus (Wilcoxon Rank Sum test, Z = 1.86, P = 0.0317), and Ma. titillans (Wilcoxon Rank Sum test, Z = 2.59, P = 0.0048).

PAGE 148

148 REFERENCES CITED Allen, R., S. K. Taylor, and E. S. Sulkin. 1970. Studies of arthropod-bor ne virus infections in Chiroptera. Am. J. Tr op. Med. Hyg. 19: 851-859. Allison, A. B., D. G. Mead, S. E. J. Gibbs, D. M. Hoffman, and D. E. Stallknecht. 2004. West Nile virus viremia in wild rock pigeons. Emerg. Inf ect. Dis. 10: 2252-2255. Anderson, R. A., and R. A. Brust. 1995. Field evidence for multiple host contacts during blood-feeding by Culex tarsalis Cx restuans and Cx nigripalpus (Diptera: Culicidae). J. Med. Entomol. 32: 705-710. Anderson, R. A., and R. A. Brust. 1996. Blood feeding success of Aedes aegypti and Culex nigripalpus (Diptera: Culicidae) in relation to de fensive behavior of Japanese quail ( Coturnix japonica ) in the laboratory. J. Vector Ecol. 21: 94-104. Anderson, J. F., T. G. Andreadis, A. J. Main, and D. L. Kline. 2004. Prevalence of West Nile virus in tree canopy-inhabiting Culex pipiens and associated mosquitoes. Am. J. Trop. Med. Hyg. 71: 112-119. Apperson, C. S., H. K. Hassan, B. A. Harrison, H. M. Savage, S. E. Aspen, A. Farajollahi, W. Crans, T. J. Daniels, R. C. Falco, M. Benedict, M. Anderson, L. McMillen, and T. R. Unnasch. 2004. Host feeding patterns of establishe d and potential mosquito vectors of West Nile virus in the easter n United States. Vector Born e and Zoonotic Dis. 4: 71-82. Appleby, P. D., J. A. Mench, and B. O. Hughes. 2004. Poultry behavior and welfare. CABI Publishing, Cambridge, MA. Bellamy, R. E., and W. C. Reeves. 1952. A portable mosquito bait-trap. Mosq. News 12: 256259. Beveroth, T. A., M. P. Ward, R. L. Lampman, A. M. Ringia, and R. J. Novak. 2006. Changes in seroprevalence of West Nile vi rus across Illinois in free-ranging birds from 2001 through 2004. Am. J. Trop. Med. Hyg. 74: 174-179. Bidlingmayer, W. L. 1966. Use of the truck trap for eval uating adult mosquito populations. Mosq. News 26: 139-143. Bigler, W. J. 1971. Serologic evidence of Venezuelan eq uine encephalitis vi rus infections in raccoons of South Central Flor ida. J. Wild. Dis. 7: 166-170. Bigler, W. J., and G. L. Hoff. 1975. Arbovirus surveillance in Florid a: Wild vertebrate studies 1965-1974. J. Wild. Dis. 11: 348-356. Bigler, W. J., E. B. Lassing, E. E. Buff, E. C. Prather, E. C. Beck, and G. L. Hoff. 1976. Endemic Eastern equine encephalomyelitis in Florida: A twenty-year analysis, 1955-1974. Am. J. Trop. Med. Hyg. 25: 884-890.

PAGE 149

149 Blackmore, C. G. M., L. M. Stark, W. C. Jeter, R. L. Oliveri, R. G. Brooks, L. A. Conti, and S. T. Wiersma. 2003. Surveillance results from the first West Nile virus transmission season in Florida, 2001. Am. J. Trop. Med. Hyg. 69: 141-150. Blackmore, J. S., L. C. Lamotte, C. A. Sooter, and P. Holden. 1962. Relationship of Culex tarsalis density to transmission ra tes of WE and SLE viruses. Mosq. News 22: 321-324. Bond, J. O. 1969. St. Louis encephalitis virus infection in man. Florida State Board of Health Monograph "Saint Louis Encepha litis in Florida" 12: 1-29. Bond, J. O., W. C. Ballard, R. E. Markush, and E. R. Alexander. 1963. The 1959 outbreak of St. Louis encephalitis in Flor ida. Florida State Board of Health Monograph "Tampa Bay Area arbovirus investigations" 5: 2-18. Bond, J. O., F. Y. Lewis, W. J. Jennings, and I. E. MacLeod. 1965. Transovarian transmission of hemagglutination inhibition antibody to St. Louis encephalitis virus in chickens. Am. J. Trop. Med. Hyg. 14: 1085-1089. Bowen, G. S. 1977. Prolonged Western equine encephaliti s viremia in the Texas tortoise ( Gopherus berlandieri ). Am. J. Trop. Med. Hyg. 26: 171-175. Bowen, G. S., and D. B. Francy. 1980. Surveillance, pp. 473-502. In T. P. Monath (ed.), St. Louis Encephalitis. American Public Health Association, Washington, DC. Broce, A. B. 1988. An improved Alsynite trap for Stable flies, Stomoxys calcitrans (Diptera: Muscidae). J. Med. Entomol. 25: 406-409. Brownstein, J. S., T. R. Holford, and D. Fish. 2004. Enhancing West Nile virus surveillance, United States. Emerg. Infect. Dis. 10: 1129-1133. Busch, M. P., D. J. Wright, B. Custer, L. H. To bler, S. L. Stramer, S. H. Kleinman, H. E. Prince, C. Bianco, G. Foster, L. R. Petersen, G. Nemo, and S. A. Glynn. 2006. West Nile virus infections projected from blood donor screening data, United States, 2003. Emerg. Infect. Dis. 12: 395-402. Byrne, R. J., and M. L. Robbins. 1961. Mortality patterns and antibody response in chicks inoculated with Eastern equine encepha lomyelitis virus. J. Immunol. 86: 13-16. Calisher, C. H., and N. Karabatsos. 1988. Arbovirus serogroups: definition and geographic distribution, pp. 19-57. In T. P. Monath (ed.), The Arboviruses: Epidemiology and Ecology. CRC Press, Boca Raton, FL. Calisher, C. H., T. P. Monath, D. J. Muth, J. S. Lazuick, D. W. Trent, D. B. Francy, G. E. Kemp, and F. W. Chandler. 1980. Characterization of Fo rt Morgan virus, an Alphavirus of the Western equine encepha litis virus complex in an unusual ecosystem. Am. J. Trop. Med. Hyg. 29: 1428-1440.

PAGE 150

150 Callahan, J. L., and C. D. Morris. 1987. Habitat characteristics of Coquillettidia perturbans in Central Florida. J. Am. Mosq. Control Assoc. 3: 176-180. Carroll, M. K., and J. A. Bourg. 1977. The night-time flight activ ity and relative abundance of fifteen species of Louisiana mosquitoes. Mosq. News 37: 661-664. Center for Disease Control and Prevention. 2003. Epidemic/Epizootic We st Nile virus in the United States: Guidelines for surveillance, prevention, and control, pp. 80, Fort Collins, CO. Center for Disease Control and Prevention. 2006a. Arboviral encephalitis cases reported in humans, by type, United States, 1964-2005. http://www.cdc.gov/ncidod/ dvbid/arbor/arbocase.htm Center for Disease Control and Prevention. 2006b. Mosquito species producing WNV positives by year. http://www.cdc.gov/ncidod/dvb id/westnile/mos quitospecies.htm Center for Disease Control and Prevention. 2006c. Vertebrate ecology. http://www.cdc.gov/ncidod/dvbid/ westnile/birdspecies.htm Chamberlain, R. W. 1958. Vector relationships of the arth ropod-borne encephalitides in North America. Ann. N. Y. Acad. Sci. 70: 312-319. Chamberlain, R. W. 1980. History of St. Loui s encephalitis, pp. 3-61. In T. P. Monath (ed.), St. Louis encephalitis. American Public Health Association, Washington D.C. Chamberlain, R. W., and W. D. Sudia. 1961. Mechanism of transmission of viruses by mosquitoes. Annu. Rev. Entomol. 6: 371-390. Chamberlain, R. W., R. K. Sikes, D. B. Nelson, and W. D. Sudia. 1954. Studies on the North American arthropod-borne encephali tides: VI. Quantitative determinations of virus-vector relationships. Am. J. Hyg. 60: 278-285. Chamberlain, R. W., W. D. Sudia, and J. D. Gillett. 1959. St. Louis encephalitis virus in mosquitoes. Am. J. Hyg. 70: 221-235. Chamberlain, R. W., W. D. Sudia, T. H. Work, P. H. Coleman, V. F. Newhouse, and J. G. Johnston. 1969. Arbovirus studies in South Florida, with emphasis on Venezuelan equine encephalomyelitis virus. Am. J. Epidemiol. 89: 197-210. Cherry, B., S. C. Trock, A. L. Glaser, L. D. Kramer, G. D. Ebel, C. Glaser, and J. R. Miller. 2001. Sentinel chickens as a surveillance tool for West Nile virus in New York City, 2000. Ann. N. Y. Acad. Sci. 951: 343-346. Coffey, L. L., A.-S. Carrara, S. Paessler, M. L. Haynie, R. D. Bradley, R. B. Tesh, and S. C. Weaver. 2004. Experimental Everglades viru s infection of Cotton rats ( Sigmodon hispidus ). Emerg. Infect. Dis. 10: 2182-2188.

PAGE 151

151 Crans, W. J. 1986. Failure of chickens to act as sentinel s during an epizootic of Eastern equine encephalitis in Southern New Jersey USA. J. Med. Entomol. 23: 626-629. Crans, W. J., and L. J. McCuiston. 1993. New Jersey's approach to encephalitis prevention. Bull. Soc. Vector Ecol. 18: 45-48. Crans, W. J., and T. L. Schulze. 1986. Evidence incriminating Coquillettidia perturbans (Diptera: Culicidae) as an epizootic vector of Eastern equine encephali tis. I. Isolation of EEEV virus from C. perturbans during an epizootic among horses in New Jersey. Bull. Soc. Vector Ecol. 11: 178-184. Crans, W. J., D. F. Caccamise, and J. R. McNelly. 1994. Eastern equine encephalomyelitis virus in relation to the avian community of a coastal cedar swamp. J. Med. Entomol. 31: 711-728. Cully, J. F., J. P. Grieco, and D. Kissel. 1991. Defensive behavior of eastern chipmunks against Aedes triseriatus (Diptera: Culicidae). J. Med. Entomol. 28: 410-416. Cupp, E. W., K. Klingler, H. K. Hassan, L. M. Viguers, and T. R. Unnasch. 2003. Transmission of Eastern equine encephalomye litis virus in Central Alabama. Am. J. Trop. Med. Hyg. 68: 495-500. Darbro, J. M., and L. C. Harrington. 2006. Bird-baited traps for surveillance of West Nile mosquito vectors: Effect of bird species, tr ap height, and mosquito escape rates. J. Med. Entomol. 43: 83-92. Davis, A., M. Bunning, P. Gordy, N. A. Panella, B. J. Blitvich, and R. A. Bowen. 2005. Experimental and natural infection of North Am erican bats with West Nile virus. Am. J. Trop. Med. Hyg. 73: 467-469. Day, J. F. 1989. The use of sentinel chickens for arboviru s surveillance in Florida. J. Fla. Mosq. Control Assoc. 60: 56-62. Day, J. F. 1991. A review of the 1990 St. Louis encephal itis virus epidemic in Indian River County, Florida. N. J. Mosq. Control. Assoc. 78: 32-39. Day, J. F. 2001. Predicting St. Louis encephalitis virus ep idemics: Lessons from recent, and not so recent, outbreaks. Annu. Rev. Entomol. 46: 111-138. Day, J. F., and D. B. Carlson. 1985. The importance of autumn rainfall and sentinel flock location to understanding the ep idemiology of St. Louis encepha litis virus in Indian River County, Florida. J. Am. Mosq. Control Assoc. 1: 305-309. Day, J. F., and G. A. Curtis. 1989. Influence of rainfall on Culex nigripalpus (Diptera: Culicidae) blood-feeding behavior in Indian River County, Fl orida. Ann. Entomol. Soc. Am. 82: 32-37.

PAGE 152

152 Day, J. F., and G. A. Curtis. 1999. Blood feeding and oviposition by Culex nigripalpus (Diptera: Culicidae) before, during, and afte r a widespread St. L ouis encephalitis virus epidemic in Florida. J. Med. Entomol. 36: 176-181. Day, J. F., and J. D. Edman. 1984. Mosquito engorgement on normally defensive hosts depends on host activity patterns. J. Med. Entomol. 21: 732-740. Day, J. F., and J. D. Edman. 1988. Host location, blood-feeding, a nd oviposition behavior of Culex nigripalpus (Diptera: Culicidae): Their infl uence on St. Louis encephalitis virus transmission in Southern Florida, pp. 1-8. In T. W. Scott and J. Grumstrup-Scott (eds.), Proceedings, Symposium: the Role of Vecto r-Host Interactions in Disease Transmission. National Conference of the Entomological Society of America, 10 December 1985, Hollywood, FL. Miscellaneous Publication 68. Entomological Society of America, Lanham, MD. Day, J. F., and A. L. Lewis. 1991. An integrated approach to arboviral surveillance in Indian River County, Florida. J. Fla. Mosq. Control Assoc. 62: 46-52. Day, J. F., and A. L. Lewis. 1992. An integrated approach to St Louis encephal itis surveillance in Indian River County, Florida. Fl a. J. Public Health 4: 12-16. Day, J. F., and L. M. Stark. 1996a. Eastern equine encephalitis transmission to emus ( Dromanius novaehollandiae ) in Volusia county, Florid a: 1992 through 1994. J. Am. Mosq. Control Assoc. 12: 429-436. Day, J. F., and L. M. Stark. 1996b. Transmission patterns of St. Louis encephalitis and Eastern equine encephalitis viruses in Florid a: 1978-1993. J. Med. Entomol. 33: 132-139. Day, J. F., and L. M. Stark. 1999. Avian serology in a St. Louis encephalitis epicenter before, during, and after a widespread epidemic in s outh Florida, USA. J. Med. Entomol. 36: 614624. Day, J. F., and L. M. Stark. 2000. Frequency of Saint Louis ence phalitis virus in humans from Florida, USA: 1990-1999. J. Med. Entomol. 37: 626-633. Day, J. F., G. A. Curtis, and J. D. Edman. 1990. Rainfall directed oviposition behavior of Culex nigripalpus (Diptera: Culicidae) and its infl uence on St. Louis encephalitis virus transmission in Indian River County, Fl orida. J. Med. Entomol.. 27: 43-50. Day, J. F., L. M. Stark, J-T. Zhang, A. M. Ramsey, and T. W. Scott. 1996. Antibodies to arthropod-borne encephalitis viruses in small mammals from Southern Florida. J. Wild. Dis. 32: 431-436. Day, J. F., E. E. Storrs, L. M. Stark, A. L. Lewis, and S. Williams. 1995. Antibodies to St. Louis encephalitis virus in armadillos from Southern Florida. J. Wild. Dis. 31: 10-14.

PAGE 153

153 Day, J. F., R. Winner, R. E. Parsons, and J. T. Zhang. 1991. Distribution of St. Louis encephalitis viral antibody in se ntinel chickens maintained in Sarasota County, Florida: 1978-1988. J. Med. Entomol. 28: 19-23. Deegan, C. S., J. E. Burns, M. Huguenin, E. Y. Steinhaus, N. A. Panella, S. Beckett, and N. Komar. 2005. Sentinel pigeon surveillance for West Nile virus by using lard-can traps at differing elevations and canopy cover cl asses. J. Med. Entomol. 42: 1039-1044. Dickerman, R., and W. Scherer. 1983. Equine herds as sentinels for Venezuelan equine encephalitis virus activity, Nicar agua 1977. Bull. PAHO 17: 14-18. Dougherty, E., and J. I. Price. 1960. Eastern encephalitis in Wh ite Pekin ducklings on Long Island. Avian Dis. 4: 247-256. Dow, R. P., P. H. Coleman, T. H. Work, and K. E. Meadows. 1964. Isolation of St. Louis encephalitis viruses from mosquitoes in Tamp a Bay Area of Florida dur ing the epidemic of 1962. Am. J. Trop. Med. Hyg. 13: 462-468. Dow, R. P., W. C. Reeves, and R. E. Bellamy. 1957. Field tests of avia n host preference of Culex tarsalis Coq. Am. J. Trop. Med. Hyg. 6: 294-303. Downing, J. D., and W. J. Crans. 1977. Ehrenberg pigeon trap as a sampler of Culex mosquitoes for St. Louis encephalitis surveillance. Mosq. News 37: 48-53. Downs, W. G., T. H. G. Aitken, and L. Spence. 1959. Eastern equine encephalitis virus isolated from Culex nigripalpus in Trinidad. Science 27: 1671. Edman, J. D. 1971. Host-Feeding Patterns of Florida Mosquitos .1. Aedes Anopheles Coquillettidia Mansonia and Psorophora J. Med. Entomol. 8: 687-695. Edman, J. D. 1974. Host-feeding patterns of Florida mosquitoes .3. Culex ( Culex ) and Culex ( Neoculex ). J. Med. Entomol. 11: 95-104. Edman, J. D., and H. W. Kale. 1971. Host behavior Its infl uence on feeding success of mosquitoes. Ann. Entomol. Soc. Am. 64: 513-516. Edman, J. D., and D. J. Taylor. 1968. Culex nigripalpus Seasonal shift in bird-mammal feeding ratio in a mosquito vector of human encephalitis. Science 161: 67-68. Edman, J. D., L. A. Webber, and H. W. Kale. 1972a. Effect of mosquito density on interrelationship of host behavior and mosquito feed ing success. Am. J. Trop. Med. Hyg. 21: 487491. Edman, J. D., L. A. Webber, and H. W. Kale. 1972b. Host-feeding patterns of Florida mosquitoes. 2. Culiseta J. Med. Entomol. 9: 429-434.

PAGE 154

154 Edman, J. D., L. A. Webber, and A. A. Schmid. 1974. Effect of host defenses on feeding pattern of Culex nigripalpus when offered a choice of blood sources. J. Parasitol.60: 874883. Eidson, M. 2001. "Neon needles" in a haystack. Ann. N. Y. Acad. Sci. 951: 38-53. Eidson, M., N. Komar, F. Sorhage, R. L. Nels on, T. Talbot, F. Mostashari, R. G. McLean, and W. N. V. A. M. S. Group. 2001a. Crow deaths as a sentinel surveillance system for West Nile virus in the No rtheastern United States, 1999. Emerg. Infect. Dis. 7: 615-620. Eidson, M., L. D. Kramer, W. Stone, Y. Hagiwar a, K. Schmit, and N. Y. S. W. N. V. A. S. Team. 2001b. Dead bird surveillance as an early warning system for West Nile virus. Emerg. Infect. Dis. 7: 631-635. Eidson, M., J. Miller, L. D. Kramer, B. Ch erry, Y. Hagiwara, and W. N. V. B. M. A. Group. 2001c. Dead crow densities and human cases of West Nile virus, New York state, 2000. Emerg. Infect. Dis. 7: 662-664. Eidson, M., K. Schmit, Y. Hagiwara, M. Anand, B. P. Backenson, I. Gotham, and L. D. Kramer. 2005. Dead crow density and West Nile virus monitoring, New York. Emerg. Infect. Dis. 11: 1370-1375. Eldridge, B. F. 1987. Strategies for surveillance, preventi on, and control of arbovirus diseases in Western North America. Am. J. Trop. Med. Hyg. 37: 77-86. Endy, T. P., and A. Nisalak. 2002. Japanese encephalitis virus: Ecology and epidemiology, Curr. Top. Microbiol. Immunol. 267: 11-48. Favorite, F. G. 1960. Some evidence of local origin of EEE virus in Florida. Mosq. News 20: 87-92. Florida Department of Agriculture and Consumer Services. 2006. Keetch Byram Drought Index (KBDI). http://flame.fl-dof.com/f ire_weather/KBDI/index.html Florida Department of Health. 2005. Surveillance and control of selected arthropod-borne diseases in Florida, pp. 85. Florida Department of Health. 2006a. Arthropod-borne encephalitis. http://www.floridacharts.com/chart s/report.aspx?domain=01&indnumber=0163 Florida Department of Health. 2006b. Weekly data for arbovirus surveillance. http://www.doh.state.fl.us/environment/c ommunity/arboviral/weekly-data.htm Foster, W. A., and E. D. Walker. 2002. Mosquitoes (Culicidae), pp. 203-262. In G. Mullen and L. Durden (eds.), Medical and Veterinary Entomology. Academic Press, Boston, MA. Gard, G. P., J. E. Shorthose, R. P. Weir, S. J. Walsh, and L. F. Melville. 1988. Arboviruses recovered from sentinel livestock in north ern Australia. Vet. Microbiol. 18: 109-118.

PAGE 155

155 Gebhardt, L. P., and D. W. Hill. 1960. Overwintering of Western equine encephalitis virus. Proc. Soc. Exp. Biol. Med. 104: 695-698. Geevarghese, G., B. H. Shaikh, P. G. Jacob, H. R. Bhat, and K. M. Pavri. 1987. Domestic pigs as sentinels to monitor the activity of Ja panese encephalitis and West Nile viruses in Kolar District, Karnataka. I ndian J. Med. Res. 86: 413-418. Gergis, N. N., and S. B. Presser. 1988. Preliminary observations of three avian surveillance systems for St. Louis encephalitis in Los Ange les and Orange Counties, California. Proc. Pap. 56th Annu. Conf. Calif. Mosq. Vect Control Assoc. 56: 54-57. Gladney, W. J., and E. C. Turner. 1970. Mosquito flight activity studies utilizing timerequipped light traps and malaise traps. Mosq. News 30: 75-80. Godsey, M. S., M. S. Blackmore, N. A. Panell a, K. L. Burkhalter, K. L. Gottfried, L. A. Halsey, C. R. Rutledge, S. A. Langevin, R. Gates, K. M. Lamonte, A. J. Lambert, R. S. Lanciotti, C. G. M. Blackmore, T. Loyless, L. Stark, R. L. Oliveri, L. A. Conti, and N. Komar. 2005a. West Nile virus epizootiology in the Southeastern United States, 2001. Vector Borne Zoonotic Dis. 5: 82-89. Godsey, M. S., R. S. Nasci, H. M. Savage, S. E. Aspen, R. King, A. M. Powers, K. L. Burkhalter, L. Colton, D. Charnetzky, S. Lasater, V. Taylor, and C. T. Palmisano. 2005b. West Nile virus-infected mosquitoes, Louisiana, 2002. Emerg. Infect. Dis. 11: 1399-1404. Gruwell, J. A., C. L. Fogarty, S. G. Bennett, G. L. Challet, K. S. Vanderpool, M. Jozan, and J. P. Webb. 2000. Role of peridomestic birds in the transmission of St. Louis encephalitis virus in Southern California. J. Wild. Dis. 36: 13-34. Guptill, S. C., K. G. Julian, G. L. Cam pbell, S. D. Price, and A. A. Marfin. 2003. Earlyseason avian deaths from West Nile virus as warnings of human in fection. Emerg. Infect. Dis. 9: 483-484. Haeger, J. S., and G. F. OMeara. 1970. Rapid incorporation of wild genotypes of Culex nigripalpus (Diptera: Culicidae) into laboratory-ad apted strains. Ann. Entomol. Soc. Am. 63: 1390-1391. Hayes, R. O. 2000. Eastern and Western Encephalitis, pp. 29-57. In J. H. Steele (ed.), CRC Handbook Series in Zoonoses. CRC Press, Boca Raton, FL. Hayes, R. O., and A. D. Hess. 1964. Climatological conditions asso ciated with outbreaks of Eastern encephalitis. Am. J. Trop. Med. Hyg. 13: 851-858. Hayes, C. G., J. B. Daniels, H. K. Maxfield, and R. E. Wheeler. 1964. Field and laboratory studies on Eastern Encephalitis in warm and cold-blooded vertebrates. Am. J. Trop. Med. Hyg. 13: 595-606.

PAGE 156

156 Henderson, J. R., N. Karabatsos, A. T. C. Bourke, R. C. Wallis, and R. M. Taylor. 1962. A survey for arthropod-borne viruses in Southcentral Florida. Am. J. Trop. Med. Hyg. 11: 800-810. Herbold, J. R., W. P. Heuschele, R. L. Berry, and M. A. Parsons. 1983. Reservoir of St. Louis encephalitis virus in Ohio bats. Am. J. Vet. Res. 44: 1889-1893. Holden, P. 1955. Transmission of Eastern equine encephalomyelitis in ring-necked pheasants. Proc. Soc. Exp. Biol. Med. 88: 607-610. Holden, P., R. O. Hayes, C. J. Mitchell, D. B. Francy, J. S. Lazuick, and T. B. Hughes. 1973. House Sparrows, Passer domesticus (L.), as hosts of arboviruse s in Hale County, Texas. 1. Field studies, 1965-1969. Am. J. Trop. Med. Hyg. 22: 244-253. Howard, J. J., J. Oliver, and M. A. Grayson. 2004. Antibody response of wild birds to natural infection with Alphaviruses J. Med. Entomol. 41: 1090-1103. Hurlbut, H. S. 1966. Mosquito salivation and virus transmission. Am. J. Trop. Med. Hyg. 15: 989-993. Jacobson, E. R., P. E. Ginn, J. M. Troutman, L. Farina, L. Stark, K. Klenk, K. L. Burkhalter, and N. Komar. 2005. West Nile virus infect ion in farmed American alligators ( Alligator mississippiensis ) in Florida. J. Wild. Dis. 41: 96-106. Jennings, W. L., A. L. Lewis, G. E. Sather, W. McD Hammon, and J. O. Bond. 1968. California encephalitis group viruses in Florid a rabbits: Report of experimental and sentinel studies. Am. J. Trop. Med. Hyg. 17: 781-787. Johnson, G. D., M. Eidson, K. Schmitt, A. Ellis, and M. Kulldorff. 2006. Geographic prediction of human onset of West Nile virus using dead crow clusters: An evaluation of year 2002 data in New York Stat e. Am. J. Epidemiol. 163: 171-180. Julian, K. G., M. Eidson, A. M. Kipp, E. Weiss, L. R. Petersen, J. R. Miller, S. R. Hinten, and A. A. Marfin. 2002. Early season crow mortality as a sentinel for West Nile virus disease in humans, Northeastern United St ates. Vector Borne Zoonotic Dis. 2: 145-155. Kale, H. W., L. A. Webber, and J. D. Edman. 1972. Effect of behavior and age of individual ciconiiform birds on mosquito feed ing success. Mosq. News 32: 343-350. Karabatsos, N., A. L. Lewis, C. H. Ca lisher, A. R. Hunt, and J. T. Roehrig. 1988. Identification of Highlands J virus from a Florida horse. Am. J. Trop. Med. Hyg. 39: 603606. King, W. 1983. The use of sentinel birds for early wa rning of mosquito-borne encephalitis in Delaware in 1981 and 1982. Proc. N. J. Mosq. Control. Assoc. 70: 137-141. Klenk, K., and N. Komar. 2003. Poor replication of West Nile virus (New York 1999 strain) in three reptilian and one amphibian sp ecies. Am. J. Trop. Med. Hyg. 69: 260-262.

PAGE 157

157 Klenk, K., J. Snow, K. Morgan, R. A. Bowen, M. Stephens, F. Foster, P. Gordy, S. Beckett, N. Komar, D. J. Gubler, and M. Bunning. 2004. Alligators as West Nile virus amplifiers. Emerg. Infect. Dis. 10: 2150-2155. Klowden, M. J., and A. O. Lea. 1979. Effect of defensive host beha vior on the blood meal size and feeding success of natural populations of mosquitoes (D iptera: Culicidae). J. Med. Entomol. 15: 514-517. Komar, N. 2001. West Nile virus surveilla nce using sentinel birds. Ann. N. Y. Acad. Sci. 251: 58-73. Komar, N., and A. Spielman. 1995. Japanese quail as sentinels for Eastern encephalitis virus. N. J. Mosq. Control. Assoc. 82: 32-36. Komar, N., D. J. Dohm, M. J. Turell, and A. Spielman. 1999. Eastern equine encephalitis virus in birds: Relative comp etence of European starlings ( Sturnis vulgaris ). Am. J. Trop. Med. Hyg. 60: 387-391. Komar, N., S. A. Langevin, S. R. Hinten, N. Nemeth, E. Edwards, D. Hettler, B. Davis, R. A. Bowen, and M. Bunning. 2003. Experimental infection of North American birds with the New York 1999 strain of West Nile virus. Emerg. Infect. Dis. 9: 311-322. Komar, N., N. A. Panella, J. E. Burns, S. W. Dusza, T. M. Mascarenhas, and T. O. Talbot. 2001. Serologic evidence for West Nile virus infection in birds in the New York City vicinity during an outbreak in 1999. Emerg. Infect. Dis. 7: 621-625. LaMotte, L. C., G. T. Crane, R. B. Shriner, and L. J. Kirk. 1967. Use of adult chickens as arbovirus sentinels. I. Viremia and persistenc e of antibody in experimentally inoculated adult chickens. Am. J. Trop. Med. Hyg. 16: 348-356. Langevin, S. A., M. Bunning, B. Davis, and N. Komar. 2001. Experimental infection of chickens as candidate sentinels for West Nile virus. Emerg. Infect. Dis. 7: 726-729. Lord, C. C., C. R. Rutledge, and W. J. Tabachnick. 2006. Relationship between host viremia and vector susceptibility for arbovi ruses. J. Med. Entomol. 43: 623-630. Lord, R. D., and C. H. Calisher. 1970. Further evidence of southw ard transport of arboviruses by migratory birds. Am. J. Epidemiol. 92: 73-78. Lounibos, P. L., and R. L. Escher. 1985. Mosquitoes associated with water lettuce ( Psitia stratiotes ) in southeastern Florida. Fla. Entomol. 68: 169-178. Marfin, A. A., L. R. Petersen, M. Eidson, J. Mi ller, J. Hadler, C. Farello, B. Werner, G. L. Campbell, M. Layton, P. Smith, E. Bresnitz M. Cartter, J. Scaletta, G. Obiri, M. Bunning, R. C. Craven, J. T. Roehrig, K. G. Julian, S. R. Hinten, D. J. Gubler, and A. C. S. Group. 2001. Widespread West Nile virus ac tivity, Eastern United States, 2000. Emerg. Infect. Dis. 7: 730-735.

PAGE 158

158 Maxwell, G. R., and H. W. Kale. 1977. Maintenance and anti-insect behavior of 6 species of ciconiiform birds in South Florida. Condor 79: 51-55. McCaig, L. F., H. T. Janowski, R. A. Gunn, and T. F. Tsai. 1994. Epidemiologic aspects of a St. Louis encephalitis outbr eak in Fort Walton Beach, Fl orida in 1980. Am. J. Trop. Med. Hyg. 50: 387-391. McLean, R. G., and G. S. Bowen. 1980. Vertebrate Hosts, pp. 381-450. In T. P. Monath (ed.), St. Louis encephalitis. American Public Health Association, Washington, DC. McLean, R. G., G. Frier, G. L. Parham, D. B. Francy, T. P. Monath, E. G. Campos, A. Therrien, J. Kerschner, and C. H. Calisher. 1985. Investigations of the vertebrate hosts of Eastern equine encephalitis during an epizootic in Michigan, 1980. Am. J. Trop. Med. Hyg. 34: 1190-1202. McLean, R. G., J. Mullenix, J. Kerschner, and J. Hamm. 1983. The House Sparrow ( Passer domesticus ) as a sentinel for St. Louis encepha litis virus. Am. J. Trop. Med. Hyg. 32: 1120-1129. Miller, D. L., M. J. Mauel, C. Baldwin, G. Burtle, D. Ingram, M. E. Hines, and K. S. Frazier. 2003. West Nile virus in farmed allig ators. Emerg. Infect. Dis. 9: 794-799. Mitchell, C. J., D. B. Francy, and T. P. Monath. 1980. Arthropod Vectors, pp. 313-380. In T. P. Monath (ed.), St. Louis Encephalitis. Amer ican Public Health Association, Washington, DC. Mitchell, C. J., C. D. Morris, G. C. Smith, N. Karabatsos, D. Vanlandingham, and E. Cody. 1996. Arboviruses associated with mosquitoes from nine Florida counties during 1993. J. Am. Mosq. Control Assoc. 12: 255-262. Mitchell, C. J., M. L. Niebylski, G. C. Smith, N. Karabatsos, D. Martin, J. P. Mutebi, G. B. Craig, and M. J. Mahler. 1992. Isolation of Eastern equi ne encephalitis virus from Aedes albopictus in Florida. Science 257: 526-527. Monath, T. P., M. S. Sabattini, R. Pauli, J. F. Daffner, C. J. Mitchell, G. S. Bowen, and C. B. Cropp. 1985. Arbovirus investigations in Arge ntina, 1977-1980 .4. Serologic surveys and sentinel equine program. Am. J. Trop. Med. Hyg. 34: 966-975. Moore, C. G., R. G. McLean, C. J. Mitchell, R. S. Nasci, T. F. Tsai, C. H. Calisher, A. A. Marfin, P. S. Moore, and D. J. Gubler. 1993. Guidelines for arbovirus surveillance programs in the United States, pp. 81, Fort Collins. Morris, C. D. 1988. Eastern equine encephalomyelitis, pp. 1-20. In T. P. Monath (ed.), The Arboviruses: Epidemiology and Ecology. CRC Press, Boca Raton, FL. Morris, C. D. 1992. Eastern equine encephalitis. J. Fla. Mosq. Control Assoc. 63: 23-34.

PAGE 159

159 Morris, C. D., W. G. Baker, L. Stark, J. Burgess, and A. L. Lewis. 1994. Comparison of chickens and pheasants as sentinels for Eastern equine encephalitis and St. Louis encephalitis viruses in Florida. J. Am. Mosq. Control Assoc. 10: 545-548. Morris, C. D., J. L. Callahan, and A. L. Lewis. 1990. Distribution and abundance of larval Coquillettidia perturbans in a Florida freshwater marsh. J. Am. Mosq. Control Assoc. 6: 452-460. Mostashari, F., M. L. Bunning, P. T. Kitsutani, D. A. Singer, D. Nash, M. J. Cooper, N. Katz, K. A. Liljebjelke, B. J. Biggerstaff, A. D. Fine, M. C. Layton, S. M. Mullin, A. J. Johnson, D. A. Martin, E. B. Hayes, and G. L. Campbell. 2001. Epidemiologic West Nile encephalitis, New York, 1999: results of a household-based seroepidemiological survey. Lancet 358: 261-264. Mount, G. A. 1970. Optimum droplet size for adult mosquito control with space sprays or aerosols of insecticides. Mosq. News 30: 70-75. Mullen, G., and L. Durden. 2002. Medical and Veterinary Entomology. Academic Press, Boston, MA. Mulrennan, J. A. 1969. Early observations during SLEV epidemics in 1959, 1961 and 1962. Florida State Board of Health Monograph "Saint Louis Encephali tis in Florida" 12: xi-xii. Nasci, R. S., and J. D. Edman. 1981. Vertical and temporal fli ght activity of the mosquito Culiseta melanura (Diptera: Culicidae) in southeas tern Massachusetts. J. Med. Entomol. 18: 501-504. Nayar, J. K. 1982. Bionomics and physiology of Culex nigripalpus (Diptera: Culicidae) of Florida: an important vect or of diseases, pp. 73, FAES T echnical Bulletin, Vero Beach, FL. Nelson, D. B., K. D. Kappus, H. T. Janowski, E. E. Buff, F. M. Wellings, and N. J. Schneider. 1983. St. Louis encephalitis Florida 1977: Patterns of a widespread outbreak. Am. J. Trop. Med. Hyg. 32: 412-416. Nemeth, N., D. Gould, R. A. Bowen, and N. Komar. 2006. Natural and experimental West Nile virus infection in five rapt or species. J. Wild. Dis. 42: 1-13. Nichols, J. B., and W. J. Bigler. 1967. Animal sentinels in a statewide arthropod-borne encephalitis surveillance program. J. Am. Vet. Med. Assoc. 151: 1767-1771. O'Bryan, P. D., and H. J. Jefferson. 1991. The year of the chicken: the good and bad of a sentinel chicken flock during the 1990 Florida SLEV epidem ic. J. Fla. Mosq. Control Assoc. 62: 59-63. Olson, J. G., W. C. Reeves, R. W. Emmons, and M. M. Milby. 1979. Correlation of Culex tarsalis population indices with the incidence of St. Louis ence phalitis and western equine encephalomyelitis in California. Am. J. Trop. Med. Hyg. 28: 335-343.

PAGE 160

160 Olson, J. G., T. W. Scott, L. H. Lorenz, and J. L. Hubbard. 1991. Enzyme-immunoassay for detection of antibodies against Eastern equine encephalomyelitis virus in sentinel chickens. J. Clin. Microbiol. 29: 1457-1461. O'Meara, G. F., and F. D. S. Evans. 1983. Seasonal patterns of abundance among three species of Culex mosquitoes in a south Florida wastew ater lagoon. Ann. Entomol. Soc. Am. 76: 130-133. Ostlund, E. N., R. L. Crom, D. D. Peders en, D. J. Johnson, W. O. Williams, and B. J. Schmitt. 2001. Equine West Nile encephalitis, Un ited States. Emerg. Infect. Dis. 7: 665669. Ott, R. L., and M. Longnecker. 2001. Statistical methods and data analysis. Thomson Learning, United States of America. Peiris, J. S., F. P. Amerasinghe, C. K. Arunagiri, L. P. Perera, S. H. Karunaratne, C. B. Ratnayake, T. A. Kulatilak a, and M. R. Abeysinghe. 1993. Japanese encephalitis in Sri Lanka: comparison of vector and virus ecology in different agro-climatic areas. Tran. R. Soc. Trop. Med. Hyg. 87: 541-548. Petersen, L. R., and A. A. Marfin. 2002. West Nile virus: A pr imer for the clinician. Ann. Intern. Med. 137: 173-179. Pilipski, J. D., L. M. Pilipski, and L. S. Risley. 2004. West Nile virus antibodies in bats from New Jersey and New York. J. Wild. Dis. 40: 335-337. Provost, M. W. 1969. The natural history of Culex nigripalpus Florida State Board of Health Monograph "Saint Louis Encepha litis in Florida" 12: 46-62. Provost, M. W. 1973. Mosquito flight and night relative hu midity in Florida. Fla. Sci. 36: 218225. Rainey, M. B., G. V. Warren, A. D. Hess, and J. S. Blackmore. 1962. A sentinel chicken shed and mosquito trap for use in encephalit is field studies. Mosq. News 22: 337-342. Ranck, F. M., J. H. Gainer, J. E. Hanley, and S. L. Nelson. 1965. Natural outbreak of Eastern and Western encephalitis in pen-raised Chuckars in Fl orida. Avian Dis. 9: 8-20. Rappole, J. H., S. R. Derrick son, and Z. Hubalek. 2000. Migratory birds and spread of West Nile virus in the Western Hemisphe re. Emerg. Infect. Dis. 6: 319-328. Reeves, W. C. 1968. A review of developments associated with the control of western equine and St. Louis encephalitis in California dur ing 1967. Proc. Pap. 35th Annu. Meet. Calif. Mosq. Control Assoc. 36: 65-70. Reeves, W. C., R. P. Scrivani, and R. E. Bellamy. 1961. Differentiation of encephalitis virus infection rates from transmission rates in mo squito vector populations. Am. J. Hyg. 73: 303-315.

PAGE 161

161 Reeves, W. C., J. M. Sturgeon, E. M. French, and B. Brookman. 1954. Transovarian transmission of neutralizing substances to Western equine and St. Louis encephalitis viruses by avian hosts. J. Infect. Dis. 95: 168-178. Reisen, W. K., J. L. Hardy, and S. B. Presser. 1992. Evaluation of domestic pigeons as sentinels for detecting arboviru s activity in Southern Calif ornia. Am. J. Trop. Med. Hyg. 46: 69-79. Reisen, W., H. Lothrop, R. Chiles, M. Madon, C. Cossen, L. Woods, S. Husted, V. Kramer, and J. Edman. 2004. West Nile virus in California. Emerg. Infect. Dis. 10: 1369-1378. Reisen, W. K., S. B. Presser, J. Y. Lin, B. Enge, J. L. Hardy, and R. W. Emmons. 1994. Viremia and serological responses in adult chickens infected with Western equine encephalomyelitis and St. Louis encephalitis viruses. J. Am. Mosq. Control Assoc. 10: 549-555. Reisen, W. K., S. S. Wheeler, S. Yamamoto, Y. Fang, and S. Garcia. 2005. Nesting Ardeid colonies are not a focus of elevated West Nile virus activity in Southe rn California. Vector Borne Zoonotic Dis. 5: 258-266. Reiter, P. 1988. Weather, vector biology, and arboviral recrudescence, pp. 245-255. In T. P. Monath (ed.), The Arboviruses: Epidem iology and Ecology. CRC Press, Boca Raton, FL. Robertson, L. C., S. Prior, C. S. Apperson, and W. S. Irby. 1993. Bionomics of Anopheles quadrimaculatus and Culex erraticus (Diptera: Culicidae) in the Falls Lake Basin, North Carolina: Seasonal changes in abundan ce and gonotrophic stat us, and host-feeding patterns. J. Med. Entomol. 30: 689-98. Rogers, A. J. 1969. Research on the control of SLEV vect ors in Florida and the Southeastern United States. Florida State Board of Health Monograph "Saint Louis Encephalitis in Florida" 12: 63-72. Rutledge, C. R., J. F. Day, C. C. Lord, L. M. Stark, and W. J. Tabachnick. 2003. West Nile virus infection rates in Culex nigripalpus (Diptera: Culicidae) do not reflect transmission rates in Florida. J. Med. Entomol. 40: 253-258. Sanders, M., A. Blumberg, and W. Haymaker. 1953. Polyradiculopathy in man produced by St. Louis encephalitis virus (S LE). South. Med. J. 46: 606-608. Sardelis, M. R., M. J. Turell, D. J. Dohm, and M. L. O'Guinn. 2001. Vector competence of selected North American Culex and Coquillettidia mosquitoes for West Nile virus. Emerg. Infect. Dis. 7: 1018-1022. Sardelis, M. R., M. J. Turell, M. L. O'Guinn, R. G. Andre, and D. R. Roberts. 2002. Vector competence of three North American strains of Aedes albopictus for West Nile virus. J. Am. Mosq. Control Assoc. 18: 284-289. SAS Institute. 2002. Version 9.1. SAS Institute, Cary, NC.

PAGE 162

162 Scott, T. W. 1988. Vertebrate Host Ecology, pp. 257-281. In T. P. Monath (ed.), The Arboviruses: Epidemiology and Ecology. CRC Press, Boca Raton, FL. Scott, T. W., and S. C. Weaver. 1989. Eastern equine encephalomyelitis virus: Epidemiology and evolution of mosquito tran smission. Adv. Virus Res. 37: 277-328. Scott, T. W., J. D. Edman, L. H. Lorenz, and J. L. Hubbard. 1988. Effects of disease on vertebrates' ability behaviorally to repel host-seeking mosquitoes, pp. 9-17. In T. W. Scott and J. Grumstrup-Scott (eds .), Proceedings, Symposium: the Role of Vector-Host Interactions in Disease Transmission. Nati onal Conference of the Entomological Society of America, 10 December 1985, Hollywood, FL. Miscellaneous Publication 68. Entomological Society of America, Lanham, MD. Scott, T. W., L. H. Lorenz, and J. D. Edman. 1990a. Effects of House sparrow age and arbovirus infection on attr action of mosquitoes. J. Med. Entomol. 27: 856-863. Scott, T. W., L. H. Lorenz, and S. C. Weaver. 1990b. Susceptibility of Aedes albopictus to infection with Eastern equine encephalomyelitis virus. J. Am. Mosq. Control Assoc. 6: 274-278. Scott, T. W., S. A. Wright, B. F. Eldridge, and D. A. Brown. 2001. Cost effectiveness of three arbovirus surveillance methods in northern California. J. Am. Mosq. Control Assoc. 17: 118-123. Senne, D. A., J. C. Pedersen, D. L. Hutto, D. J. Taylor, B. J. Schmitt, and B. Panigrahy. 2000. Pathogenicity of West Nile virus in chickens. Avian Dis. 44: 642-649. Service, M. W. 1976. Mosquito Ecology: Field Sampling Methods. John Wiley and Sons., New York. Shaman, J., J. F. Day, and M. Stieglitz. 2002. Drought-induced amplification of St. Louis encephalitis virus, Florida. Emerg. Infect. Dis. 8: 575-580. Shaman, J., J. F. Day, and M. Stieglitz. 2003a. St. Louis encephalitis vi rus in wild birds during the 1990 South Florida epidemic: The importance of drought, wetting conditions, and the emergence of Culex nigripalpus (Diptera: Culicidae) to arboviral amplification and transmission. J. Med. Entomol. 40: 547-554. Shaman, J., J. F. Day, and M. Stieglitz. 2004a. The spatial-temporal distribution of drought, wetting, and human cases of St. Louis encephali tis in South-central Florida. Am. J. Trop. Med. Hyg. 71: 251-261. Shaman, J., J. F. Day, and M. Stieglitz. 2005. Drought-induced amplification and epidemic transmission of West Nile Virus in southern Florida. J. Med. Entomol. 42: 134-141. Shaman, J., J. F. Day, M. Stieglitz, S. Zebiak, and M. Cane. 2004b. Seasonal forecast of St. Louis encephalitis virus transmission, Fl orida. Emerg. Infect. Dis. 10: 802-809.

PAGE 163

163 Shaman, J., M. Stieglitz, S. Zebiak, and M. Cane. 2003b. A local forecast of land surface wetness conditions derived from seasonal cl imate predictions. J. Hydromet. 4: 611-626. Shroyer, D. A. 1991. The 1990 Florida epidemic of St. Louis encephalitis: virus infection rates in Culex nigripalpus J. Fla. Mosq. Control Assoc. 62: 69-71. Slaff, M., and J. D. Haefner. 1985. Seasonal and spatial distribution of Mansonia dyari Mansonia titillans and Coquillettidia perturbans (Diptera: Culicidae) in the central Florida, USA, phosphate region. J. Med. Entomol. 22: 624-629. Sooter, C. A., M. Schaeffer, R. Gorrie, and A. T. Cockburn. 1954. Transovarian passage of antibodies following naturally acquired encephalitis infection in birds. J. Infect. Dis. 95: 165-167. Stamm, D. D., and R. J. Newman. 1963. Evidence of southward transport of arboviruses from the United States by migratory bird s. An. Microbiologia 11: 123-133. Stamm, D. D., R. W. Chamberlain, and W. D. Sudia. 1962. Arbovirus studies in south Alabama, 1957-1958. Am. J. Hyg. 76: 6181. Stark, L. M. and D. Kazanis. 2005. Florida Department of Hea lth, Bureau of Laboratories, Tampa: Arbovirus Surveillance: Annual Summary Report 2005. Sudia, W. D., and R. W. Chamberlain. 1964. Experimental infection of Culex nigripalpus Theobald with the virus of St. Louis en cephalitis. Am. J. Trop. Med. Hyg. 13: 469-471. Sudia, W. D., R. D. Lord, and R. O. Hayes. 1970. Collection and processing of vertebrate specimens for arbovirus studies, pp. 65, Atlanta, GA. Sudia, W. D., V. F. Newhouse, and W. A. Chappell. 1969. Venezuelan equine encephalitis virus-vector studies followi ng a human case in Dade Count y, Florida. Mosq. News 29: 596-600. Taylor, D. J., K. E. Meadows, N. J. Schneid er, J. A. Mulrennan, and E. E. Buff. 1969. St. Louis encephalitis virus and Culex nigripalpus in Florida. Florida State Board of Health Monograph "Saint Louis Encepha litis in Florida" 12: 34-45. Tiawsirisup, S., K. B. Platt, B. J. Tucker, and W. A. Rowley. 2005. Eastern cottontail rabbits ( Sylvilagus floridanus ) develop West Nile virus viremias sufficient for infecting select mosquito species. Vector Borne Zoonotic Dis. 5: 342-350. Trainer, D. O. 1970. The use of wildlife to monitor zoonoses. J. Wild. Dis. 6: 397-401. Trevejo, R. T., and W. C. Reeves. 2005. Antibody response to Culex tarsalis salivary gland antigens among sentinel chic kens in California. Am. J. Trop. Med. Hyg. 72: 481-487.

PAGE 164

164 Trock, S. C., B. J. Meade, A. L. Glaser, E. N. Ostlund, R. S. Lanciotti, B. C. Cropp, V. L. Kulasekera, L. D. Kramer, and N. Komar. 2001. West Nile virus outbreak among horses in New York state, 1999 a nd 2000. Emerg. Infect. Dis. 7: 745-747. Tsai, T. F., and C. J. Mitchell. 1989. St. Louis encephalitis, pp. 113-143. In T. P. Monath (ed.), The Arboviruses: Epidemiology and Ecology. CRC Press, Boca Raton, FL. Tully, T. N., S. M. Shane, R. P. Poston, J. J. England, C. C. Vice, D. Y. Cho, and B. Panigrahy. 1992. Eastern equine encephalitis in a flock of emus ( Dromaius novaehollandiae ). Avian Dis. 36: 808-812. Turell, M. J., D. J. Dohm, M. R. Sardelis, M. L. O Guinn, T. G. Andreadis, and J. A. Blow. 2005. An update on the potential of North Ameri can mosquitoes (Diptera : Culicidae) to transmit West Nile virus. J. Med. Entomol. 42: 57-62. Tyzzer, E. E., A. W. Sellards, and B. L. Bennett. 1938. The occurrence in nature of "Equine encephalomyelitis" in the ring-n ecked pheasant. Science 88: 505-506. University of Florida Department of Physics. 2005. Department of Physics weather station. http://www.phys.ufl.edu/weather/ Vaidyanathan, R., J. D. Edman, L. A. Cooper, and T. W. Scott. 1997. Vector competence of mosquitoes (Diptera: Culicidae) from Massach usetts for a sympatric isolate of Eastern equine encephalomyelitis virus. J. Med. Entomol. 34: 346-352. Ventura, A. K., and J. N. Ehrenhranz. 1975. Detection of Venezuel an equine encephalitis virus in rural communities of Southern Florid a by exposure of sentinel hamsters. Am. J. Trop. Med. Hyg. 24: 715-717. Villari, P., A. Spielman, N. Komar, M. McDowell, and R. J. Timperi. 1995. The economic burden imposed by a residual case of Easter n encephalitis. Am. J. Trop. Med. Hyg. 52: 813. Waage, J. K., and J. Nondo. 1982. Host behavior and mosquito feeding success an experimental study. Tran. R. Soc. Trop. Med. Hyg. 76: 119-122. Ward, M. R., D. E. Stallknecht, J. Willis, M. J. Conroy, and W. R. Davidson. 2006. Wild bird mortality and West Nile virus surveilla nce: biases associated with detection, reporting, and carcass persistan ce. J. Wild. Dis. 42: 92-106. Watson, J. T., R. C. Jones, K. Gibbs, and W. Paul. 2004a. Dead crow reports and location of human West Nile virus cases, Chi cago, 2002. Emerg. Infect Dis. 10: 938-940. Watson, J. T., P. E. Pertel, R. C. Jones, A. M. Siston, W. S. Paul, C. C. Austin, and S. I. Gerber. 2004b. Clinical characteristics and functiona l outcomes of West Nile fever. Ann. Intern. Med. 141: 360-365.

PAGE 165

165 Webber, L. A., and J. D. Edman. 1972. Anti-mosquito behavior of ciconiiform birds. Anim. Behav. 20: 228-232. Wellings, F. M., A. L. Lewis, and L. V. Pierce. 1972. Agents encountered during arboviral ecological studies: Tampa Bay area, Flor ida, 1963-1970. Am. J. Trop. Med. Hyg. 21: 201213. Williams, J. E., D. M. Watts, O. P. Young, and T. J. Reed. 1972. Transmission of Eastern (EEE) and Western (WEE) encephalitis to Bobw hite sentinels in re lation to density of Culiseta melanura mosquitoes. Mosq. News 32: 188-192. Wright, R. E., and K. L. Knight. 1966. Effect of environmental factors on biting activity of Aedes vexans (Meigen) and Aedes trivittatus (Coquillett). Mosq. News 26: 565-578. Zyzak, M., T. Loyless, S. Cope, M. Wooster, and J. F. Day. 2002. Seasonal abundance of Culex nigripalpus Theobald and Culex salinarius Coquillett in north Florida, USA. J. Vector Ecol. 27: 155-162.

PAGE 166

166 BIOGRAPHICAL SKETCH Kevin Conrad Kobylinski was born on May 9, 1981 in Frederick, Maryland, to Rae Ann Kobylinski and Edmund Arthur Kobylinski. He has two younger brothers, Patrick and William. At an early age Kevin developed a fondness of the insect kingdom, collecting locust shells and keeping spiders as pets. He l oved his grandfathers cabin in ru ral Missouri and was quite fond of walks through the woods and observing the natural world. Shortly after his eighteenth birthday Kevin left home for college at Kansas State University. He began his college career in En gineering but was not satisfied with this career path. His calling was found after enrolling in Dr. Greg Zolnerowichs introductory course, Insects and People. By the end of that semest er he had switched his degree to Wildlife Biology with a minor in Entomology. Kevin excelled in his Entomology coursework and took every class available. He volunteered to work in the Insect Zoo operated by Dr. Ralph Charlton. During the summers Kevin worked for Drs. Albe rto Broce, Thomas Janousek, and Ludek Zurek on the West Nile Virus Surveilla nce Project. This introduced Kevin to his current passion, Medical Entomology and Vector Biology. Kevin is happy working in this field because he knows that the work he does is used to benefit the world by helping people.