The relationship between palpal morphology and host-seeking behavior in adult mosquitoes (Diptera: culicidae), especiall...

MISSING IMAGE

Material Information

Title:
The relationship between palpal morphology and host-seeking behavior in adult mosquitoes (Diptera: culicidae), especially Culiseta melanura (Coq.)
Physical Description:
ix, 302 leaves : ill. ; 29 cm.
Language:
English
Creator:
Choate, Paul Merrill, 1948-
Publication Date:

Subjects

Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1989.
Bibliography:
Includes bibliographical references (leaves 242-301).
Statement of Responsibility:
by Paul Merrill Choate, Jr.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001558333
notis - AHH1977
oclc - 22608266
System ID:
AA00003764:00001

Full Text













THE RELATIONSHIP BETWEEN PALPAL MORPHOLOGY AND HOST-
SEEKING BEHAVIOR IN ADULT MOSQUITOES (DIPTERA: CULICIDAE),
ESPECIALLY CULISETA MELANURA (COQ.)




















By

PAUL MERRILL CHOATE, JR.


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

UNIVERSITY OF FLORIDA


1989
















ACKNOWLEDGEMENTS

Many people have directly or indirectly influenced the

direction and completion of this dissertation. Dr. Tom Walker

first introduced me to the complexities of insect ecology.

Dr. Reece Sailer (deceased) agreed to chair my earlier

master's degree research in taxonomy of Coleoptera. Dr. P.

J. Darlington (deceased) offered encouragement during my

studies of Carabidae. He was always quick to provide

literature and advice. Dr. George Ball (Alberta) has

continued to prod me to completion of a graduate degree,

although both he and I at one time envisioned that work in

carabid taxonomy would lead to a graduate degree in

entomology. Leon Croizat openly corresponded with me during

a time when I was wandering between total abandonment of a

career in entomology and pursuit of a graduate degree. Fellow

coleopterists have always provided support and constructive

criticism. Many enjoyable days have been spent in the field

with good friends and colleagues. Among these are Scott

Gross, Lloyd Davis, Robert Davidson, Don Wilson, Dave Brozska,

and Mike Thomas.

Although our careers have diverged in different

directions, I owe Dr. Robert Woodruff many thanks for his









encouragement from the time of my arrival at the University

of Florida. Despite our differences of opinion concerning

personal priorities and areas of research, I still owe him a

debt of gratitude. I also thank Dr. Howard Weems for his

personal and financial support for publications.

No acknowledgements are complete without thanking the

people personally responsible for providing financial and

logistical support. Dr. Donald W. Hall graciously accepted

me back into the graduate student world, agreeing to support

me both as cochairman and research advisor. Dr. Dan Kline

(USDA) provided an assistantship and found the means to keep

me headed in the right direction in spite of my affinity for

wandering. Dr. Howard Frank has enthusiastically encouraged

me to complete my graduate degree program. His continued

interest and investment of energy have been important. Dr.

Dave Carlson (USDA) graciously provided space and equipment

for my gas chromatography research.

Research projects would not succeed without the

availability of study sites. Mr. John Whitehead has

generously provided unlimited access to his Lake Butler farm.

A special thanks is given to him and his family for allowing

my unannounced visits to his farm.

Dr. J. Paul Gibbs introduced me to vector surveillance

and its related problems while working on Eastern Equine

Encephalomyelitis. The many unknown aspects of this disease

led me to work on mosquitoes and host preference.


iii









Jimmy Becnel has provided SEM expertise, darkroom

facilities, company in the field, and listened patiently to

the ideas presented here. Likewise, Bill Oldacre helped me

to understand the physics behind detection of signals, energy,

and all things mathematical of which we tend to be

conveniently ignorant. Julio Hector, chemist, computer

programmer, and friend, often worked late to help me finish

data analysis and produce the chromatograms that are used

here. Without his expertise and self sacrifice, this project

would never have been completed within the necessary time

frame. Genie Avery gave willingly of her time and computer

expertise, helping to sort and arrange this lengthy document.

Hank McKeithen willingly permitted me the use of his computer

during many unannounced visits.

Family members suffer the most and gain the least during

a graduate program. My parents have patiently waited to see

me finish, despite many indications to the contrary. I thank

them for their many years of love and support.

Our daughter Teresa had to withstand the pressure of a

graduate student/father. It wasn't easy having a tired

graduate student to come home to. Finally, my wife Angela

provided the tolerance, support, forgiveness, and love that

are vital to the success of a graduate student.


















TABLE OF CONTENTS


ACKNOWLEDGEMENTS ...................................ii

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

CHAPTERS

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

2. BIOLOGY AND SEASONAL ABUNDANCE OF CULISETA
MELANURA AT LAKE BUTLER, FLORIDA .........19

General Biology .......................... 19
Seasonal Abundance .......................23

3. VARIATION IN CUTICULAR HYDROCARBON PROFILES
IN CULISETA MELANURA .....................38

Introduction ............................ 38
Methods and Materials ....................41
Results ...................................45
Conclusions ...............................81
Discussion ................................82

4. HOST PREFERENCE IN MOSQUITOES, A REVIEW ...84

Introduction ..............................84
Literature Review ........................85
Host Preference in Culiseta melanura and
Cs. inornata ...........................106
Discussion ...............................107

5. MORPHOLOGY OF SENSORY STRUCTURES .........108

Introduction .............................108
Methods and materials ................... 128
Results ..................................128










6. LARVAL REARING TEMPERATURE AND ITS EFFECT
ON ADULT MOSQUITO MORPHOLOGY, ESPECIALLY
PALPAL OLFACTORY STRUCTURES .............185

Introduction ............................ 185
Methods and materials ...................196
Results ..................................199
Conclusions ..............................216
Discussion ...............................217

7. QUANTIFICATION OF HOST EMANATIONS, TYPES,
AND THE IMPLICATION FOR ATTRACTION OF
ADULT MOSQUITOES .........................220

Introduction ...........................220
Host emanations (especially CO2)-quantity
and detection .......................... 225
Quantification of host emanations, espec-
ially CO2 ................................227
Calculation of respired volumes of gases
and CO2 .................. ............. 229
Calculation of body surface area for var-
ious animals ........................... 232

8. DISCUSSION, SUMMARY, PROPOSAL FOR FUTURE
RESEARCH .................................236

Discussion and summary ..................236
Proposal for future research ............240

REFERENCES ........................................242

BIOGRAPHICAL SKETCH ............................... 302















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

THE RELATIONSHIP BETWEEN PALPAL MORPHOLOGY AND HOST-SEEKING
BEHAVIOR IN ADULT MOSQUITOES (DIPTERA: CULICIDAE), ESPECIALLY
CULISETA MELANURA (COQ.)

By

Paul Merrill Choate, Jr.

August, 1989

Chairman: Donald W. Hall
Major Department: Entomology and Nematology


Analysis of cuticular hydrocarbon profiles of adult

Culiseta melanura (Coquillett) from New Hampshire, Vermont

(new record), and Florida showed no evidence of cryptic

species. Variation (within site) was as great as variation

between sites. Cuticular hydrocarbon profiles are also

illustrated for Culex territans Walker, Uranotaenia

sapphirina (Osten Sacken), and Cs. melanura from New

Hampshire. Percent composition of hydrocarbons showed

seasonal variation in Cs. melanura from Lake Butler, Florida.

Larval rearing temperature is one factor contributing to

variation in hydrocarbon profiles.

Larval rearing temperatures produced intra-specific

variation in body size, palpal surface area, and the number

of sensory structures within Ae. aegypti (Linnaeus) and Cs.


vii









melanura. Palpal configurations and surface area are

presented for Cs. melanura, Cs. inornata Williston, Culex

territans, Cx. restuans Theobald, Cx. quinquefasciatus Say,

Cx. salinarius Coquillett, Cx. nigripalpus Theobald, Cx.

tarsalis Coquillett, Uranotaenia sapphirina, Psorophora ferox

(Humboldt), Ps. ciliata (Fabricius), Aedes aegypti, Ae.

albopictus (Skuse), Ae. canadensis Theobald, Ae. vexans

(Meigen), Ae. triseriatus (Say), Ae. infirmatus Dyar & Knab,

Ae. atlanticus Dyar & Knab, Ae. sticticus (Meigen), Ae.

mitchellae Dyar, Ae. sollicitans (Walker), Ae.

taeniorhynchus (Weidemann), and Orthopodomyia signifera

(Coquillett).

A correlation exists between sensory capabilities of

mosquitoes and the size range of preferred hosts as defined

by host body weight and surface area. Estimates are given

for respired volumes of gases and surface area for selected

animals. Sensory capabilities are defined as palpal surface

area containing olfactory structures/number of sensory

structures per palp, compared to the volume of air sampled

between palpi during flight. Evidence for a morphological

basis for host preference and host preference shift is

presented. Host preference is proposed to be a function of the

volume of host emanations of common compounds, rather than

specific odors. Variation in host preference is correlated

with seasonal change in adult mosquito structure. A

hypothetical cycle involving host preference shift is


viii









presented to demonstrate the possibility that a single species

such as Cs. melanura may be capable of vectoring diseases such

as Eastern Equine Encephalomyelitis from birds to mammals.















CHAPTER 1

INTRODUCTION

The structures involved in host seeking are only now

beginning to be recognized. The chemical aspects of host

detection and location are not at all understood. Preference

for different hosts is suggested by bloodmeal analysis, but

the criteria for one host being preferred remain unclear.

Females of most mosquito species will seek a blood meal

at some time during their adult life. Autogenous species may

postpone blood meals until the second gonotrophic cycle

(O'Meara, 1985). The presence or absence of autogeny within

a population may in part be due to larval nutrition and

density (Lounibos et al., 1982). Various factors such as

chemical attraction, possible host specificity, physiological

age, host availability, and environmental conditions, interact

in various combinations to accomplish host seeking and a

successful blood meal. Mosquito species vary in nutritional

requirements for egg development (Woke, 1937). The genus

Toxorhynchites has females which never take a blood meal.

Mosquitoes will feed on sugar in addition to blood, using

nectar from flowers as the sugar source (Bidlingmayer & Hem,

1973; Magnarelli, 1977b, 1980; Nasci & Edman, 1984; Nayar

& Sauerman, 1975; Philip, 1941; and Van Handel, 1972, 1985).









2

Seasonal distribution, habitat preference, and host

behavior influence host seeking and behavior. It is this

host seeking behavior that adult mosquito surveillance

attempts to exploit.

Numerous references in the literature document species

composition of mosquito collections. Many different

collecting techniques are employed, including light traps,

CO2 baited traps, animal baited traps, resting boxes, and ramp

traps (Bast & Rehn, 1963; Bellamy & Reeves, 1952; Breeland

& Pickard, 1965; Bidlingmayer & Hem, 1981; Davis, 1978; Edman

et al., 1968; Furlow & Young, 1970; Gui et al., 1942;

Gunstream & Chew, 1967; Harden et al., 1970; Harrison et al.,

1982; Hauff & Burgess, 1960; Huffaker & Back, 1943; Kinzer

et al., 1978; Meyer, 1977; Minson et al., 1970; Provost,

1959; Schreck et al., 1972; Service, 1976; Stryker & Young,

1970; Vickery et al., 1966; and Villavaso & Steelman, 1970).

Reports in these publications reveal that no single trap type

collects all species, and therefore one collection technique

will not adequately sample the mosquito fauna of a given area.

Adult surveillance assumes those species present will be

collected. During discussion of New Jersey traps for sampling

mosquito populations, Huffaker & Bach (1943) stated the

following:

It has been assumed by the great majority of
mosquito control workers using the trap that,
because of the very nature of the method, the
various species are caught in numbers proportional
to their respective occurrences in the heterogeneous
mosquito population.











This assumption is based upon fallacious grounds
the mosquitoes possess all degrees of variation
particular to common living things it would
be hard to see why it is not wholly logical to
expect that different species of mosquitoes will
exhibit marked variations in their responses .
It is, furthermore, a fundamental biological
principle that ecological variation is not a
respecter of phylogenetic ties. Hence, it may be
expected that marked differences in behavior exist
even within a genus.(p.561)

The same may be said for all current sampling techniques.

Additional "ground truthing" by larval sampling will usually

reveal the presence of other species that for various reasons

are not being collected in the traps being used. Recognition

of these difficulties results in the following observation.

We really do not understand the complex actions involved in

host seeking by even one species of mosquito.

Before we are able to analyze the behavior of a

particular species, correct recognition (identification) of

that species is essential. Unfortunately, adult mosquitoes

are very fragile insects, easily dismembered and denuded of

the very structures used to identify them. Personal

experience has revealed an alarming incidence of

misidentification of specimens used in reports of surveillance

collections. In the absence of voucher specimens, the

researcher has to consider the possibility of

misidentifications when interpreting reports in the

literature.

Several available techniques are currently being used for

species recognition. All named forms of mosquitoes for the









4

entire world are listed in Knight and Stone, 1977. Knight

(1974) reviewed the history of mosquito taxonomy in the United

States. Identification manuals for North American mosquitoes

include Darsie & Ward (1981) and Carpenter & LaCasse (1955).

Southeastern United States is covered in King et al., 1960.

These manuals include morphological keys to adults and larvae,

and are the standard identification manuals for North American

mosquito workers (Zavortink, 1974). Keys to the pupae of some

mosquitoes were presented in Tinker and Stojanovich (1962).

Additional techniques are being used to study intra-specific

variation more closely, sibling species, and species complexes

in insects. Among these are chromosomal cytogenetic studies

(Kitzmiller et al., 1967; Rao & Rai, 1987), amino acid

identification (Ball, 1952; Ball & Clark, 1953; Micks &

Ellis, 1951; Micks et al., 1966) cuticular hydrocarbon

analyses (Castner & Nation, 1984; Carlson, 1982, 1983; Carlson

& Bolton, 1984; Carlson & Walsh, 1981; Carlson & Service,

1979, 1980; Carlson & Yocom, 1986; and Milligan et al., 1986),

electrophoretic comparisons (Ayala and Powell, 1972; Makela

& Richardson, 1977),and comparison of sensory structure

arrangement on antennae in Anopheles (Ismail & Hammond, 1968).

Integration of these techniques should provide the taxonomic

detail necessary for analysis of behavior of a particular

species (Barr, 1974; Eldridge, 1974; Faran, 1979; Rogers,

1974). Interpretation of behavioral data is done with more

confidence when specimens are identified accurately.









5

Host seeking by female mosquitoes is in part dependent

upon the physiological age of the insect. Various techniques

for the determination of the physiological and chronological

age (age grading) of insects have been employed (Biscoe-

Tyndal, 1984; Christophers, 1911; Corbet, 1960, 1962;

Detinova, 1962, 1968; Hitchcock, 1968; Johnston & Ellison,

1982; Lewis, 1958; Magnarelli, 1976; Magnarelli & Anderson,

1981; Magnarelli et al., 1984; Mullens & Schmidtmann, 1982;

Rosay, 1961; Schlein, 1979; and Schlein & Gratz, 1972). If

it is possible to determine the approximate physiological age

of adult mosquitoes, then it may become possible to make

predictions concerning host seeking and population dynamics.

Field observations on the parity of biting flies have included

studies by Scholl et al., 1979; Samarawickrema, 1968; and

Magnarelli et al., 1984. Age grading of mosquitoes has been

done by ovarian examination. Mer (1936) found that ovarian

dissection could be used to recognize prehibernation females

of Anopheles elutus Edwards. Corbet (1960) determined that

adult mosquitoes parasitized by water mites were nulliparous

98% of the time. Lanciani (1979a, 1979b, 1986), Lanciani &

Boyett (1980), and Lanciani & Boyt (1977) determined that

water mite parasitism affected adult longevity.

Each physiological stage of an adult female (nulliparous,

parous, gravid, bloodfed) elicits a behavioral shift that

greatly influences (bias) interpretation of any one trapping

technique. Host-seeking females are generally those











mosquitoes that are collected in baited traps. Resting boxes,

truck traps, ramp traps, and malaise traps are assumed to be

less biased in their collections, but even these may be shown

to attract disproportionate numbers of certain physiological

states of adults. Therefore, several techniques must be

employed simultaneously, and each collection compared. Then

a more representative sample of the age structure of a given

population will be achieved.

Since most adult females will at some time seek a blood

meal, the type of blood meal sought may help explain the

functional significance of different sensory structures known

to exist on the mouthparts and antennae of biting flies,

including mosquitoes. Bloodmeal analysis reveals an apparent

wide host range. Host records range from amphibians and

reptiles to birds, fish, and mammals, including man. While a

complete review of bloodmeal records is not the intention of

this paper, the following citations will give the reader a

representation of the variety of host preference records, as

well as some of the discrepancies regarding the application

of the term host preference. Articles dealing with bloodmeal

analysis include the following: Anderson, 1967; Beir et al.,

1988; Bertsch & Norment, 1983; Boorman, 1961; Boreham & Snow,

1973; Brown, 1966; Burkot & DeFoliart, 1982; Chandler et al.,

1975; Christopher & Reuben, 1971; Crans, 1964, 1965, 1970;

Crans & Rockel, 1968; Davis, 1940; Dow et al., 1957; Downe,

1960, 1962, 1963; Edman, 1971, 1974, 1979a, 1979b; Edman &









7

Bidlingmayer, 1969; Edman & Downe, 1964; Edman & Haeger,

1977; Edman et al., 1972; Gunstream et al., 1971; Hayes,

1961; Hayes et al., 1973; Hopla, 1965; Irby & Apperson,

1988; Laarman, 1955, 1958; LeDuc et al., 1972; Magnarelli,

1977a; McClelland & Weitz, 1963; McIver, 1968; Means, 1968;

Murphey et al., 1967; Nasci, 1982b, 1984, 1985, 1986b; Nasci

& Edman, 1981a; Nolan et al., 1965; Rempel et al., 1946;

Schaefer & Steelman, 1969; Shalaby, 1969; Shemanchuk, 1969;

Smith & Weitz, 1959; Snow & Boreham, 1973; Suyemoto et al.,

1973; Takahashi & Shimizu, 1971; Tempelis, 1970, 1975;

Tempelis et al., 1967, 1970; Washino & Tempelis, 1983; Woke,

1937; and Wright & DeFoliart, 1970. Some species appear

restricted to cold-blooded vertebrates (Crans, 1970). Others

use a wide range of hosts, exhibiting a marked seasonal host

shift (Edman & Taylor, 1968). Culiseta melanura is considered

to be an avian feeder, but records exist for other hosts,

including man and horses (Edman et al., 1972; Hayes & Doane,

1958; Joseph & Bickley, 1969; Moussa et al., 1966; and

Schober, 1964). iMuch speculation exists as to the nature of

attractiveness of different hosts. The recognition of a

particular vertebrate species by a mosquito is a function of

the combined senses of olfaction, thermal reception, and

visual cues (Altner & Prillinger, 1980; Sippell & Brown,

1953). We are currently unable to state with certainty which

of these factors are most important at any particular

distance. Only a few of the many structures on the surface











of the adult mosquito have been studied. Among these are the

carbon dioxide receptors on the palpi, and thermal receptors

and other olfactory structures on the antennae. Their

chemical thresholds have not been determined, but indirect

evidence suggests that the numbers of these structures are

somehow related to the amount of chemical odor detectable by

the individual mosquito (Chapman, 1971). Variation in

perceived host preference has been considered by some to be

a reflection of host availability. However, for the present

I will consider host preference a real phenomenon, and will

examine the implication of such a phenomenon. If in fact host

preference exists, then the ability to discriminate between

hosts must exist. If this ability to discriminate exists, it

should be revealed in some variation of sensory structures

used in host seeking and location. What are the stimuli

believed to be important in host seeking? How and where are

these received?

In order to analyze seasonal host preference records for

any species of mosquito, seasonal variation in adult behavior

and morphology should be considered. Smith (1961) noted that

the site of resting places for An. gambiae Giles and An.

pharoensis Theobald were seasonally different. Michener

(1945) observed that in mosquitoes the winter individuals of

a species are often larger than the summer ones, as well as

being differently colored. He noted that overwintering forms

of An. maculipennis freeborni Aitken were larger and darker









9

than the summer ones. The same appeared true for An.

quadrimaculatus Say. Even more striking were the seasonal

morphs of Culex apicalis (= territans) and Cx. nigripalpus.

The seasonal forms of these species were sufficiently

different to pose a problem in correct identification.

Seasonal variation in size has also been reported in Cx.

tarsalis (Boch & Milby, 1981), with the smallest individuals

occurring in the hottest months. Associated with reduced size

in Cx. tarsalis was a decrease in attractiveness to traps of

the same type that were attractive during the cooler months

when larger individuals were present.

Fish (1985) noted significant variability in the size of

biting females of many vector and pest species of mosquitoes.

Citing Takahashi (1976), Baqar et al. (1980), and Grimstad &

Haramis (1984), Fish stated that it was becoming apparent that

vector competence can be affected by the size of individuals

comprising a population. He concluded that an analysis of

size variation for natural populations of vector species

should be a first step in the study of vectorial capacity in

mosquitoes. Haramis (1985) noted that larval nutrition had

a direct effect on fitness of adult mosquitoes, with small

adults having reduced survival and fecundity, but transmitting

LaCrosse virus more effectively than large mosquitoes.

Landry et al. (1988) noted significant seasonal variation

in the size of adult female Aedes triseriatus. Craig and

Vandehey (1962) noted that rearing temperatures affected









10

certain color mutations in Ae. aevypti. Zuska and Berg (1974)

documented seasonal temperature as one of the primary factors

affecting color variation in South American Tetanoceroides

(Diptera: Sciomyzidae). Rearing temperature affected the rate

of loss of general characters in Ae. nigromaculis and Culex

pipiens quinquefasciatus, with rotation of male genitalia

occurring 12 hours after emergence at 280C, but not for 51

hours at 17C (Rosay, 1961).

The ability of mosquitoes to become infected with and

transmit viruses has also been shown to be temperature-

related (Turell et al., 1985). Since the size of individual

mosquitoes is related to the seasonal temperatures, and the

ability to transmit also is temperature-related, it follows

that viral activity would be predicted to be highest during

the summer months. This is known to be true for many of the

encephalitis viruses (Hess et al., 1963). Grimstad (1983)

proposed a reduced gut barrier to viral particles in smaller

individuals as the mechanism behind higher infection rates in

smaller individuals. These same smaller individuals also were

shown to take proportionately more infectious viral particles

in a bloodmeal. Included among viral diseases transmitted by

mosquitoes is eastern equine encephalomyelitis (EEE). This

virus is believed to involve an enzootic cycle maintained

within avian hosts by one or more species of aviphilic

mosquitoes, then somehow transmitted to larger mammalian hosts

such as man and horses, where fatalities are frequent. The











mechanism behind transmission to dead-end hosts is as yet

unknown. The primary enzootic vector is believed to be Cs.

melanura, although approximately twenty species of insects

have had the virus of EEE isolated from them.

The study site for this research was 3 km east of Lake

Butler, Florida,in Union Co., on the property of John

Whitehead. This farm was chosen because of 2 confirmed cases

of EEE that occurred in 1985 and 1986. Field studies had

demonstrated a stable population of Cs. melanura on the

property. A wide variety of animals, including a herd of

cattle, horses, dogs, pigs, guinea fowl, deer, turkey and

rabbits was found within the confines of the property. Three

chicken houses contained 75,000 chickens during part of the

year. Large open tracts of pasture were surrounded by pine

woods, red-maple and gum swamps, and river flood plain.

Preliminary data and personal field observations

indicated the presence of a larger, darker winter form of Cs.

melanura which was gradually replaced by a lighter, smaller

summer form. Adult seasonal variation in size and

susceptibility to C02-baited CDC miniature light traps were

the main catalysts behind the research. If vector

surveillance for mosquitoes is so dependent upon collecting

particular species, the mechanisms behind a behavioral shift,

either in terms of host preference, or in terms of

attractiveness to traps, need to be investigated and

understood.











My original purpose for working at this study site was

to collect adult female mosquitoes for virus isolation.

Presumably those mosquitoes transmitting EEE to horses would

be collected and identified as the vector(s) by virus

isolation. Therefore, the first object of this research was

to determine the mosquito species that occurred at the study

site. A preliminary year of collection data confirmed the

existence of approximately 30 species of mosquitoes. These

included: An. crucians, An. punctipennis, An.

quadrimaculatus, Toxorhynchites rutilus, Ae. aegypti, Ae.

triseriatus, Ae. hendersoni, Ae. sticticus, Ae. mitchellae,

Ae. dupreei, Ae. atlanticus, Ae. infirmatus, Ae. vexans, Ae.

canadensis, Ae. fulvus pallens, Psorophora howardii, Ps.

ciliata, Ps. columbiae, Ps. ferox, Cx. quinquefasciatus, Cx.

salinarius, Cx. nigripalpus, Cx. pilosus, Cx. territans, Cx.

restuans, Cx. erraticus, Cs. melanura, Cs. inornata,

Uranotaenia sapphirina, Ur. lowii, Orthopodomyia signifera,

and Coquillettidia perturbans.

Of these, only eight were definitely present throughout

the entire year. These were An. crucians, An. punctipennis,

An. quadrimaculatus, Cx. quinquefasciatus, Cx. salinarius, Cx.

nigripalpus, Cx. territans, and Cs. melanura. Additionally,

Cx. erraticus, Aedes aeqypti, Orthopodomyia siqnifera, and

Aedes triseriatus may have been present all year. Since both

Aedes species and Orthopodomyia are seldom collected as

adults, their presence throughout the year was not confirmed.









13

Sampling techniques for Cs. melanura consisted of larval

collections to verify breeding, adult collections from resting

boxes (Edman et al., 1968) and CDC miniature light traps

baited with dry ice as a CO2 source. Resting box collections

were made on the same days that CDC collections were made.

Unfortunately, at the end of the seventh month of this

research the swamp being used for collections was clear-cut

for eventual planting of pine. As a result, collections that

previously were being made in wooded areas became open area

collections. I chose to leave traps and resting boxes in the

same locations for the remainder of the year. The data for

these collections are only used here to confirm the presence

of adult Cs. melanura throughout the year. No statistical

analysis of the collection data is possible due to the

complete disruption of the study site for one-third of the

collection period.

Many investigations into the occurrence of EEE have

correlated the presence of Cs. melanura with viral activity

(Bryant et al., 1973; Chamberlain, 1958; Chamberlain et al.,

1951, 1954; Dardiri et al., 1957; Dougherty & Price, 1960;

Feemster, 1938, 1957; Ferguson, 1954; Goldfield & Sussman,

1968; Grady et al., 1978; Grimstad, 1983; Hayes et al.,

1962; Jaynes et al., 1962; Kelser, 1933; LeDuc et al.,

1972, 1975; Main et al., 1979; Muul et al., 1975; Oglesby,

1948; Srihongse et al., 1980; Stamm et al., 1962; Sudia et

al., 1968; Wallis, 1959; Wallis et al., 1958, 1974; and









14

Williams et al., 1971, 1972, 1974). Some investigators feel

that the presence of this species is necessary for the virus

to cycle within resident bird populations, to later somehow

be transmitted to other hosts. One measure for estimating

potential viral activity involves monitoring populations of

Cs. melanura. I decided to monitor the seasonal occurrence

of this species at the Lake Butler site. To insure that

behavioral and structural comparison were reported accurately,

one of the first questions to be answered was what insect

species was actually being studied? A necessary part of this

study was to insure that in fact Cs. melanura was, by all

available parameters, the Cs. melanura described in the

literature. With this in mind, I obtained collections from

the type locality (New Hampshire) for comparison. No

morphological differences were apparent between New Hampshire

and Florida specimens. Since evidence exists for cryptic

species within other mosquito species complexes, I felt that

an additional comparison would reinforce my belief that only

one species was in fact involved here. Cuticular hydrocarbons

were chosen for further comparisons. These comparisons are

presented in Chapter 3. As a result of these comparisons,

several questions arose concerning variation in the adult

hydrocarbon patterns? Was this variation correlated with the

two seasonal morphs in the wild populations of Cs. melanura

at Lake Butler? If seasonal shifts in hydrocarbon patterns

occur, how might this affect the use of this technique to









15

solve taxonomic problems? How much of a difference would be

necessary to correctly interpret species differences?

Cuticular hydrocarbons constitute a part of the chemical

makeup of individual insects. Chemical composition within a

species may be environmentally changed (Bryant, 1974). I felt

that if the chemical identity of a species could be

environmentally shifted the possibility certainly must exist

for behavioral changes also being influenced by environmental

parameters. Before conducting experimental analysis of the

effect of one environmental parameter (temperature) on adult

mosquito olfactory structures, I reviewed host preference

records to search for suggestion of seasonal differences.

Host preference records are numerous in the literature.

However, most reports are based on pooled samples collected

over several years from several locations. While these reports

document host preference variation, they seldom discuss

seasonal host preference variation. By reanalyzing the data

from several different authors, trends in host preference

began to emerge, and seasonal patterns were suggested. These

patterns are discussed in Chapter 4.

The morphology of sensory structures in mosquitoes

reveals much interspecific variation. Do structural

differences affect the ability to select certain hosts? Could

the morphological variation be correlated with seasonal

behavioral changes? Could host detection capabilities be

correlated with morphology? If morphology could be correlated









16

with host preference, what effect might a change in morphology

have on host preference? What might be a practical method to

measure sensory capabilities?

Based on the possibility of morphological variation

within host detection sensory structures, I decided to look

at the variation within the overall configuration of these

same structures in several species. The actual structures

involved in host detection are illustrated and discussed in

detail to clarify the shape and configurations that occur in

mosquitoes. A thorough understanding of the structures

involved is necessary to understand host preference.

Interspecific variation in these structures is discussed in

Chapter 5. Having documented variation between species, and

using the suggestions that resulted from data presented in

Chapter 3 regarding the effect of temperature on the chemical

makeup of cuticular hydrocarbon profiles, I then compare the

effect of varied larval rearing temperature on the sensory

structures of Ae. aegypti and Cs. melanura. Aedes aegypti was

chosen because of its rapid development time, ease in rearing,

and accessibility. Wild Cs. melanura were chosen to determine

if field observed variation could be duplicated in the lab.

These results are presented in Chapter 6.

Variation within the number of sensory structures within

a species left unanswered the question of how to quantify the

effect of changes in morphology on host detection

capabilities. I decided to estimate the surface area of the









17

palpal segment that contains the CO2 receptors, and to see how

this was affected by changes in the size of the adult females

that were produced at varied temperatures. Using these

calculations, I then decided to look for correlations between

surface area and host preference. Recognizing that the

surface of the palp that contacts air for host detection

actually represents the inner half of each palp, and that the

volume of air that is sampled by a female mosquito is a

function of the volume of air contained between the palpi at

any point in time, I compared the volumes of air sampled by

various sized palpi. These comparisons are presented in

Chapter 7.

My hypothesis is that host preference records for various

mosquitoes are attributable to the morphology of the female

mosquito and to the volume of odors given off by various

animals. In order to compare these two factors, a measure of

host emanations must be estimated. I calculated volume of

respired gases for a wide range of animal species, and then

arranged them by groups according to similar amounts of

expired volume. These calculations were then compared with

reported host preference records. Three factors, host surface

area, host respired volume, and respired volume of CO2/minute,

were found to be correlated with host preference. From these

comparisons a pattern emerged that is proposed as an

explanation for host preference and as a means to reexamine

trap design.









18

The application and testing of the ideas presented here

will involve long-term studies, involving many different

disciplines. What began as a simple examination into the

behavior of a single species has raised questions concerning

the concepts of trap design and the validity of the term host

preference. I hope that the documentation and comparison

interspecificc and intraspecific) of sensory structures known

to be involved in host seeking will form the basis for

investigations into various attractants, and possibly offer

an explanation for the variation in host preferences that have

been reported in the literature. Based on the few common

emanations that occur in mammals, a design should be possible

to maximize the attraction and collection of mosquitoes that

prefer a certain host size range. This in turn should

maximize our ability to survey for vectors of diseases such

as encephalitis.















CHAPTER 2

BIOLOGY AND SEASONAL ABUNDANCE OF Culiseta melanura AT
LAKE BUTLER, FLORIDA

General Biology


Culiseta melanura occurs as far north as eastern Canada,

as far west as the Mississippi River, and as far south and

east as Texas to the Gulf of Mexico and Florida (Darsie &

Ward, 1981). This species has been a popular insect to study

since its implication as the enzootic vector of Eastern Equine

Encephalomyelitis (EEE) in Louisiana (Chamberlain et al.,

1951). Among the species of Florida Culicini mosquitoes, Cs.

melanura females are easily recognized by their uniform dark

brown color, lack of dorsal abdominal banding patterns,

extremely long curved proboscis, and a tuft of several long

setae on the underside of the wing at the anterior proximal

margin of the subcostal vein. Males are less easily

recognized, but may be identified by palpi extending beyond

the proboscis, no dorsal abdominal banding pattern, uniform

color, and the same tuft of hair on the underside of the

wings. Larvae are readily recognized in the field, being very

elongate, having a long slender siphon, and characteristically

colored with a horizontal banding pattern caused by

pigmentation of each segment. Larvae that are collected from









20

burrows are typically very pale, while those collected from

open pools are much darker.

Culiseta melanura have a characteristic egg raft unlike

any other mosquito observed in Florida (Fig. 2-1). Eggs are

attached in a circular doughnut-shaped, concave raft.

Numerous papers have been written dealing with various aspects

of the biology of this species. The reader is referred to

these for a comprehensive review: (Burbutis & Lake, 1956;

Edman et al., 1968; 1972; Favorite & Davis, 1958; Hayes, 1958;

1961; 1962; Joseph & Bickley, 1969; Lake et al., 1962; Love

& Goodwin, 1961; 1963; Mokry, 1984; Morris & Srihongse, 1978;

Morris et al., 1976; 1980; Moussa et al., 1966; Muul et al.,

1975; Nasci & Edman, 1981a; 1981b; 1984; Reeves et al., 1948;

Scott et al., 1984; Siverly & Schoof, 1962; Spielman, 1964;

Wallis, 1954; 1962; Wallis & Whitman, 1967; and Wirth, 1947).

This species has been reported from all but one county of

Florida, although voucher specimens are lacking to confirm

most records. Because of the difficulty in separating

specimens that have been damaged in trap collections, the

southern distribution of this species in Florida awaits

confirmation. Adult activity occurs throughout the year in

Florida. The number of generations per year must vary

according to seasonal fluctuations in rainfall and

temperature. A minimum number of 2 generations per year

occurs. Gravid females are present throughout the year.

Since adults are known to live for several months, and since









Fig. 2.1. Egg raft of Culiseta melanura, 4 hours old. Gravid
female collected in resting box, Lake Butler, Florida.













































































































j -











populations tend to be overlapping with little apparent

synchrony in Florida, the actual number of generations becomes

difficult to determine. There are, however, 2 distinct

seasonal morphs in Florida which replace each other. A large

dark cool weather morph appears in early winter and is the

predominant form until late spring and early summer, where it

is gradually replaced by a smaller lighter form. Replacement

is never absolute, perhaps owing to the longevity of adult

females.

Seasonal abundance

Collections were made weekly to determine seasonal

abundance at the study site. These collections included four

CDC miniature light traps baited with 5 pounds of dry ice per

trap, and resting box collections. Each CDC trap had a

resting box (Fig. 2.2) placed nearby. Trap collections were

pooled by months collected. Several notable events occurred

during the study period. The primary collection site centered

around a small cypress dome surrounded by pine flatwoods.

Typically this habitat dries completely during the early

summer months. Periods during which no water was present in

the swamp occurred in April,May, and June. Unfortunately

during November the pine woods were clear cut, exposing the

cypress and leaving the cypress dome as a small island in the

middle of a vast expanse of open field. Catches declined

dramatically following the loss of surrounding woods.

Although the site continued to be a source of Cs. melanura,

















































(0



0
4




U


-4



0
C)
0

r-I






0
0






C"
0

(n














larval populations declined steadily, and now only a few

larvae may be found there. This is perhaps due to the

unattractiveness of the open habitat. Larvae may only be

found there in stump holes (Fig. 2.3) and under root systems

(Fig.2.4). Prior to the clear-cut operations, larvae were

abundant in the open but shaded pools throughout the cypress

dome. As water levels receded, larvae were found closer and

closer to the overhangs of the hummocks that were occupied by

Vaccinium sp. Adult populations retreated to woods that were

left undisturbed by the logging operations. There adult

populations were sampled regularly by the methods described

above. Numerous burrows within these woods provided abundant

oviposition sites, and larval populations were present as long

as water levels were high enough to pool water at the ends of

the burrow systems. When water levels dropped below the level

of the burrow systems, larvae were absent. However, larvae

reappeared very soon after burrows became reflooded.

Resting box collections dropped dramatically after the

woods were cut, but C02-baited trap catch numbers rose. This

may represent more host-seeking activity into the now exposed

open areas. Previously adults may have left the swamps to

venture into the open woods in search of a blood meal.

Personal experience has demonstrated that the poorest bait

collections are made close to the oviposition sites for Cs.

melanura. Conversely, these are the preferred sites for

maximum resting box collections. These factors influence the

effectiveness of surveillance techniques.









Fig. 2.3. Rot hole at base of pine tree, leading down to
underground larval collection site of Culiseta melanura.









28








...





. ,,


A ,I I ql
.N' ;.


i Jt ,
r w


5 'jl'4 "









Fig. 2.4. Rotted stump cavity in which Culiseta melanura
adults and larvae were collected throughout the year at Lake
Butler, Florida.












31

To illustrate the seasonal collection records for Cs. melanura

at Lake Butler, Florida, data are presented in Tables 2.1 &

2.2. Table 2.1 consists of a comparison between total monthly

catches (male & female) using CDC miniature bait traps and

resting boxes. Table 2.2 is a further breakdown of the female

monthly resting box catches subdivided into physiological

states. Since the habitat was altered severely in November,

these data are presented to illustrate the fact that CO2-

baited traps alone appear to miss periods of summer activity

of Cs. melanura in Florida, whereas resting boxes may do the

same in winter. Location of these traps presumably plays a

significant role in the effectiveness of either technique.

Larvae were abundant during February 1986. The swamp

gradually dried up until in late April no surface water was

present. Egg rafts were present in March and early April

right up to the time that surface water disappeared. The

swamp remained dry for the month of May. Heavy showers

occasionally pooled water for a few days, but no permanent

water was present until June. Larvae were not present (Cs.

melanura) until July. These were early instar. Whether these

were the result of oviposition in April, or from recent

oviposition is unclear. During the drought I dug down to

water level, a depth of 30 inches. At about 10 inches the

leaf litter and detritus was possibly sufficiently moist to

sustain larvae during drought periods. Early instars were

collected from the freshly dug holes. I cannot say for


j









32

Table 2.1

Seasonal abundance of Culiseta melanura at Lake Butler, Florida
(Feb. 1986 Mar. 1987)
CDC-C02 Resting Boxes

Month Males Females Males Females

February 0 2 10 5
March 1 20 139 123
April 0 63 195 230
May 0 134 86 398
June 0 8 42 230
July 0 5 28 77
August 0 0 10 65
September 0 15 156 290
October 0 19 289 277
November* 0 2 8 9
December 0 9 68 26
January 0 77 26 9
February 0 77 13 12
March 0 178 7 3

swamp clear-cut


certain that oviposition did not occur in the holes. However,

later rearing experiments with Cs. melanura demonstrated that

even after 2 months at moderate temperatures (up to 30"C) and

in the presence of food, some larvae appeared to be still

second instar despite the fact that the majority of their

siblings had pupated. This may offer a mechanism for larval

survival during extended drought. If larval dormancy is

partially induced by high temperatures in Florida (as it

apparently is cold-induced in the north), the larvae might

survive for several months below the surface of the ground,

sustained by high humidity in moist detritus, but otherwise

remaining inactive. This is a matter that merits further

investigation.









33

One of the important factors that needs to be monitored

in female mosquitoes is their physiological state. Periods

of host-seeking activity will be followed by the presence of

gravid and ovipositing females. The timing of these

activities is better monitored by resting box collections.

Light trap collections usually consist of nulliparous females

that are presumed to be seeking hosts. Gravid and bloodfed

females are rarely attracted to baited traps. Host seeking and

bloodfeeding apparently occurred throughout the year at the

study site in Florida. Yearly resting box collections for

females are presented in Table 2. About 50% of the females

collected were graded as non-blooded and non-gravid. This

classification does not distinguish between those females that

were taking their first or second bloodmeal.

Resting box collections provided the largest collections

of Cs. melanura at the Lake Butler site. Evidence of year

round biting activity, as well as presence of males, indicates

tremendous plasticity within this species. Peaks of adult

abundance (Tables 2.1 & 2.2.) correspond to May/June and

September/October. Miniature CDC traps (Table 2.1) support

the May/June period of activity. However, the single large

collection of Cs. melanura females in May occurred on a night

in which the first rainfall in several weeks fell. This

agrees with numerous reports in the literature that Cs.

melanura adults are most active during periods of rain and

inclement weather. The November destruction of the collection









34

Table 2.2
Seasonal abundance and physiological state of Culiseta
melanura females collected from resting boxes, Lake Butler,
Fla. (Feb. 1986 March 1987). nbg=non blooded, non-gravid;
b=blooded; bg=blood-gravid; g=gravid; n=total/month;()=
percent of monthly total
Month nbc b bq g n

February 2(50) 2(50) 0 0 4
March 114(74) 15(9.7) 8(5.2) 17(11) 154
April 166(72) 2(1) 22(9.5) 42(18) 232
May 311(80) 4(1) 34(8.7) 41(10.5) 390
June 104(47) 5(2.2) 51(23) 61(27.6) 221
July 39(49) 3(3.8) 11(14) 26(33) 79
August 34(52) 8(12.3) 19(29) 4(6) 65
September 171(60) 4(1.3) 70(24) 42(14.6) 287
October 142(52) 2(1) 72(26.4) 57(21) 273
November 0 0 0 2(100) 2
December 12(50) 5(21) 3(12.5) 4(16.7) 24
January 8(62) 1(7.6) 3(23) 1(7.6) 13
February 11(92) 0 1(8) 0 12
March 2(50) 0 0 2(50) 4


site occurred at just the time when a seasonal shift should

have been most apparent. This re-emphasizes that trapping

data are biased by environmental conditions. In this example

one day of rainfall following an extended drought totally

shifted the response to baited traps. From the above data

the following points may be made. Year round activity of Cs.

melanura occurred at the Lake Butler study site. Males were

present throughout the year. There is a suggestion of

seasonal variation in the attractiveness of CO-baited traps.

Based on the resting box collections, male densities appeared

to peak one month prior to female densities. CDC collections

were greatest on nights with some precipitation. Male Cs.

melanura are not attracted to CDC CO2-baited traps in any

numbers. This fact was later determined to be a function of









35

the trap design. Simple inversion of the CDC trap (Fig. 2.5)

so that the fan pulled upwards increased all catches,

including males of other species that are normally not

collected.

During subsequent years (1988-1989) much oviposition was

noted during October, November, and December. Adults were

again present throughout the summer, and bloodfed females were

collected throughout the year. Larval activity was not

apparent during the summer months. Either larval activity

is restricted to subsurface water levels or larvae suspend

development until late summer when rains raise water levels.

Other species occurring with Cs. melanura at Lake Butler,

Florida were Cx. territans, Anopheles crucians and

Uranotaenia sapphirina. Culex territans and Ur. sapphirina

also occurred with Cs. melanura at the New Hampshire locality.

During periods of high water in summertime, Cx. pilosus

(Dyar & Knab) occurred in larger pools of open areas at the

margin of the swamp, as did various Aedes and Psorophora

species. However, within the restricted habitat of stump

holes and burrows, Cs. melanura was either the sole occupant

or occasionally shared the habitat with Cx. territans.









Fig. 2.5. Inverted CDC miniature light trap at Lake Butler,
Florida.




































i















CHAPTER 3

VARIATION IN CUTICULAR HYDROCARBON PROFILES
IN Culiseta melanura

Introduction

Cuticular hydrocarbon composition is an expression of a

genotype, thus making it a potential taxonomic character.

Lockey (1988) reviewed insect cuticular lipid composition,

stating that hydrocarbon composition is related to

taxonomically grouped species, and that closely related

species tended to have qualitatively similar hydrocarbon

patterns, with different proportions. Less closely related

species tended to have hydrocarbon patterns differing both

qualitatively and quantitatively. Within mosquitoes, species

separation has been confined to the genus Anopheles (Carlson,

1982, 1984; Carlson & Service, 1979, 1980; Milligan et al.,

1986). Despite similarities in chromatograms, species

separation within the Anopheles gambiae complex was achieved

by comparing the relative abundance of selected peaks (Carlson

& Service, 1979, 1980). Using percentage composition of

various classes of hydrocarbons, Lockey (1978, 1988) was able

to separate 3 species of Tenebrionidae.

When hydrocarbon profiles are applied to cryptic species

or to analysis of intraspecific variation, the mechanism











behind such comparisons is typically a quantitative

(statistical) comparison of peak ratios between selected

hydrocarbons. These are identified by a 4 digit number, their

Kovats Index (Kovats, 1965). Larger peaks constitute the

major percent composition of a particular sample.

Underlying statistical comparison of peak percentages or

peak ratios is the major assumption that the variation in

these values is normally distributed. In order to apply most

standard statistical comparisons, normality of distribution

of the variable of interest is assumed. Knowledge of the

normality of a sample may confirm or reject certain hypotheses

about the factor affecting the phenomenon of interest. If

non-normal distribution is found, this may indicate certain

factors affecting the variable of interest (Sokal & Rohlf,

1969). If normal distribution is assumed, predictions and

tests of hypotheses are based upon this assumption.

Statistical conclusions are only as valid as our assumptions

about the data.

Adult specimens of Cs. melanura were collected from

various localities at different times of the year. As part

of the ongoing study concerning the behavior and ecology of

hostseeking in mosquitoes, I decided to characterize Cs.

melanura by means of hydrocarbon profiles. Since the only

published data relating hydrocarbon patterns in mosquitoes was

for the genus Anopheles, I felt that the new information

provided by this technique would provide a baseline for









40

further investigations into the usefulness of this technique

for comparison of mosquito species other than Anopheles. I

was also interested in determining whether or not any evidence

existed to suggest a polytypic species. Preliminary field

observations had revealed a larger, darker winter form of this

species in Florida. To illustrate congeneric differences, I

compared Culiseta melanura chromatograms with those of Cs.

inornata. To illustrate intergeneric differences, Culex

territans and Uranotaenia sapphirina were compared from the

same habitat and location as Cs. melanura.

The type locality for Cs. melanura is New Hampshire

(Knight & Stone, 1977). Specimens were collected from New

Hampshire and Vermont (new collection record) and compared

with Florida specimens. Since one of the reported advantages

of hydrocarbon analysis is that the patterns are stable,

regardless of the age of the specimen, it was felt that a

comparison between geographic extremes of the range of this

mosquito species might reveal maximum differences (geographic)

as well as possibly revealing the existence of cryptic

species. As more and more hydrocarbon profiles were compared,

variation appeared to be at least as extreme within one site

as between geographic limits. This perplexing situation led

to the following attempt at isolating some of the sources of

variation seen in wild specimens. Before I could characterize

Cs. melanura by this technique, I felt that an attempt should

be made to explain as much of the individual variation as











possible. I present preliminary data here and suggest partial

explanations for the hydrocarbon profile variability. Also,

I discuss the implication of this variability relative to the

application of this technique to the solving of taxonomic

questions.

Methods and Materials

Wild adult Cs. melanura were collected in resting boxes

(Florida) and by CDC miniature light traps (New Hampshire &

Florida). Dry ice was added to the CDC traps in Florida, with

nightly amounts being approximately 5 pounds per trap. This

bait was suspended beside the trap, and traps were placed

about 2 feet above the ground. Because I had no previous

indication of seasonal variation in hydrocarbon patterns,

collections were made as material was needed. Specimens from

the same collection date were later pooled for analyzing

seasonal variation.

Cuticular hydrocarbons were extracted by soaking

individual mosquitoes in hexane for at least ten minutes.

This extract was then cleaned by passage through a silica-gel

hexane column. This column permitted the passage of the non-

polar hydrocarbons, but retained the polar lipids and fats.

The extracted hydrocarbons were collected into 3 ml of hexane

extract, then concentrated by evaporation with nitrogen gas.

The final extract was re-constituted and 1/10 microliter was

injected into the gas chromatograph (GC) by an on-column

injector(SGE). A Varian Model 3200 flame ionization GC









42

instrument was used. The column type was a 30 meter DB-1,

0.32 mm. diameter capillary tube. The carrier gas was helium.

The oven temperature was programmed from 60*C to 320*C at

12*C/minute for 35 minutes. The GC was coupled through a 760

series interface and a Nelson Analytical System to an IBM

PC/XT computer, an Epsom FX 80+ printer, and a Hewlett-Packard

7470A plotter for data quantification and output. Samples were

compared with alkane standards (C14 C44) for calculation of

their respective Kovats Index (KI).

Frequency distribution normality for peak percent

composition was tested using the graphic ranked deviates

(rankits) of Sokal & Rohlf (1969). Ranked deviates were

plotted for wild adult females of Cs. melanura. After

determination of non-normal distribution within the wild

sample, additional statistical comparisons were made for wild

individuals and for individuals from known environmental

conditions. Initial statistical comparisons were made between

wild New Hampshire and Florida females. Since a non-normal

distribution was suggested by the ranked deviate test, a non-

parametric ANOVA and Kruskal-Wallis test were conducted on

these data. These tests were expanded to include comparisons

between Florida samples categorized by months collected, as

well as between specimens reared at 2 different temperatures.

Discriminant analysis was performed on data from all sample

categories (New Hampshire, Florida by months, and 2 rearing

temperatures) to look for trends in similarity measured by











mean square distances. From these comparisons I was able to

correlate variation within samples with seasonality and to

suggest temperature variation as a factor influencing change

in hydrocarbon profiles. To test for the effect of

temperature upon the KI peaks of interest, 6 females from each

adult population of 15C and 30"C reared larvae were compared.

Actual percent composition for peaks with KI 3100, 3135,

3165, 3300, 3335, 3365, 4225, 4245, 4280, 4425, and 4445 were

examined initially. These 11 peaks were then made to equal

100 percent composition and adjusted percentages were

calculated for each of these peaks. These adjusted values were

used in all statistical tests, as well as in all graphs and

tables presented here.

Field-collected gravid females were placed in cages for

oviposition. These cages were plexiglass with dimensions of

45cm X 37cm X 37cm. Oviposition was at first unsuccessful

since most of the females died before laying eggs. This was

partially resolved by placing filtered water from field

collection larval sites into black oviposition jars. Some

oviposition was obtained. However, a high mortality still

occurred among gravid females. Culiseta melanura adults in

the field had been observed hanging suspended on the upper

roof of the mammal burrows that they frequented. Oviposition

cups were then modified to include a lid that covered one half

of the overall opening to the water. A higher percentage of

oviposition was achieved quickly. This technique is strongly











suggested for those interested in obtaining large numbers of

egg rafts from wild gravid females.

Egg rafts obtained in this manner were hatched in

filtered water from the Lake Butler site. Larvae were reared

in water collected from underground burrows at the same site.

These burrows consistently harbored adult and larval Cs.

melanura for more than three years. Larval rearing was

accomplished in both enamel pans and 1 liter plastic bottles.

Since the burrows are seldom if ever exposed to direct light,

larval rearing was conducted in darkened rooms at 2 different

temperatures (15"C & 30"C). Ground hog chow was added

initially to larval rearing water, but proved to foul rearing

containers quickly. Therefore, feeding consisted of the

addition of field-collected water from the same habitat,

replacing equally the water volume lost by evaporation.

Larvae reared at 30C required 2 weeks from egg hatch to

pupation, while those reared at 15*C required 4-6 weeks.

Variation in development time within the colder temperature

was quite evident. In every case, at time of initial pupation

of the cold temperature larvae there would also be larvae that

appeared to still be 2nd instar. This was rarely seen at

30*C, but did also occur. Larvae that were reared at the

warmer temperature were well synchronized.

Egg rafts were hatched at room temperature (27"C), then

split into 2 approximately equal batches to examine the effect

of temperature on larval development and hydrocarbon profiles.











One half was reared at 30*C, the other half was reared at

15C.

Results

Intergeneric differences in hydrocarbon profiles are

illustrated in Fig. 3.1. Culex territans, Ur. sapphirina,

and Cs. melanura are all quite different, both qualitatively

and quantitatively.

Intrageneric differences between Cs. melanura and Cs.

inornata are illustrated in Fig. 3.2(males) and Fig. 3.3

(females). Again, differences are readily observed between

these species. However, when the distribution frequencies of

selected KI peaks 3100, 3135, 3165, 3300, 3335, 3365, 4225,

4245, 4280, 4425, and 4445 were examined within Cs. melanura,

considerable variation was seen to exist.

Initial comparisons between specimens from New Hampshire

and Lake Butler, Florida showed similarities between the sexes

(Fig. 3.4 & Fig. 3.5). No obvious geographic variation was

suggested from the chromatograms. Comparisons between sexes

from the Florida site showed little to suggest obvious sexual

dimorphism. However, if the relative percent composition for

any of the larger peaks had been compared, it might have been

possible to separate males and females. Gas

chromatography/Mass Spectroscopy data on this await

interpretation. Examination of Fig. 3.6 shows KI 3100 peak

in males to be significantly larger than the same peak in

females. Similarly, KI 3300 peak is virtually absent in









Fig. 3.1. Cuticular hydrocarbon profiles for adult female
Culex territans, Uranotaenia sapphirina, and Culiseta melanura
from Durham, New Hampshire. Specimens were collected
September 15, 1988, near Spruce Hole bog.


























































15 20 25
Retention Time (min.)









Fig. 3.2. Cuticular hydrocarbon profiles for adult Culiseta
inornata and Cs. melanura males from Florida. Cs. inornata
specimens were field-collected as pupae, Gainesville, Florida,
January, 1989. Culiseta melanura specimens were collected as
pupae, Lake Butler, Florida, January, 1989.



























































15 20 25
Retention Time (min.)









Fig. 3.3. Cuticular hydrocarbon profiles for adult Culiseta
inornata and Cs. melanura females from Florida. Culiseta
inornata specimens were field-collected as pupae, Gainesville,
Florida, January, 1989. Culiseta melanura specimens were
collected as pupae, Lake Butler, Florida, January, 1989.






































NI


AAL A ~ Ii~ I


I I
S15 20 25
Retention Time (min.)


Cs. inornata Female
Lab


Cs. melanura Female
Lake Butler, FL


S


N
" I



S |


UULI









Fig. 3.4. Comparison between cuticular hydrocarbon profiles
of adult female Culiseta melanura from New Hampshire and
Florida.












Cs. milanura Female
New Hampshire


Cs. melanura Female
Lake Butler, FL


15 20 25 30
Retention Time (min.)









Fig. 3.5. Comparison between cuticular hydrocarbon profiles
of adult male Culiseta melanura from New Hampshire and
Florida.












Cs. melonura Male
New Hampshire


Cs. melanura Male
Lake Butler, FL


20
Retention Time (min.)









Fig. 3.6. Comparison between cuticular hydrocarbon profiles
of adult male and female Culiseta melanura from Lake Butler,
Florida.












Cs. melanura Female
Lake Butler, FL


Cs. molanuro Male
Lake Butler, FL


15 20 25 30
Retention Time (min.)











females, but present in males. The KI 3410 peak is obvious in

males, but absent in females. These differences, if

considered mathematically, might be shown to be significantly

different. Variation of this magnitude within samples of the

same species necessitated the test for non-normal frequency

distribution.

Sufficient variation existed in values for relative

percent composition for the 11 KI peaks to require further

investigation. Table 3.1 illustrates the means and standard

deviations for KI peaks from wild specimens separated by

locality and by month. Much overlap exists between

collections, but a shift in mean values is also suggested.

Two of the larger peaks (KI 3365 and 4280) were selected for

testing for normal distribution frequencies. Raw data used

for the Rankit test for normality for peaks KI 3365 & KI 4280

are presented in Table 3.2. These data are plotted in Fig.

3.7 and Fig. 3.8. As revealed by the graphs, the data are not

normally distributed. Distribution frequencies are considered

to be non-normal if the line of plotted rankit values is not

straight. If any points deviate from a perfect linear

relationship, the distribution is considered non-normal.

These same wild caught-females used in the ranked deviate

test were plotted by months of capture to see if a seasonal

shift in values in percent composition occurred. These values

are plotted and graphed in Fig. 3.9. To check for

corresponding seasonal variation in geographic location,












Table 3.1


Mean and Standard Deviation of


melanura.
KI NH
3100 Sept
AVG 0.78
STD 0.31

3135
AVG 5.64
STD 0.81

3165
AVG 3.89
STD 3.68

3300
AVG 0.45
STD 0.45

3335
AVG 5.11
STD 0.83

3365
AVG 20.43
STD 3.40

4225
AVG 7.30
STD 0.88

4245
AVG 8.91
STD 1.14

4280
AVG 5.97
STD 2.45


FLORIDA
March
2.66
5.36


4.56
0.60


1.84
0.37


2.15
3.51


5.73
0.80


April
0.96
0.51


4.13
1.47


1.40
0.72


0.37
0.29


6.22
2.94


16.53 14.24
2.71 6.95


6.22
1.13


8.76
1.17


8.50
2.06


6.60
1.12


8.42
1.97


8.20
3.24


KI values of Wild Culiseta

FL


May
2.71
2.14


5.58
1.36


2.02
0.47


3.02
2.00


7.74
2.26


17.19
3.68


6.15
0.87


6.32
1.64


4.58
2.11


August
1.44
0.55


5.47
0.92


1.05
0.37


2.35
3.66


11.80
1.28


19.59
2.14


4.71
0.88


3.71
0.84


2.71
0.89


1.93
3.17


4.79
1.33


1.63
0.63


1.80
2.74


7.24
2.92


16.36
5.02


6.12
1.20


7.34
2.35


6.64
3.30


23.65 28.47
3.27 4.67


19.40
2.41


20.98
4.14


26.06 29.12 26.61
2.64 3.19 4.25


18.61
1.66


18.05
1.86


19.55
3.12


4425
AVG
STD

4445
AVG
STD


22.16
2.19


19.36
1.74












Table 3.2. Rankit values for percent composition of KI 3365
and 4280 among wild adult female Culiseta melanura. (Bi =
Rankit value)

KI 3365 Bi KI 4280 Bi
% Comp. %Comp.
30.37 2.66 12.87 1.99
25.82 1.75 12.85 1.98
24.26 1.44 12.02 1.72
23.48 1.29 11.36 1.51
22.87 1.16 11.27 1.48
22.81 1.15 11.20 1.46
22.54 1.10 10.74 1.32
22.46 1.08 10.62 1.28
22.02 1.00 10.44 1.22
21.96 0.98 10.16 1.14
21.49 0.89 9.77 1.01
21.30 0.85 9.67 0.98
20.95 0.78 8.40 0.58
20.00 0.59 8.11 0.49
19.76 0.55 7.89 0.43
19.49 0.49 7.75 0.38
19.30 0.46 7.54 0.32
18.84 0.36 7.43 0.28
18.83 0.36 7.27 0.23
18.57 0.31 7.22 0.21
18.10 0.22 6.77 0.07
17.96 0.19 6.77 0.07
17.47 0.09 6.73 0.06
17.14 0.03 6.69 0.05
16.95 -0.01 6.67 0.04
16.54 -0.09 6.42 -0.04
16.23 -0.16 6.30 -0.07
16.20 -0.16 6.05 -0.15
16.08 -0.18 6.05 -0.15
16.05 -0.19 5.99 -0.17
15.77 -0.25 5.94 -0.19
15.34 -0.33 5.32 -0.38
14.85 -0.43 4.90 -0.51
14.47 -0.51 4.82 -0.54
14.46 -0.51 4.50 -0.64
14.19 -0.56 4.43 -0.66
13.96 -0.61 4.03 -0.78
13.77 -0.64 4.00 -0.79
13.63 -0.67 3.64 -0.91
13.60 -0.68 3.36 -0.99
13.59 -0.68 3.24 -1.03
13.06 -0.79 3.21 -1.04
12.70 -0.86 3.06 -1.09
12.62 -0.87 3.02 -1.10
12.21 -0.96 2.60 -1.23
9.43 -1.51 1.74 -1.50












Table 3.2 (cont.)

KI 3365 KI 4280
% Comp. Bi % Comp. Bi
8.17 -1.76 1.65 -1.53
6.91 -2.01 1.58 -1.55
6.27 -2.14 1.38 -1.62
5.83 -2.22 1.20 -1.67

17.01 6.53
5.03 3.19


male samples from New Hampshire and Florida were also compared

(Fig. 3.10) with females (Fig. 3.11). In all samples plotted

seasonal variation appeared in percent composition values for

KI peaks presented. Since seasonal variation occurred within

my Florida site, statistical analyses (1 Way ANOVA and

Kruskal-Wallis test) were conducted to determine whether these

observed differences were statistically significant. These

results are presented in Table 3.3. Several factors could

explain this variation. Among the possibilities are age, and

temperature. Since differences were seen in seasonal

patterns, and since I already knew of the existence of 2

seasonal morphs for Cs. melanura in Florida, an obvious factor

to test for was the effect of temperature on hydrocarbon

profiles.

Comparisons between KI 3100, 3135, 3165, 3300, 3335,

3365, 4225, 4245, 4280, 4425, and 4445 relative percent

composition for males and females reared at 15"C and 30*C are

illustrated in Fig. 3.12. The relative percent composition

of KI peaks 3335 and 3365 are higher at 30'C than 15"C, and

the reverse is true for KI peaks 4280 and 4445. Chromatograms










62

Table 3.3
ANOVA and Kruskal-Wallis results of comparisons between New
Hampshire versus Florida wild caught female Culiseta melanura
and Florida population separated by months.


NH vs FL
ANOVA
KI F P K-W
3100 1.0080 0.3204 0.0853

3135 2.9320 0.0933 0.0310

3165 13.2380 0.0007 0.0005

3300 1.8350 0.1819 0.0236

3335 4.0300 0.0503 0.0253

3365 4.6330 0.0364 0.0229

4225 6.7910 0.0122 0.0129

4245 3.2540 0.7750 0.0955

4280 0.2900 0.5928 0.5082

4425 8.0100 0.0068 0.0046

4445 0.0260 0.8726 0.7508


F
0.8700

3.4300

5.1440

2.1970

11.2890

1.8170

4.3000

16.1540

10.7910

4.3590

1.8090


FL by Months
ANOVA
P K-W
0.4649 0.0121

0.0265 0.0094

0.0044 0.0041

0.1042 0.0004

0.0001 0.0011

0.1605 0.0615

0.0105 0.0127

0.0001 0.0001

0.0001 0.0001

0.0098 0.0108

0.1619 0.2334


illustrating the effect of temperature on hydrocarbon profiles

are presented in Fig. 3.13 and Fig. 3.14. Statistical

comparisons (1 Way ANOVA and Kruskal-Wallis) between hot

(30*C) and cold (15"C) showed statistically significant

differences at the 0.05 level for 7 of the 11 peaks compared.

These results are presented in Table 3.4. Since hydrocarbon

profile variation appears to be at least in part affected by

larval rearing temperature, I then reexamined wild caught

females by month of capture.












Table 3.4
ANOVA and Kruskal-Wallis results of comparisons between
Laboratory Culiseta melanura at two different temperatures.
Hot = 30*C, cold = 15C
HOT VS COLD
ANOVA Kruskal-Wallis
KI F P
3100 5.8230 0.0250 0.0045

3135 12.5440 0.0019 0.0037

3165 0.3530 0.5588 0.6983

3300 0.5750 0.4566 0.5184

3335 44.1060 0.0001 0.0002

3365 3.8320 0.0637 0.0527

4225 8.0170 0.0100 0.0067

4245 28.1240 0.0001 0.0004

4280 10.1750 0.0044 0.0016

4425 0.9940 0.3302 0.6985

4445 6.1500 0.0217 0.0012



Discriminant analysis was applied to each of the samples to

classify specimens into seasonal categories, and to look for

trends in groupings. Table 3.5 shows how each sample is

broken down by season, illustrating areas of overlap and

separation. Table 3.6 illustrates the generated squared mean

distances for each class of specimens.














ra
r-q

'-4
0144


~44
LO
>4 c1



-29




r. 0
4






0 t


44 0

r0


4J 4-4

4.,
4JO


.4-4 0
r.M

0)4
bH
a)4




















41~.




41- 0 ()

0
.4OC




















040 (1
*rq















4.4i 4






.Q -4


04 -4 (0
'd-



r-UW
~4-

dIG PI
4 '
V4QI 4





























0
94-










0




0c
*M






*r-










E--4
-P


Cd


l ) M .
N.


CM 1O


0 0 I I
I I


sgjuUD o


rl-
0
r)












CM





0
10






GO
In




E
0

r-


.4a







(0

I-














--4


M
0 4
O
4-

0
4J N
.r-I q
r-4
AH


0 C





u2'U


4J *

-4 0
0 (0






0


U







0
(d 0





0
Q) -
M. W



0
SHC
t) 0)





Sa,
r-



0U



0 H 0)
c 0 w



r -4

* 4
c40)



a-r
a^ 3i
m +> u
^i ui
u -i )

ta r-
r0k
>i6~

























0
O









0
rT
gd
^1

O


u-i
6


0 10
6


c
0


c E

0
-W

a0
0>
Ot


S10
Jl-"
I


suoyDJ














H "-4


0





4-)



4)





0
4-41
Enc

to 0
>












41~
UO
rg(












40)
>


4r4m










.-HO

Q)
44 (d














0
.r4 4.)
4J F-

"- :j
4-)
anl
4V


'-4
~'O



















0
ctH

cI):




kHIt
10




















GN(-4 U~
rz4
ON I
>)































Q)
Cd


u






bd
U


0 U0 0 L0 0 l0 0
Ir,) to CN -


U,









---0
It)




C4

II





Ul)



x
I












00
n
















co
U,














1 N


>0
CM


o








''











x~-o
I 0
10 0


U9GOJgd G8A!DI|9














a)
HZ




0 04


p a)




00
C44

















> 1
0a(
4-)


U,4


0En
0)z4





















0rq0
4(a
4Jk























0 (04

Q)
OH



























0 0
r-I 44
0r4m


















tp -i O
or q A %
-H > =
-4OH




4.iU
HHr



U,


OH

E:



OW
r4
-H C






























p4




'U)
Pc


Co


0
0S
ctf
S.~
CO
(D
a3


0 U) 0 t 0 ') 0
f4. n t )- t-


U,)



LO



10







CD



r)


co


V-



i r
10





So
-o








U) 0


IuaGJjd GAitDI88













H




44J



0 II
41

) 0)






0
ua
zo4

0







O-














0 0
-UA








)H -H
0--4 0
.- II
o H
(Q)











.- i I- ,
(0



























H0Q
*H4-1
c





a*











,. o
Q
vl *~

4- + (
o Q 'a



*--I k -






*P fi




























-.-
,/ -"
//'














1tn
a*


















OL
S-



















--
..-

























a + X < DO
N )I I I I N W W
Ct N o x





r-
od O *


(x *-... Xb aL


'-0)0)'' 00 r)




^ I-M


C) N- v- v- v-
-~~~~ ~ Oi 01 0 1wmD-
II I r -
*^\



'^J
N O OOCO ^ O O O C O O O t-CM o^
n n S CM CM S CM ^ i ^- ^


ue8.J9d GA!IDlj8













LO
in

H





l4
0
O -







Ocl

0 0


0
r.-1







WO








r-4
0) 0)
0
*














1) OH

4J l

M 'C


0 4
S II









0) 0
U I





















0)












H4-4
( 0












0 4
'- t C
H MQ)





44 e P
Cl cr) I)
0) -1ra-l
Q< (
e'
u)
*p m n











Cr0a


















Ul)
CVJ




04


i co
rN
















C14
W

















0
cd Cu





rn



a t L to

a + 0 1


Ln 0 LO 0 n 0 Lnp

r) n NC I


IUGO.Jd GA!IDMON









Fig. 3.13. Cuticular hydrocarbon profiles of adult female
Culiseta melanura reared at 15"C and 30C, compared to wild
adult female from Lake Butler, Florida (winter specimen).












Cs. melonuro Female
Lab. 15 C


Cs. melanura Female
Lob. 30 C


e on


Cs. melanura Female
Lake Butler, FL


Retention Time (min.)









Fig. 3.14. Cuticular hydrocarbon profiles of adult male
Culiseta melanura reared at 15"C and 30"C, compared to wild
adult male from Lake Butler, Florida, (winter specimen).








79


Cs. melanura Male
Lab. 15 C












Cs. melanura Male
Lab. 30 C












Cs. melonura Male
Lake Butler, FL


S;1



8 s I



10 15 20 25 30
Retention Time (min.)












Table 3.5


Discriminant Analysis Classification Summary for Culiseta
melanura. Number of Observations and Percentages Classified
by month:


MONTH APR


APR


AUG


COLD


HOT


MAR


MAY


6
42.86


AUG COLD


1 4
7.14 28.57


0 6
0.00 100.00


1
6.67


0
0.00


HOT MAR MAY


0
0.00


0
0.00


0 11 1
0.00 73.33 6.67


1 0
12.50 0.00


1 0
8.33 0.00


1 0
10.00 0.00


0
0.00


Total 10
Per. 13.70


0
0.00


0
0.00


0
0.00


0
0.00


0
0.00


7 15
9.59 20.55


7
87.50


0
0.00


0
0.00


0
0.00


8
10.96


1 2
7.14 14.29


0
0.00


1
6.67


0
0.00


11
91.67


0
0.00


0
0.00


13
17.81


0
0.00


0
0.00


0
0.00


0
0.00


8
80.00


NH Total


0 14
0.00 100.00


0 6
0.00 100.00


1 15
6.67 100.00


0 8
0.00 100.00


0 12
0.00 100.00


1 10
10.00 100.00


0 8 8
0.00 100.00 100.00


10 10 73
13.70 13.70 100.00










81

Table 3.6

Generalized Squared Distance to each corresponding Cluster.

REG APR AUG COLD HOT MAR MAY NH
APR 0 15.76 2.77 20.63 2.82 6.70 15.01
AUG 15.76 0 20.61 7.53 23.69 14.09 42.88
COLD 2.77 20.61 0 20.19 7.91 11.83 21.02
HOT 20.63 7.53 20.19 0 33.69 24.14 54.38
MAR 2.82 23.69 7.91 33.69 0 7.20 10.14
MAY 6.70 14.09 11.83 24.14 7.20 0 11.76
NH 15.01 42.88 21.02 54.38 10.14 11.76 0

Conclusions

Hydrocarbon profiles do in fact show clear species

differences between non-closely related species. However,

within species variation is sufficient to merit further

consideration. My specimens were separable statistically by

month of collection and by rearing temperature. Both of these

results are new to science. When samples are measured by

discriminate analysis, degree of similarity is measured by

close values in squared mean distances. From this we see that

April specimens are equally close to cold (15"C) and March

specimens. May follows April which is followed by New

Hampshire, August, and finally hot (30"C). This suggests that

the variation seen in the winter and spring months is due to

cooler temperatures, and the summer-fall variation is due to

warm temperatures. Hydrocarbon profiles do not appear fixed

within a species, but are subject to significant variation due

to at least 1 factor, temperature. The warm months 8(August)

and 5(May) showed patterns similar to 30*C reared larvae, and

the colder months 4(April) and 3(March) showed patterns











similar to 15"C larvae. This suggests a seasonal shift in

hydrocarbon profile patterns that would vary in timing from

year to year depending on water level and temperature.

Discussion

Considerable variation between individuals suggests the

need for further investigation into the role of such factors

as nutrition (larval and adult), age, sex, and water quality

on the hydrocarbon profile of any particular species. Caution

is advised in interpreting differences between individuals on

a purely statistical basis. Other criteria such as structure,

behavior, seasonality, and geographic distribution need to be

considered when attempting to distinguish forms. Interaction

between these factors is likely, and needs to be addressed.

A re-examination of published articles on the use of

hydrocarbon profiles as taxonomic tools shows that variation

exhibited by specimens may not fit within the expected normal

frequency distribution pattern. This may be due to the small

sample size that was available for analysis. A 1-way ANOVA

was typically used for comparisons. This test has several

assumptions, one of which is normality of distribution.

When dealing with closely related species or suspected

siblings, sample size becomes the major factor in the ability

to separate similar forms. Since the percent composition will

be a function of quantitative differences, researchers are at

the mercy of the specimens provided and the accuracy of data

that accompanies insects. Species that are considered to be

less closely related show qualitative differences in their









83

respective GC peaks. However, closely related species appear

to show only small quantitative differences. As presented

here, sufficient variation exists within GC peaks to cause

overlap of data sets. Once the relationship between GC

profile, temperature of larval development, adult age, and

adult nutrition are understood, precise definition of a

particular species may be possible.















CHAPTER 4
HOST PREFERENCE IN MOSQUITOES, A REVIEW

Introduction


One method of discussing host preference is to analyse

the numerous articles published on the subject. I believe

that sufficient data are available in the literature to

develop an argument for host preference being other than

preference for a species of animal. This chapter contains

condensations, re-examinations, and summaries of a few

publications dealing with the subject of host preference. I

have tried to provide sufficient examples to illustrate

diversity of hosts for particular species, as well as citing

inconsistencies in applying the term host preference. This

review is presented to develop the hypothesis that host

preference involves something more than recognition of a

particular animal species. Host preference implies that the

mosquito somehow prefers certain animals for bloodmeals. If

blood alone is the primary motivating factor, then logically

almost any animal would do.

Florida has 71 known species and subspecies of

mosquitoes. Of these, 20 never bite man. This means that for

one or more reasons man is neither a suitable host nor a

preferred host for 28% of the Florida mosquitoes. For the











remaining 50 species, man is at times an acceptable host, but

no species uses man for bloodmeals to the exclusion of all

other animals .

Host preference records are usually determined by

bloodmeal analysis, preference being determined by calculated

percentages of types of animal blood meals from collections

of bloodfed female mosquitoes.

Literature Review

Crans (1965) divided mosquito populations into 4 major

categories according to their bloodmeal sources: (1) mammalian

feeders which only occasionally fed on birds; (2) avian

feeders which rarely feed on mammals; (3) general feeders

which are indiscriminate feeders on either mammals or birds;

and (4) amphibian feeders which feed primarily on cold-blooded

animals.

Rempel et al. (1946) studied the feeding habits of some

Saskatchewan Aedes mosquitoes and concluded that none of the

species studied showed a host preference. The degree to which

a species fed upon a particular host appeared to be a function

of the availability of that host. Multiple bloodmeals (from

multiple hosts) were detected in one third of the samples

studied. Examination of their data showed that for the 4

study sites Aedes spencerii (Theobald) showed average

preferences for equine (24%), human (16%), bovine (9%), and

avian hosts (9%).











Downe (1960) concluded that host preference for mammals

was correlated with body surface area of the hosts, and that

the Aedes species studied had no preference for particular

mammals. When there were 2 or more mammals close together,

larger hosts were chosen only by chance. He also noted that

bloodmeal records did not appear to be correlated with the

weights of the animals. Dow et al. (1957) reported the numbers

of Culex tarsalis (Coq.) attracted to birds to be proportional

to the size of the birds. They also stated that birds of

different species but of similar sizes attracted similar

numbers of Cx. tarsalis. Downe (1962) studied Coquillettidia

perturbans (Walker)and reported a decided preference for

birds, with some mammalian bloodmeals. He reported a

considerable number attracted to mammal hosts without feeding.

In contrast, the majority attracted to birds took bloodmeals.

Multiple feedings were common suggesting that Cg. perturbans

may be unable to complete a successful bloodmeal on a

mammalian host. This capacity for mammalian-bird multiple

feeding makes this species a prime candidate as the vector of

EEE.

Hayes & Doane (1958) reported the first record of

Culiseta melanura biting man. No bloodmeal was taken, but

skin penetration was accomplished, resulting in swelling at

the site of the bite. Culiseta melanura is considered to be

almost exclusively a bird-feeder.









87

Karstad (1961) studied reptiles in southeast Georgia.

He found snakes, turtles, and alligators to possess

significant EEE antibodies, indicating exposure to the virus.

Very few mosquitoes are thought to feed on reptiles.

DeFoliart (1967) reported Ae. canadensis readily feeding on

a variety of turtles. Crans (1964) reported the following host

bloodmeal records from New Jersey: Anopheles quadrimaculatus

preferred mammal blood, mostly deer but included man and dogs;

Cg. perturbans preferred mammals (deer & rodent), with only

a single bird record. Aedes sollicitans fed mainly on deer,

but also on larger shore birds, humans, rabbits, and pigs.

Culiseta melanura fed almost exclusively on passerine birds,

but also fed on larger birds, deer, opossum, rodents, raccoon,

and frogs. Culex pipiens was exclusively a bird feeder.

Culex salinarius fed equally on mammals and birds, and Cx.

territans fed mostly on frogs, but also on birds, rodents,

raccoon, and turtles.

Crans (1970) studied Cx. territans in New Jersey. Of 315

bloodmeals, 279 were amphibian, 19 were reptile (6 turtle, 3

snake), 6 were avian, and 2 mammalian (rabbit & rodent). Most

often the frogs being fed upon were the bullfrog Rana

catesbeiana Shaw and the green frog R. clamitans Latreille.

Other frogs included the spring peeper (Hyla crucifer Weid.),

the southern leopard frog (Rana pipiens Schreber) and the

carpenter frog (R. virgatipes Cope). Means (1968) however

reported Cx. territans biting man.









88

Murphey et al. (1967) studied 14 species of mosquitoes

and their attraction to bird, mammal, and reptile-baited

traps. Bird hosts included mallard duck, Canada goose, common

egret, barn owl, red winged blackbird, turkey vulture, and

chicken. Mammal hosts included woodchuck, muskrat, raccoon,

opossum, red fox, meadow vole, river otter, and guinea pig.

A breakdown of some of their data is presented in table 4.1.

Although various mosquito species were collected, no distinct

host preference pattern was seen by the authors. What was

apparent was that if mosquitoes came to these hosts, most

took a bloodmeal on that particular host. The diversity of

hosts that are acceptable to the species (see table 4.1) would

suggest no host preference, as reported by the authors.

However, as shown in Chapter 7, these animals have much in

common, and in fact the data suggest a preference by the

mosquitoes for amount of respired volume of gases. Not all

hosts were used for the same number of trapping nights. The

number collected in Table 4.1 represents the number collected

per night per trial for that particular host. If the total

number of mosquitoes collected per host is divided by the

number of trials for that host, the resulting number collected

per trial per host becomes mallard (110), Canada goose (117),

common egret (128), barn owl (85), red winged blackbird

(128), turkey vulture (108), chicken (164),

muskrat (168), raccoon (150), opossum (140), red fox (97),

meadow vole (144), river otter (85), and guinea pig (180).












Table 4.1

SDecies of mosquitoes attracted to various hosts


Species
Mosquitoes


Bird


Cx. salinarius


Ae. sollicitans









Ae. cantator









Ae. vexans









Cx. pipiens


Bait Host


mallard
goose
egret
owl
blackbird
vulture
chicken


mallard
goose
egret
owl
blackbird
vulture
chicken


mallard
goose
egret
owl
blackbird
vulture
chicken


mallard
goose
egret
owl
blackbird
vulture
chicken


mallard
goose
egret
owl
blackbird
vulture
chicken


Number %
/Niaht Fed


79
95
94
94
84
100
89


86
92
93
80
87
94
91


64
41
92
100
92
89
84


78
76
100
86
79
86
83


83
70
83
100
94
100
84


Mammals
Bait Host


woodchuck
muskrat
raccoon
opossum
red fox
vole
otter
guinea pig

woodchuck
muskrat
raccoon
opossum
red fox
vole
otter
guinea pig

woodchuck
muskrat
raccoon
opossum
red fox
vole
otter
guinea pig

woodchuck
muskrat
raccoon
opossum
red fox
vole
otter
guinea pig

woodchuck
muskrat
raccoon
opossum
red fox
vole
otter
guinea pig


Number %
/niaht Fed


- --' ... -Bai ----- Hos t. -- j















Table 4.1 (Cont.)


An. quad.









Coq. perturbans









An. crucians









Cx. restuans


mallard
goose
egret
owl
blackbird
vulture
chicken


mallard
goose
egret
owl
blackbird
vulture
chicken


mallard
goose
egret
owl
blackbird
vulture
chicken


mallard
goose
egret
owl
blackbird
vulture
chicken


An. punctipennis mallard
goose
egret
owl
blackbird
vulture
chicken


4
0.5
2
3
6
6
3


12
79
83
100
80
100
76


71
95
85
62
88
88
90


100
69
70
100
82
100
76


17
79
75
67
78
50
84


woodchuck 17
muskrat 21
raccoon 12
opossum 14
red fox 15
vole 21
otter 4
guinea pig 6

woodchuck 11
muskrat 14
raccoon 16
opossum 4
red fox 6
vole 5
otter 0
guinea pig 6

woodchuck 9
muskrat 13
raccoon 14
opossum 11
red fox 5
vole 19
otter 9
guinea pig 2

woodchuck 0.7
muskrat 0.5
raccoon 0.5
opossum 0.33
red fox 0
vole 0
otter 0
guinea pig 0


woodchuck
muskrat
raccoon
opossum
red fox
vole
otter
guinea pig


4.8
1
3
4
1
2
1
7


Far fewer mosquitoes were attracted to reptiles. Those species

that showed some attraction to reptiles included Cx.


81
83
87
83
87
81
50
81

80
89
93
75
100
80
0
83

83
85
84
94
40
84
78
83

66
100
100
100
0
0
0
0

68
100
42
75
100
100
100
71









91

territans, An. quadrimaculatus, Ae. sollicitans, Cq.

perturbans, Cx. pipiens, and Cx.salinarius. Culex salinarius

and Cx. pipiens failed to bloodfeed on the reptiles, while Cq.

perturbans fed on snakes but not on turtles. Aedes

sollicitans fed on a kingsnake and snapping turtle, but not

on a watersnake and box turtle. Anopheles quadrimaculatus

fed on a watersnake, snapping turtle, and box turtle. Culex

territans fed on a kingsnake, watersnake, snapping turtle, and

box turtle. The total numbers of mosquitoes attracted to

reptiles were Cx. territans (104), An. quadrimaculatus (11),

Ae. sollicitans (25), Cq. perturbans (14), Cx. pipiens (19),

and Cx. salinarius (15). Culex territans was not collected

in mammal baited traps. All of the above data were extracted

from various total numbers reported by Murphey et al.(loc.

cit.). The authors attempted to correlate mosquito species

with host preference but were unable to do so.

Edman & Taylor (1968) documented a seasonal shift from

bloodfeeding on birds to mammals in Cx. nigripalpus. Increase

in mammal-feeding started in early summer, reached a maximum

between July and October, and was followed in the fall by a

return to mainly avian hosts during winter and spring. The

shift in feeding could not be related to any shift in the

availability of hosts at either study site. Rabbits were

frequent hosts at one site but were replaced by armadillos at

the other. The authors suggested that since seasonal

abundance of hosts did not explain the shift, perhaps seasonal