. . ‘ . . I , .Jn :mdrwmu rlu , at
. . , .. : r3». Luau.
. .fiilun. "ta
. . ‘ , . . ‘ . i. ,.
;. . . V V $3559.”
I 1. , ‘ J 1,

:nHr

‘ r, .963... ‘ . ‘
, VI 4. 1 ‘2 I. ..
, I. ...
t -

 

:
.

«,7

‘r.

. in.

”waxia
s... mink x. Y
.21: a...

 

i L” .1
‘ 15:38.; £231
. at 3.1.2:}!
V ‘ triflsi‘l
.. I Iv... . ‘ ‘ ‘
... :v t ; .u. .

Jay...“ A
e:-

 

 

V t 1|
at: .. .2
i A. -I. . L.
3. - .
1“? ‘11
ii.

. \‘ 5 t
t\..vv:\slu‘(!1
ihgnvk Isl. I

 

LIBRARY
Michig'."~:a;n State
University

 

 

 

This is to certify that the

thesis entitled

BEHAVIORAL MANIPULATION OF MOSQUITO

(DIPTERA: CULICIDAE) HOST SEEKING USING ARTIFICIALLY
APPLIED WIND AND VAPOR—PHASE REPELLENT:

POTENTIAL FOR PERSONAL PROTECTION
presented by

Eric James Hoffmann

has been accepted towards fulfillment
of the requirements for

M. S . Entomology

degree in

em [ea/z,
/"

/

\.

 

 

Major professor

Date B’jd/y/

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

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

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 chlRC/DateDuepGfi-DJS

 

 

BEHAVIORAL MANIPULATION OF MOSQUITO (DIPTERA : CULICIDAE) HOST
SEEKING USING ARTIFICIALLY APPLIED WIND AND VAPOR-PHASE
REPELLENT: POTENTIAL FOR PERSONAL PROTECTION
By

Eric James Hoffmann

AN ABSTRACT OF A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Entomology

2001

James R. Miller, Ph.D.

ABSTRACT
BEHAVIORAL MANIPULATION OF MOSQUITO (DIPTERA : CULICIDAE) HOST
SEEKING USING ARTIFICIALLY APPLIED WIND AND VAPOR-PHASE
REPELLENT: POTENTIAL FOR PERSONAL PROTECTION
By

Eric James Hoffmann

Flight-limit and dilution-mediated models for wind effects on mosquito host-seeking
behavior are explored in this thesis. Field experiments on human subjects and CDC light
traps tested these models and also assessed the efficacy of DEET vapor for increasing the
protective effects of wind velocity. F an-driven application of wind and DEET vapor at 2
m/s significantly reduced mosquito orientation to, landing on, and probing of a human
subject. CDC 1i ght-trap catches of mosquitoes were also reduced by addition of wind.
This reduction fit a negative logarithmic fimction over the applied velocity range (0 to 3.7
m/s) and closely matched the theoretical dilution function as wind increases over a

constant-release source. Increasing the C02 release rate from 650 to 1,950 ml/min

increased mosquito response for all wind velocities, but had no effect on the slope of the
logarithmic response. There was no correlation between applied wind velocity and
mosquito mass. Collectively, these data strongly support a dilution and not flight limit as
the mode of action by which wind reduces mosquito host seeking. Outdoor fans as well
as vertical Or horizontal wind curtains are proposed as local protection tactics against

nuisance mosquitoes in a backyard setting.

Copyrighted by
ERIC JAMES HOFFMANN
2001

I would like to dedicate this thesis to my wife, Mandy, whose support and love are
always reminders of how wonderful life is.

iv

ACKNOWLEDGMENTS

I would like to thank my major professor Dr. James R. Miller for his unflagging
interest and encouragement, and the support of my committee members Dr. Rufus Isaacs
and Dr. Edward Walker. I also thank Earl F legler and the Michigan Department of
Natural Resources for permission to use Rose Lake facilities.

Partial funding for this research was provided by the Department of Entomology’s
Hutson Endowment for student research. Academic fellowship support was provided by

The Graduate School at Michigan State University.

TABLE OF CONTENTS

LIST OF TABLES ................................................................................. viii
LIST OF FIGURES ................................................................................. ix
KEY TO ABBREVIATIONS ..................................................................... xi
CHAPTER 1: Introduction .......................................................................... 1
Mosquito Biology .......................................................................... 1
Historical Perspective ..................................................................... 2
Mosquito Attraction via host odors ...................................................... 4

Carbon Dioxide (C02 ) .......................................................... 5
L-Lactic acid and other mosquito-attractant metabolites .................... 7
Personal Protection ........................................................................ 9
Physical Control and Environmental Parameters ........................... 10
Humidity ................................................................. 10
Temperature ............................................................. 12
Wind Velocity ........................................................... 14
Chemical Control ................................................................ 20
DEET ..................................................................... 20

Alternative Chemical Strategies ...................................... 22

CHAPTER 2: Proposed Conceptual Frameworks for Integrating Host Cues and

Physical Factors Influencing Mosquito Host Seeking ......................................... 24
Explanations for Wind- and Odor- Mediated Host Seeking ........................ 27

Wind velocity and flight limitation ............................................ 27

Stimulus concentration .......................................................... 29

Integrated wind and odor concentration model .............................. 3O

Evolutionafily stable strategy- Host odor thresholds ........................ 34

Implications and Applications- Research Questions ................................. 35

CHAPTER 3: Reduction of Mosquito Attacks on a Human Subject by a Combination

of Wind and Vapor-Phase DEET Repellent .................................................... 39
Abstract .................................................................................... 39
Introduction ................................................................................ 39
Materials and Methods ................................................................... 40

Study Site ......................................................................... 40
Spatial Layout of Test Components ........................................... 41
Wind generation .................................................................. 41
Chemical plume generation .................................................... 41
Experimental design, data collected, and analysis ........................... 43
Results ...................................................................................... 44
Discussion ................................................................................. 44

vi

CHAPTER 4: Artificially Manipulated Wind and Vapor-Phase DEET Reduce

Mosquito (Diptera :Culicidae) Catches in CDC Light Traps ................................ 48
Abstract .................................................................................... 48
Introduction ................................................................................ 49
Materials and Methods ................................................................... 51

Study site and trapping method ................................................ 51
Carbon dioxide release .......................................................... 52
Wind generation and wind velocity determination .......................... 52
DEET generation ................................................................ 54
Statistical analysis and experimental designs ................................ 54

Quantification of mosquito response to wind, C02 and DEET 54
Comparison of species responses to wind velocity and C02 56

Relationship between mosquito mass and wind velocity ......... 5 7

Results ...................................................................................... 57
Carbon dioxide release .......................................................... 57

Mosquito responses to wind, C02, and DEET .............................. 59

Comparison of Species Responses to Wind Velocity and C02 ........... 67

Mosquito Mass and Wind Velocity Relationship ........................... 67

Discussion ................................................................................. 73
Mosquito Mass and Wind Velocity Relationship ........................... 73

Negative Logarithmic Response ................................................ 73

Parallel Responses for two C02 release rates ................................ 74

Comparison of Species Responses ............................................. 74

DEET and lactic acid ............................................................ 75

Potential Practical Applications of these findings ........................... 76

Appendix 1: Record of Deposition of Voucher Specimens ................................... 82
Literature Cited ..................................................................................... 88

vii

LIST OF TABLES

Table 1: Mosquito species composition for Experiment 3 .................................... 68

Table 2: Voucher specimen data ................................................................. 83

viii

LIST OF FIGURES

Figure 1: Four graphical illustrations of mosquito decision making when host-seeking
decisions are based on instantaneous stimulus concentration and/or wind velocity. . . .25

Figure 2: General forms of the flight-limit and concentration-based models of wind-
mediated mosquito host seeking ................................................................. 28

Figure 3: Mosquito responses as C02 concentration decreases (transition 1) and as

wind velocities are subsequently decreased (transition 2) for A) flight threshold— and
concentration-mediated models and, B) the integrated wind and odor concentration
model .................................................................................................. 36

Figure 4: Experimental setup for human subj cots—based wind velocity and vapor-phase
DEET experiments .................................................................................. 42

Figure 5: Number of mosquitoes observed orienting, landing, and probing under four
fan and repellent combinations. Means followed by the same letter withing a
behavioral category are not significantly different at P < 0.05 ............................... 44

Figure 6: Experimental setup for wind velocity determination. A) Three-speed fan on
step ladder mount, B) Heavy-duty video tripod, C) Hot-wire anemometer
cl) extension arm and c2) sensor, D) Circular target with string grid ....................... 53

Figure 7: Cumulative mass loss of C02 due to sublimation in three thermos containers
at 22°C ............................................................................................... 58

Figure 8: Nmnber of female mosquitoes (combined species) captured with and
Without C02 stimulus (Experiment 1) as a function of wind velocity. Average

nightly catch was l721individuals. Individual means comprising the top line and
sharing a common letter are not different at P < 0.05 ........................................ 60

Figure 9: Number of female mosquitoes (combined species) captured with and
without C02 stimulus (Experiment 1) as a flinction of wind velocity. The dotted line

illustrates a general decay function .............................................................. 61

Figure 10: Number of female mosquitoes (combined species) captured with C02

stimulus or with no odor or light bait as a function of wind velocity. Average nightly
catch was 500 individuals. Individual means sharing a common letter were not
different at P < 0.05 ................................................................................. 62

ix

Figure 11: Number of female mosquitoes (combined species) captured with C02

stimulus as a function of wind velocity. The dotted line illustrates a general decay
function ............................................................................................. 63

Figure 12: Number of female mosquitoes (combined species) captured at two
different C02 release rates (Experiment 3) as a function of wind velocity. Average

nightly catch was 875 individuals. Individual means comprising a given line and
sharing a common letter are not significantly different at P < 0.05 ......................... 65

Figure 13: Number of female mosquitoes (combined species) captured with and

without vapor-phase DEET (Experiment 4) as a function of wind velocity. Average
nightly catch was 900 individuals. Individual means comprising a given line sharing a
common letter are not significantly different at P < 0.05 66

Figure 14: Number of female An. walkeri captured at two different C02 release rates
(Experiment 3) as a function of wind velocity ................................................. 69

Figure 15: Number of female Ae. vexans captured at two different C02 release rates
(Experiment 3) as a function of wind velocity ................................................. 70

Figure 16: Number of female Cq. perturbans captured at two different C02 release
rates (Experiment 3) as a function of wind velocity ........................................... 71

Figure 17: Scatter plots of mosquito masses as a function of applied wind velocity
A) with two C02 release rates, B) with two DEET release rates ............................ 72

Figure 18: Vertical perspective of an area of protection generated by multiple fans
collectively forming a horizontal wind curtain ................................................. 77

Figure 19: Side projection of an area of protection generated by a border of vertically
oriented wind ports forming a vertical wind curtain ........................................... 77

KEY TO ABBREVIATIONS

Abbreviations for common mosquito genera

Ae. ............. .Aedes

An. ............. Anopheles
Cq. ............. Coquillettidia
Cs. ............. Culiseta

Cx. ............. Culex

xi

CHAPTER I:
Mosquito biology, host seeking behavior, chemical cues and repellent strategies:
Literature review
MOSQUITO BIOLOGY

Mosquitoes are hematophagous insects in the order Diptera. There are
approximately 3,300 species of mosquitoes worldwide. Their habitats range from arctic
biomes to the tropical regions near the equator (Gillett 1972, Service 1980).

Mosquitoes are notorious for the females’ blood—feeding habits, both on humans
and animals. Preferred hosts range from cattle to lizards and even frogs. Female
mosquitoes bloodfeed in order to supply their eggs with the protein and other nutrients
necessary for larval development. Some species accumulate sufficient reserves through
the larval stage to allow females to mate and lay their first clutch of eggs without a blood
meal. Such autogenous egg production is useful in areas where hosts are difficult to find
or bloodfeeding itself is dangerous (O’Meara 1985). Subsequent egg production in these
species must be fueled by blood meals. Anautogenous species lack the nutrients to
provision their eggs immediately after eclosion and mating, and must take a blood meal
before all egg-laying bouts (Clements 1992). Without the need to provision eggs, male
mosquitoes do not bloodfeed. They are nectar feeders whose primary role is finding and
fertilizing the female.

Eggs are typically laid in water or in an area that will eventually be flooded by
water. Mosquito larvae (wrigglers) are completely aquatic and feed primarily on detritus,
although a few are predaceous. They obtain oxygen from the air at the surface of the

water. After four larval instars, they pupate. Unlike most insect pupae, those of

 

mosquitoes, also known as tumblers, are still active despite not being able to feed. They
are sensitive to shadows and movement and respond to apparent threats by swimming
down in the water column. After completing metamorphosis, the adult mosquito ecloses
from its pupal case at the surface of the water, and the cycle begins again (Borror et al.
1989)

Many mosquitoes overwinter as eggs, either in floodplains that will fill the
following spring, or in existing ponds. In some species, inseminated adult females
overwinter in leaf litter or other protective detritus, ready to emerge and oviposit when

the spring rains and warmth return (Gillett 1972).

HISTORICAL PERSPECTIVE

The impact of mosquitoes on humans is far from trivial. Mosquitoes are arguably
one of the most dangerous insects to human health. In addition to nuisance biting,
mosquitoes vector diseases such as malaria, yellow fever, dengue hemorrhagic fever, and
several strains of encephalitis (Fradin 1998). The debilitating effects of yellow fever and
malaria hindered colonization and development of sub-Saharan Africa for centuries
(Diamond 1998). As human population densities grew, so did the incidence of mosquito-
bome illnesses, whose influence continues to the present.

Development in the Americas has also been hindered by mosquito-borne disease.
One of the most famous cases was the Panama Canal Project of the 19008. Thousands of
workers died of mosquito-home yellow fever while digging this Pacific-Caribbean
waterway. Intensive larval and adult mosquito control measures were used throughout
the project area, without which it is unlikely the canal ever would have been finished

(Garrett 1994).

These historical anecdotes are instructive, especially since humans are still not
free of mosquito-vectored diseases. Every year, malaria alone kills an estimated three
million people worldwide (Fradin 1998). While the mortality aspect of disease vectoring
is staggering by itself, morbidity has an even farther reach. With perhaps 700 million
people being infected by a mosquito-borne disease every year (Fradin 1998), the lost
productivity and drains on the global health care system due to sublethal cases are
enormous and difficult to calculate. Malaria is again an informative case study; its
estimated social (medical and productivity) cost is over $1.8 billion per year in Africa
(Takken and Knols 1999).

Global travel elevates mosquitoes to a new and potent position intercontinental
vectors of diseases. In 1999, a strain of encephalitis known as West Nile Virus, an
arbovirus in the family Flaviviridae, was first documented in Long Island, NY and
surrounding areas. It is vectored primarily by Culex mosquitoes, and is suspected to have
been brought to the USA by ship traffic (Centers for Disease Control 2001). Over 4,323
birds were characterized as infected in the year 2000, and 74 people have been infected
since the beginning of the outbreak in the year 1999. Human cases can be lethal; already
nine deaths have been attributed to the virus. The resurgence of the disease in the year
2000 suggests that it will be a major public health concern in the United States, and
aggressive monitoring and mosquito abatement campaigns have been implemented in
eight eastern US states (Novello et a1. 2000)

Despite worldwide participation in mosquito-reduction programs, it is unlikely
that we will ever achieve high levels of regional mosquito control. Mosquito habitats are

common and range widely in size. Breeding areas may be on the scale of several square-

-.I‘\rj

 

Ev.

kilometer lakes down to cattle footprints with less than 100 ml of water. Given the
difficulty in targeted chemical control and the expense of area-wide chemical application,
we must continue to enhance existing mosquito-control programs. Areas needing further
research are the behavioral processes by which mosquitoes locate their hosts and

determining how these critical behaviors might be disrupted.

MOSQUITO ATTRACTION VIA HOST ODORS

Mosquitoes are sophisticated creatures when it comes to locating appropriate
hosts for feeding. They are thought to rely mainly on the array of odors that emanate
from their preferred host organisms rather than visual cues. Host odors form an odor
plume, which is distributed downwind of the odor source in a meandering fashion, rather
than a straight line (Murlis et al. 1992). Entry into a stimulatory odor plume elicits a
series of orientational behaviors allowing assessment of the direction from which the
odors are being released and begin upwind flight (Murlis et al. 1992). Upwind flight is
assessed primarily through optomotor response (Kennedy 1939) and will be discussed in
more detail later in this chapter. Once upwind flight is achieved, the mosquito maintains
the proper heading toward the odor source via its chemosensory and visual systems.
Upwind flight is maintained until the insect exits the plume. Once the odor plume is lost,
a mosquito begins a series of ever-broadening upwind zigzags until it reenters the plume
and can again begin directed upwind flight. Straight-line flight will often exit a
meandering plume; this “casting” behavior enables an insect to reconnect with the odor
plume with minimal lost time and energy (Murlis et al. 1992).

The major compounds attracting mosquitoes to humans are carbon dioxide (C02),

lactic acid, acetone, and fatty acids (Takken 1991, Taken and Knols 1999), although

 

some other chemicals are also reported to attract host-seeking females. Such stimuli are
reported to be most attractive when presented as pulsed releases (breathing) rather than
constant rate releases (Gillies 1980, Geier et al. 1999).

Carbon Dioxide (C02)

All living organisms emit C02 as a byproduct of respiration. It makes sense,
then, that mosquitoes would be sensitive to increases in C02 concentration and use them
as a means to locate their preferred hosts. Release rates of C02 depend on the organism;
some benchmark rates are 25, 250, and 2500 m1 COz/min for a chicken, human, and cow,
respectively (Reeves 1953). Rudolfs (1922) first demonstrated that C02 was a

stimulatory odor for mosquitoes, and in 1951, Reeves performed a series of experiments

that showed C02 was attractive to mosquitoes in a field-trapping study. While these two

research efforts are cited as the “classic” studies, many researchers have contributed to
the qualitative and quantitative research involving carbon dioxide’s attractant role in
mosquito host seeking. No credible refutations have endured, and much data have been

gathered to quantify species’ reactions to various levels of C02.
Increased release rates of C02 have varied effects depending on the mosquito

species present. Laboratory-reared Aedes aegypti females are significantly more attracted
to dry air plus C02 in comparison to their response to dry air alone (Brown et al. 1951).
Similarly, Culex tarsalis, Ae. nigromaculis, Ae. vexans, Anopheles fi'anciscanus, and An.

walkerr' are more effectively caught in field traps as C02 release rates increase (Reeves

1953, McIver and McElligott 1989). Culex and Culiseta species may be less responsive

to C02; Cs. inornata, Cs. morsitans, Cx. restuans, and Cs. pipiens showed no increase in

 

response as C02 release rates increased from 200 to 4000 mein (McIver and

McElligott 1989). Reeves’ 1953 study showed that the attractive response of two species,
Cx. thriambus and Cx. quinquefasciatus actually decreased at the highest (2500 ml/min)

C02 release rates. Kline and Mann (1998) caught fewer mosquitoes at their highest rate
(2000 ml/min C02) for Ae. dupreei, An. perplexens, Cs melanura, Cx. erraticus and

Mansonia titillans. These data suggest that individual species have a preferred stimulus

concentration and upper threshold for C02 rather than a uniform continuous positive

response for all Culicidae (Reeves 1953, Constantini et al. 1996, Dekker and Takken
1998)

While these data are generally compelling in their support of COz’s role, removal
of C02 from exhaled breath is one of the strongest tests of this compound’s effect on
mosquito host seeking. Snow (1970) filtered out the C02 from the exhaled breath of

humans, and found an 81% reduction in the number of Cx. thalassius mosquitoes

attracted relative to breath with C02. A lesser reduction occurred in An. gambiae s.s. and

An. melas, which adds further support to the idea that different species have different

C02 concentration requirements.

The trend toward increased numbers of mosquitoes attracted to traps at higher

C02 release rates may be related to increased effective distance of C02 attractant

plumes. McIver and McElligott (1989) found that the range of attractiveness for Ae.

vexans and An. walkeri increased from 3-7m to 7—1 1m with an increase of C02 from

1000 to 4000 ml/min. Gillies (1980) proposed a linear increase for the range of

attractiveness for C02 release rates up to 1000 ml/min. This model was based on four
different trapping studies using only C02 as bait.

L-Lactic acid and other mosquito-attractant metabolites

While it is effective as a stimulatory odor, C02 is not the only chemical
mosquitoes use in orienting to their hosts. C02 combined with human odor was more
effective than C02 alone (Constantini et al. 1996, Gibson et al. 1997, Dekker and Takken
1998). Price et al. (1979) concur that emanations in addition to C02 are likely to be

responsible for human attractiveness to mosquitoes. One of the isolated components of
human odor attractive to mosquitoes is L—lactic acid. Lactic acid is excreted through the
skin as a waste product of anaerobic muscle metabolism, and volatilizes into the air.
Lactic acid may act alone as an attractant (Smith et al. 1970, Geier and Boeckh 1999), or

it is possibly just a synergist with C02 (Acree et al. 1968). Mosquitoes possess grooved-

peg (A3) receptors on their antennae for this chemical, but these receptors are not also

sensitive to C02 (Davis and Sokolove 1976). Therefore, the central nervous system must
mediate any synergism between lactic acid and C02 rather than the sensory apparatus

itself. Lactic acid alone does not elicit takeoff responses in Ae. aegypti, but it does

enhance the landing and probing responses in combination with C02 (Eiras and Jepson

1991). These results suggest that lactic acid may be a required short-range cue (perhaps
an arrestant) at the very end of the host-seeking behavioral chain.

Some other chemical metabolites have been shown to either increase the number
of mosquitoes caught in traps or effect a behavioral change in the laboratory. l-Octen-3-

01 has species-dependent attraction effects. For some mosquitoes, it is both an attractant

 

on its own and a synergist with C02, while for other species, notably Cx. spp., 1-octen-3-

01 seems to have only a minor attractive effect (Takken and Kline 1989, Takken et al.
1997a, Kline and Mann 1998). Cs. melanura does not fit this trend, though, since it is the
only species that responds less to 1-octen-3-ol when the chemical is deployed at a release
rate of 2-5 mg/h (Kline and Mann 1998). Two other components of human/animal breath
have also been found to be important for host-seeking mosquitoes; acetone has an
attractive affect (Takken 1991, Takken et al. 1997a) as does butanone (Kline and Mann
1998)

There is evidence that some non-host metabolites also enhance the attractiveness
of humans to mosquitoes. One group of chemicals that has been researched is fatty acid
metabolites of bacteria. This research was spawned fiom observations of mosquito
preference for human feet during bite location assays (Knols et al. 1997a). Foot odor
resembles that of Limburger cheese, and the chemistry of these odors is very similar in
composition (Knols et al. 1997b). Wind-tunnel studies by Knols et al. (1997b) showed
that mosquitoes responded positively to the odor of Limburger cheese. This was the

documentation of mosquito attraction in the absence of C02 or a human subject. In later
field studies, worn socks proved attractive to mosquitoes with and without C02 (Kline
1998). EAG responses were significant to C5 to C14 aliphatic fatty acids (Knols et al.

1997b)

Many of the above studies examined mosquito responses to specific release rates
of particular compounds in the field or wind tunnel. One might be tempted to project
these results to human release rates and generate an all-in-one human odor for field

trapping experiments. Unfortunately, using human equivalents for compounds such as

 

V"

C02 and lactic acid has not resulted in equivalent numbers of mosquitoes caught relative

to actual human subjects. There appear to be minor components in breath and skin
secretions that together create the full stimulatory odor profile of a human. This poses a
challenge in field studies when researchers wish to attract high numbers of mosquitoes
using biologically relevant chemical release rates. The apparently simple solution of j ust
using human subjects does not solve this problem. While people are more stimulatory
than their known synthetic chemical equivalents, they are also highly variable in
attractiveness to mosquitoes (Schreck et al. 1990). This person-to—person variation is
unwelcome in field experiments, which are already heavily affected by uncontrollable
environmental factors and uneven mosquito density. The need for consistent replication
of chemical treatments (e. g. release rates) often outweighs the ideal of using humans in

mosquito choice assays.

PERSONAL PROTECTION

It is apparent that mosquitoes are sensitive to many chemicals humans release into
the environment. This ability, combined with inability to reduce our attractive odor
output, encourages humans to devise methods for interfering with the female mosquito’s
ability to effectively find hosts.

There are two primary categories for personal mosquito control. The first
category is physical control, and includes such measures as protective clothing and
netting that provide a barrier to mosquito biting. While largely effective, these measures
are often cumbersome and can be unreasonable in hot and humid conditions. Physical

control is not limited to barriers, however. Any non-chemical environmental

manipulation can be considered physical control, and mosquitoes are acutely sensitive to
abiotic factors like humidity, temperature, and wind velocity.

The second major class of personal mosquito protection is chemical control. In
areas of high mosquito pressure or when banier methods are undesirable, chemical
methods are the method of choice. There are several mosquito repellents currently
available, and all strive to create an active space around a human or area that is avoided
by host-seeking mosquitoes.

Physical Control and Environmental Parameters
Humidity

Mosquitoes show preferences for certain ranges of relative humidities. While
different species have their own optima, it is apparent that survivorship and activity
patterns are influenced by moisture. It is unclear whether biting likelihood is affected by
humidity.

Mosquito activity patterns seem to be heavily influenced by humidity regardless
of species. Grimstad and DeFoliart (1975) observed and collected 23 species of mosquito
in Wisconsin under various wind, temperature, and humidity conditions. Mosquitoes
were collected at relative humidities lower than 48% only once over a three-year study
period.

Platt et al. (1957) performed a detailed series of experiments on An.
quadrimaculatus (Say) to clarify humidity effects on this species. An experimental
chamber with relative humidity gradients ranging from 40 to 90% was used to assess
preferences, and this group found that An. quadrimaculatus preferred relative humidities

between 70 and 80 % at 15°C. One of the critical questions regarding atmospheric

10

moisture was whether mosquito response was driven by relative humidity, absolute
humidity, vapor pressure (of water), or vapor pressure deficit. The preference
experiments were performed at 5°C increments from 15 to 30°C to separate these possible
models. Platt’s group found that the preferred relative humidity was 70% at both the
highest and lowest temperatures; there was no correlation between preference and the
other measures. It is not possible for mosquito preference to be simultaneously correlated
with absolute moisture measures (humidity, vapor pressure or vapor pressure deficit) and
relative humidity. Therefore, all the absolute models are falsified. It would seem more
parsimonious for mosquitoes to assess absolute humidity in the same concentration-
dependent manner as other chemicals, but these organisms apparently possess a
mechanism to integrate the moisture content and temperature into a single measure.

Several other laboratory experiments have solidified the link between relative
humidity and general mosquito activity. In a dual-port olfactometer, An. gambiae s.s. did
not enter traps when the humidity was less than 40%. However, significantly more
mosquitoes entered ports with rising humidities than those where the humidities were
constant or falling. This result was independent of whether host odors or clean air was
used in the olfactometer (Takken et al. 1997b).

Field experiments have also borne out the humidity effect on mosquitoes. Ae.
vexans (Meigen) shows a near perfect positive correlation with humidity and activity in
the field (Platt et al. 195 8). The humidity range activating this species as assessed by
light trap catches was 60-90%. Wright and Knight found a comparable range of 40-90%
for Ae. vexans in their field experiments, and observed an activity range of 32-98%

relative humidity for Ae. trivittatus (1966).

11

Flight performance may not be affected by relative humidity in the same way that
general activity is. Female Ae. aegypti was not affected by relative humidity in sustained
flight-mill experiments (Rowley and Graham 1968). Similarly, biting activity may not be
correlated with relative humidity, as Lumsden (1947) found no significant effect of
humidity on biting likelihood in wind tunnel experiments.

In the field, humidity influences mosquito biting and flight indirectly by
modifying the behaviors that precede active host seeking. With high surface area:
volume ratios, mosquitoes are ill-equipped to manage rapid moisture loss in dry
conditions.

With respect to personal probability of receiving mosquito bites, humidity is an
informative factor. But humidity alone is not an appropriate candidate for active
manipulation of these pests. The extremely dry or humid conditions required to achieve
repellency are impractical with respect to both cost and human comfort.

Temperature

Temperature dramatically affects mosquitoes, as it does with most flying insects.
Low temperatures reduce flight muscle efficiency and inhibit proper function of many
physiological systems (Chapman 1998). Insects that fly at low temperatures are likely to
exhibit warm-up behavior, such as shivering of thoracic flight muscles to endothennically
increase flight muscle temperatures above low ambient levels.

Lower temperatures are required to reduce activity of therrnoregulating
mosquitoes. Grimstad and DeFoliart found that it took temperatures below 10°C to

reduce mosquito nectar-feeding activity for 23 Wisconsin species of mosquito (1975).

12

 

California Cx. tarsalis (Coquillett) initiated flight at 13°C and began biting at 15°C
(Bailey et al. 1965).

While mosquitoes may be physiologically able to host seek at low temperatures,
individual species have preferred temperature ranges for activity, probably driven by the
mosquito’s physiology and also host availability. Upon examining two North American
temperate species Wright and Knight (1966) found Ae. vexans has a preferred range of
16-27°C and Ae. trivittatus had a preferred range of 18-29°C. Lumsden (1947) found that
Ae. aegypti was most likely to feed at 35°C.

When preferred temperature ranges are exceeded, mosquitoes lack any
mechanism for cooling beyond that of shade seeking. Hypertherrnia in insects can be
remediated with evaporative cooling, but few insects possess the excess water to allow
for this behavior (Chapman 1998). Thomson’s 1938 work with Cx. fatigans illustrates
how physiological status shapes mosquito responses to high temperature. Females at
various stages of reproduction were presented with temperature gradients of 1 — 5°C
increments over a range of 10—30°C. Blood-fed, hungry, and newly-emerged mosquitoes
responded slightly differently to these temperature gradients, but there was a general
avoidance of temperatures that approached 30°C. These results were corroborated by
later work on An. quadrimaculatus. Behavioral alterations were observed at 36-38 °C
and there was only 57% survivorship after 20 minutes at 40°C for this species (Platt et al.
1957)

Temperature manipulation is a potential candidate for personal protection.
Mosquitoes cannot escape the lower and upper temperature thresholds for biological

processes, and heat is easier to manipulate than humidity. The major fault, however, is

13

 

that temperatures necessary to deter mosquitoes are quite high (or low) for human
comfort. Another obstacle is the technical challenge of generating heat. Increasing
temperatures from 15 to 21°C at a distance of 1.5 m required a 220-volt heater/fan
combination in a laboratory test (Hoffmann and Miller, unpublished). Such equipment
represents a significant capital investment to which must be added the cost of power
consumption.

While current technology makes temperature modification expensive, the
possibility of behavioral control via temperature modification should not be dismissed.
Eventually, more efficient and targeted heaters could become available for behavioral
manipulation of insects. Since changing the ambient temperatures also changes the
relative humidity, this strategy has the potential of affecting two parameters that
influence host seeking with one manipulation.

Wind Velocity

While temperature and humidity may individually have limited use in personal
mosquito control, wind velocity has potent behavioral effects on mosquitoes at levels
quite acceptable for human comfort. Wind is one of the most spatially and temporally
variable of the abiotic factors mosquitoes encounter. Ambient breezes are transitory by
nature, and interrnittency is increased by the influence of terrain elevation and
obstructions. Such features generate wind shadows and wind breaks where wind speeds
are much lower than for open areas (Bidlingrnayer 1985, Bidlingrnayer et al. 1985).

A discussion of wind velocity would be incomplete without addressing
contemporary theory on mosquito flight and the physics of wind. Mosquito (and other

insects’) navigation in wind is thought to be driven largely by optomotor responses

14

(Kennedy 1939, Klassen and Hocking 1964, Snow 1980, Gillies and Wilkes 1981).
Optomotor theory argues that there is a preferred rate at which the ground images move
across the insect’s eyes (perceived ground speed) and that air speed, altitude, and/or
direction will be adjusted to maintain this preferred rate. While ground speed is a fixed
measure, perceived ground speed shifts relative to altitude. At higher altitudes, ground
features appear to pass beneath the insect more slowly, even though actual ground speed
may be unchanged. Airspeed is yet another complicating issue, as that measure is based
on how fast the insect is moving relative to the wind being encountered. If a mosquito is
flying at 0.5 m/s with no wind, its air- and ground speeds are both 0.5 m/s. With a 0.5
m/s headwind, however, the mosquito’s ground speed drops to 0 m/s (hovering) despite
the energy output for 0.5 m/s of forward motion. In such cases, a mosquito must reduce
altitude, fly faster, or fly at an angle to the wind to maintain the ideal perceived ground
speed.

At some point, a mosquito will be unable to satisfy both upwind flight and
apparent ground speed criteria. Given the goal-oriented nature of host seeking,
mosquitoes faced with excessive wind will likely desist or take refuge from the wind.
Such a wind refuge is the boundary layer, a cushion of air, typically up to 2 m altitude, in
which wind speeds are significantly less than those at higher altitudes (Taylor 1974). In
fact, one sees 50% reductions in effective wind speeds at a height of 0.25 m above the
ground (Snow 1982). The reduction is due to fluid viscosity and friction over the surface
(V ogel 1981), and, the thickness of the boundary layer is inversely related to wind speed

(Snow 1977, Vogel 1981).

15

 

Boundary layer use by mosquitoes has been demonstrated through simultaneous
suction trapping at various heights. Snow (1977) found that flight heights and wind
speed are inversely related. Significantly more An. melus (Theobald) were caught lower
to the ground at winds in excess of 0.8 m/s (Snow 1977). This does not imply that
mosquitoes do not fly above the boundary layer in high wind conditions. It merely
demonstrates a population-level probability for how individuals are distributed under
various wind conditions (Bidlingrnayer 1974). There are situations, such as dispersal,
where flight above the boundary layer is preferred. The higher winds make host
orienting nearly impossible but allow for rapid and efficient downwind flight to new
feeding, mating, or oviposition areas (Snow 1977, Snow 1979, Service 1980, Snow
1982). Upon reaching an appropriate location, a mosquito could drop back into the
boundary layer and resume orientation behaviors in more favorable wind conditions.

In addition to providing a refuge from wind’s disrupting effects, the boundary
layer may also play a role in upwind navigation in the absence of visual cues. Despite
strong evidence against flight in the absence of visual cues (Daykin et al. 1965), Gillett
(1979) suggests that a mosquito at level, sustained flight within the boundary layer can
assess the direction fi'om which the wind is originating by merely changing altitude. If
the mosquito is flying downwind, and lowers its flight elevation, it will encounter an
apparent deceleration, as the boundary layer reduces the tailwind. An assumption is
made that the change in velocity would be sensed by inertial receptors. Similarly, if the
mosquito is flying upwind and lowers its altitude, its impeding headwind is reduced and
it senses positive acceleration. This model assumes relatively constant winds at a given

altitude; but even with this limitation, it is a feasible mechanism for upwind flight in the

16

 

 

FT

absence of visual cues and optomotor input. It also relaxes some of the light intensity
requirements intrinsic to optomotor-based host-seeking models. It should be noted that
upwind orientation does not equal upwind flight. A mosquito in complete darkness may
be oriented upwind, but it has no way of assessing whether it is making headway relative
to the ground or is actually experiencing negative displacement as the headwind pushes it
backwards. While this inertial model is theoretically possible, blinded insects are not
successful in wind-tunnel flight (J.Miller, personal communication).

While the boundary layer may be a partial refuge from wind, its protection is not
complete. Mosquitoes almost always engage headwinds while host seeking, and they are
not equipped to counter high wind velocities. Lewis and Taylor (1967) developed a
relationship between mean wind speed and insect size based on their experimental as well
as literature data for these parameters. Their theoretical values of about 1 m/s for
mosquitoes are home out well in field experiments. Wind velocity limits have been
established for several American and African mosquitoes. Wind velocities of over 0.9
m/s reduce flying and biting of both Ae. vexans and Ae. trivittatus (Wright and Knight
1966). Cx. neavei and Cx. poicilipes are also affected by higher wind velocities; suction
trap catches were consistently lower at wind speeds above 1 m/s compared to speeds
below this level (Snow 1979). Bite-count experiments in Gambia, West Africa showed
mean flight speed estimates for Ae. melas and Cx. thalassius to be 1.12 and 1.14 m/s,
respectively (Snow 1980). These two species were caught at higher wind velocities, but
numbers were greatly reduced at wind velocities above these calculated mean flight
speeds. For North American species, Grimstad and DeFoliart (1975) found that winds

over 2.8 m/s halted nectar feeding in 23 common Wisconsin mosquito species. The

17

general range for successful orientations (as assessed by trap catches) seems to be around
0.8 m/s; trap catches are drastically reduced by winds exceeding this rate. An exception
appears to be arctic mosquitoes, some of which have an incredibly high flight capacity --
up to 8 m/s (Neilson and Neilson 1953, Service 1980). However, this thesis focuses
primarily on temperate North American species, thought to share more modest flight
capacities.

Reductive effects of wind are not restricted to velocities approaching the absolute
flight limit for mosquitoes. Bidlingrnayer el al. (1985) found that while winds below
0.24 m/s had no discemable effect on the number of mosquitoes caught in suction traps,
winds between 0.25 m/s and 0.49 m/s reduced catches by 75%. In their collections of 23
species of Wisconsin mosquitoes, Grimstad and DeFoliart found velocities of about 1.5
m/s reduced nectar feeding in some species (1975).

Bidlingrnayer, Day and Evans published one of the most comprehensive studies
of mosquito response to ambient wind in 1995. Suction trap collections and average
wind velocities were automatically taken every 15 min rather than at typical nightly
intervals. Trap catch for all species declined as wind velocity increased; winds of 0.5 m/s
reduced overall numbers caught by 50% relative to numbers caught at 0.25 m/s (the
lowest velocity class measured). 75% reductions were seen at velocities of 1.0 m/s.
While this study did not encounter wind velocities below 0.25 m/s, the authors found no
evidence of a velocity plateau below which there is no significant difference in catches.
They posit that even winds of 0.1 m/s would reduce mosquito catches when compared to
perfectly cahn conditions. Another important finding of this work was that mosquito

response to wind velocity appears to be logarithmic (Bidlingmayer et al. 1995) rather

18

r

 

than linear or quadratic. Previous studies had simply lacked the rigorous sampling design
required to adequately describe the relationship mathematically.

In addition to their robust conclusions on ambient wind effects, Bidlingrnayer et
al. ( 1995) also suggest three possible causal mechanisms of reduced mosquito catch at
higher wind velocities. First is the impact on optomotor response resulting in either
lower flight height or turning downwind in order to maintain a preferred apparent ground
speed. This model predicts a higher density of mosquitoes in the boundary layer at
higher wind speeds, and thus higher trap catches up to the velocity limit. This model was
inconsistent with the Bidlingmayer et al. (1995) findings; traps within the boundary layer
caught fewer mosquitoes as wind velocity increased.

A second model focuses on the direct effects of wind velocity on the mosquito.
Higher winds reduce the maximum recruitment rate to the traps over time, although the
traps may still be fully attractive if operating over long periods. A third model involves
the impact of energy reserves of mosquitoes. The most efficient time to fly would be at
low wind velocities, and the largest percentage of the population would be able to host
seek. Host seeking would also occur at higher velocities, but only by mosquitoes having
the energy reserves to overcome the physical flight banier. The latter two models are not
exclusive, and both may be involved in reducing trap catches of mosquitoes at higher
wind velocities. (Bidlingrnayer et al.1995).

While field studies of ambient wind are useful in addressing wind as a monitored
factor, there are few studies where wind velocity is directly manipulated in field settings.
Gillies and Wilkes (1981) tried this approach with a wind tunnel in Gambia, West Africa.

They used a calf as odor bait and a grid to electrocute mosquitoes that successfully

l9

 

navigated the tunnel. Catches of Mansonia species fell sharply above 0.5 m/s and
became negligible above 1.8 m/s. While the restricted entrance of this wind tunnel
limited the host-seeking responses of the mosquitoes, it was a first step in showing that
wind velocity could be manipulated in the field with meaningful results.

Wind has behavioral activity on mosquitoes at levels comfortable to humans, and
is inexpensive to generate. Thus, it is an appropriate option to manipulate for mosquito
deterrence. However, there is an appreciable gap in knowledge of mosquito response to

manipulated wind in outdoor settings frequented by humans.

CHEMICAL CONTROL
DEE T

Chemical repellents have become a mainstay for easy and inexpensive mosquito
control. The most commonly used chemical mosquito repellent in the United States is
DEET (MN-diethyl-m-toluamide or, alternatively, diethyl methyl benzamide). DEET is a
potent mosquito repellent first marketed in 1956. It is considered the most effective
broad-spectrum repellent, working not only against mosquitoes, but also on ticks,
deerflies, and sandflies. An estimated 200,000,000 people worldwide and one-third of
the US population use DEET every year (US Environmental Protection Agency 1980),
which amounts to billions of applications. DEET has an effective repellent duration in
excess of 5 h, and various studies using highly concentrated (>90%) formulations see
greater than 90% effectiveness 8 h after application (Schreck and McGovern 1989, Chou

et al. 1997, Frances et al. 1999). DEET has an oral LD50 of ~ 2000 mg/kg and an

inhalation LC50 of 5.95 mg/L.

20

 

DEET’s mode of action upon mosquitoes is unclear. Davis and Sokolove (1976)
found grooved-peg (A3) lactic acid sensilla on the antennae of Ae. aegypti and
determined through neurophysiological recordings that DEET inhibited the firing of these
neurons, or reduced impulses in an already active cell. DEET may also affect other
sensilla on the mosquito antenna, as evidenced by DEET’s interference with sensitivity to
ethyl proprionate, an oviposition stimulant (Kuthiala et al. 1992). These lines of evidence
have generated a series of pore-blocking models for DEET action (Wright 1975, McIver
1981), which unfortunately suffer from inconsistencies with experimental results. Insects
encountering a repellent stream show increased turn angles and often exit the stream
immediately (Daykin et al. 1965, Davis and Bowen 1994). From these data, the insect
must perceive the repellent as a negative stimulus. Ifthe pore-blocking model were
correct, then the mosquito should have continued on through the plume - oblivious to the
attractant (or repellent) odors (Davis 1985). McIver’s 1981 analgesic model does have
some subtleties that bear pointing out, even if the data are not corroborating. DEET may
interact with membrane components, causing a general modification in the sensory
pattern. Alterations in mosquito behavior may stem from this distorted pattern (McIver
1981)

DEET’s action as an inhibitor does have some support, however. Recent
olfactometer experiments failed to show repellency (displacement away fiom the source)
under the experimental conditions (Dogan and Rossignol 1999, Dogan et al. 1999).
Regardless of its mode of action, there is general agreement that DEET interacts in some
manner with the mosquito nervous system, be it inhibition of sensors, stimulation of

sensors, or some across-fiber behavioral modification.

21

Despite these recommending traits, there is a strong public misperception that
DEET is a toxic chemical that should be used only if absolutely necessary. DEET has a
remarkably strong safety record after almost 50 years of intensive worldwide use. Toxic
reactions to DEET are extremely rare; since 1956, only 14 individuals have been
diagnosed with encephalopathy (neurotoxicity) associated with DEET. Three of these
individuals died; the others made a full recovery (Osimitz and Grothaus 1995). From
1986 to 1989, The American Association of Poison Control Centers reported taking calls
for 9,086 cases of DEET exposures. Upon follow-up of these cases, 5 patients reported
major effects and one patient had died (after intentionally ingesting 802 of DEET)
(Osimitz and Grothaus 1995). These toxicological data compare favorably with
alternative “natural oil” repellents, which have themselves been implicated in at least 3
human deaths since 1990 (Goodyer and Behrens 1998).
Alternative Chemical Strategies

While the toxicity of DEET is very low, it does have some negative physical
attributes. It is a plasticiser, and can cause damage to synthetic fabrics and finishes.
This, coupled with a chemical smell and fears of toxicity, has contributed to mistrust of
DEET. The desire for less harsh repellents has spurred governments and corporations to
seek longer-lasting and more broadly-effective compounds that meet the requirements of
the concerned end user. One of the new classes of chemical mosquito repellents is
piperidine, which is being tested as a replacement for DEET in US military applications.
Several piperidine compounds have been tested (Yap et a1. 1998, Debboun et a1. 1999,
Frances et al. 1999); one of the most promising is KBR 3023 (Bayer AG, Germany).

This compound is remarkably non-toxic; its oral and dermal LD50 toxicities are 4,743

22

and over 2,000 mg/kg body mass, respectively (Yap et al. 1998). KBR 3023 is under
consideration for EPA registration for consumer use and is likely to be available in the
United States in the next few years (Spagnoli, personal communication).

While chemical repellents are widely used with well-established safety profiles, it
is apparent that there is a popular mistrust of dermal chemical applications for mosquito
repellency. Oily feel, chemical smell, and uninformed fear reinforce public demand for
“natural” alternatives to the various chemicals that are used for personal and area
protection fi'om mosquitoes. As a result, we see continued demand for bug-zappers,
ultrasonic emitters, and home-remedies, despite limited evidence of their efficacy. The
need remains for alternative mosquito management measures that are effective,
inexpensive, and inoffensive to even the most sensitive individuals.

This thesis research was motivated by this need for such alternatives. Mosquito
behavior is markedly affected by abiotic factors that can be manipulated to reduce
mosquito presence or ability to host seek. There are significant gaps in research
involving artificially-applied wind, and the field affects of this factor on mosquito
behavior need to be quantified. The objectives of this research were to: 1) further
quantify the relationship between mosquito host seeking and artificially-manipulated
wind velocity, and 2) provide background data potentially supporting the development of

wind application as a strategy for protecting humans from mosquitoes.

23

 

1‘.“

CHAPTER 11
Proposed Conceptual Frameworks for Integrating Host Cues and Physical Factors

Influencing Mosquito Host Seeking.

This chapter explores theoretical concepts on decision processes mosquitoes
might use when finding hosts and proposes models of mosquito host seeking. In Figure
l, I present graphical representations of two factors that are reported to influence host
seeking, wind velocity and stimulus concentration. Factors limiting mosquito host-
seeking ability are proposed, and their interrelationships are expressed quantitatively.

Much mosquito behavioral data have been collected that suggests increasing wind
velocity limits mosquito orientation to otherwise stimulatory host cues. In calm
conditions, this flight limit would be equivalent to the maximum possible ground speed,
but when ambient wind velocities exceed this velocity, reported to be between 1-2 m/s,
host seeking is supposedly not possible. A refinement of this approach takes into account
the idea of “preferred optomotor response” (Kennedy 1939, Snow 1980, Gillies and
Wilkes 1981). If there were in fact a preferred rate of ground passage below the
mosquito, the flight limit would be less than the theoretical maximum so the mosquito
could maintain that preferred rate in the face of headwinds. A mosquito may “decide”

whether or not to host seek in the presence of a stimulus like C02 based simply on

whether or not ambient wind conditions are below this flight limit (Figure 1A). In
acceptable wind conditions, the mosquito would fly until it finds the host, loses the

plume, or succumbs to exhaustion.

24

/ No oriented flight D Upwin d flight fig: Super tical flight
W’V/

00000000

 

   

 

JStImuusonICc etttmtiott—> StiuusonmlC tration->
A)Physi cialflghtthre sohld B)Physi

limit.

citalflgh tlimit andu urppc C02

 

 

 

C) Physical flight limit and lower C02 D) Physi

 

 

 

 

 

citsoalflghthrehldwitha dec
thre soshldwithu SUCTP criticalflight ntiofunc nuand urppc rCOzthresohld
Figure e-zosll(AC)Fu urngraphica cailllus oons oosfm osuiq odec onmakingwhen
hos seekinecrgd sear eeobasd ntrintsleotantan ouss stimulusco sorrticetrao nand/o
wintdsv eloc ci.ty

25

C02 receptors on the maxillary palpi of Aedes aegypti exhibit a biphasic dosage-
dependent response that initially increases then eventually decreases with increased C02

concentration (Davis and Bowen 1994). These observations lead to a refinement of the
basic model (Figure 1B) such that a mosquito will host seek in winds below the physical

flight limit but only at C02 concentrations under the neurological dosage limit.
Grant et al. (1995) described minimum activation concentrations of C02 required
to elicit host-seeking behavior in mosquitoes at 150 — 300 ppm C02. With C02 present

in the atmosphere at 300 — 400 ppm without host respiration, a mechanism to screen

relevant concentrations (exhaled human breath contains 45,000 ppm C02 (Gillies 1980)

from background “noise” is necessary. It is also possible that given a sufficiently high

C02 concentration, mosquitoes will attempt to fly upwind regardless of the amount of

forward progress. While such supercritical flight has not been demonstrated in insects

thus far, I wish to point out its possibility. Upon sensing a C02 concentration above the

minimum threshold, a mosquito might orient upwind and continue to fly — even if the
mosquito were losing ground in the effort (Figure 1C).

Such absolute chemical maxima and minima may be too simplistic. A more
sophisticated design would use a gain function that incorporates some measure of cost-
benefit analysis for a given combination of wind velocity and stimulus concentration
(Figure 1D). The cost is the amount of energy required to travel to the host, and the
benefit is the complete blood meal. A mosquito would be more likely to orient to a host

in high wind conditions under high C02 conditions compared to low C02. Given the

26

 

diluting effect of wind, sensing high C02 would indicate very close proximity to the host

and a higher chance of a positive investment return.
Explanations for Wind and Odor-Mediated Host Seeking

While conceptual illustrations of possible behavioral outcomes are interesting,
they lack the potency to explain “how” or “why” the variables involved would exert their
effect. I explore three possible explanations for mosquito observed host-orienting
behavior relative to wind velocity and stimulatory host odors and describe tests that could
be conducted to determine their validity. The first two of these involve absolute
measures of environmental conditions, while the third is an energetics-based model that
focuses on available resources and perceived distance to target (host). An additional
evolutionary model is also proposed to address mosquito host odor thresholds.
Wind Velocity and Flight Limitation

The first model, favored by current mosquito literature, focuses on the mosquito’s
flight limit as the dominant factor influencing mosquito host seeking. This model

corresponds to Figure 1C and predicts that for any appropriate release rate of C02, a

mosquito will fly toward the host. If mosquito host seeking is driven exclusively by wind
velocity, then mosquito responses in a wind tunnel or trapping study should exhibit little
response change up to the flight limit, after which a dramatic decline would be evident
(Figure 2). Both “maximum flight velocity” and “optomotor response” considerations
would yield this rapid decline. If mosquito host seeking were shaped entirely by flight
limits, then mosquito response would be unaffected by increased in-air concentrations of

C02 at a given wind velocity.

27

 

cub
L
l

------- Flight limit model

.3
N
n

—— Concentration -based model

Mosquito Response

 

 

 

 

 

U U

0 ix 2x 3x 4 x
Wind Velocity (m/s)

Figure 2: General forms of the flight-limit and concentration-based models of
wind-mediated mosquito host seeking.

28

5X

If the flight-limit model were correct, a positive relationship between mosquito
mass and wind velocity would be expected; larger mosquitoes would be capable of
negotiating higher velocities (Lewis and Taylor 1967). In addition to the ability to
generate greater locomotory force, larger mosquitoes are likely to have higher energy
reserves and greater muscle mass, supportive attributes when undertaking upwind flight
in demanding winds. In addition to weaker flight, smaller mosquitoes may be more
unstable than larger ones in turbulence because of decreased inertia.

Stimulus Concentration

Rather than forcefully influencing mosquito host-seeking behavior, it is possible
that wind simply dilutes stimulus concentration. If stimulus concentration governs
mosquito host seeking, then host seeking will take place whenever appropriate

concentrations of C02 are present, other conditions permitting. Outcomes of decision-

making so influenced could be illustrated by Figure 1C.

This model’s uniqueness lies in the predicted response-curve shape for mosquito
trap catch vs. wind velocity. It is well known and readily calculable that concentration of
a chemical emitted into air from a point source and at a constant rate over time
diminishes geometrically with wind velocity; each doubling of wind velocity halves
concentration of the chemical. Moreover, this dilution is steepened by turbulence, which
increases more than linearly with wind velocity. With this in mind, the idealized shape of
a mosquito response curve in a trapping study would be a rapid drop in trap catch at low
wind velocities and an asymptotic approach to zero as wind velocities further increased
(Figure 2). In reality, the response level with no wind is finite and depends on the release

rate from the source. Furthermore, turbulence and diffusion would modify the shape of

29

this curve, resulting in a steeper initial drop at low velocities, and a more rapid approach

to background levels of C02 at higher wind velocities. With all other aspects being

equal, increasing initial stimulus release rates from the source would increase the number
of mosquitoes at a given wind velocity relative to a lower release rate. In contrast to the
flight-limit model, a positive relationship between body mass and wind velocity is judged
unlikely in a host-seeking system relying only on the concentration of attractants.
Integrated wind and odor concentration model

Upon encountering a host-odor plume, a mosquito may choose whether or not to
orient to that plume and fly towards the source host. This choice could be based on the
determination of distance to a host and whether or not energy stores are available to
successfully anive at the host.

Mosquito upwind flight may also be influenced by the likelihood and payoff of a
blood meal against the energy needed to get to the host. High concentrations of

excitatory odors (C02, lactic acid, etc.) would tip the balance toward investing energy in

host seeking, while inhibitory factors such as wind would make such efforts too costly (or
even impossible) with respect to time and energetics.

How might a mosquito assess distance to a target that is unknown and unseen? I
propose an Integrated wind and odor concentration model by which mosquitoes might
determine the distance to the host. According to this model, one behavioral requirement
and two physical abilities have to be integrated. First, a mosquito must have a
programmed response to a particular host-odor release rate. Second, the mosquito must

possess the ability to perceive in-air host odor concentrations. Finally, the mosquito must

30

have the ability to determine wind velocity with respect to the ground. The integration of
these elements is probably governed by across-fiber patterning (Chapman 1998).

There is evidence that mosquitoes possess all of the required sensory abilities to
perform the required integration. Insects as a group are well known to have dosage-
response curves to stimuli, indicating that they are able to quantitatively assess chemical
concentration rather than mere presence or absence. Many species of mosquitoes have
preferred hosts (Clements 1992) and may have selected responsiveness to a particular
range of host odors. Mosquitoes also exhibit an acceptable range (either innate or
learned) of wind-assessing flight behaviors. The capacity to determine wind velocity is
evidenced by the behavioral response of lower flight altitude and/or landing in high wind
conditions (Grimstad and DeFoliart 1975, Gillett 1979, Snow 1979). This adjustment in
flight speed is likely associated with the optomotor response-linked, preferred rate at
which objects are perceived as passing under the mosquito. At maximum flight output,
the ground should be passing at a particular rate. If the ground is passing at a slower-
than-preferred rate, the mosquito is likely encountering a headwind. I am not suggesting
that a mosquito can determine the precise velocity of the wind. The requirement of this
model is more of a general perception of ambient wind speeds and the ability to integrate
that wind speed with the detected concentration of host odors in making some
determination of distance to the odor source.

Such integration would hinge on some sort of dilution computation; however, no
math is required. The mosquito is determining distance in the same way you or I
determine how close you are to a barbecue without seeing it. We have an expectation of

what a barbecue smells like at close range, and some impression for how the intensity of

31

the odor decreases over distance. Thus you can assess how close you are to the barbecue
without any mathematical equations. I suggest that incorporating the wind velocity
component is not an exceptional feat, either. Ifit is a windy day, we could probably
make some compensation for the dilution effect.

This is a time-averaged model, and focuses more on dilution than the favored

models proposed for insect (primarily Lepidoptera) pheromone plumes. C02 has a much

lower molecular mass than the typical insect sex-attractant pheromone, making it a more
mobile molecule in air (Chapman 1998). Mosquitoes are also considered to orient at
much shorter ranges (under 15m) than Lepidoptera (Edman 1979).

A simple equation for calculating in-air concentration of a host odor is:

[C] = R *f(d)

Vw

Where:

[C] is the in-air concentration

R is the release rate ml/min (a constant)
f(d) is some diffusion/dilution coefficient dependent on the distance from the point source

Vw is the wind velocity m/s (a constant)
VW is in the denominator because concentration of a chemical from a constant

point source is inversely related to the wind velocity- assuming no diffusion.
The practical application of this general concept would be arranging the equation
to determine, “What ‘d’ corresponds to the measured concentration [C] and wind

velocities Vw given the expected release rate R?” For the mosquito seeking a blood meal

32

or a human seeking a barbecue, the value need not be to the millimeter, but close enough
as to make a determination of the amount of time required to get to the odor source.

Minimum time to the source is represented by the equation:

T = D
Vm ‘ Vw
Where:

T is the time required to get from the current position to the odor source
D is the perceived distance to the odor source

Vm is the maximum ground speed of the mosquito
Vw is the ambient wind speed that the mosquito is flying against

This equation assumes that the mosquito is taking the most direct route to the odor
source, and is affected only by direct headwinds. After making the determination of time
to the odor source, the mosquito can assess the energy expenditures required for such a
sustained flight at maximum velocity and the status of physiological reserves.

This model corresponds to the behavioral predictions illustrated in Figure 1D.

For a given wind velocity (under any flight maximum), the mosquito response depends

on what the instantaneous concentration of C02 is, and visa versa. Since the shape of the

gain function is unknown, it is not possible to postulate the form of a response curve if
mosquito trap catch were plotted against wind velocity. Mosquito mass would likely
prove a significant factor in upwind success; the cost-benefit approach is closely tied to
energy reserves.

While time-to-host is linked to energy consumption, it also extracts other costs

from the insect not explored in this model. Exposure to predation and other mortality-

33

inducing events is time-related, and an insect may become dehydrated during flight
(Brantjes 1981).
Evolutionarily Stable Strategy - Host Odor Thresholds

Perhaps mosquitoes are not using energetics to determine whether or not to fly
upwind towards an odor source at all. There may be a threshold concentration of host
odors above which a mosquito will orient to a host regardless of the energetic cost (to
exhaustion) until a) bloodfeeding occurs or b) the host odor is lost. This threshold would
be above ambient odor concentrations and would imply a high probability that the
mosquito would encounter the host before running out of energy. This approach is more
evolutionary than the energetics model and places less of a burden on the mosquito to
have a distance/concentration function imbedded in its neural circuitry.

Mosquitoes with a low threshold for stimulatory host odors are likely to orient to
more distant hosts than mosquitoes with higher thresholds. Since they are flying farther
to their hosts (or flying in higher headwinds), these less-selective mosquitoes use a higher
proportion of their energy reserves for flight than the more selective elements of the
population. With high levels of energy being devoted to host-orienting and upwind
flight, less energy can go to egg production and maturation. While individual fitness may
not be compromised by longer flights, fecundity in such individuals is almost certainly
lower.

Individuals with a very high threshold for the appropriate host odors also suffer
reproductive loss. Given the rapid drop in vapor concentration with distance from the
source, a high-threshold mosquito has to be very close to a host to elicit the change from

ranging to host-seeking behavior. High-threshold mosquitoes would spend most of their

34

energy in ranging behavior (appetitive flight), searching for an odor plume. Actual host-
orienting and upwind flight would be rare, since the volume of air containing the
minimum odor concentrations is low. These individuals would be sacrificing fecundity
for flight in much the same way that low-threshold individuals do.

This evolutionary model suggests that a genetically stable odor concentration
threshold exists such that, on average, reproductive success is maximized. Mutations for
new thresholds would need to be accompanied by co—occuning mutations in other
sensory abilities or flight capacity in order to comprise a stable portion of the population.
Implications and Applications - Research Questions

Efficient testing of the first three physical models would involve experiments

where wind velocity, C02 release rate, and in-air C02 concentrations could be measured

and manipulated and mosquito responses recorded. The fourth, evolutionary, approach
suffers a common limitation of evolutionary theory- testability. It is nearly impossible to

test an entire population of mosquitoes for a common C02 threshold, and tests of a

limited number of individuals would be inappropriate. It would be impossible to
discriminate whether individuals departing from the mean were refutations of the model,
less-fit individuals, or a mosquito that has additional physiology (or behavior) to

compensate for the higher or lower C02 thresholds.

The physical models could be first tested against their predicted responses to

wind velocity and C02 concentration. The first two models fit figure 3A, while the cost-

benefit approach is represented by figure 3B. A wind tunnel could be used to fly

mosquitoes in various wind/C02 conditions. One could find an instantaneous C02

concentration at which a mosquito could orient and fly upwind. The C02 concentration

35

could then be dropped to the point where the mosquito no longer exhibited host-seeking

behavior (Figure 3A and 3B, transition 1). While C02 concentration was held constant,

wind velocity would then be reduced until the mosquito again began host—seeking

behaviors (Figure 3A and 3B, transition 2). For all winds and C02 concentrations below

physiological maxima, wind and concentration models would Show no recovery of host-
seeking behavior as wind velocity is dropped. Conversely, the cost benefit model
predicts a changeover from upwind flight, to no oriented flight, and back to oriented

flight as the wind/C02 levels transition from point one to two to three.

If it is determined that there is no recovery of host-seeking behavior with

decreased wind velocity and a fixed C02 concentration, then it comes time to separate

the concentration- and flight limited- mediated models. This could be done in the wind
tunnel as well, although a field study of baited light traps and artificially applied wind
would be effective.

Mosquitoes would be assayed for their response to a fixed C02 release rate (not

in-air concentration) as winds transitioned from no applied wind to velocities at which
mosquito orientation behaviors ceased. The experiment would be repeated for increasing

C02 release rates up to the physiological maximum where C02 becomes repellent or

anesthetic.
If the response curves showed a dramatic decrease at a particular wind velocity

and no difference at varied C02 release rates, then the dilution model would be refuted.

In contrast, the dilution model predicts an exponential-type decay curve with an
asymptotic approach to zero as wind increases and increased responses to higher release

rates at a given velocity.

37

In addition to exploring the practical possibilities for using wind to reduce
mosquito attacks on humans in the backyard setting, the research of this thesis was also
slanted toward differentiating between the various models whereby these effects are

mediated.

38

CHAPTER III
Reduction of Mosquito Attacks on a Human Subject by Combination of Wind and

Vapor-Phase DEET Repellent

Abstract

In a central Michigan wetland setting, electric fan-generated wind at 2 m/s plus
DEET vapor at 800 rig/L air significantly reduced mosquito behaviors of orientation,
landing, and probing in response to a human subject. The reductions relative to no
applied wind and no DEET were 74, 75, and 70%, respectively. The DEET effect was
significant as revealed by a wind/DEET interaction. We suggest a combination of
directed wind and volatile repellent might be developed as a mosquito deterrent strategy

for the backyard setting.

Introduction

It is well established that environmental factors significantly impact the ability of
mosquitoes to find and feed upon their hosts. For example, humidities of 70-80 percent%
and temperatures between 37-40°C promote peak activity in female Anopheles
quadrimaculatus (Say) (Platt et al. 1957). Optimal flight temperatures for Aedes aegypti
center around 21°C (Rowley and Graham 1968), while Takken et al. report that rising
humidity is a key stimulus for An. gambiae host seeking (1997b). Wind velocity is also a
major factor impacting mosquito attraction. Bidlingrnayer et al. (1995) documented trap
catch reductions of 50% at ambient wind velocities of 0.5 m/s, and further decreases as
wind velocity approached or exceeded mosquito air speeds. As estimated mosquito flight

speeds in still air range from 04-16 m/s (Bidlingrnayer 1985), the potential for

39

controlling mosquitoes by manipulating wind velocity appears real but perhaps
underappreciated.

Currently, dermal application of chemical repellents is the mainstay of personal
mosquito control. MN—diethyl-m-toluamide (DEET) formulations remain the most
effective and commonly used of the available repellents (Cockcrofi et al. 1998, Fradin
1998). DEET has been extensively reviewed, and Osirnitz and Grothaus (1995)
evaluated its toxic risk as “very low.” Despite high levels of effectiveness and strong
safety record, many people still object to dermal application of (DEET) for protection
from mosquito bites. Oily feel, chemical smell, and fear of toxicity are reasons cited by
those who avoid the products (Fradin 1998). DEET’s functionality as a vapor-phase
repellent, rather than a topical contact repellent, has not been established.

Given the published data on mosquito flight and our experiences with mosquito
behavior, this study was performed to determine whether: 1) artificially generated wind
might effectively deter nuisance mosquitoes, and 2) adding repellent vapor to this moving

air would augment a wind effect.

Materials and Methods

Study site

All trials were conducted in a 1,000 m2 meadow on the Rose Lake Wildlife
Research Area in Clinton County, Michigan (Peacock Rd, 1 km north of 1-69). The test
site was surrounded by trees and shrubs and was less than 100 m from standing and
flowing water sources. Mosquitoes were present and abundant; the most common
species collected during this study were An. punctipennis, Ae. vexans, Ae. stimulans, and

Ae. trivittatus.

40

Spatial layout of test components

The layout of the experiment is illustrated in Figure 4. A 40.6 cm x 50.8 cm air
filter was placed behind a variable-speed electric floor fan and the wind generated by the
fan was aimed at a human subject seated 1.6 m downwind of the fan/ filter combination.
An observer/assistant was present but seated outside of the plume area to assist with data
collection and to monitor for overt reactions to the chemical plume or overexposure to
mosquito bites. Both the human subject and observer wore protective clothing including
mosquito netting headgear to conform to Michigan State University human subjects
committee approval guidelines. The approval also stipulated that the number of human
subjects be highly limited until the value of this approach is assessed by peer review.
Wind generation

Wind was generated using one of two identical Cyclone 25 cm radius, 3-speed
electric floor fan (Lasko, West Chester, PA). An Onan 4ocycle gasoline-powered
generator (Briggs & Stratten; Milwaukee, WI) delivered electrical power (120V, 12.5
amps), and was situated at least 15 m from the test subject. The “low” fan setting was
used; at 1.6 m downwind the velocity was 0.9 i 0.2 m/s SD as determined by a hot-wire
anemometer (Series 471, Dwyer Instruments, Inc.; Michigan City, IN). Separate fans
were used for DEET treatment and the no-repellent “controls.”
Chemical plume generation

MN-diethyl-m-toluamide (DEET) was applied to the El Flow 11 air filter
(Flanders Precisionaire, Inc; St. Petersburg, FL) at 2.4 mg/cmz, for a total of 4.75 g per
filter. Lower dosages were tried initially, but appeared to offer little benefit. The

formulation used was Ben's 100 95% DEET (Tender Corporation; Littleton, New

41

 

42

Figure 4. Experimental setup for human subjects-based wind velocity and vapor-phase DEET experiments.

Hampshire) applied via a finger pump. DEET evaporation rates were calculated
gravimetrically to be 470 ug/cmZ/hr. The filter area was 1,981 cmz, yielding a
maximum loss of 0.91 g/h of DEET. This information in combination with
approximations of plume volume yields an estimation of DEET vapor concentration at
800 ug/L air. This concentration was detectable by human smell, though it was not
objectionable.

Experimental design, data collected, and analysis

A 2 x 2 factorial design was used, with the factors being fan on vs. off and
repellent present vs. absent. Each of the four treatment combinations was tested for 10
min. intervals separated by a 5-min. setup period between treatments. Treatments were
blocked by one-hour periods; no more than two blocks were completed per day. Data
collected were visual counts of numbers of mosquitoes orienting to, landing on, and
probing the subject’s exposed forearms.

Mosquitoes were judged Orienting when <1 m from the subject and directing
flight towards or station-keeping near the subject. Random flight through the test area
occurred rarely. Landing was scored when a mosquito alighted on the subject’s exposed
forearms, and Probing required lowering the proboscis as if to pierce the skin.
Mosquitoes were allowed to freely orient and land, but as soon as they initiated probing
behavior, and usually before biting, they were aspirated into a vial for subsequent
identification. Nine complete blocks of this experiment were accumulated for analysis

between July 1-29, 1999. Data were transformed by (loglo + 0.5) and analyzed by l-way

ANOVA using PROC GLM (SAS Institute 1996). Both main effects and interactions

were evaluated.

43

Results

Mosquito pressure was high for most of the blocks and provided adequate data for
analysis. A maximum of 46 probing mosquitoes was counted in 10 min and the average
was 6.3 i 9.4 SD. Averages for orienting and landing were 13.1 i 18.3 and 8.6 :I: 12.9,
respectively. With both DEET-present and DEET—absent treatments included in the
analysis, turning the fan on significantly reduced the number of mosquitoes orienting (F
= 8.29; d.f. = l; P = 0.008), landing (F = 9.66; d.f. = l; P = 0.005), and probing (F =
14.45; d.f. = 1; P = 0.001) (Figure 5). There was no significant effect of repellent alone
on mosquito orienting (P = 0.18), landing (P = 0.26), or probing (P = 0.48). DEET was
not expected to have a statistical main effect due its inability to envelope and protect the
human subject without the fan creating a vapor plume.

However, there was a significant interaction between the fan and repellent for
orienting (P = 0.050) and landing (P = 0.027) behaviors. The interaction was not
pronounced for the probing behavior (P = 0.061). The mean number of mosquitoes was
lower for the fan-on, repellent-present combination than for the fan—on, repellent-absent
combination for orienting, landing and probing.

Discussion

Administration of DEET repellent to the wind stream generated by an electric fan
was relatively simple to accomplish. While it did not completely eliminate visits from
mosquitoes, fan-generated wind plus DEET did confer appreciable protection at a wind
flow judged “comfortable” by the subject. The significant interaction between the F an
and DEET factors is attributable to the action of the fan delivering DEET vapor to the

human subject. This action also explains the fan’s main effect, which is only apparent

44

35 Dian-off, repellent present
Efan-off, repellent-absent

3° Ifan-on, repellent-absent

25 Ifan-on, repellent-present

Mean number of mosquitoes +I-
Standard error

 

orienting landing probing

Figure 5: Number of mosquitoes observed orienting, landing, and probing under
four fan and repellent combinations. Means followed by the same letter within a be-
havioral category are not significantly different at P < 0.05.

45

when there is DEET in the wind stream (Figure 5). Although the fan-on, DEET-absent
treatment was not significantly different from the fan-off, repellent-absent treatment in
this experiment, preliminary tests of higher fan velocities strongly suggested that higher
winds without DEET conferred substantial protection from mosquitoes.

Wind likely has at least three impacts on mosquito host seeking efficiency. First,
increased wind velocity reduces the absolute rate of mosquito recruitment to an upwind
host (Bidlingrnayer et al. 1995). A second impact is based on optomotor theory
(Kennedy 1939, Klassen and Hocking 1964, Snow 1980, Gillies and Wilkes 1981) and
suggests that mosquitoes have preferred rates of ground passage and will cease flying if
perceived ground speed drops below tolerable rates; wind has a direct effect. A third
possible impact of wind is dilution of host-stimulus chemicals. A given rate of

emanating host volatiles (such as C02 and lactic acid) would become less stimulatory as

wind velocity increased because the concentration, and thus dosage delivered to
chemosensory organs, would fall. This experiment did not seek to quantify the
contribution of each of these impacts, and these models are addressed experimentally in
Chapter IV.

Chemical suitability is an issue to be addressed for this as well as any repellent
studies. Suitability is based on several factors, not least of which is the toxicity profile of
the chemical. While DEET has an admirable safety record (Robbins and Chemiak 1986,
Osimitz and Grothaus 1995, Goodyer and Behrens 1998), careful attention must be paid
to dosage. We estimated maximum uptake for a 10-minute exposure in this study could
have reached 100 mg if the subject, inhaling 13 liters of plume air per minute, absorbed

all of the volatized chemical inhaled. By comparison, Robbins and Chemiak (1986)

46

analyzed available DEET application data and estimated that an Everglades park worker
averages 4.25 g of derrnally applied DEET per day, while the average DEET user in the
general population was found to apply 1.65 g/d. Given a mean cutaneous absorption rate
of 5.6% (Osimitz and Grothaus 1995), the worker would be absorbing 238 mg of DEET
per day and an average user 92 mg/d. The rate of DEET used in the current study was
high compared to these measures. If this wind-vaporized DEET use were to be pursued,
the actual absorbed dosage would need to be quantified beyond these conservative
estimates; it is judged likely that the subject of this study absorbed only a fraction of the
inhaled DEET. The toxicity profile for inhaled chemical should also be determined. It is
possible that there are different physiological effects of DEET vapor compared to dermal
applications.

While DEET may not be the most suitable fan-propelled vapor-phase repellent,
we suggest that this strategy has a place in personal protection from mosquitoes. Electric
fans are readily available, comfortable, and could serve a dual purpose of cooling and

pest deterrence on warm evenings. Tests of this nature should be expanded to other

repellents, including the piperidine compound Autan® (Bayer AG) and some of the
essential oils. Benefits of fan-delivered chemicals must be carefully scrutinized in the

light of toxicity profiles.

47

CHAPTER IV
Artificially Manipulated Wind and Vapor—Phase DEET Reduce Mosquito

(Diptera: Culicidae) Catches in CDC Light Traps

Abstract:
In a central Michigan wetland setting, artificially manipulated wind velocities reduced

mosquito catches in CDC light traps with C02 released at 650 or 1,950 ml/min. This

function was negatively logarithmic over the range of velocities tested (0 m/s to 3.7 m/s
or 0 to 8.3 mph). This reduction closely matches expected the dilution function as wind

increases over a constant release source. Tripling the C02 release rate consistently

doubled the number of mosquitoes caught across the range of wind velocities tested.

Addition of DEET vapor (910 mg/h) at 650 ml/min C02 significantly (P < 0.001)

reduced the number of mosquitoes caught relative to equivalently moving air without
repellent. Analysis of responses to wind by Aedes vexans, Anopheles walkeri, and
Coquillettidia perturbans does not support the hypothesis that different species respond
differently to wind. There was no relationship between mosquito body mass and either

applied wind velocity (P = 0.25) or C02 release rate (P = 0.7836) across the range of

wind velocities tested. Collectively, our data suggest that applied wind diminishes
mosquito response primarily by diluting attractants rather than by exceeding the flight
capability of mosquitoes. Outdoor fans as well as horizontal or vertical wind curtains are

proposed as local deterrent tactics for nuisance mosquitoes in the backyard setting.

48

Introduction

Environmental factors such as temperature, humidity, and wind strongly influence
mosquito activity levels (Thomson 1938, Kennedy 1939, Platt et a1. 1957, Lewis and
Taylor 1967). Temperature and humidity effects have received considerably more
research attention than has wind. Mosquitoes rely heavily on wind-generated, host-odor
plumes to find blood meals, but velocities greater than 03-10 m/s can reduce successful
orientations supposedly because of mosquitoes’ weak flight capabilities (Gillies and
Wilkes 1981, Bidlingrnayer 1985, Bidlingrnayer et al. 1985, Bidlingrnayer et al. 1995).
The highest wind velocities at which orientation to hosts or flowers has been documented
range from 0.8 to 2.8 m/s for subarctic species (Grimstad and DeFoliart 1975,
Bidlingrnayer et al. 1995). An arctic species, Aedes cantons, has been observed station-
keeping in winds up to a remarkable 8 m/s (Nielsen 1953). When wind velocities
increase to levels above optimal ground speeds (Kennedy 1939), mosquitoes cease flying
or fly closer to the ground and take advantage of the boundary layer, where wind
velocities are substantially lower (Vogel 1981, Snow 1982).

Information is scant for mosquito flight responses to manipulated wind velocities
under field conditions. Using a live calf in an open-ended wind tunnel, Gillies and
Wilkes (1981) found that catches of Mansonia and Anopheles mosquitoes on the African
savannah began to decline at wind speeds greater than 0.5 m/s and nearly ceased above
1.4 m/s. This study neither precisely quantified the relationship between wind velocity
and mosquito response, nor fully explored the potential of artificially-generated wind as a

tool for deterring or otherwise manipulating mosquito behavior.

49

The intent of this study was to use electric fans to manipulate wind velocity and
assess the practical potential of wind as a protectant from mosquitoes. These studies used

carbon dioxide (C02) as an attractant and, in one test series, DEET was added to the
wind. C02 is one of the best-known chemical cues mediating mosquito host-finding

(Gillies 1980, Takken 1991, Takken and Knols 1999). It is suspected to be both a

behavioral activator and a true kairomone (Gillies 1980). C02 alone is effective as bait

in behavioral experiments and for population monitoring (Reeves 1953, McIver and

McElligott 1989, Mboera et al. 1997); human equivalencies (ca. 250 ml/min C02) are

relatively easy to obtain. The effective active space (linear distance from release point)

of C02 increases linearly with increasing release rates up to 1000 mein (Gillies 1980).

MN-diethyl-m-toluamide (DEET) is a widely used and potent topical chemical
used in the management of the biting behaviors of mosquitoes and other human-biting
insects. In addition to its repellent value in topical applications, there is evidence that it
functions in the vapor phase. DEET-impregnated gauze in an olfactometer was sufficient
to significantly inhibit mosquitoes fiom orienting to a lactic acid bait (Dogan et al. 1999).

DEET inhibits neurons sensitive to lactic acid, another important attractant
chemical for mosquitoes (Davis and Sokolove 1976). By inhibiting lactic acid
perception, DEET may break the suggested chain of chemically-mediated behaviors that
lead to a successful blood meal (Davis 1985, Dogan et al. 1999). It is unclear whether
DEET acts on other sensory structures in addition to lactic acid sensors.

While mosquitoes as a group are clearly influenced by environmental factors and
physiological status, it is thought likely that there is intraspecific variation in preferred

flight speeds (Snow 1980), responses to temperature (Wright and Knight 1966), and C02

50

dosages (Reeves 1953, McIver and McElligott 1989, Constantini et al. 1996, Dekker and
Takken 1998).
The specific objectives of the current study were to: 1) quantify mosquito

response to fan-manipulated wind velocity and two release rates of C02, 2) determine

whether wind-reduction of mosquito catch could be enhanced by vapor-phase DEET, 3)

determine whether different species of mosquito have different response profiles to C02

and wind velocity and, 4) attempt to differentiate between the flight-limit and dilution-

based models as explanations for the mode of wind action upon mosquito orientation.

Materials and Methods.
Study site and trapping method

Experiments were conducted between June 3 and September 9, 2000, in the center
of an open grass field (ca. 5 ha) at the Rose Lake Wildlife Research Area in Clinton
County, Michigan. The experimental area, adjacent to two remote storage buildings, was
bordered by woodland and less than 1,600 m from standing water habitats. These

habitats ranged from small woodland pools less than 3 In in diameter to a large lake in

excess of 2.5 kmz. Mosquito populations were high throughout the season. Moderate
temperatures and consistent rains kept mosquito breeding sites productive. Adults were
actively host seeking throughout the experimental area, as evidenced by persistent
attempts to bloodfeed from the investigators during all tests.

Because of the impracticality of deploying equivalently attractive humans (Price
et al. 1979, Schreck et a1. 1990) for prolonged field tests, we used Centers for Disease

Control (CDC) light traps (John W. Hock; Gainesville, FL) with and without C02 bait to

quantify mosquito orientation success. Steel support poles held" the traps 1.5 m above

51

ground height in two parallel rows with approximately 12 m between each trap. Each
trap intake fan and 1.4 W light (when used) was powered by a 12-volt rechargeable gel-
cell battery.
Carbon dioxide release

For treatments using host odors, a thermos (ca. 2L) containing 600 g dry ice was

inverted above the CDC light trap. As the dry ice sublimed, C02 gas spilled from the
spigot. The release rate of C02 was determined gravimetrically afier placing

approximately 600 g of dry ice into each of three different therrnoses for six h, the
planned field experimental interval.
Wind generation and velocity determination

Artificially-generated wind was produced by 51 cm-diam, three-speed floor fans
(Model WC-2000, Lakewood Engineering & Mfg. Co.; Chicago, IL). Desired at-trap
wind velocities were obtained by aiming fans at differing power settings from various
distances. Fans powered from 120-volt outlets in the adjacent storage buildings were
mounted on 0.6 m stepladders for stability and support. 16—Gauge, outdoor-use,
extension cords carried power from multiple outlets to the fans with little voltage drop.

Average wind velocities for various power settings and distances were determined
by aiming the fan at a circular (43 cm radius) string grid with nine intersections (Figure
6). Wind velocities were taken at each intersection and the four cardinal points with a
hot-wire anemometer (Model HHF5 1 , Omega Engineering, Inc.; Stamford, CT) mounted
on a heavy video tripod for stability and averaged to obtain the mean velocity for a given

distance and power setting.

52

C 43 cm

\. °\

 

 

 

 

 

 

 

Figure 6: Experimental setup for wind velocity determination. A) Three-speed
fan on step ladder mount, B) Heavy-duty video tripod, C) Hot-wire anemometer
c1) extension arm and c2) sensor, D) Circular target with string grid.

53

DEE T Generation

DEET vapor was generated by spraying 4.3 g technical grade DEET (Fluka;
Milwaukee, WI) onto a 38 cm x 51 cm E-Z Flow fumace filter (Flanders Precisionaire,
Inc; St. Petersburg, FL) and affixing it to the back of fans aimed at CDC traps. DEET
was vaporized as a fan drew air through the impregnated filter and propelled it toward the
trap. The combination of fan, light trap, thermos, and DEET screen (when used) was
termed a trapping station.

Each nightly trap run (block) lasted from ca.1730- 2330 EDT. Runs were set up
only when ambient winds were non-existent to <1 m/s and predicted to remain so. It was
accepted that transient breezes above 1 nr/s might occur. These were viewed as affecting
the performance of those traps with the lowest applied velocities for only short times
during an otherwise acceptable experimental period. Temperatures were above 10°C for
all trapping periods. During experimental periods, I was attentive to weather conditions
in the test vicinity to assure appropriate experimental parameters were maintained. On
three occasions, tests were abandoned and results disqualified because of excessive wind
and/or rain. Trap collection bags were immediately sealed when tests were terminated
and trapped mosquitoes frozen at — 40°C within 8 h of collection. Counting and
taxonomic determination were performed under a dissecting microscope.

Statistical Analysis and Experimental Designs

Quantification of mosquito response to wind, C02, and DEE T

To correct for variance-mean dependence, count data were transformed using a

log10 (number of mosquitoes caught + 0.5) before analysis. ANOVA was performed

54

using PROC GLM in SAS (V8) and regression statistics were generated by ANCOVA
using PROC GLM (SAS Institute 1996).

Four winflelocities with and without C02 (Experiment 1)

Eight trapping stations were arranged in two parallel rows of four. The
experimental design was a blocked, randomized 4 x 2 factorial design with applied wind

velocity and C02 release rate as the factors. The experiment was blocked by night to

account for uncontrolled effects of varying temperature, humidity, and mosquito

population density. Applied wind velocities were 0, 1.3, 1.6, or 2.4 m/s. C02 release

rates were 0 or 650 ml/min. Trapping stations were in the same location each night, but

the treatment combination assignments to individual stations were randomized before

each block. Eight blocks were completed.

_A_ssessment of mosquito background densitiesgnd high velocity response (Experiment 2)
The spatial arrangement was similar to Experiment 1 except that only seven

stations were used. Instead of a factorial design, a randomized complete block design

was employed with five applied wind velocities at 650 ml/min C02 and two wind

velocities with unbaited (no light or odor stimulus) traps. Baited wind velocities were 0,
1.1, 2.0, 2.5, or 3.7 m/s, and unbaited velocities were 0 or 3.7 m/s. Eight blocks were
completed. Unbaited traps were used as negative controls to measure the background
level of mosquitoes flying through the experimental area. While the unbaited traps
caught non-orienting mosquitoes, the baited traps sampled actively orienting females.
Baited trapping stations with mosquito catches at or below those of the unbaited traps
imply that the applied wind treatments for these traps negated the attractants. This is a

reasonable measure of the degree of personal protection afforded by a particular

55

treatment, since it establishes the number of mosquitoes in the environment that would be
encountered by chance alone.

Five wind velocities and two C02 release rates (Experiment 3)

This experiment used 10 stations in two parallel rows of five. The design was a

blocked 5 x 2 factorial with applied wind velocity and C02 release rate as factors. The
wind velocities applied were 0, 1.0, 1.1, 2.1, or 2.7 m/s. C02 release rates were 650 or

1,950 mein. The higher release rate was achieved by hanging three therrnoses from the
support pole. Nine blocks were completed.
Five wind velocities withfiand without DEET vapor (Experiment 4)

The test of DEET vapor used 10 stations in two parallel rows of five. A blocked,
5 x 2 factorial design was used with applied wind velocity and DEET vapor as factors.
The wind velocities applied were 0, 0.3, 0.4, 1.4, or 2.4 m/s. The slight difference
between these velocities and those of Experiment 2 resulted from addition of the fumace
filters to the back of the fan. DEET was either applied to the filter attached to the back of
the fan, or a filter was left untreated. Fans used for DEET dispersion were never used for
non-DEET treatments. DEET-present and no applied wind treatment was created by
deploying a DEET-impregnated filter 2 m from the CDC light trap. All of the trapping

stations were baited with light and 650 mein C02. Seven blocks of this experiment

were completed.

Comparison of species responses to wind velocity and C O 2
Variation in mosquito responses was assessed using loglo- transformed counts

from Experiment 3 after species identifications. Regressed responses to wind were

compared using PROC GLM AN COVA in SAS. Only slopes were compared, since the

56

intercepts would vary with species’ population densities, which were unknown and not
likely to be equivalent.
Relationship between mosquito mass and wind velocity

Using mean mosquito mass/ trapping station samples from Experiments 2 and 3,
ANOVA procedures in PROC GLM of SAS (V8) (SAS Institute 1996) were used to

determine whether mosquito responses to wind velocity and C02 were influenced by

mosquito mass. Regressed responses to wind were compared using PROC GLM

ANCOVA in SAS.

Results

Carbon dioxide release
The average loss of C02 was 74.4 g/h (range of 67.3 to 81.8 g/h) (Figure 7). R2-

values exceeded 0.99 for each thermos, confirming a linear (thus constant) release rate
over the chosen time interval. Meeting the criterion of constant release rate vastly
simplified interpretations of experimental results.

The volume of C02 corresponding to this loss can be calculated with the
relationship PV = nRT: where P is the air pressure (atrn), V is the volume of C02 (L), n
is the moles of C02 (1.69 mol), R is the constant 0.0821 L atm/ K mol, and T is the
temperature (K). The air temperature in the laboratory was measured as 22°C (294K) and
the pressure was approximated at 1 atrn. The calculated mean volume of C02 lost was
680 ml/min for the three containers (range 615 to 742 ml/min C02). Since the

atmospheric pressure was approximated, it was deemed appropriate to approximate the

release rate of an individual thermos as 650 ml/min C02.

57

 

500

 

 

 

450 o Thermos 1 Mass lost (9) = 81.8 g/h ' (h) o
a 400 . n Thermos 2 Mass lost (9) = 74.2 g/h ' (h) . U
,3 350 - A Thermos 3 Mass lost (9) = 67-3 slh " (h) , ‘
a 300 - ' - '
E 250 - . .
(E 200 - '
'5' 150 -
5 100 - '
U 50 - a -
0 . . f t 1
0.00 1.00 2.00 3.00 4.00 5.00 6.00
Elapsed Time (h)

Figure 7: Cumulative mass loss of C02 due to sublimation in three thermos contain-
ers at 22°C.

58

Mosquito Responses to Wind, C 0 2, and DEE T
Four wind velocities with and without C02: Experiment 1

Addition of C02 significantly increased the number of mosquitoes caught in the
light traps (Figure 8, Figure 9) (F = 289.43; d.f. = 1; P < 0.0001). Conversely, applied
wind velocity significantly reduced the number of mosquitoes caught (F = 5.76; d.f. = 3;

P = 0.020). The log-transformed regression equation (Figure 8) for the COz-baited traps
was log10 (catch + 0.5) = - 0.44 (wind velocity (m/s)) + 2.93 and there was a significant
interaction between wind velocity and C02 (F = 3.80; d.f. = 3; P = 0.016) on the number

of mosquitoes captured. The slope of the linear relationship between wind velocity and

log-transformed trap catch was significant (t = -5.46; d.f. = 1; P = 0.001) for the C02-

baited data but not for the unbaited traps (P = 0.691). Untransformed nightly means with

a best-fit logarithmic curve are shown in Figure 9.

Assessment of mosquito bac_kground den_sities agd high velocitv responie f Experiment 2)
The slope for log-transformed mosquito trap catch vs. wind velocity was

significantly different fi'om zero (t = -9.63; d.f. = l; P = 0.001) for those traps employing

650 ml/min C02 as a stimulatory odor (Figure 10). The equation for this relationship is
log10 (catch + 0.5) = - 0.67 (wind velocity (m/s)) + 2.45 with an x-intercept of 3.7. There

was no significant difference between the baited and unbaited traps at 3.75 m/s (Tukey’s
HSD: P = 0.05); at this velocity, mosquito catch was reduced to background levels.
Unbaited traps (0 and 3.7 5 m/s) were also not significantly different from one another,
which supports the assumption that higher wind velocity did not measurably impair trap
efficacy. Untransforrned nightly means with a best-fit logarithmic curve are shown in

Figure 11.

59

35 v k we Leakage 8: 03 8:2
:oEEoo a wetmnm can we: a8 05 mummies—co $808 13239: 256365 35; we? :28 .329: owfio>< 5623 CE? .«o

5:05 a we 2 808535 3385 Nov 3055 pee 5m? BEES 938% 355803 39332: page.“ we LBEZ ”w Semi

£5 base» 255 8:24
o.m 9N o.~ m... 04 md

- r n n n p

o_
o

 

9.
o

‘0.
O
as -/+ (9'0 + tramp ounbsow) 01301

 

A a

owl

o_
,_

I
‘0.
‘—

oowt

l
o.
N

 

Mosquito Catch
J.
61

~00 ESE: o D

 

coo—.1

o_
on

N00 5:58 owe O

 

 

l
‘0.
or)

60

doses—fl Home Epsom e 33.52:: on: peace 2:. 5:022, .653 we

3.5 £83» 255 3:93

 

 

mam N 3 P no o
In . 0.1.0 . . U o
Trill-..
lift/m . com
tili/lt/r/ T DOV
x ..... / . coo
, . o8
.. 83
¥ some
Noo Eases o e
. 83
N8 Ease o3 D
.. o8.

 

 

 

~53er a me 2 825338 2:285 NOD 82E? can 55, Cebuano Amomooam poems—=03 393338 0380.“ me 39:52 ”a 2:me

as 'I+ “0190
orgnbsow Anqfim ueew

61

.36 v m 3 “Santa. 8: 0.63 coau—
coEEoo a mam—mi 2:88 3:229: £13239: com me: :88 EEME omEo>< .3623 9:3 me .8323.“ a 3 AN Eon—toga

:3 Em: .5 Bee o: 53, .8 magnum N00 .23 B538 @66on 3585.3 323688 038$ .«o c3832 ”2 onE

3.5 ESE 255 E39...
o.v 0m oh 9N 0N mé o; md o.o

       
     

 

 

 

3mm 0 o "01

.. an

. md m

to m

o7 . o; m

.w b

MI“. I . m.

C m P m

o 3

m 2:. . 3. m

m. u.
0 .

M E: 8 .Noo easeo D at m m .M

S

831 60 Ease o8 o . o.m (h

./

. S

m m H

 

62

doze—Em >83 355m a 9:823.— 25 333 2:. .5622,

2:3 .«o 5:83 a 8 Am EoEtomxm—v 3.25% N00 :33 vegan”. 60.30on 35933 823308 Bafia mo “3:52 H: 8:me

3.5 £83» 35» 8:9?

 

 

v m.m m mN N mé _. md 0
Fl]. . 51,3338! . . . n 0
viii.
.. X, . 2: M
III! E
Ill 3 u
/, ... com 2 m
a/ ms
[/1 u m
11 com H .M
I n
m... m
, . 8
x1. oov .m
60 case o8 O ,, n...
.. com 0

 

 

 

com

63

Five wind velocities and two C02 release rates (Experiment 3)
Tripling the C02 release rate significantly and consistently increased (F = 15.79;

d.f. = 1; P < 0.0002) while increasing wind velocity significantly decreased (F = 65.56;

d.f. = 4; P < 0.0001) the number of mosquitoes caught under either C02 regime (Figure

12). There was no significant interaction between these two factors. Regression analysis
yielded no significant difference in either the slope (P = 0.21) or y-intercept (P = 0.98)

for the two C02 release rates. The equation for the 650 ml/min C02 release rate was:
loglo (catch + 0.5) = - 0.67 (wind velocity (tn/5)) + 2.13. This equation yields an x-

intercept of 3.2, which is interpreted as the wind velocity (m/s) at which the trap catch is

reduced to zero. The 1,950 ml/min C02 release rate yielded an equation of loglo (catch +

0.5) = - 0.67 (wind velocity (m/s)) + 2.45 with an x-intercept of 3.7.
Five wind velocities with and without DEET vapor (Experiment 4)

As in the previous experiments, increasing wind velocity significantly reduced the
number of mosquitoes caught (F = 39.81; d.f. = 4; P < 0.0001), as did the addition of
DEET vapor (F = 12.15; d.f. = 1; P = 0.001). The interaction between the two factors
was nearly significant (P = 0.0563). The slopes for each of the treatment groups (Figure
13) were significantly different fi'om zero (t = -5.91; d.f. = 1; P < 0.0001) but not one
another (P = 0.1595). The y-intercept of the DEET treatment regression was not
significantly different from that of the no-DEET line (P = 0.16). The regression equation

for the no-DEET response is: log10 (catch + 0.5) = - 0.49 (wind velocity (m/s)) + 1.96
with an x-intercept of 4.0. When DEET was present, the equation was: log10 (catch + 0.5)

= - 0.73 (wind velocity (tn/5)) + 1.88 with an x-intercept of 2.6.

64

.mod v m 3 Event? basemfiwmm 8: 2a .532 .8588
m wet—Em 98 0:: 53m a wcmmmafieo 2808 :32th $1333.65 m5 m9» :88 3E3: om80>< .5622, 2:? me 8:05
a 3 Am 808535 .888 0322 Nov Eo..o&% 95 an @0598 60.28% 85958 323688 295a .«o 52:52 ”E 0.5mm”—

35 £83» 255 8:93.

 

 

... m.N
NOD 233$ 36 .

coo? 1 Nov case 82 4 . o.m

 

 

 

O.m m.N O.N m.—. 0.? m6 o.o

J- D D b b b ooo
. .m
4 a .3 m
m
or- .c 4 a .oe m
.w as b
t n
a . .3 W.
0 4 4 $3 3
.m 87 3 m
m. a u.
o +
M .0
(S\
+
./
S
3

m.m

65

.36 v R an EBabE bygoamcwfi 8a an $52 :2:
.88 a wgnm 0:: 52w 5 mammtafioo 282: :32sz £132sz ooa 33 :23 baa—mm: owEo>< .3622, 9:3 mo 5:03
a mu Av 268:3me Emma awash—$3? :55; can .33 @2393 $30on 3:580“; 883688 038$ no 69852 ”2 053”—

3.5 £83» .53 3:9?
oh mN o.~ me o; md o.o

 

or:

 

on:L

 

Mosquito Catch

83 .. $3 + N8 5:35 e3 0 . 9m
60 Ease o8 O

 

 

as -/+ (9'0 + Imus onnbsow) 0130']

 

md

66

Comparison of Species Responses to Wind Velocity and C 0 2

Sixteen species were present in the light trap samples for Experiment 3 (Table 1).
Voucher specimens are deposited in the Entomology Museum, Michigan State University
(Appendix 1). Collectively, An. walkeri, Ae. vexans and Coquillettidia perturbans made
up over 92% of the identified individuals. I considered these three species sufficiently
abundant to permit statistical analysis of response patterns on an individual species basis.
The other species had either very low counts (1-5 per block) or were present in too few
blocks for appropriate analysis. Figures 14-16 illustrate the responses to applied wind

and C02 for each of these three species. Within-species comparisons of the slope
relationships for the two C02 release rates were not significantly different for An. walkeri

(P = 0.577), Ae. vexans (P = 0.746), or Cq. perturbans (P = 0.275). Between-species

comparisons of linear-response slopes at each C02 release rate also yielded no significant

differences.
Mosquito Mass and Wind Velocity Relationship
In Experiment 3, mass did not significantly influence (Figure 17A) mosquito

response to C02 (P = 0.783) or applied wind velocity (P = 0.22). The slopes of the
regressions (not shown in Figure 17) for either C02 release rate were not significantly

different from zero (P = 0.2545) (Figure 17A). Likewise, There was no significant
relationship between DEET and mosquito mass (P = 0.13) or applied wind velocity and
mosquito mass (P = 0.4409) in Experiment 3 (Figure 173). The slopes of the regression
were not significantly different from zero (P = 0.437). Some error may have been

introduced by damage to the mosquitoes in the handling stages of counting and

67

Table 1. Mosquito species composition for Experiment 3

 

 

Mosquito Genus/ Species % of Total Mosquitoes Captured 1
Aedes
canadensis (Theobald) 0.09
cinareus (Meigen) 0.01
sticticus (Meigen) 0.03
stimulans (Walker) 0.48
trivittatus (Coquillett) 0.20
vexans (Meigen) 35.35
Anopheles
punctipennis (Say) 1.70
quadrimaculatus (Say) 2.78
walkeri (Theobald) 39.10
Culex
pip/rest 2 0'28
salinarius (Coquillett) 0-34
Culiseta
inornata (Williston) 0-15
melanura (Coquillett) 0-01
morsitans (Theobald) 0-29
Coquillettidia pertubans (Walker) ”-34
Damaged 1.85

 

1. Total number of mosquitoes was 7,919.
2. Culex pipiens (Linneus) and Cx. restuans (Theobald) are best distinguished by egg
raft type. We lacked this information, and these species were grouped together.

68

 

7‘ .- ¢
i

5623 we?» we
:285: a me Q 508:3me 83.. ammo—2 NOD 389$“. 95 8 3593 thS: .2V no 883308 oREoM .«o eons—=2 ”3 oEmE

3.5 £85» 255 333.

m md N m... _. md o

I D D D I

 

°.
0

I
‘9.
O

U
0.
P

 

oT .

Mosquito Catch

I
0.
N

8 T Ncu case o8 9
Nov 5:58 82 <

 

as -/+ (9'0 + “9193 onnbsow) 01301

 

 

‘0.
N

69

.5323 9:3
mo eouoeé m 8 Am geoEton—xmv moan.— 0328 Nov 2.20%? 025 3 @0530 338: .mV .«o $33308 0380.“ no H3852 ”2 onE

@E E83» 255 82.5.4.
m ON N m. _. v md o

_ b p L p -

 

0.
o

t . no

0
F
l
I
O
‘—

 

fi
‘0.
‘—

Mosquito Catch
4
4

, .
q
N

o9 1 N8 55:6 one 0
Nov 58:8 $2 4

 

as -/+ (9'0 + Imus ounbsow) 01301

 

 

‘0.
N

70

5623 we? .3 so:
.33 a me An Eofitoquv magma ammo—B NOD Heouobmv 625 an BEES 28?:th $9 no moose—omega 038$ me 89:52 ”3 onE

 

3.5 563$ 2:3 8:92 3.
m ¢ on N 3 2 no
r .1 p L p p 0.0

/

I
‘0.
0

OF... 4

'7
O
F

 

I
'0.
‘-

 

4

o_
N

8 7 N8 55%. o8 0
N8 58:8 82 4

 

 

Mosquito Catch
.0. .
N
as -/+ (9'0 + llama ounbsow) ("301

71

1.4-

 

 

 

 

 

 

1.2 - A .
1 - o
A 8
0.8 EE Q g
0.6 :
’53 0-4 o 650 ml/min C02
E 9
2.; 0.2 A 1950 ml/min 002
g 0 I I I U - ‘ ‘ww-ijhmww '- I U
,3 o 0.5 1 1.5 2 2.5 3
3-
8 1
E 0.9 B O
a” 0.8 0
3 0.7 o
< 0.6 g 8 g
0.5
0.4 1 .. M, ~_,_,___,p_,__,,_ .
3? o 650 ml/min €02
. ‘l '
0.1 q . 0 650 ml/min C02+DEET ?
o . . Q . f w p
o 0.5 1 1.5 2 2.5 3

Applied Wind Velocity

Figure 17 A and B: Scatter plots of mosquito masses as a function of applied wind

velocity A) with two C02 release rates (Experiment 3), B) with two DEET release
rates (Experiment 4).

72

transferring to storage containers. However, there did not appear to be a difference in
damage intensity between larger and smaller specimens that would negate an authentic
relationship between body mass and flight capacity.

Discussion
As previously established for ambient winds (Wright and Knight 1966, Snow

1980, Bidlingrnayer 1985, Bidlingrnayer et a1. 1995), wind velocity as manipulated in the
current study strongly reduced mosquito orientation. Moreover, my collective data allow
differentiation between two possible modes of wind action: dilution vs. flight limitation.
Despite the broad acceptance that mosquito orientation is limited by flight capacity, the
data herein support only dilution, based on the following lines of evidence:
Mosquito Mass and Wind Velocity Relationship

If the flight-limit model were in play under my test conditions, a positive
correlation of body mass would be expected vs. wind velocity, as supported by Lewis and
Taylor’s (1967) analysis for flying insects generally. As a robust 14,173 mosquitoes
generated this outcome, it seems unlikely that a real effect was missed. The lack of mass
effect is an outcome consistent with the stimulus dilution model, where response would
be shaped by stimulatory factors, which are independent of body size.
Negative Logarithmic Response

The relationship between trap catch and wind velocity was negatively logarithmic
for all baited treatments (Figures 8 - 12). This is to be expected when viewed in the light
of a dilution-mediated model of mosquito host seeking. The concentration of a chemical
emitted into air from a constantly-releasing point source is readily calculable. It
diminishes geometrically with wind velocity; each doubling of wind velocity halves the

concentration of the chemical. The general form of this equation is y = l/x where y is the

73

stimulus concentration at a given wind velocity (x). Additional dispersion of the
chemical due to turbulence and diffusion would cause the concentration to fall faster than
the theoretical values of this equation, and the resulting relationship would appear more
like a logarithmic function.

Patterns in my data seem completely explained by stimulus dilution. Trap catches
consistently decayed more than geometrically over all applied wind velocities, and, there
was no evidence of response truncation at the highest wind velocities. Such a truncation
would be expected if mosquitoes were approaching their flight limit. Likewise, dilution
effects appear to explain a preponderance of the negative logarithmic response
documented in previous studies involving ambient winds (Bidlingrnayer et al. 1995).

Parallel Responses for Two C02 release rates
The remarkably parallel responses to wind velocity with two C02 release rates
(Figure 12) further support the dilution model. Tripling the release rate of C02

consistently doubled the number of mosquitoes caught for all applied winds (up to 2.7
m/s) with no evidence of truncation at the highest velocity. If these velocities approached
a flight limit, the responses for the high and low release rates should have converged at
the higher wind velocities.
Comparison of Species Responses

As evidenced by the uniformity of the log-transformed plots of trap catch, most of
the mosquito species in this study responded similarly to applied wind velocity.
However, the general response pattern of Cq. perturbans did differ from that of either Ae.

vexans or An. walkeri. Whereas the latter two species had very linear (and remarkably

74

 

 

parallel) responses to the two C02 release rates, Cq. perturbans’ responses for the two

release rates converged between 2 and 2.5 m/s.

This response could be viewed as evidence for a physical flight limit for this
species at ca. 2.5 m/s with responses below this limit being driven by stimulus
concentration.

These data for mosquito mass, logarithmic response, and response to elevated

release rates of C02 lead us to conclude that general mosquito response to hosts is not

driven by flight limits and preferred optomotor response, but rather by the stimulus
concentration that results from wind- and distance-mediated dilution. It is quite possible
that the mosquito species encountered here could successfully host seek at much higher
velocities than those experimentally applied, provided that the instantaneous

concentrations of C02 remained elevated and in the physiologically active range.

Previous flight estimates of mosquitoes have all been taken from unbaited suction traps
or experiments with one release rate. Maximum flight estimates in the literature of 1-2
m/s may widely underestimate these insects’ flight capacity. More accurate mosquito
flight capacities could be experimentally determined using protocols similar to those

employed in this study, but using higher C02 release rates and correspondingly increased

wind velocities.
DEE T and Lactic Acid

Since lactic acid was not included in my experiments, DEET’s effect in reducing
trap catch (Figure 13) contradicts previous conclusions that DEET’s repellent effect is
derived solely from inhibition of lactic acid detection (Davis and Bowen 1994, Dogan et

al. 1999). DEET is known to have physiological effects on lactic acid-stimulated neurons

75

(Davis and Sokolove 1976, Davis 1985, Davis et al. 1987), but this chemical’s action
appears to be more complex than eliciting host-seeking insensitivity to lactic acid. The
physiological point of action may remain the lactic acid sensor (and/or other sensory
organs), but the behavioral response might be decreased upwind displacement or
inhibition of landing behavior. These effects are consistent with the increased turn angles
for mosquito flight paths quantified for DEET vs. DEBT-free stimulus plumes (Davis and
Bowen 1994).

Potential Practical Applications of These Findings

Artificial wind either alone or in conjunction with vapor-phase DEET shows
promise as a means of personal protection from mosquitoes. Chapter 3 established this
effect for a human subject, and the current study demonstrated the reductive potential of
wind as a function of the applied wind velocity. For all mosquito species encountered in
this study, wind velocity alone uniformly and dramatically reduced captures of
mosquitoes in traps. Such generalized effect across species is encouraging since, in the
backyard setting, some 10 different species could be present at a given time in some
locations. Fortunately, wind velocities substantially reducing mosquito orientation are
comfortable for people. Thus, wind alone fiom an out-door fan could be a useful tool in
non-chemical protection from these pests.

By employing a coordinated array of fans at ground height, it may be possible to
establish a zone of protection (Figure 18) larger and thus more functional than that
offered by a single fan. Arrangements like that of Figure 18, where the wind stream
would not be aimed directly at the target of protection, would permit use of wind

velocities in excess of those judged comfortable by humans. Deflection and dilution of

76

 

   
    
    

Ocillating
Fan
Protection

 

Figure 18: Vertical perspective of an area of protection generated by multiple fans
collectively forming a horizontal wind curtain.

“News

é . ‘ Area of protection

IIIIIIIIIIIHKG>
*IHIIIIIIIIIEG
(IIIIHHIIHES
tlllllllllllll:
'IIIIIIIIIIIE'EG
‘llllllllllllfl
‘llllllllllll'fl

1‘ = Wind directed upward

Figure 19: Side projection of an area of protection generated by a border of vertically
oriented wind ports forming a vertical wind curtain.

77

odor plumes by these crosswinds from appropriate fans would likely present a strong
challenge to orienting mosquitoes. It has been suggested that mosquitoes host seek
primarily within the boundary layer (Snow 1979), and that they reduce flight altitude in
the face of high wind velocities (Gillett 1979). Both of these behaviors would be
exploited by a horizontal wind curtain, as mosquitoes would naturally be drawn to lower
altitudes rather than flying over the barrier. With host-odor directionality disrupted, a
mosquito at the barrier threshold would probably lose contact with the plume entirely. A
horizontal wind curtain might be effective without vapor-phase repellent.

Vertical wind curtains might also be feasible. These could be generated using
multiple fans aimed upward or upward-vented air from a single large fan (Figure 19).
Mosquito reduction via this method would stem largely from the wall-like nature of the
wind curtain, dilution of host stimuli, and the loss of the odor plume’s directionality as
attractant chemicals pass into the upward air currents. Admittedly, the mosquito
populations would need to be high and/or dangerous to justify the likely expense of a
vertical wind curtain.

Improved release of vapor-phase repellent is also fertile ground for additional
investigation. An automated-release system that dosed vapor into the wind plume would
be far more efficient than the primitive release method described in this study. I did
attempt to use Michigan State University’s MicrosprayerTM (Isaacs et al. 1999)
technology to deploy vapor phase DEET, but the dispensed aerosol was not sufficiently
fine to stay suspended and envelope a target area in vapor. However, Microsprayer—type
devices could be improved to function as automated repellent delivery systems for area

repellency when combined with fans. Special care should be taken to protect foliage

78

from repellent aerosol droplets, though; preliminary experiments with DEBT and

Microsprayers resulted in significant phytotoxicity on grass and garden plants that were
under or downwind of the release device. Alternative repellents such as Autan® or active

botanicals should be tested for their effectiveness as wind driven vapors in addition to
dermal applications.

Irrespective of whether these particular wind- or wind/chemical—based proposals
to protect humans in the back-yard setting prove feasible, the data reported here will
hopefully stimulate and anchor additional exploration of the potential for mosquito
manipulation by air -- a medium so common and available it may have long been

overlooked.

79

 

APPENDIX

80

APPENDIX 1

Record of Deposition of Voucher Specimens

81

'r-HL‘1

 

APPENDIX 1

Record of Deposition of Voucher Specimens*
The specimens listed on the following sheet(s) have been deposited in the named
museum(s) as samples of those species or other taxa, which were used in this research.
Voucher recognition labels bearing the Voucher No. have been attached or included in
fluid-preserved specimens.
Voucher No.: 2001-07
Title of thesis or dissertation (or other research projects):
Behavioral Manipulation of Mosquito (Diptera : Culicidae) Host Seeking Using
Artificially Applied Wind and Vapor-Phase Repellent: Potential for Personal Protection

Museum(s) where deposited and abbreviations for table on following sheets:

Entomology Museum, Michigan State University (MSU)

Investigator’s Name(s) (typed)
Eric J ames Hoffmann

 

 

Date 8/16/01

*Reference: Yoshimoto, C. M. 1978. Voucher Specimens for Entomology in North
America.
Bull. Entomol. Soc. Amer. 24: 141-42.

Deposit as follows:
Original: Include as Appendix 1 in ribbon copy of thesis or dissertation.

Copies: Included as Appendix 1 in copies of thesis or dissertation.
Museum files.

Research project files.

This form is available from and the Voucher No. is assigned by the Curator, Michigan
State University Entomology Museum.

82

Appendix 1

Table 2: Voucher specimen data

4 Pages

e1of

Pag

Sun.

650

 

\8N 8% R. iii -\.
as. > Em
3.8.55 8 m 898.2 a... c. .388

.2 mcoEBoam 3E. o>onm 05 cozooom

 

5B In 200

 

 

 

 

so- EON .02 520:0)

ccmEtoI mecca... 05
83.5 @952 £23.69...
5338: a $35 .3368 can.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:ws. N 88 g 82.
cm 83: .o m .5: 3. __2m .82 .8m
8.52.2, 9.8 8oz .oo 8.56 220.195. 88.88 882
:ms.
N 88. .t 22.
:92 F 88 .2 95..
um .53: .o m .5: cm :5 .82 .8m
3.6.5.. 9.8 8cm .8 8.5.0 220.10.: 828.6 882
:ms. m 88 .F .882
:ms. F 88 .t 22.
:92 F 88 .2 82,
3. :23 .o m .5: cm :5 .82 .8m
9.2.2. 9.8 8cm .8 8.5.0 22910.: 8.8888 88<
m m s
m m w .r. M ugh. % W m s 3.383 new :96. 850 .5 3.88
U e P m u u o. m. w 9 com: .0 8.02.8 mcoE_ooam .8 San .38.
M m w 0 M M cu. N a Fm.
co 59:32

83

 

0.00 .2050

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

50.... 95
830022 30.00.25
2.825 990 80.8.2 9.. 5 .089.
.0. 0:00.803 00.0.. 0>000 0... 0020001 0:00.50... 0050.. 00m
8-88 82 889$ .89... .9982 0.0.8.085
2.000000: h....u....00...... 000.2%... 00....
S
e
. g
\l a
1 w P
.m n 4_ 00.2 o. 88 .N 95..
d (m f B. 8.3... m 8: um :20 .82 .80
m. 2 o 98...... 9.8 88. .00 8.5.0 22050.5. 9.99. 882
e
M. m 2_ :02 m 9.2 .3 2....
T m. 8-. .o z 8: am 888.. .82 .80
mm 85.5.2. 9.9. 83. .8 8.5.0 220.10.... 8.8.2.. 882
00.2 N 88 .N. 2....
00.2 N SON .. 95..
um 8...: .o m .5: 3. =20 .82 .80
952.2. 9.3 8cm. .8 8% 220.20.... 08.8...» 882
m e m r m is e m. m 03.00000 0:0 0x .0 0 .0 00.000
W m m. e aluu Nu. m m. w w 000.. .0 00.02.00 0:05.000... .0. 0.00 .000. c 0. 5 . m.
M m w m A A P N m E
”.0 .0252

 

 

 

84

 

200 .2050

 

 

 

 

 

 

 

4 Pages

Appendix 1
of

Table 2 (cont'd).

Page 3

8.0:... amp
5:002). 30.0525
2.82:: 90.0 80.8.5. 9.. 5 .88..
.2 0:05.0000 00.0.. 0>000 0... 0020000. 50502.. 0050.. 00m
5988 .02 .9.9...> 69.2. .0082 0.8.8.8....
3.000000: .. 0.00:0 .0:0_._000 00:.
:05. 2 88 .5: 2....
:0 8...: .o w .5: :0 8.0 .82 .80
98...... 9.9. .80 .8 8.5.0 220.10.... 59.0.... 8.9.82
:05. m 88 .3 .882
:05. F 88 .5 .882
:0 8...: .o w .5: :0 :90 .8280
. 9......2. 9.9. 080 .8 8.5.0 220.10.... 8.8.8888 8.9.082
:05. N 88 .5. 2....
:05. 0 88 .5 2....

:0 8...: .o m .5: :0 :20 .8280
05.5.2. 9.9. 080 .8 8.5.0 220.105.

 

 

 

 

 

 

 

 

 

:05. N 8.: .0. 2....
:05. . 8.: .N: 2....
09. .o z .5: :0 888.. .8280
98...... 9.9. E0 8.5.0 2.0210... 0.89.88... 8.8082
m m m a .2... h. m m .w s 02.00000 0:0 :98. .050 .0 00.0000
seP .n m m o. w 9 0:.0 0000050500005 0 00 .
mmmdmmmmMmEB 8... . .93....
.50 805:2

 

 

 

 

85

 

0.00 .2050

 

0.00005. 30.00.25

 

0.0m:

 

2.082.: 90.0 80.8.5. o... c. .88..

 

.2 000020000 00.0.. 0>000 00. 0020000.
.380 .oz 8823

0000.001 000.00 05
300.02 0..200..00>0.
3.000000: 0. 0.0000 .000...000 00:.

 

 

Appendix 1

Table 2 (cont'd).
Page 4 of 4 Pages

P

88 .: .80....

88 .R 2....
00 8...: .o m 8: ..0 .20 .82 .80
9.0.5.. 9.3 080 .8 8.0.0 25.10....

080 .R 2....

88 .. 08..
00 8...: .o m .5: :0 8.0 .8... .80
9.0.5.. 9...... 080 .oo 8.....0 5.0.10.5.

88 .R 2....
00.. .o z .5: 00 888.. 8.4. .80

00000.00. 200.30

0.050.... 0.00.200

 

 

 

 

 

 

 

 

 

 

 

 

020...... 00.04 000ml. .00 020:0 24.910.5— 00.00530 0.0.00
S
r 0 0 e m m 09.88.. 08
...n. w. w Way m. N w 000: .0 00.00.60 000020000 .2 0.00 .000: 00x0. .050 .0 00.000w
t a g
0 A A P N L E
.0. 0.02

 

86

LITERATURE CITED

87

 

 

LITERATURE CITED

Acree, F ., Turner, R. B., Gouck, H. K., Beroza, M., and Smith, N. 1968. L-Lactic Acid:
A mosquito attractant isolated from humans. Science 161, 1346-1347.

Bailey, S. F ., Eliason, D. A., and Hoffmann, B. L. 1965. Flight and Dispersal of the
mosquito Culex tarsalis Coquillett in the Sacramento Valley of California.
Hilgardia 37, 72-113.

Bidlingrnayer, W. L. 1974. The influence of envionmental factors and physiological stage
on flight patterns of mosquitoes taken in the vehicle aspirator and truck, suction,
bait and New Jersey light traps. Journal of Medical Entomology 11, 119-146.

Bidlingrnayer, W. L. 1985. The measurement of adult mosquito population changes-
some considerations. Journal of the American Mosquito Control Association 1,
328-348.

Bidlingrnayer, W. L., Evans, D. G., and Hansen, C. H. 1985. Preliminary study of the
effects of wind velocities and wind shadows upon suction trap catches of
mosquitoes (Diptera: Culicidae). Journal of Medical Entomology 22, 295-302.

Bidlingrnayer, W. L., Day, J. F., and Evans, D. G. 1995. Effect of wind velocity on
suction trap catches of some Florida mosquitoes. Journal of the American
Mosquito Control Association 11, 295-301.

Borror, D. J ., Triplehom, C. A., and Johnson, N. F. 1989. An introduction to the study of
insects, Harcourt Brace, New York.

Brantjes, N. M. B. 1981. Wind as a factor influencing flower-visiting by Hadena bicruris
(Noctuidae) and Deilephila elpenor (Sphingidae). Ecological Entomology 6, 361-
363.

Brown, A. W. A., Sarkaria, D. S., and Thompson, R. P. 1951. Studies on the responses of
the females Aedes aegypti mosquito. Part I.-The search for attractant vapours.
Bulletin of Entomological Research 42, 105-115.

Centers for Disease Control. “Questions and Answers about West Nile Virus- CDC
DVIBD.” [online] available <http://www.cdc.gov/ncidod/dvbid/
westnile/q&a.htm.> June 19, 2001.

Chapman, R. F. 1998. The Insects: Structure and Function, Cambridge University Press,
Cambridge, UK.

Chou, J. T., Rossignol, P. A., and Ayres, J. W. 1997. Evaluation of commercial insect

repellents on human skin against Aedes aegypti (Diptera: Culicidae). Journal of
Medical Entomology 34, 624-630.

88

Clements, A. N. 1992. The biology of mosquitoes, Vol. 1. Chapman & Hall, New York.

Cockcrofl, A., Cosgrove, J. B., and Wood, R. J. 1998. Comparative repellency of
commercial formulations of deet, pennethrin and citronellal against the mosquito
Aedes aegpti, using a collagen membrane technique compared with human arm
tests. Medical and Veterinary Entomology 12, 289-294.

Constantini, C., Gibson, G., Sagnon, N., della Torre, A., Brady, J ., and Coluzzi, M. 1996.
Mosquito responses to carbon dioxide in a West African Sudan savanna village.
Medical and Veterinary Entomology 10, 220-227.

Davis, E. E. and Sokolove, P. G. 1976. Lactic acid-sensitive receptors on the antennae of
the mosquito, Aedes aegypti. Journal of Comparative Physiology A 105, 43-54.

Davis, E. E. 1985. Insect repellents: Concepts of their mode of action relative to potential
sensory mechanisms in mosquitoes (Diptera: Culicidae). Journal of Medical
Entomology 22, 237-243.

Davis, E. E. and Bowen, M. F. 1994. Sensory physiological basis for attraction in
mosquitoes. Journal of the American Mosquito Control Association 10, 316-325.

Daykin, P. N., Kellogg, F. E., and Wright, R. H. 1965. Host-finding and repulsion of
Aedes aegypti. The Canadian Entomologist 97, 239-263.

Debboun, M., Strickman, D., Klein, T. A., Glass, J ., Wylie, E., Laughinghouse, A.,
Wirtz, R., and Gupta, R. K. 1999. Laboratory evaluation of A13-37220, A13-
35765, CIC-4 and DEET repellents against three species of mosquitoes. Journal
of the American Mosquito Control Association 15, 324-347.

Dekker, T. and Takken, W. 1998. Differential responses of mosquito sibling species
Anopheles arabiensis and An. quadriannulatus to carbon dioxide, a man or a calf.
Medical and Veterinary Entomology 12, 136-140.

Diamond, J. 1998. Guns, Germs and Steel, W.W. Norton & Company, New York.

Dogan, E. B., Ayres, J. W., and Rossignol, P. A. 1999. Behavioral mode of action of
deet: Inhibition of lactic acid attraction. Medical and Veterinary Entomology 13,
97-100.

Dogan, E. B. and Rossignol, P. A. 1999. An olfactometer for discriminating between
attraction, inhibition, and repellency in mosquitoes (Diptera: Culicidae). Journal
of Medical Entomology 36, 788-793.

Edman, J. D. 1979. Orientation of some Florida mosquitoes (Diptera: Culicidae) toward
small vertebrates and carbon dioxide in the field. Journal of Medical Entomology
15, 292-296.

89

 

Eiras, A. E. and J epson, P. C. 1991. Host location by Aedes aegypti (Diptera: Culicidae):
a wind tunnel study of chemical cues. Journal of Entomological Research 81,
15 1-1 60.

F radin, M. S. 1998. Mosquitoes and mosquito repellents. Annals of Internal Medicine
128, 931-940.

Frances, S. P., Cooper, R. D., Popat, S., and Sweeney, A. W. 1999. Field evaluation of
the repellents DEET, CIC-4, and A13-37220 against Anopheles in Lae, Papua

New Guinea. Journal of the American Mosquito Control Association 15, 339-341.

Garrett, L. 1994. The coming plague, Penguin Books, New York.

Geier, M. and Boeckh, J. 1999. A new Y-tube olfactometer for mosquitoes to measure
the attractiveness of host odours. Entomologia Experimentalis et Applicata 92, 9-
19.

Geier, M., Bosch, 0. J ., and Boeckh, J. 1999. Influence of odour plume structure on
upwind flight of mosquitoes towards hosts. The Journal of Experimental Biology
202, 1639-1648.

Gibson, G., Constantini, C., Sagnon, F., della Torre, A., and Coluzzi, M. 1997. The
responses of Anopheles gambiae, and other mosquitoes in Burkina Faso, to C02 -
the start of a search for synthetic human odour. Annals of Tropical Medicine and
Parasitology 91, S 123-8 124.

Gillett, J. D. 1972. The Mosquito: Its life, activities and impact on human aflairs,
Doubleday & Co., Inc., Garden City, New York.

Gillett, J. D. 1979. Out for blood: Flight orientation up-wind in the absence of visual
clues. Mosquito News 39, 221-229.

Gillies, M. T. 1980. The role of carbon dioxide in host-finding by mosquitoes (Diptera:
Culicidae): a review. Bulletin of Entomological Research 70, 525-532.

Gillies, M. T. and Wilkes, T. J. 1981. Field experiments with a wind tunnel on the flight
speed of some West African mosquitoes (Diptera: Culicidae). Bulletin of
Entomological Research 71, 65-70.

Goodyer, L. and Behrens, R. H. 1998. Short report: The safety and toxicity of insect
repellents. American Journal of Tropical Medicine and Hygeine 59, 323-324.

Grant, A. J ., Wigton, B. E., Aghajanian, J. G., and O'Connell, R. J. 1995. Electrophysical
responses of receptor neurons in mosquito maxillary palp sensilla to carbon
dioxide. Journal of Comparative Physiology A 177, 389-396.

90

 

Grimstad, P. R. and DeFoliart, G. R. 1975. Mosquito nectar feeding in Wisconsin in
relation to twilight and microclimate. Journal of Medical Entomology 11, 691-
698.

Isaacs, R., Ulczynski, M., Wright, B., Gut, L., and Miller, J .R. 1999. Performance of the
microsprayer, with appication for pheromone-mediated control on insect pests.
Journal of Economic Entomology 92(5), 1 157-1164.

Kennedy, J. S. 1939. The visual responses of flying mosquitoes. Procedings of the
Zoological Society of London, Series A 109, 221-242.

Klassen, W. and Hocking, B. 1964. The influence of a deep river valley system on the
dispersal of Aedes mosquitoes. Bulletin of Entomological Research 55, 304.

Kline, D. L. 1998. Olfactory responses and field attraction of mosquitoes to volatiles
from Limburger cheese and human foot odor. Journal of Vector Ecology 23, 186-
194.

Kline, D. L. and Mann, M. O. 1998. Evaluation of butanone, carbon dioxide, and 1-
octen-3-ol as attractants for mosquitoes associated with north central Florida Bay
and cypress swamps. Journal of the American Mosquito Control Association 14,
289-297.

Knols, B. G. J ., Takken, W., Cork, A., and De J ong, R. 1997a. Odour-mediated, host-
seeking behaviour of Anopheles mosquitoes: a new approach. Annals of Tropical
Medicine and Parasitology 91, SI 17-Sl 18.

Knols, B. G. J ., van Loon, J. J. A., Cork, A., Robinson, R. D., Adam, W., Meijerink, J .,
De Jong, R., and Takken, W. 1997b. Behavioral and electrophysiological
responses of the female malaria mosquito Anopheles gambiae (Diptera:
Culicidae) to Limburger cheese volatiles. Bulletin of Entomological Research 87,
1 5 1 -1 59.

Kuthiala, A., Gupta, R. K., and Davis, E. E. 1992. Effect of the repellent Deet on the
antennal chemoreceptors for oviposition in Aedes aegypti (Diptera: Culicidae).
Journal of Medical Entomology 29, 639-643.

Lewis, T. and Taylor, L. R. 1967. Introduction to Experimental Ecology, Academic
Press, New York.

Lumsden, W. H. R. 1947. Observations on the effect of microclimate on biting by Aedes
aegypti (L.) (Dipt., Culicidae). Journal of Experimental Biology 24, 361-373.

Mboera, L. E. G., Knols, B. G. J ., Takken, W., and della Torre, A. 1997. The response of
Anopheles gambiae SJ. and A. funestus (Diptera: Culicidae) to tents baited with
human odour or carbon dioxide in Tanzania. Bulletin of Entomological Research
87 , 1 73-1 78.

91

McIver, S. B. 1981. A model for the mechanism of action of the repellent deet on Aedes
aegypti (Diptera: Culicidae). Journal of Medical Entomology 18, 357-361.

McIver, S. B. and McElligott, P. E. 1989. Effects of release rates on the range of
attraction of carbon dioxide to some southwestern Ontario mosquito species.
Journal of the American Mosquito Control Association 5, 6-9.

Murlis, J ., Elkinton, J. S., and Cardé, R. T. 1992. Odor plumes and how insects use them.
Annual Review of Entomology 37, 505-532.

Nielsen, E. T. and Nielsen, H. T. 1953. Field observation on the habits of Aedes
taeniorhynchus. Ecology 34, 141-156.

Novello, A., White, D., Kramer, L., et. al. 2000. Update: West Nile Virus activity -
Eastern United States, 2000. Morbidity and Mortality Weekly Report 49, 1044-
1047.

O'Meara, G. F. Ecology of autogeny in mosquitoes. Lounibos, L. Philip, Rey, Jorge R.,
and Frank, J. Howard. 459-471. 1985. Vero Beach, Florida, Florida Medical
Entomology Laboratory. Ecology of mosquitoes: Proceedings from a workshop.

Osimitz, T. G. and Grothaus, R. H. 1995. The present safety assessment of DEET.
Journal of the American Mosquito Control Association 11, 274-278.

Platt, R. B., Collins, C. L., and Witherspoon, J. P. 1957. Reactions of Anopheles
quadrimaculatus Say to moisture, temperature and light. Ecological Monographs
27, 303-324.

Platt, R. B., Love, G. L., and Williams, E. L. 1958. A positive correlation between
relative humidity and the distribution and abundance of Aedes vexans. Ecology
39, 167-169.

Price, G. D., Smith, N., and Carlson, D. A. 1979. The attraction of female mosquitoes
(Anopheles quadrimaculatus (Say)) to stored human emanations in conjunction
with adjusted levels of relative humidity, temperature and carbon dioxide. Journal

of Chemical Ecology 5, 383-395.

Reeves, W. C. 1951. Field studies on carbon dioxide as a possible host stimulant to
mosquitoes. Proceedings of the Society for Experimental Biology and Medicine
77, 64-66.

Reeves, W. C. 1953. Quantitative field studies on a carbon dioxide chemotropism of
mosquitoes. American Journal of Tropical Medicine and Hygeine 2, 325-331.

Robbins, P. J. and Cherniack, M. G. 1986. Review of the biodistribution and toxicity of
the insect repellent MN-diethyl-m-toluamide (DEET). Journal of Toxicology and
Environmental Health 18, 503-525.

92

 

Rowley, W. A. and Graham, C. L. 1968. The effect of temperature and relative humidity
of the flight performance of female Aedes aegypti. Journal of Insect Physiology
14, 1251-1257.

Schreck, C. E., Kline, D. L., and Carlson, D. A. 1990. Mosquito attraction to substances
from the skin of different humans. Journal of the American Mosquito Control
Association 6 , 406-410.

Schreck, C. E. and McGovern, T. P. 1989. Repellents and other personal protection
strategies against Aedes albopictus. Journal of the American Mosquito Control
Association 5, 247-250.

SAS Institute. 1996. SAS/STAT user ’3 guide, version 6.12, 4th ed. SAS Institute, Cary,
NC.

Service, M. W. 1980. Effects of wind on the behaviour and distribution of mosquitoes
and blackflies. International Journal of Biometeorology 24, 347-353.

Smith, C. N., Smith, N., Gouck, H. K., Weidhaas, D. E., Gilbert, 1. H., Mayer, M. S.,
Smittle, B. J ., and Hofbauer, A. 1970. L-lactic acid as a factor in the attraction of
Aedes aegypti (Diptera: Culicidae) to human hosts. Annals of the Entomological
Society of America 63, 760-770.

Snow, W. F. 1970. The effect of a reduction in expired carbon dioxide on the
attractiveness of human subjects to mosquitoes. Bulletin of Entomological
Research 60, 43-48.

Snow, W. F. 1977. The height and direction of flight of mosquitoes in West African
savanna, in relation to wind speed and direction. Bulletin of Entomological
Research 97, 271-279.

Snow, W. F. 1979. The vertical distribution of flying mosquitoes (Diptera: Culicidae)
near an area of irrigated rice-fields in the Gambia. Bulletin of Entomological
Research 69, 561-571.

Snow, W. F. 1980. Field estimates of the flight speed of some West African mosquitoes.
Annals of Tropical Medicine and Parasitology 74, 239-242.

Snow, W. F. 1982. Further observations on the vertical distribution of flying mosquitoes
(Diptera: Culicidae) in West African savanna. Bulletin of Entomological Research
72, 695-708.

Takken, W. and Kline, D. L. 1989. Carbon dioxide and 1-octen-3-ol as mosquito
attractants. Journal of the American Mosquito Control Association 5, 311-316.

Takken, W. 1991. The role of olfaction in host-seeking of mosquitoes: a review. Insect
Scientific Applications 12, 287-285.

93

 

Takken, W., Dekker, T., and Wijnholds, Y. G. 1997a. Odor-mediated flight behavior of
Anopheles gambiae Giles sensu stricto and An. stephensi Liston in response to
C02, acetone and 1-octen-3-ol (Diptera: Culicidae). Journal of Insect Behavior
10, 395-407.

Takken, W., Knols, B. G. J ., and Otten, H. 1997b. Interactions between physical and
olfactory cues in the host-seeking behaviour of mosquitoes: the role of relative
humidity. Annals of Tropical Medicine and Parasitology 91, $119-$120.

Takken, W. and Knols, B. G. J. 1999. Odor-mediated behavior of afrotropical malaria
mosquitoes. Annual Review of Entomology 44, 13 1-157.

Taylor, L. R. 1974. Insect migration, flight periodicity and the boundary layer. Journal of
Animal Ecology 43, 225-238.

Thomson, R. C. M. 1938. The reactions of mosquitoes to temperature and humidity.
Bulletin of Entomological Research 24, 125-140.

US. Environmental Protection Agency. 1980. N,N-diethyl-m-toluarnide (DEET)
Pesticide Registration Standard (EPA-540/RS-81-004).

Vogel, S. 1981. Life in moving fluids: the physical biology of flow. Willard Grant Press,
Boston, Mass.

Wright, R. H. 1975. Why mosquito repellents repel. Scientific American 233, 104-111.

Wright, R. E. and Knight, K. L. 1966. Effect of environmental factors on biting activity
of Aedes vexans (Meigen) and Aedes trivittatus (Coquillett). Mosquito News 26,
595-578.

Yap, H. H., Jahangir, K., Chong, A. S. C., Adanan, C. R., Chong, N. L., Malik, Y. A.,
and Rohaizat, B. 1998. Field efficacy of a new repellent, KBR 3023, against
Aedes albopictus (Skuse) and Culex quiquefasciatus (Say) in a tropical
environment. Journal of Vector Ecology 23, 62-68.

94

 

mumalllilljllflllljllllflllllul