|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
DEVELOPMENT OF A PRACTICAL TECHNIQUE FOR SAMPLING THE
AFROTROPICAL MALARIA VECTORS Anopheles gambiae S.L. AND An. funestus
JUSTIN ERIC HARBISON
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
This document is dedicated to my parents, Kent and Judy Harbison.
I would like to thank my parents, Kent and Judy Harbison, for their unwavering
support. I thank Dr. Jonathan Day, Dr. Dan Kline, Dr. Roxanne Rutledge-Connelly, and
Dr. Sandra Allan for their help in Florida. I thank Dr. Richard Mukabana, Dr. Evan
Mathenge, Dr. Bart Knols, Dr. Gerry Killeen, and Dr. Ulrike Fillinger for their
willingness to help out an American student. I thank everyone else back in Kenya not
only for helping me complete this project, but for their willingness to teach me about
their great country. I thank Debbie Hall for all the administrative help. I thank Marinela
Capanu for statistical assistance and Scott Weihman for the use of his photos. Finally,
special thanks go to Camille Francisco for her undying patience while I was in Africa.
TABLE OF CONTENTS
A C K N O W L E D G M E N T S ................................................................................................. iii
LIST OF TABLES ........... .. ............. ................................ vi
LIST OF FIGURES ......... ..................................... ....... vii
A B STR A C T ..................... ................................... ........... ................. viii
1 IN TR OD U CTION ............................................... .. ......................... ..
A approaches to M alaria Control. ...................................................... .....................1
H history of M alaria Control ..................................................................................... 2
Future of M alaria C control ............................................... ...... .............. ...............
Role of Rural Housing in Malaria Transmission...................................................3
Sam pling M ethods for M alaria Vectors Indoors ........................................ ...............4
Indoor Sampling of Resting Mosquitoes ........................................ ...............5
2 LITER A TU R E REV IEW ............................................................. ....................... 7
H historical O overview .................... .................................. .... ........ ................. 7
Im portance of the D disease W orldw ide ................................. ..................................... 8
D description of D disease Transm mission .................................. ............... ............... 8
Im portance of M alaria in A frica............................................................................... 10
Im portant V ector Species in Africa................................ .................................... 11
S u m m a ry .......................................................................................................1 2
3 M ATERIALS AND M ETHOD S ........................................ ......................... 14
Site Descriptions ...................................... .......................... ............... 14
Survey of Indoor Resting Habitats of M osquitoes ....................................................14
Representative Accuracy of Hand Collection........................................ ...............17
Development of Cloth Indoor Resting Box in Semi-Field Conditions..................... 18
Comparison of Accuracy of Resting box/Resting Net to Hand Collection ..............19
Development of an Alternative to Cloth Box in Semi-Field Conditions....................22
F ield T rials ............................................... .... ......... ...... 2 2
S e m i-F ie ld ...................................................................... 2 3
4 RESULTS AND DISCU SSION ........................................... .......................... 24
Results .................. ............................. ............ ...............24
Survey of Indoor Resting Habitats of Mosquitoes ............................................24
Representative Accuracy of Hand Collection ..................................................25
Development of Cloth Indoor Resting Box in Semi-Field Conditions ..............25
Comparison of Representative Accuracy of Resting Box/Resting Net to Hand
C collection ................. ...................................... .. ... .... ................. 26
Development of an Alternative to Cloth Box in the Semi-Field.........................28
D iscu ssion ..... ............... ....................................32
Indoor Resting Places of Mosquitoes................... .. ......... .............32
Com prison of Sam pling Techniques ...................................... ............... 33
Sum m ary ............. ............................................... .... ... ...... .. 35
5 NOTES ON THE DEVELOPMENT OF A PRACTICAL Anopheles GRAVID
T R A P ................................................................................3 7
Introduction .................................... .. ....... ..................... 37
Prelim inary Sem i-Field Trials in Florida......................................... ............... 38
Sem i-F ield T rials in K eny a .............................................................. .....................40
Field Trials in K enya .................................. .. ... ... ..... ............ 41
Results ........... ............. ........ .................................42
F lo rid a T ria ls ................................................................................................. 4 2
K en y a T rials ................................................................4 2
D isc u ssio n ............................................................................................................. 4 2
L IST O F R E FE R E N C E S ............. ................. ..................................................45
B IO G R A PH IC A L SK E T C H ....................................................................................... 50
LIST OF TABLES
4.1. M ean area and height of dwellings in study area.....................................................25
4.2 Percent of total actual number of mosquitoes (437) collected from each location......26
4.3 Total actual number of mosquitoes collected at each height.......................................26
4.4 The mean percent of mosquitoes recaptured in semi-field conditions using a resting
box either treated or untreated with deltamethrin. ............. ................................... 26
4.5 Comparison of actual number of mosquitoes collected by the three methods tested. .27
4.6 Comparison of the accuracy (percent of the overall number of mosquitoes found in
each dwelling) over the 27 test dates for the three methods tested ........................29
4.7 Comparison of time taken to complete sampling using a resting box (cloth or
cardboard) and resting net to hand collection. ................................. ............... 29
4.8 Actual number of An. gambiae mosquitoes caught by method, including sex and
gonotrophic status, during the 27 test days of the Latin Square. ..........................30
4.9 Actual number of An. funestus mosquitoes caught by method, including sex and
gonotrophic status, during the 27 test days of the Latin Square .............................30
4.10 Actual number of Culicine mosquitoes caught by method, including sex and
gonotrophic status, during the 27 test days of the Latin Square. ...........................31
4.11 Actual number of unidentified anopheline mosquitoes caught by method,including
sex and gonotrophic status, during the 27 test days of the Latin Square. ...............31
4.12 Mean number of mosquitoes caught daily by a resting basket placed in a house in
the village of Lwanda. ........................ ...... ................ ..1... ... 31
4.13 Comparison of the resting basket to the cloth resting box of the total percentage of
mosquitoes recaptured in semi-field conditions ................................. ............... 31
4.14 Mean percent of mosquitoes recaptured by each density used for both the resting
box and resting basket. ........................... .................... .. .. .. .... ........... 32
LIST OF FIGURES
3.1 Map of Kenya showing the study sites. .......................... ..................... 15
3.2 Dwellings in Lwanda. A) Single family structure B) Multiple family structure.........16
3.3 Objects in dwellings on which mosquitoes can be found resting.............................17
3.4 Components of the experimental tests A) Resting box hung in experimental hut. B)
Experimental hut in modified greenhouse. C) Modified greenhouse. ...................20
3.5 Treatments used in the Latin Square. A) Cloth resting box hanging in dwelling B)
Cardboard box hanging in dwelling C)Diagram of a resting net spread out in a
dwelling D) Expert collecting mosquitoes by aspiration. ......................................22
3.6 Alternative to the cloth resting box. A) Cloth resting box and resting basket B)
Resting basket hanging in dwelling. ............................................. ............... 23
4.1 Layout of dwellings. A) Single Family B) Multiple Family ....................................25
5.1 Design of oviposition trap. A) Full view. B) Collecting cartridge ...........................39
5.2 Modifications to design. A) with clothes hangers B) with black mesh C) with vial
w ith sugar solution. ....................... ...................... ................... .. .....40
5.3 Trap designs in Kenya. A) Mbita (two variations) B) Emergence C) Florida D) Cone
(tw o variations). .................................................... ................. 43
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science
DEVELOPMENT OF A PRACTICAL TECHNIQUE FOR SAMPLING THE
AFROTROPICAL MALARIA VECTORS Anopheles gambiae S.L. AND An. funestus
Justin Eric Harbison
Chair: Jonathan Day
Major Department: Entomology and Nematology
Malaria, transmitted by Anopheline mosquitoes, is not only the most important
insect-borne disease but one of the top three infectious diseases in the world. Resting
boxes are a well-established method for sampling Anopheline mosquitoes. Few resting
box designs, however, are used indoors. A practical method, utilizing cloth resting boxes
and "resting nets" for sampling indoor malaria vectors of Africa, was developed in semi-
field (a modified greenhouse) and field conditions. In semi-field conditions, a cloth
resting box recaptured 36.1 9.9% (mean SE) of three different densities of An.
gambiae s.s. (Giles) females of varying gonotrophic status. The accuracy method then
was compared to the collection of mosquitoes using an oral aspirator in a rural Kenyan
village. The developed method caught a significantly higher percentage (P=0.05) of
resting mosquitoes than hand catches with an oral aspirator and required slightly less time
to complete (8.0 + 3.9 minutes versus 15.0 + 0.0 minutes). A "resting basket" was also
developed as an even more practical alternative to the cloth resting box. The resting
basket was tested under the same semi-field conditions as the resting box. No significant
difference (P=0.47) was found in recapture rates between the resting basket and box,
suggesting that it could be used as an alternative to the cloth resting box. The materials
used in the method, which requires little training to implement, can be easily obtained in
rural settings where malaria is of great concern.
Approaches to Malaria Control.
Historically, there are two general approaches (vertical and horizontal) to control of
diseases such as malaria. Vertical approaches call for centralized national programs that
act as an independent entity in national health care systems. Horizontal approaches
involve broad based local programs emphasizing basic needs such as health education,
safe water, and adequate food supply (Tan et al. 2003). Positive aspects to vertical
approaches are a government commitment to standardized health care and attention given
to the control of all aspects of the disease (such as prevention and treatment). Also, due
to clear hierarchical structures typically found in such programs, there are fewer
organizational complications compared to those observed in decentralized systems
(Kroeger et al. 2002). Vertical programs, however, are criticized for focusing solely on
one health problem and failing to build local capacities with wider health benefits (Tan et
al. 2003). In some cases vertical programs must be initiated by larger international
organizations when national governments do not have the needed resources.
Horizontal control programs are strongly based in decentralized health systems and
in community mobilization. They are geared to become more focused to the needs of
each specific region. Problems that arise from horizontal approaches are that often
decentralized health systems do not have the resources and support held in vertical
programs, and in many rural regions people cannot access basic health facilities (Deressa
et al. 2003, Garfield 1999, Killeen et al. 2002).
History of Malaria Control
In the 1950s, the World Health Organization (WHO) realized the severity of
malaria in terms of human morbidity and mortality and the resulting reduction of
agriculture and industry on a global scale. In 1956, the Global Malaria Eradication
campaign was launched. The campaign was a relatively worldwide vertical effort based
on interrupting transmission through indoor residual spraying (and other mosquito control
measures), management of marshes, creation of agricultural land, and modification of
human lifestyle (i.e., window screens, air conditioning, and television). The success of
the National Malaria Eradication Program in the United States was, in part, due to
vertical disease programs from the Tennessee Valley Authority and the Communicable
Disease Center (previously the Office of Malaria Control in War Areas) (CDC 2004a,
CDC 2004b). Malaria eradication was realized in much of the world's temperate areas,
but failed or was not attempted in many tropical areas. In Africa, eradication was only
attempted in Zimbabwe, South Africa, and Ethiopia since malaria was considered too
great a problem in the other countries (Kager 2002). In response to the anticipation that
national primary health care systems would take over comprehensive malaria information
systems, aiding in management of malaria, the vertical approach programs of the WHO
were dismantled in the 1980s. International interest and funding in malaria research
waned during this time. Since then malaria transmission in many of the affected regions
has increased dramatically (Gilles 2002).
Future of Malaria Control
In the past 15 years, global efforts have been made to combat the increasing
morbidity and mortality associated with the disease. Initiatives such as Roll Back
Malaria (RBM), Multilateral Initiatives on Malaria (MIM), and the African Initiative on
Malaria (AIM) have been created to reestablish malaria as a global priority, to increase
support from wealthier nations, and to develop sector-wide approaches to malaria control.
This global push to create regional programs involves not only peripheral health care
systems, but educating and empowering community members of the affected areas
(Conteh et al. 2004, Deressa et al. 2003, Magnussen et al. 2001). Because such
horizontal approaches still have shortcomings, it is likely that large scale malaria control
will only be achieved through the integration of aspects from both horizontal and vertical
programs (Kroeger et al. 2002).
Role of Rural Housing in Malaria Transmission
The physical condition of rural houses plays an important part in the epidemiology
of malaria, especially when transmitted by endophagic and endophilic vectors like
Anopheles gambiae (Giles). High levels of malaria transmission are usually associated
with vectors that prefer to feed on humans indoors (Harwood and James 1979).
Successful eradication of malaria in many regions of the world is due in part to mosquito-
proofing houses (i.e., screened windows and air conditioning). Many people living in
regions where malaria causes the highest morbidity and mortality do not have access to
Rural houses and villages (especially those of poor construction) are considered the
main foci for malaria transmission in these regions (Muirhead-Thomson 1982, Konradsen
et al. 2003, Fullerton and Bishop 1933, Gamage-Mendis et al. 1991). In Kenya, Githeko
et al. (1994) and Bogh et al. (1998) found that 74% and 99%, respectively, of An.
gambiae sensu lato (Giles) collected resting indoors had fed on human blood. Open eaves
provide a ready entry point into rural houses and increase the chance of finding higher
densities of mosquitoes resting indoors (Lindsay et al. 2003, Schofield and White 1983).
Houses built without ceilings also have been shown to increase human exposure to
malaria vectors. (Lindsay et al. 1995, Schofield and White 1983). Houses also provide
suitable habitats for vectors of human filariasis (Culex spp), yellow fever, and dengue
(Aedes aegypti L.).
Sampling Methods for Malaria Vectors Indoors
Although there are various methods for sampling the eggs and larvae of
mosquitoes, the most commonly used approach is to sample adults (Service 1993). There
exist numerous methods to sample adults, especially host-seeking females. Many of the
techniques used to sample mosquitoes seeking human blood are conducted indoors or in
artificial shelters. Such methods include bednet traps, hand net collections, drop net
collections, and aspirating mosquitoes attempting to bite a sleeping human subject
(Service 1993). However, ethical issues arise when studies attempt to quantify the
numbers of mosquitoes actively trying to bite humans. Studies involving methods such
as counting the number of mosquitoes landing on a human (human landing catches) and
human-baited traps can potentially expose human subjects to malaria through a bite from
an infected mosquito. These studies are often conducted during the time of day when
many female mosquitoes are actively seeking blood. Ways to circumvent using human
subjects include the use of carbon dioxide and light traps, as well as animal baited traps.
The accuracy of these methods varies, but their cost and practicality are often beyond the
means of community level control programs in developing countries. Also, mosquitoes
may be attracted to mechanical devices for reasons other than host-seeking and
mosquitoes attempting to feed on animals may not behave in the same manner as
Collecting adult resting mosquitoes, however, is a method that is considered to be
more representative of mosquito populations than collections made by trapping adults in
flight. This is because a wider range of mosquitoes, in terms of bloodmeal status, age,
sex, and gonotrophic cycle, are taken in resting collection (Service 1993). Resting
collections are often made from outdoor natural shelters, such as vegetation and tree-
holes, but a few important vector species can be found in man-made structures. Because
of this, sampling a population of vectors resting indoors can be valuable to surveillance
and control programs (WHO 1992, Lindblade et al. 2000).
Indoor Sampling of Resting Mosquitoes
Hand catches with a plastic tube or mechanical aspirator and Pyrethrum Spray
Catches (PSC) are two commonly used methods for collecting mosquitoes resting indoors
(WHO 1992). However, depending on the ability and willingness of the collector, human
error can reduce the accuracy of counts made by hand catches. Pyrethrum spray catches
are not only expensive but can cause unnecessary exposure to chemicals. A wide variety
of artificial resting places have been tested to sample outdoor resting mosquitoes, but
relatively few artificial resting places have been used indoors. Those tested indoors have
seen limited success (Service 1993, Komar et al.1995, Smith 1942). Yasuno et al. (1977)
tested the use of plywood boxes as a method to sample indoor mosquitoes, but found the
boxes worked only in high densities of mosquitoes and low humidities. Resting boxes
made of cardboard and black muslin cloth caught 30-60% of all Aedes aegypti collected
indoors by hand catches (Edman et al. 1997). Das et al. (1997) developed an insecticide-
impregnated fabric (IIF) trap for use indoors, but this is an impractical tool in much of the
rural settings in Africa where many of the materials needed for construction are not
readily available. In Kenya, Sexton et al. (1990) used a 5 x 6 ft reed ceiling mat from
which to make weekly hand catches of indoor resting densities of Anophelines but did
not test this method against other kinds of indoor resting collections
Malaria has long been a plague of mankind. Evidence of this is seen in the high
host specificity of the four malaria species that infect humans, suggesting a long
association between humans and the parasites (Gilles 2002). Archeological evidence
suggests that human malaria existed in the eastern Mediterranean region as early as the
beginning of the Neolithic Period (9,500 BC) (Harwood and James 1979). In the fifth
century, Hippocrates was the first to describe the clinical picture and some complications
of malaria. The English word "malaria" is derived from the Italian mal aria meaning
"bad air." The French word "paludisme" and the Spanish "paludismo" also refer to
malaria and come from the Latinpalus meaning "swamp." Malaria, paludisme, and
paludismo have all originated from the idea that malaria was the result of inhaling "bad
air" from swamps (Foster and Walker 2002, Gilles 2002, Harwood and James 1979).
The first description of the malaria parasites in human red blood cells was from the
French army surgeon Laveran in 1880. Seventeen years later Ronald Ross, a physician in
India, discovered a developing form of malaria parasite in the body of an infected
mosquito, greatly aiding the understanding of malaria transmission. In the late 1950s the
WHO launched a global campaign that successfully eradicated malaria in much of the
world's temperate zones including parts of North America, Europe, and Australia.
Control of malaria was not as successful in many tropical areas. In fact, in the past
twenty-five years there has been a significant increase of malaria incidence in tropical
areas such as southeastern Asia and tropical Africa. In 1998, the WHO stated malaria to
be returned to its list of top priorities and introduced a new initiative called, "Roll Back
Malaria." The success of this program has yet to be evaluated and reported.
Importance of the Disease Worldwide
Today, malaria is the most important insect-borne disease in public health (Durden
and Mullen 2002, Foster and Walker 2002). Malaria, along with AIDS and tuberculosis,
are the three most important infectious diseases in the world. Malaria affects around 40%
of the world's population, much of which live some of the poorest countries in the world
(WHO 2004, WHO 1995). In 1995, countries with intense malaria transmission had
income levels equal to roughly a third of those in countries without malaria (Gallup and
Sachs 2001). In Africa alone, it is estimated that the annual costs of malaria exceed US
$2 billion (WHO 2000). Malaria not only affects poor people, but it keeps them poor.
The disease is found mainly in subtropical and tropical climates. In the past,
malaria was more ubiquitous and was found in temperate regions including areas in the
United States (Gilles 2002, Honigsbaum 2001). In the early 1900s, 6 to 7 million cases of
malaria were reported in the continental United States annually (Harwood and James
1979). Now malaria is responsible for more than 300-500 million illnesses and a least
one million deaths each year (WHO 2004, WHO 1995).
Description of Disease Transmission
Human malaria is caused by one of four species of parasite in the genus
Plasmodium, family Plasmodiidae, suborder Haemosporiidae, order Coccidae (Sinden
and Gilles 2002, Marquadt et al 2000). The life cycle of Plasmodium species is complex,
including an exogenous phase (a cycle in mosquitoes) and an endogenous phase (a cycle
in humans) (Foster and Walker 2002). The four species of Plasmodium infecting humans
are P. falciparum, P. vivax, P. ovale, and P. malariae. These parasites are transmitted to
humans by mosquitoes of the genus Anopheles. The first sexual developmental stage of
the parasites, the gametocytes, are ingested along with asexual stages by the mosquito in
the bloodmeal of an infected human. After fertilization occurs a diploid zygote is formed.
The zygote (or ookinete) moves through the peritrophic membrane of the mosquito gut.
Depending on the species of parasite, 8 to 16 days later the ookinete ruptures, releasing
sporozoites that make their way to the mosquito's salivary glands where they can be
passed on to another human upon a subsequent bloodmeal (Sinden and Gilles 2002,
Foster and Walker 2002).
The severity of disease is dependent on the species of parasite and the general
health and immune status of the infected person (WHO 1995). Cases of mixed infections
of two or more Plasmodium species are not uncommon (Sinden and Gilles 2002).
Nonfatal infections that are left untreated can last more than five months depending on
the immune system of the individual (Foster and Walker 2002). Common symptoms are
febrile paroxysm (violent attack of fever), myalgia (muscle pain), headache, nausea,
diarrhea, and vomiting.
In areas of stable malaria transmission (little variation in transmission over several
years) some degree of acquired immunity is common in adults who survived bouts of
malaria as children. This means that infected adults may be asymptomatic or exhibit only
slight clinical symptoms. Effects of immunity to malaria on individual are; 1) prevention
of infection with the same species of parasite, 2) reduction in parasite multiplication, 3)
destruction of parasite, and 4) aid in tissue repair (Marquadt et al. 2000). This immunity
is no doubt beneficial to the individual, but typically in such cases proper treatment is not
sought and the infected immune person can serve as a suitable host for the disease. In
areas of unstable malaria transmission (much variation in transmission over several
years), epidemics become a problem since people in these regions aren't exposed to the
disease long enough to develop immunity.
Children under the age of five and pregnant women are the most vulnerable to
malaria infection. Mortality in children results from three main presentations of malaria.
One being an acute infection (often presented as seizures or coma) killing the infected
child quickly. The second is the development of severe anemia from repeated infections
of malaria parasites. The final is a low birth weight often the consequence of the mother
being infected with malaria during her pregnancy. This last presentation is a major risk
factor during the child's first month of life. Malaria can also cause children to become
more susceptible to other infections including respiratory illnesses, diarrhea, and other
common childhood illnesses. In regions of unstable malaria transmission, pregnant
women are at extremely high risk of maternal and perinatal death. In regions of stable
malaria transmission, infection is usually asymptomatic due to some acquired immunity,
but commonly causes severe maternal anemia and babies with a low birth weight. A
curious occurrence that results from malaria during pregnancy is that there is often a
heavy infection of parasites concentrated in the placenta, impairing fetal nutrition. The
reason for this is poorly understood (Shulman and Dorman 2002).
Importance of Malaria in Africa
Of at least one million annual deaths attributed to malaria, 90% occur in sub-
Saharan Africa with the majority (90%) of these deaths occurring in children (WHO
2003). For many of the sub-Saharan countries more than a quarter of hospital admissions
are due to malaria. Data suggests that the number of malaria cases have increased over
the last decade (WHO 2004). Plasmodiumfalciparum is most common in Africa and is
responsible for the majority of malaria deaths worldwide (WHO 2003).
The appearance of a P. falciparum infection can range from almost asymptomatic,
to an acute febrile, illness, to severe life-threatening cerebral malaria sometimes resulting
in coma. Cerebral malaria can also cause anemia (lower than normal amount of
hemoglobin or red blood cells), hemoglobinuria (presence of red blood cells in the urine),
jaundice, hypoglycemia (low blood sugar), renal dysfunction (improperly functioning
kidneys), psychosis, shock, and pulmonary edema (fluid in the lungs) (Warrell 2002).
Important Vector Species in Africa
The most important and well-studied vectors of malaria are found in the An.
gambiae complex. The complex consists of seven described species: An. gambiae sensu
strict (s.s.), An. arabiensis, An. merus, An. melas, An. quadriannulatus, An.
quadriannulatus species B, andAn. bwambae (Service 2002). Their behavior, vector
status, and distribution differ in various aspects. Anopheles quadriannulatus and An.
quadriannulatus species B feed mainly on cattle and are not vectors of malaria.
Anopheles bwambae is a rare mosquito found only to breed in the Semiliki forest in
Uganda and is not considered an important malaria vector. Anopheles merus is a malaria
vector found in lagoons and mangrove swamps along the coast of West Africa. This
species breeds only in salt-water. Anopheles melas is considered to be the East African
equivalent to An. merus breeding in salt-water lagoons and swamps.
The two most important species in the complex are An. arabiensis and An. gambiae
s.s. The first is found in regions of dry savannah. The second is found in more humid
climates and readily becomes anthropophilic and endophilic (preferring to feed on
humans both outdoors and indoors and to rest indoors). Because most of its time is spent
indoors close to people, it is considered a highly efficient vector for malaria. Anopheles
arabiensis also exhibits some endophily but will also rest and feed outdoors, sometimes
feeding on domestic animals. Often An. gambiae s.s. will feed at least two times before
beginning its gonotrophic cyle (Foster and Walker 2002). The extra blood meals act as a
substitute for sugar (Foster and Walker 2002). Anopheles gambiae s.s. is found in almost
all sub-Saharan countries. Collectively all the species in the An. gambiae complex are
referred to as An. gambiae sensu lato (s.l).
The most important vector in Africa after An. gambiae s.s. and An. arabiensis is
An. funestus. It has a widespread distribution south of the Sahara and feeds
predominately indoors. The larvae are found in more permanent waters associated with
vegetation such as swamps and marshes. Anophelesfunestus feeds both indoors and
outdoors and prefers to rest indoors after feeding.
Malaria has been, and continues to be, one of the world's most important diseases.
Despite extensive attempts at eradication and control, the malaria situation has worsened
over the past two decades and parasites are re-emerging in areas where they had been
successfully or nearly eradicated, such as the Republic of Korea, Iraq, and Turkey (WHO
2000). Because malaria affects many of the world's poorest nations, control of the
disease has been difficult due to a lack of a solid health infrastructure. Sub-Saharan
Africa is most affected by disease and is home to An. gambiae, one of the world's most
efficient malaria vectors. It is considered a highly efficient vector because it prefers to
feed on humans and will take multiple blood meals (Foster and Walker 2002, Service and
African malaria vector control and research programs often focus on sampling adult
populations of An. gambiae s.l. (Giles), An. funestus (Giles), and other important vectors.
A major constraint in many developing countries where malaria is rampant is the lack of
funds available to such programs. Commonly used methods for sampling adult
mosquitoes include hand catches with oral or mechanical aspirators, Pyrethrum Spray
Catch, and CDC light traps. These methods can become costly and labor-intensive,
especially for community-based malaria control programs. The goal of this study was to
develop a technique for sampling adult malaria mosquitoes for use in research and
community control programs. To accomplish this, the following five objectives were
1. Test the efficacy of simple resting boxes for sampling An. gambiae s.s. mosquitoes
resting indoors in semi-field conditions (screen-walled greenhouse). The design of
the semi-field test structure is described by Mathenge et al. (2002).
2. Identify the preferred natural indoor resting sites for common malaria vectors, An.
gambiae s.l. and An. funestus.
3. Assess the representative accuracy of the sampling of mosquito specimens through
hand catches by an experienced collector using an oral aspirator.
4. Compare the accuracy of sampling common malaria vectors using a technique
utilizing indoor resting boxes and bednets to 15 minutes of hand catches by an
experienced collector using an oral aspirator.
5. Test the efficacy of resting baskets for sampling An. gambiae mosquitoes indoors
in semi-field conditions (screen-walled greenhouse).
MATERIALS AND METHODS
All study sites were located on the shores of Lake Victoria in the Suba District,
within Nyanza Province, Western Kenya (altitude 1100-1300m)(Fig. 3.1). All semi-field
trials were conducted at the International Centre for Insect Physiology and Ecology's
(ICIPE) Biological Station in the town of Mbita Point. Fields trials were conducted in the
nearby village ofLwanda Nyamasare. The two sites are within approximately 9 km of
each other. The area has two rainy seasons from March to June and October to
December with an average annual rainfall of 700-1800mm. Average temperatures range
from 160 to 340 Celsius. Malaria is considered holoendemic in the region. Suba District
is home to around 156,000 people with the majority of these people belonging to the
patrilineal Luo ethnic group. Luos make a living primarily through fishing and
subsistence agriculture. Description of the area is given in Mathenge et al. (2002),
Geissler et al. (2000), Okech et al. (2003), and Knols et al. (2002).
Survey of Indoor Resting Habitats of Mosquitoes
From March 27th to April 18th2004 at varying times of the day (8:25 to 16:45 hr.),
thirty-two dwellings were visited for inspection by a local expert mosquito collector with
eight years experience in collecting resting mosquitoes with a plastic tube oral aspirator.
This method for sampling adult mosquitoes resting indoors, often referred to as hand
collection or hand catches, is a widely used method and provides a suitable technique to
quantify the resting site of each collected mosquito (Service 1993, WHO 1992).
Suba Disrict /
e kaL Victoria
bita (ICIPE) KENYA
-'-- -- /r y
.. 4 0/'
/... -. |
J '--" 3^-
J /- '' .
4 / 0d
4 ___ 0 __4 __ aKoftr
Figure 3-1. Map of Kenya showing the study sites.
The technician thoroughly searched each dwelling (Fig. 3.2) collecting mosquitoes
and aspirating them into paper cups covered with netting. Searching involved not only
walking around the dwelling and collecting visible mosquitoes, but also looking under
and behind furniture, behind curtains, around and under pots and pans, etc. Cups
containing live collected specimens were placed in a sealable container with cotton balls
soaked with ethyl acetate to prepare the specimens for processing. Nine dwellings were
revisited five to ten days after the first collection. Two separate rooms, each of similar
size to a single dwelling, were checked in the same large house. The technician was given
as much time needed to collect all mosquitoes seen. The time spent collecting at each
house was noted.
Dwellings were made of wood and mud with a thatch or a corrugated iron sheet
roofs or made completely out of wood and corrugated iron sheets. All dwellings were
built without ceilings with the underside of the roof exposed. Dwellings were all
rectangular except for one circular house. The lengths of two adjacent walls inside all
rectangular dwellings were measured to find the total area inside. For the circular
dwelling, the diameter was measured to find the area. The height from the floor to the
highest point of all dwellings was also measured from the inside.
IA ) B)
Figure 3-2. Dwellings in Lwanda. A) Single family structure B) Multiple family
The height where each mosquito was collected was recorded as high, medium, or
low. A mark of "high" required the technician to lift his arm above his shoulder (roughly
161 to 240 cm from the ground). "Medium" was designated as any collection made from
the technician's shoulders to his waist (approximately 81 to 160 cm from the ground).
Mosquitoes caught below the waist of the technician, causing him to bend over or kneel,
were designated as "low" (0 to about 80 cm from the ground). The object (wall, jug,
pots, hanging shirt, etc.) the mosquito was collected from was also recorded. The objects
on which the mosquitoes were collected were later divided into 5 different categories
2. plastic (plastic items such as bottles and basins)
3. cloth (items such as clothes and bed-nets)
4. furniture (more permanent items less likely to be moved around such as a bed, a
cabinet, chairs, etc.)
5. temporary (found on items likely to be disturbed or moved such as lanterns or
Finally, the number of all mosquitoes that were collected in an area deemed
"hidden" or protected was noted (Fig. 3.3). These categories are similar to the studies
completed by Pal et al. (1960) and Wattal and Kalra (1960).
Figure 3-3. Objects in dwellings on which mosquitoes can be found resting.
Representative Accuracy of Hand Collection
Twenty-one dwellings of similar size and make were visited by the expert collector
and the investigator from April 1st to April 16th 2004. The walls were constructed of
either wood and mud or wood and corrugated iron sheets. The roofs were made from
either corrugated iron sheets or thatch. All dwellings were without ceilings. Fourteen
were in multiple family units and seven were single family units. Mosquitoes were
collected by hand using a plastic tube oral aspirator by the collector until an entire search
of the house was completed (no more mosquitoes found). The time needed to hand
collect all mosquitoes was noted. The time of day the collection took place ranged from
8:25 to 16:30 hr. Immediately after finishing a search of the dwelling, a Pyrethrum Spray
Catch (PSC) was performed. A commercial-grade pyrethroid insecticide (Doom Fast
KillTM, manufactured by Mortein, Australia) in 300g/494 mls containers were substituted
for pyrethrum because it could be easily purchased and was safer (higher dilution of
chemicals than the industrial grade) to use for the investigator and families living in the
dwellings. The active ingredients of the insecticide were d-phenothrin (Pyrethroid) 1.0
g/kg and imiprothrin (Pyrethroid) 0.4 g/kg. For each dwelling the combined number of
mosquitoes collected from the PSC and hand collection was found giving a total
complete catch for each dwelling. From this combined number, the percentage collected
by hand was found.
Development of Cloth Indoor Resting Box in Semi-Field Conditions
Semi-field trials were conducted from April 3rd to June 15th, 2004 at the ICIPE,
Mbita Point Biological Station at Mbita Point, western Kenya (000 25'S, 340 13'E). A
30cm x 30cm x 30cm cloth resting box, similar to the design described by Crans (1989)
was tested for its ability to sample indoor resting mosquitoes. The box was made of a
plain 2 cm thick galvanized wire frame with blue cotton cloth. Blue cloth was sewn to
cover the outside of the box with black cotton cloth sewn to cover the inside. The
attractiveness of dark colors to mosquitoes is discussed in Bidlingmayer (1994). A flap
made of mosquito netting (mesh size 196) with a sleeve was sewn to the top of the box.
This was closed to facilitate capture when the number of mosquitoes in the box was
The box was hung approximately 50 cm from the ground in the opposite right
corner from the door of an experimental hut made of plywood (3.2m x 2.7m x 1.7m)
inside a modified screen-walled 11.4m x 7.1 m x 4.2m greenhouse (Cambridge Glass
House Co. Ltd., U.K.) (Fig. 3.4). Both hut and greenhouse are described by Mathenge
(2002). The box was hung from the far right corner because it was the opposite corner
from the door, a large source of light in the hut. Each night one of three different
densities of female An. gambiae s.s. mosquitoes from Mbita strain colony at ICIPE (low
= 50, medium = 100, and high = 200, as in Mathenge et al. 2002) was released into the
greenhouse at approximately 21:00 hr. For each density, mosquitoes were allowed to
feed on 10% sugar solution for at least 24 hours after emergence. Half of the females
were not allowed a bloodmeal and half were allowed to feed on blood for ten minutes
daily for three nights prior to release. This adult rearing protocol is approved by the
Kenya Medical Research Institute and the Kenyan National Ethical Review Committee
(2001). The resting box was checked hourly from 7:00 to 12:00 hr and again at 15:00 hr.
All mosquitoes caught in the box were counted and removed each time. The number of
mosquitoes collected each hour was noted. After the final collection of the day, two
technicians searched and collected the remaining females in the greenhouse and then
disposed of them. One similarly designed resting box treated with deltamethrin (the cloth
soaked in 2 liters of water with a 25% m/m concentration) was also tested in preliminary
tests but was discarded when almost 15% fewer mosquitoes were caught. The
deltamethrin was chosen for testing because it could be purchased nearby and in
preliminary trials some of the released mosquitoes were found dead in the box making
collection of mosquitoes easier.
Comparison of Accuracy of Resting box/Resting Net to Hand Collection
A comparison of accuracy of the collection methods was performed in Lwanda
Nyamasare village from April 13th to June 30th 2004.
A ,i B.
,- ." *'. C )
Figure 3-4. Components of the experimental tests A) Resting box hung in experimental
hut. B) Experimental hut in modified greenhouse. C) Modified greenhouse.
Preliminary trials suggested that a double size blue bednet (Supanet TM) hung from the
highest point inside a dwelling and then spread out and tied to the top part of the eaves
could facilitate in capturing resting mosquitoes out of reach from the technician
performing hand catches. To aid in collection this resting net could be lowered within
reach of the technician when needed. All dwellings were made with mud walls; two had
thatched roofs, the other had a roof composed of corrugated iron sheets. Three dwellings
of similar size and make were chosen to compare the efficacy of three methods of indoor
resting collection in a Latin Square design. The three collection methods compared were
1. A cloth resting box (described in previous semi-field trials) and a resting net
2. A plain cardboard box of similar size to the cloth resting box (42 cm by 21.5 cm by
25 cm) and resting net
3. Aspiration for 15 minutes using a plastic tube aspirator.
The time (15 min) chosen for hand collection was based on the results from the
previous survey and recommended in WHO (1992). When trials involved hand
collection, the resting net was tied in a knot and placed out of the way for the
homeowner. Both the cardboard box and bednets (for resting nets) are items that can
easily be found in Lwanda. The blue cloth resting box was made from materials found in
Mbita (12 km away). Immediately after mosquito collections at each dwelling, a PSC
was performed (as described previously). Following the completion of spray catches at
each house the collection methods were rotated. Dwellings were tested every 3 days to
allow ample time for the insecticide sprayed during spray catches to dissipate. The
number of days was chosen based on preliminary data from the Latin square. Trials
showed that numbers of mosquitoes caught by PSC increased from the previous spray
catch after two days. The sex and gonotrophic status of all captured mosquitoes was
noted. The categories for gonotrophic status (bloodfed, half gravid, gravid, and unfed)
was taken from Service and Townson (2002). The mosquitoes collected were identified
as either An. gambiae s.l., An. funestus, Culicine, or unidentified Anopheline. The Latin
square was run for a total of 27 experimental days in 3 complete rotations. The time
taken to complete sampling of a dwelling using the resting box (cloth or cardboard) with
a resting net was also noted.
Figure 3-5. Treatments used in the Latin Square. A) Cloth resting box hanging in
dwelling B) Cardboard box hanging in dwelling C)Diagram of a resting net
spread out in a dwelling D) Expert collecting mosquitoes by aspiration.
Development of an Alternative to Cloth Box in Semi-Field Conditions
To evaluate the possibility of using a wicker basket (30 cm tall with an opening of
28 cm in diameter) with an attractive black cotton cloth lining inside (Bidlingmayer
1994) as method to sample indoor resting mosquitoes, a preliminary field trial was run
from April 23rd to June 30th 2004 in the village of Lwanda Nyamasare (Fig. 3.6). The
'resting basket' was hung approximately one meter from the floor in a corner of a
bedroom in a single family dwelling. The opening was faced toward the middle of the
room. The dwelling had walls made of mud and stick with an iron corrugated sheet roof.
Two to five people slept there nightly. Collections from the basket were made once at
approximately 10:00 hr at least every three days. The number of mosquitoes collected
From May 15 to July 13, 2004 trials were run in semi-field conditions using a
resting basket of equal size and make to the one tested in the field. The basket was hung
approximately 50 cm from the ground placed in the far right corer opposite the door of
an experimental hut made of plywood (3.2 m x 2.7 m x 1.7 m) (Fig. 3.6) inside a
modified screen-walled 11.4 m x 7.1 m x 4.2 m greenhouse (Cambridge Glass House Co.
Ltd., U.K.). Both hut and greenhouse are described in Mathenge et al., (2002). Each
night one of three different densities of female An. gambiae s.s. mosquitoes (low = 50,
medium = 100; and high = 200, as in Mathenge et al. 2002) from the ICIPE Mbita strain
colony was released into the greenhouse at approximately 21:00 hr. The method for
testing the recapture rate of the resting basket was the same as the previously described
trials using the cloth resting box. For both low (50 females) and high (200 females)
densities, ten trials were run while thirty-two trials were run for the medium (100
r_| "" .
Figure 3-6. Alternative to the cloth resting box. A) Cloth resting box and resting basket
B) Resting basket hanging in dwelling
RESULTS AND DISCUSSION
Survey of Indoor Resting Habitats of Mosquitoes
Dwellings typically held one family. Some dwellings were part of a single large
structure which consisted of at least three or more dwellings (rooms with doors to the
outside) connected under the same roof (Fig. 4.1).
All style of dwellings in the study area remained quite constant among single and
multiple family structures. Dwellings consisted of a rectangular room divided by a
hanging cloth sheet, reed mat, or mud wall. The sleeping area was always on the
opposite side of the divider from the door. Structures containing a single family tended
to be larger in area and had a higher roof than dwellings in multiple family structures
In the dwellings mosquitoes were collected off a variety of different substrates with
around a quarter of those found in areas hidden from plain view (Table 4.2). The category
"Hidden (protected)" was noted in addition to the substrate. For example, a mosquito
collected from aplastic item was found in an area on the item that was "hidden" from
plain view of the collector. A General Linear Model procedure using a Wilcoxon rank
sum test and a Kruskal-Wallis test was performed to analyze the percent of mosquitoes
caught at low, mid, or high heights in the dwellings (SAS 2001). There was no significant
difference between the heights the mosquitoes were found (Table 4.3).
or reed mat O
Corrugated iron 0
sheet oor Reed mat, cloth,
I or mud divider
*Most collection done
on this side of divider,
away from light of door*
' pots, etc.
Table and chairs
Figure 4-1. Layout of dwellings. A) Single Family B) Multiple Family
Table 4-1. Mean area and height of dwellings in study area
Structure Mean area ( Std) Mean Height ( Std) Number
Single Family 14.4 + 4.2 m2 3.4 0.4 m2 13
Multiple Family 9.5 2.5 m2 2.7 0.4 m2 14
Representative Accuracy of Hand Collection
Out of the 531 mosquitoes caught in 31 dwellings by both hand collection and the
PSC performed afterwards, hand collection caught 32.2% of the total of the two methods.
Development of Cloth Indoor Resting Box in Semi-Field Conditions
Trials using a treated box were stopped once it became apparent that the percent of
the mosquitoes released that were captured was less than the untreated. The mean
percentages for each of the densities tested on the untreated box were similar (Table 4.4).
The total mean percent of mosquitoes recaptured from all trials of the untreated box
(Table 4.4) was similar to the percent caught by hand collection in 31 dwellings (32.2%).
Table 4-2. Percent of total actual number of mosquitoes (437) collected from each
Furniture (more permanent) 30%
Temporary items 8%
Plastic items 4%
Hidden (protected) _26%*
*The category "Hidden" was noted separately.
Table 4-3. Total actual number of mosquitoes collected at each height. Note: Numbers
followed by the same letter are not significantly different (P<0.05), df = 2,
Chi-square = 0.97.
Height dwelling where collected (in cm) Number of mosquitoes caught
0 to 80 "Low" 176a
81 to 160 "Medium" 111a
161 to 240 "High" 119a
Table 4-4. The mean percent of mosquitoes recaptured in semi-field conditions using a
resting box either treated or untreated with deltamethrin. (Mean Standard
So Density-50 Density-100 Density-200 Total
Type of box
T ffemales females females percentage
39.4% + 10.6 38.3% + 8.9 30.7% + 8.5 36.1% + 9.9
N=10 N=10 N=10 N=30
15.3% + 3.0 25.3% + 10.9 27.3% + 6.2 22.6% + 9.6
N=3 N=3 N=3 N=9
Note: (N=number of trials performed). *Trials with the treated box were stopped early
once the it was shown that the it caught fewer mosquitoes than the treated.
Comparison of Representative Accuracy of Resting Box/Resting Net to Hand
The total number of mosquitoes collected during the Latin Square experiment using
each method (hand collection, cloth resting box and resting net, and cardboard box and
resting net) was calculated (Table 4.5). For each method, the total number of mosquitoes
caught by a method was added to the numbers caught by Pyrethrum Spray Catch (PSC)
performed after the same method giving an actual overall total (Table 4.5). The percent
of the overall total number of mosquitoes collected using each method was found by
dividing the total actual number mosquitoes caught a method by the overall total number
of mosquitoes caught for each method (Table 4.5)
Table 4-5. Comparison of actual number of mosquitoes collected by the three methods
Hand Collection Blue resting box and Cardboard box and
resting net resting net
mosquitoes caught 1166 1020 1060
(PSC + method)
caught using 166 282 256
Percent of overall
mosquitoes 14.2% 27.6% 24.1%
Note: (N=number of trials performed).
The cloth resting box/resting net caught 1.7 times the number of mosquitoes and
almost twice the percentage of the total mosquitoes as hand of collection (Table 4.5).
The cardboard box/resting net caught 1.5 times the number of mosquitoes and almost
10% more of the percentage of the total mosquitoes as hand collection (Table 4.5). The
percentages captured by each of the methods on each test day were transformed to
achieve linear model assumptions by:
where x is the percentage of mosquitoes captured by a method on each test day. The
results were then analyzed using the Tukey Multiple Comparison Procedure (SAS 2001).
The catches from the cloth resting box/resting net method were significantly different
than hand collection (significance was set at P<0.05) (Table 4.6). There was no
significant difference between the cloth resting box/resting net and cardboard box/resting
net methods nor between the cardboard box/resting net and hand collection. The time
taken to complete sampling using a resting box/resting net method was slightly shorter
than the 15 minute time suggested by WHO (1992) for hand collection (Table 4.7).
The results of the identification to species, sex, and gonotrophic status showed An.
gambiae s. 1. to be the most collected of the four types of mosquitoes (An. gambiae s.l.,
An. funestus, Culicine, and an unidentified anopheline) identified from the Latin Square
catches (Tables 4.8-4.11). The highest numbers of An. gambiae s.l. mosquitoes found in
the dwellings were bloodfed (882 mosquitoes) and unfed females (884 mosquitoes)
(Table 4.8). This trend was also seen with An. funestus (Table 4.9). For both Culicine
and unidentified Anopheline mosquitoes, males were the most collected in the dwellings
(Tables 4.10 and 4.11)
Development of an Alternative to Cloth Box in the Semi-Field
The resting basket caught a mean of 4 mosquitoes daily in one dwelling in the field
(Table 4.12). A General Linear Model procedure was performed on the percentages of
mosquitoes recaptured using a Type III sum of squares test for the total percent of
mosquitoes recaptured (SAS 2001). There was no significant difference (P<0.05) in the
total percent recaptured between the two methods (Table 4.13). There was no interaction
between the density of mosquitoes and the methods used but a difference was found
among the densities of mosquitoes. For further analysis, a Tukey-Kramer adjustment for
multiple comparisons was performed on the pooled (resting box and resting basket)
percent of mosquitoes recaptured by each density (SAS 2001). Results showed a
significant difference between low and high densities (Table 4.14
Table 4-6. Comparison of the accuracy (percent of the overall number of mosquitoes
found in each dwelling) over the 27 test dates for the three methods tested.
Methods followed by the same letter are not significantly different (P<0.05).
Note: F= 3.37, df= 2, 50.
Date Hand Collectiona
Mean percentage for a
each method SE
Note: F= 3.37, df= 2, 50
Table 4-7. Comparison of time taken to complete sampling using a resting box (cloth or
cardboard) and resting net to hand collection.
Mean time taken to
Number of Time suggested by WHO
complete sampling using
trials using a resting box complete sampling using (1992) to complete a search
d r g n resting box and resting net with hand collection
and resting net with hand collection
+ Standard Error
31 8.0 3.9 min 15 min
Actual number of An. gambiae mosquitoes caught by method, including sex
na d gonotrophic status during the 27 te e
Method Bloodfed Gravid af Unfed Male Total
Cloth box 18 4 8 7 9 46
Cardboard 6 4 4 3 0 17
nd 60 21 10 30 12 133
Resting 163 53 26 103 35 380
PSC 635 265 138 741 243 2022
Total 882 347 186 884 299 2598
Note: For the resting net method, the numbers from both nets used in the Latin Square
were calculated together.
Table 4-9. Actual number of An. funestus mosquitoes caught by method, including sex
and gonotrophic status, during the 27 test days of the Latin Square
Method Bloodfed Gravid af Unfed Male Total
Cloth box 1 0 3 0 0 4
Cardboard 0 0 2 0 0 2
nd 4 1 2 1 1 9
Resting 7 2 5 2 2 18
PSC 26 16 8 29 14 93
Total 38 19 20 32 17 126
Note: For the resting net method, the numbers from both nets used in the Latin Square
were calculated together.
Table 4-10. Actual number of Culicine mosquitoes caught by method, including sex and
gonotrophic status, during the 27 test days of the Latin Square.
Bloodfed Gravid Hf Unfed Male Total
Cloth box 0 1 0 0 0 1
Cardboard 0 0 0 0 0 0
and 0 0 0 0 1 1
Resting 0 2 2 2 3 9
PSC 8 27 4 17 24 80
Total 8 30 6 19 28 91
Note: For the resting net method, the numbers from both nets used in the Latin Square
were calculated together.
Table 4-11. Actual number of unidentified anopheline mosquitoes caught by method,
including sex and gonotrophic status, during the 27 test days of the Latin
Method Bloodfed Male Unfed Total
Cloth box 0 0 0 0
Cardboard box 0 0 1 1
Resting Net 0 0 1 1
PSC 4 15 5 24
Total 5 15 7 27
Note: For the resting net method, the numbers from both nets used in the Latin Square
were calculated together.
Table 4-12. Mean number of mosquitoes caught daily by a resting basket placed in a
house in the village of Lwanda.
Mean number of mosquitoes caught Number of days tested
3.9 3.3 33
Table 4-13. Comparison of the resting basket to the cloth resting box of the total
percentage of mosquitoes recaptured in semi-field conditions.
Type of trap Total mean percent of mosquitoes
Typerecaptured ( SE)
_________J __________________recaptured ( SE )____
33.8% + 8.9a
36.1% + 9.9a
Note: N=Number of trials run for each method, percentages followed by the same letter
are not significantly different (P<0.05), F= 0.51, df = 1, 76)
Table 4-14. Mean percent of mosquitoes recaptured by each density used for both the
resting box and resting basket.
Density Mean Percent of Mosquitoes Recaptured
Note: N=Number of trials run for each method, percentages followed by the same letter
are not significantly different (P<0.05), F= 41.47, df = 1, 76)
Indoor Resting Places of Mosquitoes
The most important malaria vector in Africa, An. gambiae, will readily rest in the
dwellings of rural villages (Service and Townson 2002). A common method for
sampling populations of these resting mosquitoes is a search of a dwelling by a trained
technician using an oral aspirator (Service 1993, WHO 1992). A properly trained
technician performing the search is essential as the mosquitoes are difficult to find and
catch. This idea is reinforced by the results of this study. Since it is also generally
thought that both male and female mosquitoes are attracted to darker areas indoors, one
would expect to find mosquitoes resting at low heights around the dwelling since they
would be further away from a main light source, the eaves. Although a larger percent of
mosquitoes were found at heights from 0 to 80 cm, the number of catches was not
significantly different from those higher than 80 cm. About a quarter of the mosquitoes
(26%) were also found in an area not in plain view of the collector. The results of this
study would suggest that it would be difficult to accurately pinpoint an area in a dwelling
to focus the aspiration as mosquitoes. The mosquitoes were found on a broad variety of
substrates at various heights with many resting in areas that would require the collector to
search under, above, or around household items. Many times during the study the
collector needed to move furniture, pots, clothes, and other household items to be able to
reach the mosquitoes. This emphasizes the need for adequate training for anyone
attempting to sample resting mosquitoes by hand collection using an oral or mechanical
Comparison of Sampling Techniques
The local expert technician had eight years experience with hand catches, likely
more than most people attempting that method. This would suggest a high quality of
collection in comparison to other hand collectors. Since the expert was given as much
time needed to complete a search of each dwelling during the initial part of the study, it
was assumed that the majority of the mosquitoes collected in the PSC, performed after
the search, were resting above the reach and out of the sight of the collector. Since the
expert caught only approximately a third (32%) of the total number found, it was decided
that any sampling method developed should account for mosquitoes resting on the
exposed underside of roofs (made of thatch or corrugated iron sheets) as this would be a
likely area to find them (Lindsay et al. 1995, Schofield and White 1983).
Resting boxes have been a well established method for sampling mosquitoes
(Service 1993). Few, however, have been used indoors. The results from the semi-field
trials suggest that the cloth resting box could collect close to the same percentage of
mosquitoes (36.1%) as the expert hand collector in the field (32.2%). However, it is
likely that a cloth resting box placed near a similar location in a dwelling could not
accurately sample mosquitoes that prefer to rest in the underside of the roof. This
assumption is based on the idea that since the expert technician was not able to sample
mosquitoes resting so high, a stationary trap would also have the same difficulty. For this
reason, a resting net was tested with the cloth resting box. The results of the Latin Square
trial suggest that the cloth resting box/resting net method could catch a significantly
higher percentage of the total population resting in a dwelling than that of an expert
performing hand catches using an oral aspirator.
To increase the practicality of the method the resting basket was tested in the semi-
field. Because there was no significant difference in recapture rates between the resting
basket and the cloth resting box, the resting basket could be an acceptable alternative to
the cloth resting box. The cloth resting box (or resting basket) and resting net method has
a number of advantages over hand collection:
1. It can more accurately sample the population resting in a dwelling (a higher
percentage of mosquitoes will be collected).
2. The materials needed for construction are inexpensive and can be easily found in
many rural areas as opposed to the time and money to needed to train and/or hire a
technician to perform hand catches.
3. It requires less training, not only because mosquitoes are easier to see in the black
cloth of the box and in the bright blue of the net, but also because the investigator
only needs to search a set area (the trap and the net) rather than the whole house.
4. The time needed to complete a search is shorter than 15 minutes, which is
commonly assigned for hand catches (WHO 1992).
The use of resting boxes and nets also has advantages over the PSC technique.
Although a greater percentage of mosquitoes can be caught with spray catches, the spray
technique is too expensive and impractical for use on a regular basis for local control
programs. Pyrethrum spray catches also expose the investigators and the members of the
tested household to potentially harmful chemicals that may take several days to dissipate
(Service 1993). This study showed that cloth resting boxes or resting baskets used with
resting nets can be a practical method for local malaria control programs to sample indoor
Malaria continues to be one of the world's most important human diseases.
Although, has been eradicated in some areas, it continues to be a threat to 40% of the
world's population. Many of those people affected by the disease live in some of the
poorest nations in the world. The lack of resources and infrastructure has produced major
obstacles to malaria control in those countries. Because malaria affects many people
living in impoverished nations, many of the more expensive and technical methods
utilized by developed countries for mosquito and disease control are not feasible.
An initiative to create sustainable malaria control programs at a community level
has been implanted by international organizations such as the WHO (2000). As local
malaria control programs become more accepted, technology is needed to keep such
programs sustainable. Developing new methods and tools for malaria surveillance and
control, which are both user-friendly and economically viable, are essential for this to
happen. Because the most important vector of malaria in Africa, An. gambiae s.l., readily
rests indoors after feeding, developing practical ways to sample populations of those
mosquitoes would be valuable.
The sampling technique tested in this study, was developed specifically for malaria
control programs based at community levels. The materials needed for the technique can
be found inexpensively in rural villages where community malaria control programs are
run. The technique provides a more user-friendly alternative to both PSC and hand
catches, both of which are commonly used in malaria control programs and research
(Service 1993, WHO 1992).
Members of control programs sampling with the new technique have only to collect
from a defined area (box or basket and net) compared to hand catches where a complete
search of a dwelling is required. Mosquitoes can also be seen easier on the black cloth in
the box or basket and on the blue cloth of the net than on the walls, thatch, and other
items in the dwellings. Since mosquitoes can be found and collected more easily than
hand catches, less training and experience is needed to effectively sample a dwelling.
Although sampling with PSC captures most, if not all, mosquitoes in a dwelling,
the cost of the materials needed (insecticide, sheets, protection against chemicals) is often
beyond the means of community malaria control programs. The use of such a method
could not be sustained for very long due to its expense and the constant need for
insecticide. The safe use and storage of the chemicals needed for PSC is also an
important issue for the safety of the people doing the sampling and the homeowners.
Since no chemicals are needed with the new technique, it can be performed safely by
members of control programs with no danger to the homeowners. Because the materials
needed for the new technique can be inexpensively purchased once, it can be used
Further research should focus on developing more inexpensive and practical
methods of malaria surveillance and control that are accessible to the people facing the
greatest disease load. Such research should be broadly based in multiple disciplines due
the complexity of the malaria problem. Things like access to healthcare, economic
stability, and cultural behavior also play important roles in the transmission of malaria
and need to be considered. Even though much of the world's malaria control stems from
international organizations, a larger global effort is needed to have a lasting effect on
NOTES ON THE DEVELOPMENT OF A PRACTICAL ANOPHELES GRAVID TRAP
The ability to sample the populations of host-seeking mosquitoes is essential to the
evaluation of malaria control programs. The exposure to potentially infectious bites on
people sampling host-seeking mosquitoes makes many such sampling methods ethically
questionable. As an alternative to sampling host-seeking adults, gravid traps can provide
similar data (reproductive capacity of mosquito populations and the dispersal patterns of
mosquitoes as they seek larval habitats after biting humans) without the risks of exposure.
Such information is of great importance to vector control programs as many programs
focus on and have seen success with larvicide application in potential or known larval
Gravid traps that have been tested and used in research and mosquito control
programs, sample only Culicine mosquitoes (Service 1993, Ritchie et al. 2004). Many of
these traps focus specifically on collecting gravid Culex mosquitoes (Service 1993,
Mboera et al. 2000). Although a variety of methods exist to sample the eggs of malaria-
carrying Anopheline mosquitoes, no technique has been devised to effectively trap and
sample these mosquitoes solely when they are gravid (Service 1993).
Gravid traps, like those used to sample Culicine mosquitoes, would become
problematic in many malaria endemic regions because they require a significant degree of
skill and maintenance to ensure they are working correctly. The commonly used, CDC
Gravid Trap (J. W. Hock Co. Gainesville, FL.), requires a net collection bag, motorized
trap with a fan, pan with water, a battery, and a battery charger. They would also be quite
costly to programs, particularly in African countries where resources available for control
are usually very limited. This study was undertaken to develop a much more practical
gravid trap that effectively samples egg-laying adults. A simple trap would have a
number of advantages over present traps:
1. It would not require skilled personnel for sampling mosquitoes
2. It would be far less expensive than current gravid traps
3. Data collected from traps will most likely reflect a more realistic composition and
behaviour of the egg-laying mosquito population since no artificial attractants will
4. It could be easily replicated and therefore be more likely to be used in mosquito
surveillance programs; since collected adults and eggs would be disposed, it could
directly aid malaria control programs.
The objective of this investigation was to develop a practical gravid trap using
materials readily available in rural areas where malaria is endemic. Such a trap would be
beneficial to malaria control programs at a community level where resources are often
Preliminary Semi-Field Trials in Florida
Semi-field trials were conducted weekly from June 8th to October 1st 2004 at the
United States Department of Agriculture's (USDA) Center for Medical, Agricultural, and
Veterinary Entomology (CMAVE) located in Gainesville, Florida. Traps were tested
singly in screened cages (2m x 2m x 2m). Prior to placement of the traps, 70 gravid
Anopheles albimanus females were released into each cage and allowed to acclimatize for
3-4 hours. The females were left in each cage with a trap for 24 to 48 hours after which
the number of females caught was noted. After collections were made the cage was left
open for at least four days to release any females left in the cage.
One design was tested and modified from oviposition basins from CDC Gravid
Traps (J. W. Hock Co. Gainesville, FL.), plastic funnels, and plastic 40 ounce peanut
butter jars. All designs were spray-painted with a matte finish primarily of black and
white (Fig. 5.1). Spray-painted traps were left outside for at least six weeks prior to
testing to remove any odors. Some variations of trap designs had modifications such as
the addition of black plastic mesh, the presence of cotton soaked in sugar water, and
various amounts of hay infusion (Fig. 5.2). All designs tested had approximately 2 liters
of well water placed in the basin.
collecting cartridge Holes poked through
jar to allow for air flow
with pieces of basin Ild
Larce funnel (black
outsidewhite inside) *
Plastic basin (from CDC gravid trap)
painted black on outside and
white on inside A) B
Figure 5-1. Design of oviposition trap. A) Full view. B) Collecting cartridge
Vial holding cotton soaked
in 10% sugar solution
Funnel held up with clothes hanger wire A) Blak plastic esh B) C)
Figure 5-2. Modifications to design. A) with clothes hangers B) with black mesh C) with
vial with sugar solution.
Semi-Field Trials in Kenya
From February 17th to March 11th 2004, 1 to 3 color variations of 4 trap designs
were tested daily in a screened cage (3m x 3m x 3m) at the ICIPE, Mbita Point Biological
Station at Mbita Point, western Kenya. A black plastic bucket (35cm x 30cm) was placed
in a hole in the middle of the screened cage. The bucket was placed so that the top was
even with the ground level. Some of the soil removed from the hole was placed back
into the bucket and water from Lake Victoria (of varying amounts depending on the trap
design) was also placed in the bucket. The water was changed daily. Traps were made
of 2cm thick plain galvanized wire, white mosquito netting and either black or blue
colored cotton cloth. Daily, fifty gravid An. gambiae s.s. females were collected from the
ICIPE Mbita strain colony and released into the screened cage at 6 pm. The following
day from 6:30 to 12:00 hr, traps were checked for any females. Froml4:00 to 16:00 hr.
the remaining females were aspirated out of the screened cage. Following colony
protocol all gravid females had been allowed to blood feed on the investigator's arm for
ten minutes daily for three consecutive days prior to their release. A colony egg trap was
placed in the colony cage from which the tested gravid females were taken to note the
ovipositional success of that batch of females.
The first experimental day, no trap was placed on the bucket and eggs were noticed
the following morning. On February 21st the bucket was removed from the screened
cage, fresh lake water placed in it and then positioned in a sunlit area for 6 hours to
"cook." The bucket was checked for any mosquito eggs that may have been laid by wild
mosquitoes and then returned to the screen cage prior to the release of that day's colony
mosquitoes. This "cooking" was done for each of the remaining trials. On February 23rd
the soil in the bucket was replaced by mud from a known Anopheles larval site in Mbita
that was not subject to larval control. On March 5th, dark blue and black cotton cloth was
draped on the outside of the screened cage and water was poured on it prior to release of
mosquitoes to increase the humidity. This was again continued daily. On March 7th,
filter paper used in the colony egg traps was attached to the lower part of one of the traps
and positioned half in the water to entice the females into ovipositing. On March 8th, the
screened cage was moved into a modified greenhouse (Mathenge et al. 2002). On March
12th a miniature version of the black trap (Fig. 5.3) was placed in a colony cage over a
standard egg trap used for maintaining colonies. Only one gravid female was captured
during all semi-field trials.
Field Trials in Kenya
From February 19t to March 17th 2004 field trials of the traps designs were
conducted. Field sites were chosen in the Kamasengre district of Rusinga Island, about
12 km away by road from Mbita. The trap bucket was placed in a hole dug in the ground
approximately 4m from a single family dwelling. Soil from the hole and fresh lake
water were added to the bucket. Four days later mud from a nearby larval site was placed
in the bucket instead of soil from the hole. The design placed on or in the bucket was
checked daily at approximately 7:00 hr. A week after these preliminary trials two more
sites were added; one had a bucket placed approximately a meter from a large larval site
(about 10m long) and another placed far from any larval sites or homes. This location
was considered a control. Mud from the larval site was placed in both of these new
buckets as well as lake water. The lake water was changed daily. Trap designs were
rotated daily among the three field sites and the semi-field site. On March 17th the field
sites were closed.
Out of 31 traps tested from June 8th to October 1st, 14 traps captured mosquitoes.
Out of those 14 traps, an average of 2.36 1.74 mosquitoes was caught. The mean
recapture ( SE) rate of the mosquitoes released for the 14 traps was 3.3 2.3%.
In the semi-field experiment only one gravid female was captured. This was by the
emergence design with black cloth. Only one mosquito (a Culicine) was caught in trials
in the field. The emergence design with blue cloth placed near the larval habitat caught
The description of larval habitats for An. gambiae s.l. has been well studied
(Warrell and Gilles, 2002). Larvae are typically found in sunlit temporary pools often
associated with humans. They develop in roadside ditches, wheel ruts, hoof prints from
wild and domestic animals, concrete holes, and village pots. Larvae can even to survive
close to five days on damp soil (Koenraadt et al. 2003).
Entry hole diameter
3 cm with a cinch
Bucket tie to close the hole
35 cm Mud from larval habitat
Lake water Sleeve Is 6.5 cm
Lake watery for removal diameter
changed daily of caught
Bucket diameter: 30 cm mosquitoes A)
13cm 9 cm 52.5 cm 13 cm
62.5 cm- ..-Diameter. 19.5 cm
Entry /7 cn
44 cm cloth
Figure 5-3. Trap designs in Kenya. A) Mbita (two variations) B) Emergence C) Florida
D) Cone (two variations).
Most larval sites contain fresh water without much debris. Preliminary trials conducted
by other investigators in Mbita Point, Kenya showed that An. gambiae will readily
oviposit in plastic tubs and buckets used for washing clothes and commonly found
around the town and nearby communities (Pers.comm. U. Filinger, University of
Durham, England). Although the identification of larval sites can be accomplished
without much difficulty, the identification of oviposition attractants has yet to be
One of the obstacles faced by testing a practical gravid trap out in the field is that
because An. gambiae will lay eggs in such a wide variety of habitats that are commonly
found in great numbers around the area, the trap would need to be at least as attractive to
females as the natural habitats. By placing a trap over a bucket or other potential larval
habitat the site no longer becomes sunlit and therefore probably less attractive. Further
investigations into oviposition chemical and visual attractants would be needed to
overcome the deterrent of shade created by the trap.
A problem faced by semi-field experiments is that gravid colony females that have
emerged from the same batch of colony pupae may not oviposit at the same time if at all.
If fifty gravid females of the same age and exposed to the same amount of blood are
released into a semi-field cage only a small number of females if any may actually try to
oviposit. More time spent rearing test mosquitoes should be taken to insure that
conditions are created to enhance the number gravid females that attempt to oviposit each
Although a practical method for sampling gravid Anopheles has yet to be
developed, such a tool would be of great use to malaria control programs. A practical
gravid trap could be a safer alternative to other sampling methods involving chemicals or
sampling host-seeking mosquitoes. Identifying oviposition attractants for Anopheles
mosquitoes could not only aid the development and use of gravid traps but provide
insight to the ecology of the mosquito and create more options for mosquito monitoring
LIST OF REFERENCES
Bidlingmayer, W.L. 1994. How mosquitoes see traps: role of visual responses. J. Am.
Mosq. Control Assoc. 10(2):272-279.
Bogh, C., Pedersen, E.M., Mukoko, D.A., and J.H. Ouma. 1998. Permethrin-
impregnated bednet effects on resting and feeding behaviour of lymphatic filariasis
vector mosquitoes in Kenya. Med Vet Entomol. 12: 52-59.
Centers Disease Control and Prevention (CDC). April 23, 2004a. CDC's originsand
malaria. www.cdc.gov/malaria/history/eradicationus.htm Accessed September
Centers for Disease Control and Prevention (CDC). April 23, 2004b. Eradication of
malaria in the United States (1947-1951).
www.cdc.gov/malaria/history/eradicationus.htm Accessed September 2004.
Conteh, L., Sharp, B., Streat, E., Barreto, A., and S. Konar. 2004. The cost and cost-
effectiveness of malaria vector control by residual insecticide house-spraying in
southern Mozambique: a rural and urban analysis. Trop. Med. Int. Health.
Crans, W. J. 1989. Resting boxes as mosquito surveillance tools. from Proceedings of
the Eighty-Second Annual Meeting of the New Jersey Mosquito Control
Association, Inc., pp.53-57. http://www.rci.rutgers.edu/-insects/restbox.htm
Accessed September 2004.
Das, P.K, Sivagnaname, N., and D.D. Amalraj. 1997. A comparative study of a new
insecticide-impregnated fabric trap for monitoring adult mosquito populations
resting indoors. Bull. Ent. Res. 87: 397-403.
Deressa, W., Ali A., and F. Enqusellassie. 2003. Self-treatment of malaria in rural
communities, Butajira, southern Ethiopia. Bulletin WHO 81(4): 261-268.
Durden L. and G. Mullen. 2002. "Introduction." Medical and Veterinary Entomology.
Mullen, G. and L. Durden Eds. Academic Press. San Diego. pp. 1-13.
Edman, J., Kittayapong, P., Linthicum, K., and T. Scott. 1997. Attractant resting
boxes for rapid collection and surveillance ofAedes aegypti (L.) inside houses. J.
Am. Mosq. Control Assoc. 13(1):24-27.
Foster, W. and E. Walker. 2002. "Mosquitoes (Culicidae)." Medical and Veterinary
Entomology. Mullen, G. and L. Durden Eds. Academic Press. San Diego. pp. 203-
Fullerton, H., and E. Bishop. 1933. Improved rural housing as a factor in malaria
control. South. Med. J. 26(5):465-8.
Gallup, J., and J. Sachs. 2001. The economic burden of malaria. Am. J. Trop. Med.
Hyg. 64(1, 2) S: 85-96.
Gamage-Mendis, A., Carter, R., Mendis, C., De Zoysa, A., Herath, P., and K.
Mendis. 1991. Clustering of malaria infections within an endemic population: risk of
malaria associated with the type of housing construction. Am. J. Trop. Med. Hyg.
Garfield, R. 1999. Malaria control in Nicaragua: social and political influences on
disease transmission and control activities. Lancet 354: 414-18.
Geissler, P., Nokes, K., Odhiambo, R., Aagard-Hansen, J, and J. Ouma. 2000.
Children and medicines: self-treatment of common illnesses among Luo
schoolchildren in western Kenya. Soc. Sci. Med. 50: 1771-1783.
Gilles, H. 2002. "Historical outline." EssentialMalariology 4th ed. Warrell, D.and
H.Gilles eds. Oxford University Press Inc. New York. pp. 1-7.
Githeko, A., Service, M., Mbogo, C., Atieli, F., and F. Juma. 1994. Origin of blood
meals in indoor and outdoor resting malaria vectors in western Kenya. Acta Trop.
Harwood, R. and M. James. 1979. Entomology in Human and Animal Health. 7th ed.
Macmillan Publishing Co. New York.
Honigsbaum, M. 2001. The Fever Trail. Farrar, Straus, and Giroux. New York.
Kager, P. 2002. Malaria control: constraints and opportunities. Trop. Med. Int. Health
Killeen, G., Fillinger, U., Kiche, I., Gouagna, L., and B. Knols. 2002. Eradication of
Anopheles gambiae from Brazil: lessons for malaria control in Africa? Lancet.
Infect. Dis. 2(6): 18-27.
Kenya Medical Research Institute. 2001. Approval from the 87th meeting of the Kenya
Medical Research Institute and National Ethical Review Committee for screen-
house mosquito rearing and semi-field experimental environments for risk-free
malaria transmission studies. Nairobi, Kenya.
Koenraadt, C., Paaijmans, K., Githeko, A., Knols, B., and W. Takken. 2003. Egg
hatching, larval movement and larval survival of the malaria vector Anopheles
gambiae in desiccating habitats. Malaria J. 2:20.
Konradsen, F., Amerasinghe, P., Van Der Hoek, W., Amerasinghe, F., Perera, D.,
and M. Piyaratne. 2003. Strong association between house characteristics and
malaria vectors in Sri Lanka. Am. J. Trop. Med. Hyg. 68(2):177-181.
Knols, B., Njiru, B., Mathenge, E., Mukabana, W., Beier, J., and G. Killen. 2002.
MalariaSphere: A greenhouse-enclosed simulation of a natural Anopheles gambiae
(Diptera: Culicidae) ecosystem in western Kenya. Malaria J. 1:19.
Komar, N., Pollack, R., and A. Spielman. 1995. A nestable fiber pot for sampling
resting mosquitoes. J. Am. Mosq. Control Assoc. 11(4):463-467.
Kroeger, A., Ordonez-Gonolez, J., and A. Avina. 2002. Malaria control reinvented:
Health sector reform and strategy development in Colombia. Trop. Med. Int. Health
Lindblade, K., Walker, E., and M. Wilson 2000. Early warning of malaria epidemics in
African highlands using Anopheles (Diptera: Culicidae) indoor resting density. J.
Med. Entomol. 37(5):664-674.
Lindsay, S., Jawara, M., Paine, K., Pinder, M., Walraven, G. and P. Emerson 2003.
Changes in house design reduce exposure to malaria mosquitoes. Trop. Med. and
Int. Health 8(6) :512-517.
Lindsay, S., Armstrong, J., Amstrong Schellenberg, J., Zeiler, H., Daly, R., Salum,
F., and H. Wilkins. 1995. Exposure of Gambian children to Anopheles gambiae malaria
vectors in an irrigated rice production area. Med. Vet. Entomol. 9:50-58.
Magnussen, P., Ndawi, B., Sheshe, A., Byskov, J., and K. Mbwana. 2001. Malaria
diagnosis and treatment administered by teachers in primary schools in Tanzania.
Trop. Med. Int. Health 6(4):273-279.
Marquadt, W., Demaree, R., and R. Grieve. 2000. Parasitology and Vector Biology.
Academic Press. San Diego.
Mathenge, E., Killeen, G., Oulo, D., Irunga, L., Nfegwa, P., and B. Knols. 2002.
Development of an exposure-free bednet trap for sampling afrotropical malaria
vectors. Med. Vet. Entomol. 16: 67-74.
Mboera, L., Takken, W., Mdira, K. and J. Pickett. 2000. Sampling gravid Culex
quinquefasciatus (Diptera: Culicidae) in Tanzania with traps baited with synthetic
oviposition pheromone and grass infusions. J. Med. Entomol.37(1):172-176.
Muirhead-Thomson, E. 1982.Behaviour patterns of blood-sucking flies. Pergamon
Okech, B., Gouagna, L., Killen, G., Knols, B., Kabiru, E., Beier, J., Yan, G., and J.
Githure. 2003. Influence of sugar availability and indoor microclimate on survival of
Anopheles gambiae (Diptera: Culicidae) under semifield conditions in western
Kenya. J. Med. Entomol. 40 (5): 657-663.
Pal, R., Nair, C., Ramalingham, S., Patil, P., and B. Ram. 1960. On the bionomics of
vectors of human filariasis in Ernakulam (Kerala), India. Ind. J Malariol. 14: 595-
Ritchie, S., Long, S., Smith, G., Pike, A., and T. Knox. 2004. Entomological
investigations in a focus of dengue transmission in Cairns, Queensland, Australia,
by using sticky ovitraps. J. Med. Entomol. 41(1):1-4.
SAS Institute. 2001. SAS system for Windows, release 8.2. SAS Institute. Cary, NC.
Schofield, C. and G. White. 1983. Engineering against insect-borne diseases in
thedomestic environment: House design and domestic vectors of disease. Am. J.
Trop. Med. Hyg. 78: 285-292.
Service, M. 1993. Mosquito Ecology. Field Sampling Methods. 2nd Ed. ElsevierScience
Publishers, Essex, UK.
Service, M. 2002 "Characteristics of some major Anopheles vectors of human malaria."
EssentialMalariology 4th ed. Warrell, D.and H.Gilles eds. Oxford University Press
Inc. New York. pp. 326-332
Service, M and H. Townson. 2002. "The Anopheles vector." Essential Malariology
4thed. Warrell, D.and H.Gilles eds. Oxford University Press Inc. New York. pp. 59-
Sexton, J., Ruebush, T., Brandling-Bennett, A., Breman, J., Roberts, J., Odera,
J., and J. Were. 1990. Pyrethrin-impregnated curtains and bed-nets prevent malaria in
western Kenya. Am. J. Trop. Med. Hyg. 43:11-18.
Shulman C. and E. Dorman. 2002. "Clinical features of malaria in pregnancy."
Essential Malariology 4th ed. Warrell, D.and H.Gilles eds. Oxford University Press
Inc. New York. pp. 218-235
Smith, G.E. 1942. The keg shelter as a diurnal resting place of Anopheles
quadrimaculatus. Am. J. Trop. Med. 58: 307-316.
Tan, D., Upshur, R., and N. Ford. 2003. Global plagues and the Global Fund:
Challenges in the fight against HIV, TB, and malaria. BMC Int.Health Human
Warrell, D. 2002. "Clinical features of malaria." Essential Malariology 4th ed. Warrell,
D.and H.Gilles eds. Oxford University Press Inc. New York. pp. 191-218.
Wattal, B. and N. Kalra. 1960. Studies on Culicine mosquitoes: Preferential indoor
resting habits of Culexfatigans Wiedmann, 1828, near Ghaziabad, Utter Pradesh.
India. Ind. J Malariol. 14: 605-616.
World Health Organization (WHO). 1992. Entomological Field Techniques for
Malaria Control: Part 1 Learner's Guide. World Health Organization, Geneva.
World Health Organization (WHO) 1995. Technical Report Series 857. Vector control
for malaria and other mosquito borne diseases. World Health Organization,
World Health Organization (WHO). 2000. Technical Report Series 892. WHO Expert
Committee on Malaria. World Health Organization, Geneva.
World Health Organization (WHO). 2003. Africa malaria report 2003. World Health
World Health Organization (WHO). 2004. Fact sheet 94, Roll Back Malaria InfoSheet.
World Health Organization, Geneva.
http://www.who.int/mediacentre/factsheets/fs094/en/print.html fact sheet. Accessed
Yasuno, M., Rajagopalan, P.K., Kazmi, S.J., and G.C. La Brecque. 1977. Seasonal
change in larval habitats and population density of Culexfatigans in Delhi villages.
Ind. J. Med. Res. 65: 52-64.
Justin Eric Harbison was born in St. Paul, Minnesota, on September 1, 1977, to
Kent and Judy Harbison. He graduated from St. Paul Academy in St. Paul, Minnesota, in
1996. Justin graduated from Macalester College in May 2000 with a Bachelor of Arts
degree. Prior to pursuing his master's degree at the University of Florida, Justin spent
nine months as a field technician in the Organization for Tropical Studies' Palo Verde
field station in the Guanacaste region of Costa Rica.