Development of a semi-field system for contained field trials with Aedes aegypti in southern Mexico.
ABSTRACT Development of new genetic approaches to either interfere with the ability of mosquitoes to transmit dengue virus or to reduce vector population density requires progressive evaluation from the laboratory to contained field trials, before open field release. Trials in contained outdoor facilities are an important part of this process because they can be used to evaluate the effectiveness and reliability of modified strains in settings that include natural environmental variations without releasing mosquitoes into the open field. We describe a simple and cost-effective semi-field system designed to study Aedes aegypti carrying a dominant lethal gene (fsRIDL) in semi-field conditions. We provide a protocol for establishing, maintaining, and monitoring stable Ae. aegypti population densities inside field cages.
- SourceAvailable from: Carlo Costantini[show abstract] [hide abstract]
ABSTRACT: To obtain information on adult populations of Afrotropical malaria vector mosquitoes, mark-release-recapture experiments were performed with Anopheles females collected from indoor resting-sites in a savanna area near Ouagadougou, Burkina Faso, during September 1991 and 1992. Results were used to estimate the absolute population densities, daily survival rates, and dispersal parameters of malaria vectors in that area. In 1991 a total of 7260 female Anopheles were marked and released, of which 106 were recaptured in the release village and 6 in the neighbouring villages, a total recapture rate of 1.5%. The following year 13,854 female Anopheles were released and 116 recaptured in Goundri and 8 in the neighbouring villages, a total recapture rate of 0.9%. Recaptures were found in three of eight villages near Goundri. Nearly all of the recaptured mosquitoes were An gambiae s.l. Of these, molecular determination revealed that An.gambiae s.s. and An.arabiensis were present in a ratio of approximately 2:3. Two simple random models of dispersal were simulated and the parameters of the models determined by searching for the least-squared fit between simulated and observed distributions. The mean distance moved by individual mosquitoes, estimated in this way, ranged 350-650 m day-1, depending on the model and the year considered. Population densities were estimated using the Lincoln Index, Fisher-Ford and Jolly's methods. The estimates of population size had high standard errors and were not particularly consistent A "consensus' value of 150,000-350,000 mosquitoes is believed to apply for the An.gambiae s.l. female population. Survival was estimated to be 80-88% per day.Medical and Veterinary Entomology 08/1996; 10(3):203-19. · 2.21 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Although variation in mortality is considered by virtually all vector-borne disease specialists to be one of the most important determinants of an arthropod's capacity to transmit pathogens, the operational assumption often is that insect vector mortality is independent of age. Acceptance of the non-senescence assumption leads to the erroneous conclusion that mosquito age is unimportant, results in misleading predictions regarding disease reductions after vector control, and represses study of other aspects of mosquito biology that change with age. We brought large-scale laboratory life table techniques (N > 100,000) to bear on the question of age-dependent mortality in the mosquito vector of dengue virus, Aedes aegypti. Mortality was highly age dependent in both sexes. Mortality was low at young ages (< 10 days old), steadily increased at middle ages, and decelerated at older ages. A newly derived age-dependent model of pathogen transmission shows the importance of young mosquitoes and population age structure to transmission dynamics. Departure from the age-independent mortality paradigm encourages research on overlooked complexities in mosquito biology, the need for innovative methods to study mosquito population dynamics, and the need to study age-dependent changes for an accurate understanding of mosquito biology and pathogen transmission.The American journal of tropical medicine and hygiene 02/2007; 76(1):111-7. · 2.53 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Aspiration collections of adult Aedes aegypti (L.) were made weekly from inside and outside of houses for 3 yr in a rural Thai village (n = 9,637 females and n = 11,988 males) and for 2 yr in a residential section of San Juan, Puerto Rico (n = 5,941 females and n = 6,739 males). In Thailand, temperature and rainfall fell into distinct seasonal categories, but only temperature was correlated with fluctuations in female abundance. Average weekly temperature 6 wk before mosquitoes were collected and minimum weekly temperature during the week of collection provided the highest correlations with female abundance. Accounting for annual variation significantly improved Thai models of temperature and mosquito abundance. In Puerto Rico, temperature, but not rainfall, could be categorized into seasonal patterns. Neither was correlated with changes in female abundance. At both sites the vast majority of females were collected inside houses and most contained a blood meal. Most teneral females were collected outside. Wing length—an indicator of female size—and parity, egg development or engorgement status were not correlated, indicating that feeding success and survival were not influenced by female size. At both sites, females fed almost exclusively on human hosts (≥96%), a pattern that did not change seasonally. In Puerto Rico more nonhuman blood meals were detected in mosquitoes collected outside than inside houses; no such difference was detected in Thailand. Gut contents of dissected females indicated that females in the Thai population had a younger age distribution and fed more frequently on blood than did Ae. aegypti in Puerto Rico. Our results indicated that aspects of this species' biology can vary significantly from one location to another and 1 yr to the next.Journal of Medical Entomology 12/1999; 37(1):77-88. · 1.86 Impact Factor
Am. J. Trop. Med. Hyg., 85(2), 2011, pp. 248–256
Copyright © 2011 by The American Society of Tropical Medicine and Hygiene
During the past decade, containment of genetically modified
(GM) arthropods has become an important issue among peo-
ple working in the field of vector-borne diseases. Development
of genetically manipulated strains of arthropods unable to
transmit pathogens 1 or carrying sterility genes 2 includes assess-
ing the effects of genetic modifications on vector biology in
semi-field conditions. 3, 4 Proper trials in semi-field systems,
as defined by Ferguson and others, 5 are essential to investi-
gate possible changes in fitness and/or behavior of genetically
modified vectors that may affect their compatibility with the
natural environment and their competitiveness with wild-
type conspecifics. 6 If not detected, fitness costs and behavioral
changes could cause genetic control strategies to fail. 7
To test transgenic insect technologies and effector genes in
semi-field conditions, it is important to meet certain require-
ments, 5, 8 such as, 1) developing semi-field structures of suitable
size for target organisms that assure containment and prevent
accidental release; 2) locating structures in an ecologically iso-
lated area where the target vector and pathogen are already
present; 3) reproducing as closely as possible all essential
ecological conditions for the vector at the field site (i.e., tem-
perature, humidity, solar radiation, wind, availability of larval
development sites, mating sites, adult refugees, and food for
immatures and blood hosts for adult females); 4) having the
capacity to establish and maintain a stable, local vector pop-
ulation through several overlapping generations (i.e., caged
populations should simulate free-ranging target populations
as much as possible during GM arthropod evaluations); and 5)
fine tuning methodology for measuring changes in population
density (i.e., detect increases or decreases in the number, gen-
otype and/or phenotype through time). Having facilities that
are well characterized and validated procedures for establish-
ing and maintaining a local population in a semi-field system
for overlapping generations will allow researchers to test and
modify, as necessary, the design and use of those facilities for
evaluation of GM mosquitoes.
Herein, we report the design of a semi-field system and a
protocol for its use. Our aim was to establish and maintain
a stable wild-type Aedes aegypti population in an enclosed
outdoor environment near Tapachula, Mexico, as a first step
in conducting contained field trials with Ae. aegypti . In sub-
sequent experiments we will evaluate Ae. aegypti carrying a
dominant lethal gene (fsRIDL) 2 (OX3604C) for suppression
of local, wild-type mosquito populations in field cages similar
to those we used in this study.
MATERIALS AND METHODS
Study area. Our study was carried out on a plot of land
(14°51′41²N, −92°21′15²W ) referred to hereafter as the “field
site.” The land was located 11.2 km south-east from the center
of Tapachula in the village of El Zapote and consisted of a
4.5 ha flat, rural area with 1.5 ha cultivated in cashew trees. The
rest of the plot is grassland used for grazing cattle. Surrounding
land is mainly cultivated with mango trees, soya beans, corn,
and banana trees. Climate in the area is characterized by a
rainy season during the summer months (May to October)
with an average of 2,100-mm rainfall and a dry season during
the winter months (November to April) with an average of
Cage design and maintenance. Our semi-field system con-
sisted of a 2.5 × 5.5 × 2 m (w × l × h) tent cage ( Figure 1A and D )
made of white tricot mesh, reinforced at angles and seams
with white fabric ( Figure 1B ). It had four single zipper doors,
one per side, which allows access to work in the internal space
and entry into the cage through the side (where we observed
the smallest number of mosquitoes resting on the tent walls).
Entry through the side door minimized the possibility for
mosquitoes to escape. The tent cage was supported by a steel
frame designed for an outdoor canopy (model 1020-8, Galaxy
Orion, China, Figure 1C ), measured 3.0 × 6.0 × 3.0m (w × l × h)
that had a white plastic roof, which partially protected the
tent cage from rain and direct sunlight. Twelve windows in the
canopy, six on each side of the roof, facilitated air circulation
between the tent and the roof ( Figure 1D ) and decreased wind
pressure on the roof itself. The canopy and tent cage were
fastened on a wooden platform (8.0 × 5.0 × 0.8 m, Figure 1C ).
The platform was fastened to a metal frame that elevated the
Development of a Semi-Field System for Contained Field Trials with
Aedes aegypti in Southern Mexico
Luca Facchinelli ,* Laura Valerio , J. Guillermo Bond , Megan R. Wise de Valdez , Laura C. Harrington ,
Janine M. Ramsey , M. Casas-Martinez , and Thomas W. Scott
Department of Entomology, University of California, Davis; Istituto Pasteur – Fondazione Cenci Bolognetti, University of Rome “Sapienza,” Rome,
Italy; Centro Regional de Investigación en Salud Pública (CRISP), Instituto Nacional de Salud Pública (INSP), Tapachula, Chiapas, Mexico;
Department of Microbiology, Immunology, and Pathology, Colorado State University, Fort Collins, Colorado; Department of Entomology, Cornell
University, Ithaca, New York; Fogarty International Center, National Institutes of Health, Bethesda, Maryland
Abstract. Development of new genetic approaches to either interfere with the ability of mosquitoes to transmit den-
gue virus or to reduce vector population density requires progressive evaluation from the laboratory to contained field
trials, before open field release. Trials in contained outdoor facilities are an important part of this process because they
can be used to evaluate the effectiveness and reliability of modified strains in settings that include natural environmen-
tal variations without releasing mosquitoes into the open field. We describe a simple and cost-effective semi-field system
designed to study Aedes aegypti carrying a dominant lethal gene (fsRIDL) in semi-field conditions. We provide a protocol
for establishing, maintaining, and monitoring stable Ae. aegypti population densities inside field cages.
*Address correspondence to Luca Facchinelli, Department of
Entomology, University of California, 1 Shields Avenue, Davis, CA
95616. E-mail: firstname.lastname@example.org
A SEMI-FIELD SYSTEM FOR CONTAINED FIELD TRIALS WITH AEDES AEGYPTI
entire structure 80 cm above the ground and avoided flood-
ing during the rainy season. A white plastic sheet covered the
floor of the cage to ensure a tidy, easy to clean environment.
The whole structure was isolated from the ground by 15 ant
traps (i.e., metal buckets filled with water and soap; Figure 1E )
under the legs of the platform, which was intended to minimize
ground access to the cage by arthropods and small verte-
brates. A plastic sheet laid on the ground below the platform
( Figure 1E ) prevented grass and other plants from growing
under the cage and eventually touching the platform, which
could be a means for arthropods and small animals to climb
onto the platform and eventually into the cage. The following
items were placed inside the cage to provide refugee for
mosquitoes ( Figure 1D and F ): a potted plant ( Dieffenbachia
spp.), an open black closet made of wooden boxes (100 × 30 ×
140 cm), two clay pots 10 L each containing 3 cm of soil, two
stacks of clay bricks (50 × 50 × 50 cm) with an open inner space
where mosquitoes could rest ( Figure 1D and F ), one dispenser
holding water-soaked cotton, and two dispensers holding 10%
sugar solution soaked cotton that were changed every second
day, and one data-logger (Hobo Pro v2 temp/RH, Onset
Computer Corporation, Bourne, MA) hanging in the middle
of the cage so that it was protected from direct sunlight and
could record temperature and humidity every 30 min. Clay
resting sites were moistened every second day and the plant
was watered as needed.
We built two identical tent cages, labeled cage A and cage
B. Cages were partially protected from sunlight by shade cre-
ated by nearby cashew trees. Because the semi-field systems
were used to contain locally derived Ae. aegypti , we did not
include vestibules, a double layer of mesh for cage walls or
equipment like air curtains as a means to increase contain-
ment. 4 Cage tests were performed over an 11-month period
from November 2008 to September 2009 ( Table 1 ).
Aedes aegypti strain. In our trials we used mosquitoes from
an Ae. aegypti colony established at the Centro Regional de
Investigación en Salud Pública (CRISP) insectary in Tapachula
during 2007. Aedes aegypti were colonized from eggs collected
in El Zapote using 20 ovitraps. Subsequently, eggs were
collected weekly in El Zapote and larvae were reared at the
CRISP insectary. All adults that developed from those eggs
(an average of 60 per week), were added to the colony to limit
inbreeding effects. Larvae reared in the insectary and during
experiments were fed a diet of laboratory rodent food (ground
laboratory rodent food by LabDiet, Richmond, IN) and adult
females imbibed blood from a rabbit (UC Davis Animal Care
Figure 1. Cage design: ( A ) field cage; ( B ) tricot mesh reinforced at critical points with white fabric; ( C ) steel frame of an outdoor canopy
located on a metal platform covered with wooden boards; ( D ) 12 windows, 6 on each side of the roof facilitate air flow are denoted by arrows;
( E ) metal containers filled with water and soap under legs of the platform to partially isolate the structure from the ground and blue plastic sheet
under the platform to prevent plant growth; and ( F ) refugia to create places where adults can rest inside the tent cage.
FACCHINELLI AND OTHERS
and Use protocol 15653). Adult males and females had access
to water and a 10% sugar solution.
Temperature and relative humidity. Temperature and rel-
ative humidity (RH) are important environmental param-
eters that affect adult mosquito biology. 9 In semi-field
systems, particular attention must be paid to fluctuations in
temperature and RH values so that they are similar to those
recorded in the open field surrounding the cages. Dataloggers
were used to compare temperature and RH in cage A (the
first cage built) to the open field outside the cage from June
23, 2008 to July 21, 2008 (rainy season). The datalogger inside
the cage was hung from the roof in the middle of the tent. The
other external datalogger was located less than 2 m from the
cage in a shaded place, protected from rain, direct sunlight,
and at the same height as the one located inside the cage.
During a 28-day period dataloggers recorded temperatures
and RH at the same time every 5 minutes for a total of
8,064 measurements. During the 76 days from 3 March to
7 May 2009 (dry to rainy season) temperature and RH were
simultaneously recorded (every 30 min) in cages A and B to
confirm that environmental conditions in the two cages were
Field cage studies. Field experiments ( Table 1 ) are divided
into 1) pre-trial tests of key aspects of Ae. aegypti biology and
2) a definitive trial, when we attempted to stabilize the density
of an Ae. aegypti population in the cages.
Pre-trial tests. Pre-trial tests were carried out to obtain data
on Ae. aegypti population dynamics, adult production rate, and
immature development time in the semi-field structure.
Population dynamics. Two pre-trial tests were carried out
between November 2008 and February 2009 to investigate
whether the cages were suitable to host Ae. aegypti populations
and to obtain data on population dynamics in the cages.
During each pre-trial test, 300 2-day-old males and females
(150 of each sex), were released into each cage. A restrained
rabbit was provided daily for half an hour as a source of blood
(UC Davis Animal Care and Use protocol 15653). Six water
holding containers that were different in shape, size, color, and
material were lined with paper strips, filled with purified water
to have a total liquid high of about 10 cm, placed into each
cage, and checked daily. Each day paper strips with eggs were
removed and containers were refreshed with purified water
and new paper strips.
The number of mosquitoes resting on the internal walls of
each cage, females feeding on the rabbit and eggs laid on the
paper strips were counted daily. The data were used as indirect
measures of Ae. aegypti population dynamics. The two pre-trial
tests ended after different time intervals. The first trial started
during November 2008 and ended on Day 34. The second trial
started during February 2009 and ended on Day 27.
Female mortality dynamics. At the end of each trial, surviv-
ing mosquitoes were collected using backpack aspirators
(model 1412, John W. Hock Company, Gainesville, FL ), killed
by freez ing, counted, and sexed. To confirm our data on
female mortality through time, three additional tests were
per formed as described previously during the rainy season
(June–September 2009). Adults were introduced into cages
as larvae and completed their larval development in the field
cages. Six hundred first instar larvae were seeded in three
mesh-covered buckets containing 2 L of water (200 larvae
per container) in each field cage. Adults emerging from the
containers were released into the cages until 300 males and
females (150 of each sex) per cage were reached. Adults were
maintained as described previously until they were collected.
Adult mosquitoes were collected on Days 2, 10, and 15. To
estimate the female mortality trajectory, nonlinear regression
analysis was applied to the number of surviving females
collected at each time interval.
Adult production rate. To maintain a stable population inside
the cages, it is necessary to introduce mosquitoes at certain
time intervals to compensate for adult mortality. Among
several different release approaches (i.e., release of adults,
immature stages, or eggs) we based this study on eggs. By
seeding mesh-covered containers inside cages with a certain
number of recently oviposited eggs, we could obtain adults
that experienced the same environmental conditions as larvae
and control the exact number and sex of adults produced in
the containers; freshly emerged adults were counted and sexed
before introducing them into each cage.
To apply this introduction method during the definitive
trial, it was necessary to determine the number of adults
produced after seeding a known number of eggs into a con-
tainer and the minimum and maximum larval development
times. For this purpose, eight 7-L plastic buckets were filled
with 2 L of purified water, seeded with a total of 300 embryo-
nated eggs on paper strips, and containers were covered with
a mesh to prevent deposition of additional eggs by wild mos-
quitoes. The buckets were placed on a table 2 m from the
field cages in a shaded area. Larval food was added ad libi-
tum . Paper strips were removed from the buckets on Day 5
to avoid mold development. Larvae were filtered and buckets
were refreshed every fifth day with clean water and food to
prevent excessive fungal and bacterial growth in larval rearing
water. Development times from egg to adult emergence were
recorded. Adults emerging from containers were aspirated out
of containers, counted daily, and sexed. The trial ended when
no larvae were detected in any bucket.
Definitive trial. On the basis of results from pre-trial tests,
the following protocol was developed and tested from 3 March
to 17 May 2009 as a means of maintaining stable Ae. aegypti
population densities inside field cages.
(1) A cohort of 300 male and female (150 of each sex) 2-day-
old Ae. aegypti was introduced into each tent cage.
Mosquitoes had access to water and 10% sucrose solution
(both renewed every 2 days). Once a week, a restrained
Schematic timeline field cage trials
Nov 08 Dec 08Jan 09Feb 09 Mar 09 Apr 09May 09Jun 09Jul 09 Aug 09 Sep 09
Female mortality dynamics
Adult production rate
A SEMI-FIELD SYSTEM FOR CONTAINED FIELD TRIALS WITH AEDES AEGYPTI
rabbit was introduced into each cage to provide a blood
meal for females.
(2) Two days after blood feeding, three oviposition containers
(2, 6, and 11 L, which corresponded to those containers
that collected the highest number of eggs during pre-trial
tests) were lined with filter paper and placed into each
cage. Containers were removed after 48 hrs. Oviposited
eggs were brought back to the CRISP laboratory, counted,
dried for 24 hrs, and kept in a labeled zip-lock bag.
(3) The number of mosquitoes resting on the internal walls of
the cages was recorded daily between 8:00 AM and 10:00
AM. The number of females taking a full blood meal was
visually estimated while the rabbit was in the cage. The
number of eggs produced weekly was recorded.
(4) Each week 300 eggs in a mesh-covered, 7-L plastic con-
tainer filled with 2 L of water and food ad libitum were
placed into the cage. All 300 eggs that had been produced
the previous week by mosquitoes in one cage went back
into the same cage the following week. To minimize using
eggs laid by one or a few females, we cut small pieces of
paper containing a maximum of 30 eggs from different
paper strips until 300 eggs were obtained from Week 2 on.
During Week 1, eggs from the insectary colony were used.
(5) Adults emerging from the egg container were counted,
sexed, and released into their respective cage. On the basis
of pre-trial tests, we assumed that the development time
from egg to adult took a maximum of 2 weeks and, there-
fore, on a weekly basis there were two containers in each
cage that asynchronously produced adults. Containers
were removed from cages after 14 days. In a few cases lar-
vae (always < 5) were still developing and were discarded.
(6) After 11 weeks, the trial ended and all mosquitoes in each
cage were collected using back pack aspirators, brought
back to the insectary, killed by freezing, counted, and sexed.
One of the key objectives of our study was to develop meth-
odology for detecting population density changes in the field
cages. Although measures of population density could be
achieved using Back Pack Aspirators or BG-Sentinel Traps,
which were specifically developed to collect Ae. aegypti adults, 10
we tried to use approaches that would not damage mosquitoes
or adversely affect their behavior or survival. Our aim was to
develop a methodology for maintaining stable populations
(R o ~ 1), without perturbing mosquito activity or survival, so
that in subsequent trials we will be able to confidently detect
fluctuations in population density through time that is attrib-
utable to the introduced genetically modified mosquitoes.
Statistical analysis. SPSS 15.0 (SPSS Inc., Chicago, IL) was
used for statistical analysis.
Temperature and Relative Humidity (RH). Mann-Whitney
U tests were performed to compare temperature and RH
between indoor and outdoor environments and between
cages A and B. Spearman rank correlation tests were used to
compare the number of adults resting on the internal walls
of cages to temperature and RH, which were recorded when
mosquitoes were counted.
Pre-trial tests. Population dynamics data from the first 27 days
were not normally distributed and could not be normalized
with log transformation. The nonparametric Mann-Whitney
U test was used, therefore, to determine differences between
replicates in distributions of adults, blood-fed females, and
eggs. We tested for a difference in egg production between
cage A and B during November’s pre-trial test using the χ 2 test.
Kruskal-Wallis tests were performed to investigate differences
in distributions of adults, fed females, and eggs between the
two pre-trial tests. Linear regression analysis was applied to
the mean number of adults counted daily in both trials and
to the number of females collected in cages at different time
intervals. The latter analysis was used to build best fit curves
describing adult and female mortality dynamics. A Spearman
rank correlation test was used to compare the distribution of
the daily mean number of eggs collected with the expected
number of females obtained from the curve describing female
mortality dynamics in pre trials.
Definitive trial. The Mann-Whitney U test was used to
compare the mean number of adults produced in containers in
cage A versus cage B. Linear regression analysis was applied
to the number of eggs collected weekly from cages A and B.
RESULTS AND DISCUSSION
Costs. All components used to build the semi-field system
were purchased in Tapachula and only the outdoor canopy
was produced outside of Mexico. Wooden boards for the floor,
fabric and tricot for the cage walls were easily obtained local
materials that helped minimize costs. The most expensive
component (75% of the total cost) was for the metal frame of
the wooden platform, which was caused by the relatively high
price of steel and labor cost. Each semi-field system cost the
equivalent of US $2,500.
Temperature and RH. Open field versus cage. Temperature
ranged outdoors between 21.2 and 38.1°C, median = 24.8°C,
and inside cages between 21.2 and 39.4°C, median = 24.9°C.
There were no significant differences in temperatures between
environments inside and outside of the cage ( U = 32,726,092,
P = 0.48). During the same period RH ranged between 41.1%
and 97.4% (outdoors) and 41.0% and 96.8% (indoors). The
RH was slightly higher outside (median = 92.5%) than inside
the cage (median = 91.8%) and the difference was statistically
significant ( U = 35,785,316.5, P = 0.002).
Cage A versus B. During the definitive trial we did not
detect a significant difference in temperature between cages
(range: cage A 15.7 to 41.3°C and cage B 15.6 to 41.7°C;
median temperature cage A = B = 26.5°C, U = 6,542,179.5,
P = 0.19). During the same period, RH ranged from 23.5% to
95.8% in cage A and 22.1% to 98.2% in cage B. Median RH
was slightly higher in cage B (78.0%) than cage A (75.8%).
The difference was statistically significant (Mann-Whitney
U test: U = 6,894,466.5, P < 0.001).
The statistically significant differences in RH, but not in
temperature, between inside and outside cage environments
and between cage A and cage B are likely because of the equa-
tion for calculating RH-based ambient temperature ( T ) and
dew point temperature ( D ): (RH = 100 * [( D + d )/( T + d )] a *
10 b [1/( D + d )−1/( T + d )] where “ a ,” “ b ,” and “ d ” are constants). 11 In this
model the relationship between temperature and water vapor
content in the air is exponential. This means that when baro-
metric pressure and absolute humidity are constant, for each
increase in temperature there is a relatively large decrease in
RH. The small differences in temperature recorded during our
study can, therefore, lead to statistically significant differences
in RH that are not likely to be biologically meaningful. We
concluded that temperature and RH for our cage design is a
reasonable approximation to an open, natural habitat.
FACCHINELLI AND OTHERS
The first topic that arose before starting to populate the
cages concerned the appropriate Ae. aegypti population den-
sity per cage. Ferguson and others 5 state that “There are no
general guidelines for the appropriate size of such a unit, but
ideally it should be large enough to sustain a population of
similar density to that encountered in the target environment
for numerous generations.” Unfortunately, this is not feasi-
ble in the case of Ae. aegypti unless a semi-field system is the
size of a small village, because Ae. aegypti adults are typically
present in natural environments at low densities, 12 which we
confirmed for communities in the vicinity of our study area
(Bond JG and others, unpublished data ). To have popula-
tions of sufficient size to obtain statistically valid results
(i.e., to avoid populations so small that outcomes of experi-
ments occur by chance), Ae. aegypti densities in semi-field
systems must be artificially increased. Data from previous
studies on Ae. aegypti population management in large indoor
insectaries at Colorado State University (Wise de Valdez MR,
personal communication), indicate that experimental popu-
lations fluctuating between 300 and 500 individuals per cage
of similar size to those used in this study is appropriate for
meaningful experimental purposes.
Pre-trial tests. Population dynamics. During each of the two
pre-trial tests, Mann-Whitney U tests did not detect differences
between replicates (cage A versus cage B) in the distribution
of adults (November U = 433.5, P = 0.53; February U = 281,
P = 0.21), blood-fed females (November U = 581.5, P = 0.14;
February U = 464.5, P = 0.24), or eggs (February U = 302.5,
P = 0.85). The only significant difference was in egg production
between cage A and cage B during November ( U = 566.0, P =
0.02), when females in cage A produced fewer eggs compared
with those in cage B (16,311 versus 25,422, χ 2 = 1006.1, degrees
of freedom [df] = 1, P < 0.01). This appears to be caused by
unrecognized factor(s) that influenced oviposition in cage A.
Data from cage A during November, therefore, were excluded
from further statistical analysis because egg production was
considered an outlier. Inclusion of adult and fed female
distributions from cage A did not alter overall results.
During pre-trial tests between November and February
no significant differences were observed in the distribution
of adults (Kruskal-Wallis: H = 2.55, df = 2, P = 0.28) or eggs
(Kruskal-Wallis: H = 0.85, df = 2, P = 0.65). The distribution
of blood-fed females between the two trial periods were not
comparable (Kruskal-Wallis: H = 10.28, df = 2, P < 0.01) and
didn’t provide useful information to understand Ae. aegypti
population dynamics. Blood-fed females were, therefore,
excluded from further statistical analysis. Figure 2 shows pop-
ulation dynamics based on the daily mean number of adults
counted resting on cage walls during both pre-trial tests. The
best fit curve for adult density based on this data was a loga-
rithmic function y = −37.0ln( x ) +123.8, R 2 = 0.93. According
to this model, at Day 1 ( x = 1) only 123 mosquitoes would
have survived inside the cage, which corresponds to a 41% sur-
vival rate after just 24 hours post release into the cage. After 1
week the number of adults would drop to 52 (17.3%) and then
slowly decrease until Day 29 when the population in the cage
would be extinct. This is inconsistent with our results for max-
imum female lifespan (33 days during November pre-trial)
and highlights limitations in our ability to visually count rest-
ing mosquitoes. After release and blood feeding, many adults
appeared to seek out humid and dark locations to rest. When
they were in those locations it was impossible for us to accu-
rately observe and count them. In addition, male lifespan is
expected to be shorter than that of females. 13 Consequently,
it would be best to count the two genders separately, but it
was not feasible for us to accurately distinguish between males
and females while counting them through the mesh. For these
reasons, we conclude that the number of adults resting on the
internal walls of a cage is not a dependable means of measur-
ing adult population dynamics.
Female mortality dynamics. More reliable information was
obtained from collecting adult mosquitoes. The number of
females collected inside the cage when pre-trial tests were
interrupted (130, 48, 30, 24, and 0 at Day 2, 10, 15, 27, and
34, respectively) was modeled using a logarithmic function
( y = −43.7ln( x ) + 156, R 2 = 0.97) to describe female mortality
through time ( Figure 3 ). According to this equation, on Day 1
( x = 1) 125 females would be alive inside the cage (83.3%
survival rate), 71 females would be alive on Day 7 (47.3%),
and then their number would slowly decrease until Day
35 when all mosquitoes in the cage would be dead. These
results are in agreement with what we observed in field cages.
Consequently, this function was used to calculate the number
Figure 2. Observed daily mean number (bars represent standard
deviation) of adults resting on the internal walls of cages during popu-
lation dynamics pre-trial tests (300 adults released on Day 0).
Figure 3. Female mortality dynamics described by the logarith-
mic function y = −43.7ln( x ) + 156. Diamonds represent the number of
females collected in the cages when pre-trial tests were interrupted.
A SEMI-FIELD SYSTEM FOR CONTAINED FIELD TRIALS WITH AEDES AEGYPTI
of recently emerged females we needed to introduce into each
cage (79 during the first week) to account for mortality.
Even though mortality was not constant during the first few
weeks of the definitive trial, to develop a standardized pro-
tocol we decided to introduce a constant number of females
weekly until the population stabilized. Because we assumed
a 1:1 sex ratio, we decided to introduce a similar number of
females and males into field cages each week.
After calculating the replacement rate, we sought indica-
tors of adult abundance that could be used to detect fluctu-
ations in population density. The indicator we selected was
the trend in the daily mean number of eggs produced during
pre-trial tests ( Figure 4 ). The mean number of eggs laid per
day in cages steadily decreased through time according to the
expected number of females during the previous day, which
was calculated from the mortality regression curve; the two
datasets were significantly correlated (Spearman rank corre-
lation test: r = 0.69, P < 0.01). It was not feasible to correlate
eggs produced in a short temporal window (i.e., a few days) to
the number of females because of variation in the length of
gonotrophic cycles and to differences in oviposition behavior.
Length of a gonotrophic cycle can differ among individuals 14
and females can lay eggs over a series of days in different ovi-
position sites. 15 These complications can create fluctuations in
daily measures of egg production and make it difficult to ret-
rospectively estimate over a short temporal window the num-
ber of ovipositing females and by extension the total number
of females present in a cage. Our results do indicate, however,
that eggs laid over a relatively long period of time, such as 4
weeks during pre-trial tests, is a reasonable indirect measure
of fluctuations in female population density in our semi-field
Adult production from eggs and immature development time.
The average number of males and females produced from
containers located close to the cages was 133 ± 15 (SD) and
124 ± 11 (SD), respectively. This means that in a time frame
of 14 days, about 85% of seeded eggs developed into adults.
Development time required from egg to adult ranged from 8 to
15 days; 90% of males emerged between Day 8 and 9, and 90%
of females emerged between Day 9 and 11. More than 99% of
adults emerged within 14 days. These results were useful for
calculating the number of eggs (181) that should go back into a
cage during the first week of the trial to compensate for death
predicted by the female mortality curve (79 females). Because
we expected that fluctuations of mosquito density caused by
environmental factors at a field site could be high, we decided
to increase the number of adults that we introduced into
cages to avoid abrupt population reductions. For this reason,
each week we added 300 eggs to containers in each cage (see
Material and Methods, point 4).
Trends in female density were estimated on the basis of
daily mortality of introduced females (150 at Day 0), followed
by the daily mean number of females emerging from contain-
ers. Figure 5 shows results from a simulation in which female
numbers decreas from Day 0–7 and then slowly increase until
the population density stabilizes on Day 42, after which pop-
ulation density fluctuates between a minimum of 134 and a
maximum of 190 females.
Definitive trial. Adults produced in cages. During the
definitive trial, eggs seeded in containers inside cages A and
B produced females per week (mean of 87 [95% CI = 80.80–
93.20], U = 25, P = 0.95), which were significantly lower than
those obtained during the pre-trial test (mean of 124 [95%
CI = 116.38–131.62]). The only difference between the pre-
trial test and the definitive trial was that in the first case
larvae were reared in the field, close to cages in a completely
shaded area. During the definitive trial direct sunlight hit the
containers located inside the cages for ~2 hours daily. Because
air temperature was not statistically different between the
indoor and outdoor environment, sunlight may have adversely
affected egg hatch rate and/or larval survival.
According to the simulation described previously, lower
production of adults will lead to different expected fluctua-
tions in female density compared with those showed in Figure
5 . In this case, once stabilized, expected mosquito density will
fluctuate between 94 and 125 females.
Population dynamics. Linear regression analysis indicated
that the number of eggs collected weekly in cage A decreased
over time ( y = −28.17 x + 2,264.65, r 2 = 0.66, P < 0.01) until the
end of the trial, indicating that the population in cage A did
not stabilize ( Figure 6 ). This conclusion was supported by the
low number of adults collected in cage A at the end of the trial
(48 females and 34 males). Conversely, in cage B ( Figure 6 )
the number of eggs collected decreased only slightly over time
as illustrated with the horizontal regression line ( y = −7.09 x +
1744.45, r 2 = 0.049, P = 0.51). The number of females collected
in cage B was close to the lower number expected based on
the simulation, 85 and 94, respectively. We suspect that the
Figure 4. Mean number of eggs produced daily in population
dynamics pre-trial tests.
Figure 5. Simulation of expected female population dynamics in
a field cage. A female mortality curve was applied to the number of
females that were introduced into a cage on Day 0 and then to the
daily mean number of females that emerged weekly from containers
inside the cage.
FACCHINELLI AND OTHERS
difference in population dynamics between cage A and B was
caused by damage to the plastic roof of cage A. On Day 41,
during a storm, cashew fruits fell onto and tore the plastic roof,
which could not be repaired. For the remainder of the trial, the
damaged roof of cage A allowed direct sunlight and rainfall
into the tent cage.
Another potential complication affecting adult survival and
egg production was predation. Even though we attempted to
isolate cages from mosquito predators, when disassembled at
the end of the trial, refugia in both cages contained Ae. aegypti
legs and wings, something that is typical of predation by ants.
This indicates that, although our study was carried out in con-
fined and partially isolated cages, predation was probably an
issue that to some extent influenced adult survival. We were
not able to determine if the remains of the mosquitoes we
found were caused by scavenged carcasses or insects that were
killed by a predator. Adding vestibules with sleeved access to
cages may help reduce predator entry into cages.
Visual counts of adults resting on the internal walls were
not a useful indicator of population dynamics. This conclusion
was confirmed by the observation that the number of adults
counted resting on cage walls was inversely correlated to tem-
perature recorded just before starting the count (Spearman
rank correlation test: cage A r = −0.41, P < 0.01; cage B r =
−0.24, P < 0.05). Unexpectedly, no correlation was found with
RH (Spearman rank correlation test: cage A r = 0.18, P =
0.14; cage B r = 0.04, P = 0.70), although temperatures and
RH recorded in cages were inversely correlated (Spearman
rank correlation test: cage A r = −0.85, P < 0.01; cage B r =
−0.87, P < 0.01). This suggests that the lower number of mos-
quitoes resting on cage walls was a function of temperature
(i.e., lower densities at higher temperatures) and indicates that
the tendency to seek dark and perhaps cooler places as the
resting sites was influenced more by temperature than by RH.
In future studies, earlier inspection during cooler times of day
(e.g., 7:00–8:00 am ) should be explored to increase the oppor-
tunity to observe and count adults.
According to the female mortality curve, the correspond-
ing daily mean survival rate over a 33-day period was 0.88 ±
0.07 (SD). This is comparable with Ae. aegypti survival rates
estimated by McDonald 16 and Trpis and others 17 during mark
release recapture experiments; i.e., 0.89. Our results indicate,
however, that survival for Ae. aegypti females in our semi-
field system is age dependent. The highest daily survival rate
recorded (> 0.92) was between Day 8 and 18. Other authors
have similarly discussed the issue of age-dependent mosquito
survival. In a laboratory study in which they monitored sur-
vival of > 100,000 Ae. aegypti , Styer and others 18 reported age-
dependent changes in mortality. It was low at young ages (< 10
days old), steadily increased at middle ages, and decelerated
at older ages. Similarly, Ae. aegypti survivorship during mark-
recapture studies was not constant over the recapture period
and decreased with age. 19– 23
Results from our trials are consistent with use of our semi-
field system design and procedures protocol for contained
field trials with Ae. aegypti . Our cage design is intended for use
in trials with a self-limiting system; i.e., fsRIDL. For trials that
require assessment of gene flow (i.e., transgene drive systems)
investigators will need to determine if modifications in con-
tainment features or cage structure are needed; i.e., capacity
to expand or collapse connections between cages to simulate
varying barriers to movement among populations.
Our methods for calculating female mortality and the num-
ber of new females that should be introduced into a field
cage per unit time allowed us to maintain a reasonably sta-
ble mosquito density, which will be important for assessing
population reduction and transgene spread strategies. Egg
production through time was a good proxy to detect fluctua-
tions in population densities. The ability to monitor changes in
population densities could be improved if a method is devel-
oped to sample adults directly without adversely affecting
their fitness. Temperature inside cages was equivalent to that
recorded outside. Although RH was different between indoor
and outdoor environments and between the two cages, those
differences were < 7% and, thus, we did not consider them
a biological impediment to field trials. Temperature and RH
are the most important environmental parameters to monitor
Figure 6. Regression line y = −28.17 x + 2,264.65 representing the egg production trend in cage A (left), and regression line y = −7.09 x + 1,744.45
representing the egg production trend in cage B (right) during definitive trial, with 95% confidence intervals.
A SEMI-FIELD SYSTEM FOR CONTAINED FIELD TRIALS WITH AEDES AEGYPTI
and to attempt to keep under control so that cages are suit-
able for establishing and maintaining Ae. aegypti populations
through time. In future studies, these parameters should be
compared with those of Ae. aegypti resting sites within the
If future research with GM mosquitoes requires double
layers of mesh for containment purposes, we expect that it
will affect airflow through cages and thus may increase inter-
nal temperature and alter humidity, compared with the sin-
gle layer of mesh we used. In those cases, investigators may
need to modify cage design to account for environmental
differences inside cages compared with natural Ae. aegypti
Cages should be built in an open space that minimizes or
eliminates the risk of falling objects damaging the structure.
Shade, which is important to prevent cages from overheating,
can be provided with shade netting or other structural features
that preserve desired environmental attributes without risking
damage to the integrity of the structure.
Blood meals should be provided to females at least twice a
week. This will allow at least two ovipositions per week and,
thus, a larger number of data points that can be used to calcu-
late and assess fluctuations in population density.
Although experiments were conducted in structures
designed to be contained environments, invasion by predators
can be an issue affecting target species mortality. For example,
small airborne spiders can land on the cage and ants can for-
age or start new colonies in cages, even if attempts are made
to isolate the cages from ants. Because the use of any insecti-
cide must be avoided, in a few months cages located in a tropi-
cal environment can become a favorable habitat for predators
that exploit prey enclosed in a limited setting. Predation can
be minimized by carefully cleaning cages before starting an
experiment and regularly checking them, and using a platform
to support cages for evidence of predator invasion. In our tri-
als, although very few ants were detected in the cages during
normal trial activities, at the end of the trial ant colonies were
discovered concealed between wooden boards on the plat-
form, inside one flowerpot, and under the white plastic sheet
covering the floor of a tent cage.
Environmental heterogeneity in cages is an important aspect
for semi-field trials, because it gives Ae. aegypti the possibility
to exploit different microhabitats in an enclosed environment.
However, we emphasize that the more heterogeneous the cage
environment, the more difficult it becomes to visually inspect
the cages and detect mosquito predators . We suggest using
simple resting sites such as black plastic buckets and avoiding
structures or material because that could possibly create an
environment in which predators can hide and establish popu-
lations; i.e., potted plants.
The semi-field system we built and used during the trial is
cost-effective and has low maintenance costs, ~US $300 per
year. The plastic roof should be changed after 18 months and
the tricot mesh cage after 10 months of use because plastic and
fabric deteriorate in tropical areas caused by exposure to sun-
light and high humidity.
With the addition of a vestibule, openings with sleeves to
work inside the cage without having to enter the cage, and
the use of proper security operation procedures, our semi-
field system is suitable for contained field trials with geneti-
cally engineered mosquitoes, like those bearing sterility genes
(i.e., RIDL) that do not require high security facilities.
Received July 27, 2010. Accepted for publication January 25, 2011.
Acknowledgments: We are grateful to Juan Carlos Joo Chang, Luís
Antonio García Rodas, Crystian Hidalgo Citalán Uriel, Hugo Cigarroa
de Los Reyes, and Nallely Sofía Maza Ramos for technical assistance.
Financial support: This research was supported by funds from the
Regents of the University of California from the Foundation for
the National Institutes of Health through the Grand Challenges in
Global Health Initiative (GC7 #316) and by a Pasteur Institute-Cenci
Bolognetti Foundation grant to Laura Valerio.
Disclosure: This research benefited from discussions with working
groups in the Research and Policy for Infectious Disease Dynamics
(RAPIDD) program of the Science and Technology Directorate,
Department of Homeland Security, and the Fogarty International
Center, National Institutes of Health.
Authors’ addresses: Luca Facchinelli and Laura Valerio, Department
of Entomology, University of California, Davis, CA, E-mails: lfacchi
email@example.com and firstname.lastname@example.org . J. Guillermo Bond, Janine
M. Ramsey, and M. Casas-Martinez, Centro Regional de Investigación
en Salud Pública (CRISP), Instituto Nacional de Salud Pública
(INSP), Chiapas, Mexico, E-mails: email@example.com , firstname.lastname@example.org ,
and email@example.com . Megan R. Wise de Valdez, Department of
Microbiology, Immunology, and Pathology, Colorado State University,
Fort Collins, CO, E-mail: firstname.lastname@example.org . Laura C.
Harrington, Department of Entomology, Cornell University, Ithaca,
New York, E-mail: email@example.com . Thomas W. Scott, Department
of Entomology, University of California, Davis CA, and Fogarty
International Center, National Institutes of Health, Bethesda, MD,
E-mail: firstname.lastname@example.org .
1. Franz AW , Sanchez-Vargas I , Adelman ZN , Blair CD , Beaty BJ ,
James AA , Olson KE , 2006 . Engineering RNA interference-
based resistance to dengue virus type 2 in genetically modified
Aedes aegypti . Proc Natl Acad Sci USA 103: 4198 – 4203 .
2. Alphey L , Nimmo D , O’Connell S , Alphey N , 2008 . Insect popula-
tion suppression using engineered insects . Adv Exp Med Biol
627: 93 – 103 .
3. Scott TW , Takken W , Knols BG , Boëte C , 2002 . The ecology of
genetic modified mosquitoes . Science 298: 117 – 119 .
4. Benedict M , D’Abbs P , Dobson S , Gottlieb M , Harrington LC ,
Higgs S , James A , James S , Knols B , Lavery J , O’Neill S , Scott
TW , Takken W , Toure Y , 2008 . Guidance for contained field tri-
als of vector mosquitoes engineered to contain a gene drive sys-
tem: recommendations of a Scientific Working Group . Vector
Borne Zoonotic Dis 8: 127 – 166 .
5. Ferguson HM , Ng’habi KR , Walder T , Kadungula D , Moore SJ ,
Lyimo I , Russell TL , Urassa H , Mshinda H , Killeen GF ,
Knols BG , 2008 . Establishment of a large semi-field system for
experimental study of African malaria vector ecology and con-
trol in Tanzania . Malar J 20: 158 .
6. Helinski ME , Knols BG , 2008 . Mating competitiveness of male
Anopheles arabiensis mosquitoes irradiated with a partially or
fully sterilizing dose in small and large laboratory cages . J Med
Entomol 45: 698 – 705 .
7. Scott TW , Rasgon JL , Black WC , Gould F , 2006 . Fitness studies:
developing a consensus methodology . Bridging Laboratory and
Field Research for Genetic Control of Disease Vectors . Knols
BG , Frontis LC , eds. Wageningen , The Netherlands , 171 – 181 .
Available at : http://library.wur.nl/frontis/disease_vectors/16_scott
.pdf. Accessed February 2011 .
8. Scott TW , 2005 . Containment of arthropod disease vectors . ILAR
J 46: 53 – 61 .
9. Kessler S , Guerin PM , 2008 . Responses of Anopheles gambiae,
Anopheles stephensi, Aedes aegypti , and Culex pipiens mosqui-
toes (Diptera: Culicidae) to cool and humid refugium condi-
tions . J Vector Ecol 33: 145 – 149 .
10. Williams CR , Long SA , Russell RC , Ritchie SA , 2006 . Field effi-
cacy of the BG-Sentinel compared with CDC Backpack
Aspirators and CO2-baited EVS traps for collection of adult
Aedes aegypti in Cairns, Queensland, Australia . J Am Mosq
Control Assoc 22: 296 – 300 .
FACCHINELLI AND OTHERS
11. Parish OO , Putnam TW , 1977 . Equation for the Determination of
Humidity from Dewpoint and Psychrometric Data . Washington
DC : National Aeronautic and Space Administration (NASA),
Technical Note TN D-8401 .
12. Scott TW , Morrison AC , Lorenz LH , Clark GG , Strickman D ,
Kittayapong P , Zhou H , Edman JD , 2000 . Longitudinal studies
of Aedes aegypti (Diptera: Culicidae) in Thailand and Puerto
Rico: population dynamics . J Med Entomol 37: 77 – 88 .
13. Muir LE , Kay BH , 1998 . Aedes aegypti survival and dispersal
estimated by mark-release-recapture in northern Australia .
Am J Trop Med Hyg 58: 277 – 282 .
14. Reiter P , 2007 . Oviposition, dispersal, and survival in Aedes aegypti :
implications for the efficacy of control strategies . Vector Borne
Zoonotic Dis 7: 261 – 273 .
15. Colton YM , Chadee DD , Severson DW , 2003 . Natural skip ovipo-
sition of the mosquito Aedes aegypti indicated by codominant
genetic markers . Med Vet Entomol 17: 195 – 204 .
16. McDonald PT , 1977 . Population characteristics of domestic Aedes
aegypti (Diptera: culicidae) in villages on the Kenya Coast I. Adult
survivorship and population size . J Med Entomol 14: 42 – 48 .
17. Trpis M , Hausermann W , 1986 . Dispersal and other population
parameters of Aedes aegypti in an African village and their
possible significance in epidemiology of vector-borne diseases .
Am J Trop Med Hyg 55: 1263 – 1279 .
18. Styer LM , Carey JR , Wang JL , Scott TW , 2007 . Mosquitoes do
senesce: departure from the paradigm of constant mortality .
Am J Trop Med Hyg 76: 111 – 117 .
19. Reisen WK , Mahmood F , Parveen T , 1980 . Anopheles culicifa-
cies Giles: a release-recapture experiment with cohorts of
known age with implications for malaria epidemiology and
geneti cal control in Pakistan . Trans R Soc Trop Med Hyg 743:
307 – 317 .
20. Constantini C , Li S , della Torre A , Sagnon N , Coluzzi M , Taylor
CE , 1996 . Density, survival and dispersal of Anopheles gambiae
complex mosquitoes in a West African Sudan savanna village .
Med Vet Entomol 10: 203 – 219 .
21. Haramis LD , Foster WA , 1983 . Survival and population density of
Aedes triseriatus (Diptera: Culicidae) in a wood lot in central
Ohio, USA . J Med Entomol 20: 391 – 398 .
22. Harrington LC , Vermeylen F , Jones JJ , Kitthawee S , Sithiprasasna
R , Edman JD , Scott TW , 2008 . Age-dependent survival of the
dengue vector, Ae. aegypti , demonstrated by simultaneous
release and recapture of different age cohorts . J M Entomol 45:
307 – 313 .
23. Harrington LC , Edman JD , Costero AC , Clark GG , Kittayapong P ,
Scott TW , 2001 . Analysis of survival of young and old Aedes
aegypti (Diptera: Culicidae) from Puerto Rico and Thailand.
2001 . J Med Entomol 38: 537 – 547 .