Environmental factors associated with the distribution of Anopheles arabiensis and Culex quinquefasciatus in a rice agro-ecosystem in Mwea, Kenya.

Abstract

Studies were conducted between May and June, 2006 to investigate the environmental factors affecting the distribution of An. arabiensis Patton and Culex quinquefasciatus Say in Mwea, Kenya. The sampling unit comprised all non-paddy aquatic habitats and ten randomly selected paddies and canals located within a 200 m radius from the periphery of the study site. Thirteen physico-chemical variables were recorded for each sampling site in each sampling occasion and a sample of mosquito larvae and other aquatic invertebrates collected. The non-paddy aquatic habitats identified included pools and marshes. Morphological identification of 1,974 mosquito larvae yielded four species dominated by Cx. quinquefasciatus (73.2%) and An. arabiensis (25.0%). Pools were associated with significantly higher Cx. quinquefasciatus larval abundance and less diversity of other aquatic invertebrates compared with other habitat types. In contrast, the abundance of An. arabiensis did not differ significantly among habitat types. Culex quinquefasciatus habitats had higher water conductivity and exhibited a higher abundance of other aquatic invertebrates than An. arabiensis habitats. Chi-square analysis indicated that the two species were more likely to coexist in the same habitats than would be expected by chance alone. Anopheles arabiensis larvae were positively associated with dissolved oxygen and adults of family Haliplidae and negatively associated with emergent vegetation and Heptageniidae larvae. Culex quinquefasciatus larvae were positively associated with dissolved oxygen, total dissolved solids, Chironomidae larvae, and Microvelidae adults and negatively associated with emergent vegetation. These findings suggest that both biotic and abiotic factors play a significant role in niche partitioning among Cx. quinquefasciatus and An. arabiensis, a factor that should be considered when designing an integrated vector control program.

Full text

Page 1

56?Journal of Vector Ecology June?2008
Environmental factors associated with the distribution of Anopheles arabiensis and
Culex quinquefasciatus in a rice agro-ecosystem in Mwea, Kenya
Ephantus J. Muturi1, Joseph Mwangangi2,4, Josephat Shililu2,3, Benjamin G. Jacob1, Charles Mbogo4,
John Githure2, and Robert J. Novak1
1Department of Medicine, William C. Gorgas Center for Geographic Medicine, 206C Bevill Biomedical Research Building,
845 19th Street South, University of Alabama at Birmingham, Birmingham, AL 35294, U.S.A.
2Human Health Division, International Centre of Insect Physiology and Ecology, Nairobi, Kenya
3Department of Zoology, Jomo Kenyatta University of Agriculture and Technology, Nairobi, Kenya
4Centre for Geographic Medicine Research-Coast, Kenya Medical Research Institute, Kilifi, Kenya
Received 6 July 2007; Accepted 3 December 2007
ABSTRACT: Studies were conducted between May and June, 2006 to investigate the environmental factors affecting the
distribution of An. arabiensis Patton and Culex quinquefasciatus Say in Mwea, Kenya. The sampling unit comprised all non-
paddy aquatic habitats and ten randomly selected paddies and canals located within a 200 m radius from the periphery of the
study site. Thirteen physico-chemical variables were recorded for each sampling site in each sampling occasion and a sample
of mosquito larvae and other aquatic invertebrates collected. The non-paddy aquatic habitats identified included pools and
marshes. Morphological identification of 1,974 mosquito larvae yielded four species dominated by Cx. quinquefasciatus
(73.2%) and An. arabiensis (25.0%). Pools were associated with significantly higher Cx. quinquefasciatus larval abundance and
less diversity of other aquatic invertebrates compared with other habitat types. In contrast, the abundance of An. arabiensis
did not differ significantly among habitat types. Culex quinquefasciatus habitats had higher water conductivity and exhibited
a higher abundance of other aquatic invertebrates than An. arabiensis habitats. Chi-square analysis indicated that the two
species were more likely to coexist in the same habitats than would be expected by chance alone. Anopheles arabiensis larvae
were positively associated with dissolved oxygen and adults of family Haliplidae and negatively associated with emergent
vegetation and Heptageniidae larvae. Culex quinquefasciatus larvae were positively associated with dissolved oxygen, total
dissolved solids, Chironomidae larvae, and Microvelidae adults and negatively associated with emergent vegetation. These
findings suggest that both biotic and abiotic factors play a significant role in niche partitioning among Cx. quinquefasciatus
and An. arabiensis, a factor that should be considered when designing an integrated vector control program. Journal of
Vector Ecology 33 (1): 56-63. 2008.
Keyword Index: An. arabiensis, Cx. quinquefasciatus, environmental factors, integrated vector management, rice agro-
ecosystem, Kenya.
INTRODUCTION
Mosquito species differ in the type of aquatic habitats
they prefer for oviposition based on location, the physico-
chemical condition of the water body, and the presence
of potential predators (Shililu et al. 2003, Piyaratnea
et al. 2005). Physico-chemical factors that influence
oviposition, survival, and the spatio-temporal distribution
of important disease vector species include salts, dissolved
organic and inorganic matter, degree of eutrophication,
turbidity, presence of suspended mud, presence or absence
of plants, temperature, light and shade, and hydrogen
ion concentration (Mogi 1978, Amerasinghe et al. 1995,
Gimnig et al. 2001). Understanding how these factors affect
the distribution of a particular vector species and how they
influence larval abundance is an essential component of
larval biology and of great importance in the design and
implementation of integrated vector management plans.
Several studies have examined the relationship between
habitat characteristics and larval abundance. In Sri Lanka,
Anopheles culicifacies Giles was positively associated with
light and vegetation and negatively associated with the
presence of potential predators, while Anopheles varuna
Iyengar was positively associated with a variety of aquatic
fauna (Piyaratnea et al. 2005). In Venezuela, salinity
and dissolved oxygen were associated with the spatial
distribution of Anopheles aquasalis Curry and Anopheles
oswaldoi Peryassu (Grillet 2000). In Orange County, CA,
U.S.A., the distribution of Culex tarsalis Coquillett was
significantly associated with the percent cover by Typha spp.
root masses and Typha spp. stem density per square meter
(Walton et al. 1990). Culex quinquefasciatus Say larvae,
in Peninsular Malaysia, were most abundant in polluted
drains containing 1.0 to 2.0 g/liter of dissolved oxygen, 1.0-
2.4 g/liter of soluble reactive phosphate, and 0.1-0.9 g/liter
of ammoniacal nitrogen (Hassan et al. 1993).
In Africa, similar studies with malaria vectors An.
gambiae Giles s.s and An. arabiensis Patton have yielded
variable results. While some studies failed to detect any
significant relationship between An. gambiae s.l. and

Page 2

Vol.?33,?no.?1?Journal of Vector Ecology 57
environmental variables (Minakawa et al. 1999), others
have reported significant relationships. For example, in
Eritrea, An. arabiensis was associated with shallow, clean
water and sunlit habitats (Shililu et al. 2003) as were those
reported for An. gambiae s.s in western Kenya (Munga et
al. 2005). Fluctuations in physico-chemical factors of the
rice field environment during the course of the rice growing
cycle also impacted significantly on temporal distribution
and abundance of An. arabiensis (Muturi et al. 2007a).
Despite the importance of Cx. quinquefasciatus in
transmission of Bancroftian filariasis in Kenya (Mwandawiro
et al. 1997), little is known about its larval ecology. Few
studies have reported a significant association between the
genus Culex and environmental factors such as pH, canopy
coverage, debris coverage (Minakawa et al. 1999), and
decaying organic matter (Asimeng and Mutinga 1993), with
little attempt to assess the effect of these factors on species
population dynamics (Muturi et al. 2007a).
A preliminary study aimed at understanding the
biology of anophelines prior to implementation of a malaria
vector control program in Mwea Irrigation Scheme, Kenya,
revealed Cx. quinquefasciatus to be a predominant nuisance
species and a potential vector of filariasis and arboviruses in
the area (Muturi et al. 2006). Further studies revealed that
this species thrives in a variety of aquatic habitats including
rice fields, canals, seepage areas, ditches, marshes, pits, and
temporary pools (Muturi et al. 2007a,b). The objective of this
study was to characterize the physico-chemical and biotic
factors of these mosquito larval habitats and to identify the
factors that influence the abundance and distribution of
An. arabiensis and Cx. quinquefasciatus in diverse aquatic
habitats.
MATERIALS AND METHODS
Study area
The Mwea Rice Irrigation Scheme is located 100 km
northeast of Nairobi in a riverine plain southeast of Mount
Kenya at an altitude of about 1,200 m. A full description
of the area is given by Muturi et al. (2006). The study area
has two annual rainfall seasons, the long rains in April/May
and the short rains in October/November. The average
annual rainfall, temperature, and relative humidity is 950
mm (range: 356-1,626 mm), 21.3°C (range: 16.0-26.5° C),
and 59.5% (range: 52-67%), respectively. The 1999 Kenya
National census estimated the number of individuals in
this area to be 150,000 and the number of households to be
25,000. The Rice Scheme covers an area of about 13,640 ha,
more than 50% of which is used for irrigated rice cultivation
and the remaining area is used for subsistence farming,
grazing, and community activities.
Kangichiri village, located within the scheme was
selected for this study. The village has approximately 150
homesteads with approximately 650 residents. More than
90% of the houses have mud walls with iron roofing. Cows,
goats, chickens, and donkeys are the primary domestic
animals kept in the village. Their sheds are generally located
within 5 m of most houses. More than 75% of the village land
is under rice cultivation and human habitation occupies the
remaining area with less than 10% utilized for growing a
variety of vegetables and bananas. The typical rice cultivation
cycle includes a land preparation–transplanting period
(July–August), a growing period (August–November), and
a post-harvest period (November–December).
Sampling of mosquito larvae and non-mosquito
invertebrates
Mosquito larval samples were collected over a six-week
period between May and June, 2006. The collections were
done in all non-paddy aquatic habitats and ten randomly
selected paddies and associated canals located within a
200 m radius of the village. The habitats were sampled
for mosquito larvae twice per week and up to 20 dipper
samples, depending on the size of the aquatic habitat, were
taken using a standard mosquito dipper (350 ml). Mosquito
larvae were sorted by genus and the 3rd and 4th instars were
preserved in 100% ethanol. The larvae were later identified
microscopically to species using the taxonomic keys of
Hopkins (1952) and Gillies and Coetzee (1987).
The non-mosquito aquatic invertebrates collected in the
dipper samples, along with mosquito larvae at each habitat,
were also preserved in 100% ethanol and later identified
to family using taxonomic keys of Merritt and Cummins
(1996). The number of individuals of each family identified
were counted and recorded.
Larval habitat characterization
Aquatic habitats were classified based on their size and
appearance. A small excavation filled with water was defined
as a pool and a low-lying wet land with grassy vegetation
was defined as a marsh. An irrigated or flooded field where
rice is grown was classified as a paddy and the long and
narrow strip of water made for paddy irrigation was defined
as a canal. The environmental variables recorded from each
sampling site when each sample was collected included;
water depth, turbidity, salinity, total dissolved solids (TDS),
pH, temperature, conductivity, dissolved oxygen, distance
to the nearest house, floating, emergent and submerged
vegetation cover, and habitat type. Distance of the sampling
site to the nearest house was measured with a tape when
it was shorter than 50 m. When the distance exceeded 50
m, it was estimated visually. Emergent vegetation cover was
defined as the proportion of the water surface area that was
covered by emergent vegetation. Floating vegetation cover
was estimated as the proportion of the water surface area
that was covered by floating vegetation. Any plant below the
water surface was classified as submerged vegetation, and the
proportion of water surface area covered by this vegetation
was estimated. Water turbidity was estimated visually
against a white background and classified as either clear, less
turbid, or turbid. The pH, conductivity, dissolved oxygen,
and temperature were measured using a hand-held YSI
650 Multi-Parameter Display System (YSI Environmental,
YSI Incorporated, Yellow Springs, OH. U.S.A.). Salinity
and TDS were measured using field hand-held equipment
YSI EC 300 (YSI Environmental, YSI Incorporated, Yellow

Page 3

58?Journal of Vector Ecology June?2008
Springs, OH, U.S.A.).
Data analyses
Data were analyzed using SPSS version 11.5 (SPSS,
Inc., Chicago, IL, U.S.A.) and SYSTAT version 11 (SYSTAT
Software Inc, San Jose, CA, U.S.A.) statistical packages.
The relative abundance of mosquitoes was expressed as
the number of mosquito larvae and pupae per 20 dips
because larval and pupal counts were low. The degree of
association between Anopheles and Culex larvae in the
aquatic habitats was tested by chi-square. The differences
in larval counts among habitat types were compared
by repeated measures analysis of variance (ANOVA).
Variation in habitat characteristics based on the presence
or absence of An. arabiensis and Cx. quinquefasciatus was
compared using ANOVA test for continuous variables
and chi-square for categorical variables. Where significant
differences were obtained in ANOVA test, the means were
separated by Tukey’s HSD test. Forward multiple regression
analysis was used to obtain the best predictor variables
explaining the abundance of the mosquito larvae. Statistical
analysis was done after a log transformation log10 (n+1) of
larval abundance values to normalize the distribution and
minimize the standard error (SE).
RESULTS
Species composition and abundance of mosquitoes and
other aquatic invertebrates
In total, 528 collections were made at 44 sampling
sites, representing four habitat types; paddies (n=10), canals
(n=10), pools (n=16), and marshes (n=8). During the 528
sampling visits, the sampling sites contained water on 348
visits. Anopheles larvae were collected in 176 samples, 71
of which had Anopheles alone. Culex larvae were found in
149 collections and 44 of these collections were exclusively
Culex sp. Combinations of both Anopheles and Culex larvae
were found in 105 collections. Chi-square analysis indicated
that Anopheles and Culex larvae were more likely to co-exist
in the same sampling site than would be expected by chance
alone (χ2 = 46.21, P < 0.01). The mean number of Anopheles
larvae collected was 4.23 (SE = ± 1.01) per 20 dips and the
mean number of Culex larvae was 12.63 (SE = ± 2.91) per 20
dips. Pupal counts averaged 1.91 (SE = ± 0.75) per 20 dips.
A total of 1,974 larvae was examined microscopically
and identified morphologically to species. The collections
yielded four mosquito species dominated by Cx. quinque-
fasciatus (73.2%) and An. arabiensis (25.0%). The other two
species were Culex annulioris Theobald (1.3%) and Culex ti-
gripes Grandpre and Charmoy (0.5%). Anopheles arabiensis
and Cx. quinquefasciatus were represented in all four habitat
types and Cx. annulioris and Cx. tigripes were found only in
pools and canals (Table 1). Repeated measures ANOVA and
Tukey’s HSD tests revealed that An. arabiensis larval abun-
dance did not vary significantly among habitat types (F =
0.72, df = 3, 11, P > 0.05), but Cx. quinquefasciatus larval
abundance was significantly higher in pools compared with
the other habitat types (F = 6.32, df = 3, 11, P < 0.05).
Table 2 shows the distribution of other aquatic
invertebrates in different habitat types. Thirteen insect
families belonging to five orders were collected (n =
955). These were mainly dominated by Dytiscidae
(35.5%), Ephemeridae (29.4%), Belostomatidae (8.2%),
Hydrophilidae (5.9%), and Notonectidae (4.7%). Mites
(Acari) and snails (Mollusca) were also encountered
occasionally and together with the other insect groups, they
accounted for the remaining 16.4% of the total collection.
Paddies and canals were the most diverse in terms of the
number of insect families and other aquatic arthropods
collected, whereas pools were the least diverse.
Larval habitat characteristics
The physico-chemical characteristics of the four habitat
types are represented in Table 3. Overall, paddies were
located the furthest distance from the nearest homestead
and were also characterized by low values of TDS, high
amounts of emergent and floating vegetation cover, and
cooler turbid waters. Canals were deep habitats with less
turbid waters, low values of TDS, less emergent vegetation
cover, and high amounts of floating vegetation cover.
Marshes were deep, clean water habitats closer to human
habitation and contained low amounts of floating vegetation
cover, high amounts of emergent vegetation cover, and high
TDS values. Pools were the shallowest, warmest and most
turbid habitats with small amounts of floating and emergent
vegetation cover and high TDS values.
The characteristics of An. arabiensis and Cx.
quinquefasciatus larval habitats are summarized in Table
4. Culex quinquefasciatus larval habitats differed from An.
arabiensis larval habitats in water conductivity and the
average number of other aquatic invertebrates collected
in the habitats. Culex quinquefasciatus habitats had higher
water conductivity values and higher abundance of other
aquatic invertebrates compared with An. arabiensis habitats
(F = 4.562 and 9.950, df = 3, 114, P < 0.05). Habitats with
both species differed from habitats with either species in the
presence of emergent vegetation (less common in habitats
with both species) and distance to the nearest house
(habitats with both species were closer to human habitation,
F = 5.497 and 8.553, df = 3, 209, P < 0.05).
Relationship between mosquito larval abundance and
environmental variables
Forward multiple regression analysis was used to analyze
the effect of various parameters on the relative abundance of
immature stages of An. arabiensis and Cx. quinquefasciatus.
The model included 26 variables; 13 physico-chemical
variables and 13 insect families that were identified in the
samples. Significant models explaining 30.1% and 48.9%
of the relative abundance of Cx. quinquefasciatus and An.
arabiensis, respectively, were fitted by the regression model.
Seven of the 26 parameters were significantly associated
with the relative abundance of immatures of at least one of
the two mosquito species. These included dissolved oxygen,
emergent vegetation, and TDS among the physico-chemical
parameters and Haliplidae, Heptageniidae, Chironomidae,

Page 4

Vol.?33,?no.?1?Journal of Vector Ecology 59
Table 1. Species composition and abundance (counts/20 dips) of mosquito species collected in different habitat types in
Kangichiri village, Mwea, Kenya.
Habitat typeNo. of sites
Cx. quinquefasciatus An. arabiensisCx. annuliorisCx. tigripes
Paddy
Canal
Pool
Marsh
Total
10
10
16
8
44
2.33 ± 1.15
8.43 ± 4.71
36.35 ± 12.62
0.88 ± 0.88
9.39 ± 2.64
3.11 ± 1.71
2.27 ± 1.04
7.04 ± 2.27
0.31 ± 0.31
3.28 ± 0.96
NF0.09 ± 0.04
NF
0.08 ± 0.05
NF
0.06 ± 0.02
0.62 ± 0.62
0.08 ± 0.05
NF
0.16 ± 0.15
NF: Indicates that the species was not found in respective habitat type.
Table 2. Relative abundance (counts/20 dips) of non-mosquito aquatic invertebrates in different habitat types in Kangichiri
village, Mwea, Kenya.
Habitat type
Canal
0.14
0.00
2.49
0.11
0.35
0.19
1.59
0.19
0.19
0.16
0.32
0.54
0.08
0.05
0.00
Insect family
Hydrophilidae
Haliplidae
Dytiscidae
Curculionidae
Libellulidae
Coenagrionidae
Ephemeridae
Heptageniidae
Chironomidae
Notonectidae
Corixidae
Belostomatidae
Microvellidae
Snails
Mites
Paddy
0.31
0.08
2.47
0.09
0.12
0.03
2.20
0.16
0.07
0.29
0.24
0.52
0.04
0.01
0.03
Pool
0.85
0.12
1.65
0.00
0.00
0.04
0.65
0.00
0.58
0.15
0.00
0.20
0.00
0.00
0.00
Marsh
0.38
0.00
1.25
0.56
0.00
0.13
2.50
0.06
0.13
0.81
0.06
0.19
0.00
0.13
0.00
Total
0.36
0.06
2.19
0.13
0.14
0.08
1.82
0.13
0.19
0.29
0.20
0.51
0.04
0.03
0.01
and Microveliidae among the potential competitors. The
relative abundance of An. arabiensis larvae was negatively
associated with emergent vegetation and Heptageniidae
larvae and positively associated with dissolved oxygen and
Haliplidae adults. Cx. quinquefasciatus larval abundance
was negatively associated with emergent vegetation and
positively associated with TDS, dissolved oxygen, larvae of
Chironomidae, and adults of Microvelidae (Table 5). The
other environmental variables were excluded in the model
because they had weaker associations with mosquito larval
abundance.
DISCUSSION
Results of this study demonstrated co-existence of
An. arabiensis and Cx. quinquefasciatus larvae in the same
aquatic habitats at rates greater than would be expected by
chance alone. This is in agreement with results from previous
work (Minakawa et al. 1999, Fillinger et al. 2004). Although
both species were associated with emergent vegetation cover
and dissolved oxygen, they differed in their association with
the other five of the seven significant variables. Anopheles
arabiensis larvae were positively associated with adults of
family Haliplidae and negatively associated with larvae
of family Heptageniidae, and Cx. quinquefasciatus was
positively associated with TDS, Chironomidae larvae, and
Microveliidae adults. These findings suggest that both
biotic and abiotic factors play a significant role in niche
partitioning among Cx. quinquefasciatus and An. arabiensis,
which should be considered when designing an integrated
vector control program.
Mosquitoes use chemical and biological cues to detect
the presence of larval competitors and avoid ovipositing
in such habitats (Blaustein and Kotler 1993). Members
of the family Heptageniidae feed on detritus (Lamp and
Britt 1981) which also constitute the food for mosquito

Page 5

60?Journal of Vector Ecology June?2008
Table 3. Average values (± SE) of the environmental variables used to characterize the aquatic habitats in Kangichiri village,
Mwea, Kenya.
Variables
Habitat type
Paddy
49.20 ± 2.02
Canal PoolMarsh
9.38 ± 0.43Distance (m) 33.92 ± 2.9530.54 ± 2.40
Emergent vegetation cover (%)38.76 ± 3.9530.30 ± 2.6425.23 ± 3.5651.88 ± 8.32
Floating vegetation cover (%)20.35 ± 3.618.11 ± 2.571.15 ± 0.850.00 ± 0.00
Submerged vegetation cover (%)0.37 ± 0.200.03 ± 0.03 0.04 ± 0.040.38 ± 0.38
Depth (cm)7.04 ± 0.2810.24 ± 1.07 6.19 ± 0.569.94 ± 1.26
Salinity (ppt)17.08 ± 1.87 23.79 ± 2.5721.75 ± 4.0626.38 ± 7.31
Dissolved oxygen (Mg/L)5.62 ± 1.52 2.78 ± 0.623.52 ± 0.951.86 ± 0.77
pH7.16 ± 0.067.09 ± 0.056.86 ± 0.07 7.10 ± 0.09
Conductivity (µs) 152.17 ± 10.41 134.91 ± 14.23202.57 ± 26.57 172.84 ± 30.39
Temperature (°C)27.08 ± 0.71 27.52 ± 1.229.44 ± 0.67 24.07 ± 0.58
Total dissolved solid (ppt)0.10 ± 0.01 0.08 ± 0.010.18 ± 0.02 0.23 ± 0.04
Invertebrate total6.63 ± 0.706.41 ± 0.93 4.69 ± 0.876.19 ± 0.99
Turbidity 1.99 ± 0.081.78 ± 0.13 2.38 ± 0.151.25 ± 0.11
Table 4. Average values (± SE) of the measured environmental factors in aquatic habitats with either An. arabiensis or Cx.
quinquefasciatus alone, both species, or none of the mosquito species in Kangichiri village, Mwea, Kenya.
Cx. quinquefasciatus
only
44
An. arabiensis
only
71
BothNeither
No. of collections 105128
Distance (m) 41.92 ± 4.5646.11 ± 4.6032.87 ± 3.67 37.62 ± 2.12
Emergent vegetation (%) 41.17 ± 6.14 43.72 ± 7.3922.09 ± 4.7136.88 ± 2.92
Floating vegetation (%) 10.00 ± 7.6910.61 ± 6.839.61 ± 4.51 13.11 ± 2.44
Submerged vegetation (%)0.17 ± 0.170.67 ± 0.550.30 ± 0.26 0.15 ± 0.11
Depth (cm) 9.25 ± 1.668.28 ± 1.18 8.70 ± 1.24 7.59 ± 0.36
Salinity (ppt) 22.35 ± 3.2221.64 ± 4.0317.33 ± 3.00 20.72 ± 2.05
Dissolved oxygen (Mg/L)7.23 ± 5.228.23 ± 3.419.20 ± 3.02 1.97 ± 0.33
pH 6.80 ± 0.11 6.90 ± 0.17 6.85 ± 0.087.20 ± 0.03
Conductivity (μs) 192.54 ± 39.00 106.82 ± 20.13153.98 ± 23.63 164.97 ± 9.72
Temperature (°C) 26.64 ± 0.95 27.88 ± 1.9429.88 ± 1.9526.65 ± 0.46
TDS (ppt) 0.14 ± 0.030.11 ± 0.020.10 ± 0.020.13 ± 0.01
Other invertebrates 10.67 ± 1.90 4.61 ± 0.94 5.35 ± 0.746.15 ± 0.57
Habitats with clear water (%)25.027.839.132.1

Page 6

Vol.?33,?no.?1?Journal of Vector Ecology 61
larvae. Our findings seem to indicate that gravid females
of An. arabiensis would avoid ovipositing in habitats where
members of family Heptageniidae are present, presumably
to avoid direct competition. Kramer and Garcia (1989)
observed a positive relationship between mosquito larvae
and other predators. However, none of the insect families
that were significantly associated with mosquito larvae
in the current study is known to compete with or prey
on mosquito larvae. Because most insect predators are
generalized in their diet (Mogi 1978) they may prey on
individuals across these families. Our findings suggest that
habitats in which mosquito larvae are abundant are also
attractive to other aquatic invertebrates that are preyed
upon by similar predators. Interestingly, the most abundant
insect families were not significant predictors of mosquito
larval abundance, an indication that other factors besides
numbers are important in determining the outcome of
interspecific interactions. Although our sampling may have
been biased in collection of some aquatic invertebrates, these
findings underscore the need to understand how mosquito
larvae interact with other aquatic insects when formulating
a vector control program with the aim of preserving the
natural sources of mosquito larval mortality.
Emergent vegetation is known to have direct effects
on some mosquito species by obstructing gravid females
from ovipositing and supporting a greater diversity of
aquatic predators (Rajendran and Reuben 1991, Grillet
2000). Emergent vegetation may also reduce the amount
of sunlight reaching the water surface, resulting in lower
temperatures. Reduced temperatures cause a decline in
microbial growth upon which mosquito larvae depend
on and increases the larval development time exposing
them to greater risks of contact with potential predators
and competitors (Ramachandra-Rao 1984). To avoid these
difficulties, An. arabiensis is known to prefer open sunlit
habitats without vegetation for oviposition (Gimnig et al.
2001, Shililu et al. 2003) and the pattern appears to be the
same for Cx. quinquefasciatus.
Several studies have examined the influence of dissolved
oxygen concentration on the abundance of Anopheles spp.
and Culex spp. with contradicting results. Grillet (2000)
reported a positive association between dissolved oxygen
and the abundance of An. oswaldoi and Amerasinghe et al.
(1995) reported a negative association between dissolved
oxygen and An. culicifacies. Similarly, Senior-White (1926)
reported Cx. vishnui from hyper-oxygenated water, a finding
that has recently been confirmed by Sunish and colleagues
(2006). However, it remains unknown whether it is oxygen
per se or an associated physico-chemical or abiotic factor
that influences the abundance of mosquito species. Sunish
et al. (2006) suggested that high algal productivity and
associated photosynthesis is responsible for high dissolved
oxygen concentrations in aquatic habitats, thereby favoring
higher survival of mosquito larvae. In the current study,
we did not quantify the amount of algal growth, but
considering that the majority of the sampled habitats had
Azolla sp. and other floating and submerged vegetation
which consists mainly of algal biomass, it is likely that algal
productivity resulted in higher concentration of dissolved
oxygen, thereby promoting larval productivity. TDS, which
is the sum of all dissolved organic, inorganic, and suspended
solids in water, was also a significant factor in productivity
of Cx. quinquefasciatus. In most areas of its distribution,
Cx. quinquefasciatus prefer habitats rich in dissolved matter
(Hassan et al. 1993) and such habitats tend to have high
TDS.
Habitats in which both An. arabiensis and Cx.
quinquefasciatus co-existed differed from those that
harbored either of the two species. Habitats with both
species were closer to human habitation and were less
likely to have emergent vegetation than those with either
of the two species. The results also indicated that Cx.
quinquefasciatus was able to thrive in habitats with higher
density of other aquatic invertebrates compared with An.
Species Environmental variables CoefficientsP values
An. arabiensis
Dissolved oxygen (mg/l)0.679 0.000
Haliplidae0.212 0.001
Heptageniidae-0.2440.001
Emergent vegetation -0.1470.013
Cx. quinquefasciatus
Chironomidae0.382 0.000
Microveliidae0.2630.000
Dissolved oxygen (mg/l) 0.1830.009
Emergent vegetation -0.1730.015
Total dissolved solids (ppt)0.144 0.041
Table 5. Multiple regression analysis of the abundance of An. arabiensis and Cx. quinquefasciatus in relation to measured
biotic and abiotic characteristics.

Page 7

62?Journal of Vector Ecology June?2008
arabiensis, suggesting that this species was more tolerant
to interspecific associations among aquatic invertebrates
than An. arabiensis. Considering that emergent vegetation
was a significant predictor of both An. arabiensis and Cx.
quinquefasciatus, the results suggest that the two species are
likely to exploit the habitat characteristics that are strongly
associated with either species. Gimnig et al. (2001) observed
a similar trend between An. arabiensis and An. gambiae s.s.
This phenomenon is likely to benefit both species particularly
in the more complex habitats such as paddies and canals
where the diversity of other aquatic invertebrates, including
potential predators, was great.
Although the complex community structure of the rice
fields and canals does not support higher mosquito density
as do the less complex habitats such as pools, they are
capable of holding water for a longer period than the pools,
making them important larval habitats over time. Previous
studies have shown that a large number of low density, but
continuously productive, habitats contribute more to the
density of adult mosquito populations than single high
density larval habitats (Grillet 2000). This underscores the
need to exhaustively target all the larval habitats regardless
of their larval densities.
The specific cues that trigger oviposition behavior in
mosquitoes are largely unknown. It has been suggested
that a large number of variables may be correlated with
other characteristics that act as cues for ovipositing females
(Gimnig et al. 2001). Because of methodological and logistical
difficulties, the current study had several limitations. First,
it is possible that the estimated environmental variables
were correlated with other factors that were not considered
in this study. Secondly, the study was conducted over a six-
week period after which time all the aquatic habitats dried
up. Although this design provided essential information on
environmental factors that regulate mosquito population
in temporary aquatic habitats, this period was too short to
establish the temporal changes in larval habitats. Thirdly, it
was also problematic to compare the productivity of larger
habitats such as paddies, canals, and marshes with the smaller
pools due to variation in the number of dips collected. To
standardize the results, we estimated the larval abundance
as larval counts per 20 dips. Although this method is widely
used in larval sampling, it may have overestimated the larval
counts in small pools. Finally, the diversity of other aquatic
invertebrates was estimated from the dipper samples and
this may have underestimated both the species diversity
and abundance. A more comprehensive study targeting a
large number of variables and a more elaborate method of
estimating the diversity of other aquatic invertebrates over a
longer period of time is warranted.
In summary, this study has established the habitat
characteristics of Cx. quinquefasciatus and An. arabiensis
in a rice agro-ecosystem in Mwea Kenya. The study has
established that although both species may exploit the same
habitats for larval development, they respond differently to
habitat factors resulting in niche partitioning. These results
provide useful information that can be used as a guide for
designation and implementation of an integrated vector
control program.
Acknowledgments
We are grateful to Prof. Christian Borgemeister,
Director General, International Centre of Insect Physiology
and Ecology, for his strong support in this project. We
acknowledge the data collection provided by Simon Muriu,
Peter Barasa, Enock Mpanga, James Wauna, Peter M.
Mutiga, William Waweru, Nelson Maingi, Martin Njigoya,
Paul K. Mwangi, Christine W. Maina, Isabel M. Marui,
Gladys Karimi, Irene Kamau, Susan W. Mugo, Nicholus
Gachoki, Charles Kiura, and Naftaly Gichuki. This research
was supported by NIH/NIAID grant # U01A1054889
(Robert Novak).
REFERENCES CITED
Amerasinghe, F., N. Indrajith, and T. Ariyasena. 1995.
Physico-chemical characteristics of mosquito breeding
habitats in an irrigation development area in Sri Lanka.
Ceylon J. Sci (Biological Sciences). 24: 13-29.
Asimeng, E. and M. Mutinga. 1993. Effect of rice husbandry
on mosquito breeding at Mwea Rice Irrigation Scheme
with reference to biocontrol strategies. J. Am. Mosq.
Contr. Assoc. 9: 17-22.
Blaustein, L. and B. Kotler. 1993. Oviposition habitat
selection by the mosquito, Culiseta longiareolata:
Effects of conspecifics, food and green toad tadpoles.
Ecol. Entomol. 18: 104-108.
Fillinger, U., S. Sonye, G.F. Killeen, B.G. Knols, and N.
Becker. 2004. The practical importance of permanent
and semipermanent habitats for controlling aquatic
stages of Anopheles gambiae sensu lato mosquitoes:
operational observations from a rural town in western
Kenya. Trop. Med. Intern. Hlth. 9: 1274–1289.
Gillies, M.T. and M. Coetzee. 1987. A supplement to the
Anophelinae of Africa south of the Sahara (Afro-
tropical region). Johannesberg. Pub. South Afri. Inst.
Med. Res. 55: 1-143.
Gimnig, J., M. Ombok, L. Kamau, and W. Hawley. 2001.
Characteristics of larval anopheline (Diptera: Culicidae)
habitats in Western Kenya. J. Med. Entomol. 38: 282-
288.
Grillet, M. 2000. Factors associated with distribution of
Anopheles aquasalis and Anopheles oswaldoi (Diptera:
Culicidae) in a malarious area, northeastern Venezuela.
J. Med. Entomol. 37: 231–238.
Hassan, A., V. Narayanan, and M. Salmah. 1993.
Observations on the physico-chemical factors of
the breeding habitats of Culex quinquefasciatus Say,
1823 (Diptera: Culicidae) in towns of north western
Peninsular Malaysia. Ann. Med. Entomol. 2: 1-5.
Hopkins, G. 1952. Mosquitoes of the Ethiopian Region: Larval
Bionomics of Mosquitoes and Taxonomy of Culicine
Larvae. Adlard and Son Ltd, London.
Kramer, V. and R. Garcia. 1989. An analysis of factors
affecting mosquito abundance in California wild rice

Page 8

Vol.?33,?no.?1?Journal of Vector Ecology 63
fields. Bull. Soc. Vector Ecol. 14: 87-92.
Lamp, W.O. and N.W. Britt. 1981. Resource partitioning
by two species of stream mayflies (Ephemeroptera:
Heptageniidae). Great Lakes Entomol. 14: 151-157.
Merritt, R. and K. Cummins. 1996. An Introduction to the
Aquatic Insects of North America, 3rd ed. Kendall/Hunt
Publishing Company, Iowa.
Minakawa, N., C. Mutero, J. Githure, J. Beier, and Y. Guiyun.
1999. Spatial distribution and habitat characterization
of anopheline mosquito larvae in western Kenya. Am.
J. Trop. Med. Hyg. 61: 1010-1016.
Mogi, M. 1978. Population studies on mosquitoes in the
rice field are of Nagasaki, Japan, especially on Culex
tritaeniorhynchus. Trop. Med. 20: 173-263.
Munga, S., M. Noboru, Z. Guofa, J.B. Okeyo-Owuor,
K.G. Andrew, and Y. Guiyun. 2005. Oviposition site
preference and egg hatchability of Anopheles gambiae:
effects of land cover types. J. Med. Entomol. 42: 993-
997.
Muturi, J., J. Shililu, B. Jacob, J. Githure, W. Gu, and R.
Novak. 2006. Mosquito species diversity and abundance
in relation to land use in a riceland agroecosystem in
Mwea, Kenya. J. Vector Ecol. 31: 129-137.
Muturi, E.J., J.M. Mwangangi, J. Shililu, S. Muriu, B. Jacob,
E.W. Kabiru, W. Gu, C. Mbogo, J. Githure, and R.
Novak. 2007a. Mosquito species succession and the
physicochemical factors affecting their abundance in
rice fields in Mwea, Kenya. J. Med. Entomol. 44: 336-
344.
Muturi, E., J. Shililu, W. Gu, B. Jacob, J. Githure, and R.
Novak. 2007b. Larval habitat dynamics and diversity
of Culex mosquitoes in rice agro-ecosystem in Mwea,
Kenya. Am. J. Trop. Med. Hyg. 76: 95-102.
Mwandawiro, C., Y. Fujimaki, Y. Mitsui, and M. Katsivo.
1997. Mosquito vectors of bancroftian filariasis in
Kwale district Kenya. East Afri. Med. J. 74: 288-293.
Piyaratnea, M.K., F.P. Amerasinghe, P.H. Amerasighe, and
F. Konradsen. 2005. Physico-chemical characteristics
of Anopheles culicifacies and Anopheles varuna breeding
water in a dry zone stream in Sri Lanka. J. Vector Borne
Dis. 42: 61–67.
Rajendran, R. and R. Reuben. 1991. Evaluation of the
water fern Azolla microhilla for mosquito population
management in the rice-land agroecosystem of south
India. Med. Vet. Entomol. 5: 299-310.
Ramachandra-Rao, T. 1984. The Anophelines of India Revised
edition. ICMR, New Delhi.
Senior-White, R. 1926. Physical factors in mosquito ecology.
Bull. Entomol. Res. 16: 187-248.
Shililu, J., G. Tewolde, S. Fessahaye, S. Mengistu, H. Fekadu,
Z. Mehari, G. Asmelash, D. Sintasath, G. Bretas, C.
Mbogo, J. Githure, E. Brantly, R. Novak, and J. Beier.
2003. Larval habitat diversity and ecology of anopheline
larvae in Eritrea. J. Med. Entomol. 40: 921-929.
Sunish, I., R. Reuben, and R. Rajendran. 2006. Natural
survivorship of immature stages of Culex vishnui
(Diptera: Culicidae) complex, vectors of Japanese
Encephalitis Virus, in rice fields in Southern India. J.
Med. Entomol. 43: 185-191.
Walton, W., E. Schreiber, and M. Mulla. 1990. Distribution
of Culex tarsalis larvae in a freshwater marsh in Orange
County, California. J. Am. Mosq. Contr. Assoc. 6: 539-
543.