Intra-population plasticity of Anopheles darlingi ’ s ( Diptera , Culicidae ) biting activity patterns in the state of Amapá , Brazil

Objective To assess the variation in Anopheles darlingi’s biting activity compared to An. marajoara in the same locality and to biting activity data from other regions. Methods Using human bait, eight observations of the biting activity of An. darlingi and An. marajoara were carried out during 1999 and 2000 in the municipality of São Raimundo do Pirativa, state of Amapá, Brazil. Each observation consisted of three consecutive 13-hour collections, close to full moon. There were shifts of collectors in the observation points and nocturnal periods. Results An. darlingi revealed considerable plasticity of biting activity in contrast to An. marajoara, which showed well-defined crepuscular biting peaks. No significant correlation between density and biting activity was found, but a significant correlation existed between time and proportional crepuscular activity, indicating underlying ecological processes not yet understood. Two of the four available data sets having multiple observations at one locality showed considerable plasticity of this species’ biting patterns as well. Conclusion Intra-population variation of biting activity can be as significant as inter-population variation. Some implications in malaria vector control and specific studies are also discussed.


INTRODUCTION
Knowing the biting activity patterns of malaria vectors is essential for accurately planning vector control actions.The biting behavior of Anopheles darlingi Root, 1926, main malaria vector in Amazon region, has been studied extensively throughout the last six decades.Great differences in biting activity patterns have been reported throughout the species' geographic extent, from very pronounced almost exclusively crepuscular to almost exclusively nocturnal activity. 4e variation of biting activity patterns contrasts to those reported for most other Anopheline mosquitoes in the region, that most often follows a crepuscular biting pattern. 8,13,14This fact, in addition to other biological variation, has originated the hypothesis that An. darlingi is a species complex.Species complexes are common within the Nyssorhynchus subgenus of Anopheles.Of currently recognized or possible species complexes, 11 An.albitarsis, An. triannulatus, An. oswaldoi, and An.nuneztovari, only An. nuneztovari is known for its marked differences in biting activity patterns. 13Genetic studies of An. darlingi have shown so far that this is a monotypic species. 11And there is no evident geographical distinction of biting activity patterns.
The present study examines the plasticity of An. darlingi's biting activity at one locality to assess whether the type of variation of biting behavior found at one site is present in the whole range.Another important Anopheles species, An. marajoara, collected at the same site and that shows a distinct crepuscular biting behavior, was used as a comparison to An. darlingi.

Study area
The biting activity observations were carried out at the locality of São Raimundo do Pirativa, municipality of Santana, state of Amapá, Brazil (0°2' N, 51°15' W).Amapá has an equatorial hot and humid climate with temperatures varying between 22-32 °C.The rainy season extends from January until July with a mean monthly rainfall of 2,100 mm.The dry season extends from August until December with a mean monthly rainfall of 178 mm.The average relative humidity is 85%.The study site lies along the Matapí River in a flooded forest area ("várzea") with tidal influence.It is a small riverine community of approximately 200 inhabitants.Their main economic activities are fishing, small-scale extraction of forest products, small-scale agriculture, and per diem work in neighboring buffalo farms.Malaria transmission is intense and year-round and both Plasmodium falciparum and P. vivax are notified as vectors by the local health services.Plasmodium malariae occurs as well, evidenced through population immunological examination (36% IgG antibody prevalence) and CS-rate studies of the Anopheles species (Branquinho and Voorham*).
A recent study looking for the An.albitarsis complex in an area close to the study site revealed only the presence of An. marajoara.Therefore, analysis of the available data will be carried out assuming a one-species setting, and this species name will be used throughout the text.This species has recently been shown to be a potentially important vector, (Conn 5 and Branquinho and Voorham*) in the state of Amapá.

Data collection and processing
Along seven consecutive months, from May to October 1999, and once in March 2000, three consecutive night collections were performed close to full moon.Local community adults were hired as mosquito collectors.These people have already been exposed to the circulating Plasmodium parasites, and were offered chemoprophylaxis according to the Ministry of Health guidelines.Mosquitoes landing on the bare legs of three human collectors outdoors were caught before biting and kept in separate collection cups for each observation hour.The observation night started 30 minutes before sunset and ended 30 minutes after sunrise.Since the study site is situated almost at 0° latitude, there were 13 hours of observation per night.There were two collection shifts: the first 6 hours of nighttime followed by the last 7 hours.Over the three nights of observation, there was a rotation of collectors regarding their shift periods and observation points.Immediately after the completion of an hour of observation, the species of collected mosquitoes were identified, and stored on silica gel for further infectivity studies, until 1 AM.Processing continued the next morning.
Data obtained from three consecutive nights of observation were lumped into a biting activity data set, thus correcting for eventual individual differences between collectors and collecting spots.Three-night observation is referred by a week number, the March 2000 observation is referred as a continuation of the 1999 week-count.
In May 1999 an age-composition study was done with the collected An. darlingi from the first 7 hours of the night, using parity observations following Detinova.

RESULTS
The results are summarized in Table 1 and show that An. darlingi (74.7%) and An.marajoara (22.7%) were the most abundant species collected.Anopheles darlingi showed a sharp increase in density by the end of the rainy season, while An.marajoara density was fairly constant throughout the study.
The observed biting activity patterns are shown in Figure 1 for An.darlingi and in Figure 2 for An.marajoara.A large variability in biting activity over the eight observations was observed for An.darlingi, contrasting to more stable patterns seen for An.marajoara.While the latter continuously showed a well-pronounced peak during the first three hours of the night, the former species varied considerably in its activity pattern.In addition, An. darlingi showed a more pronounced nocturnal activity than An.marajoara.
Two potential causes for the observed plasticity of An. darlingi were examined.First whether relative  Intra-population plasticity of An. darlingi.Voorham J density contributed to the variation.Results showed that there was no significant relationship between density and the proportional contribution of biting during the first three hours for An.darlingi (p=0.82) and An.marajoara (p=0.93).
Second whether there was a time of year effect on the collection.A significant positive correlation between week-number (i.e.time) and the proportional contribution of biting during the first three hours can be observed for An.darlingi (r=0.85,p=0.007) but not for An.marajoara (r=0.46,p=0.25).
The single parity observation was performed using 311 An. darlingi specimens divided over the first seven hours of the night, and showed no significant differences of parity-rate between the hours (χ 2 , p=0.07), and a mean value of 66%.

DISCUSSION
Overall biting activity patterns of An. darlingi in the study is similar to the results reported for the state of Amazonas, Brazil, 6 French Guiana, 9 Colombia,* Suriname, 7 and Venezuela.**Although these regions are located in the northern part of the continent, it does not imply a geographical determination of biting activity patterns.For instance, in the state of Rondônia, Brazil 8 , occasionally this species has shown very little crepuscular activity, and in several localities in the northern state of Pará, Brazil 14,15 -relatively close to the study area -, this species has shown pronounced crepuscular biting activity.
Few studies have been conducted on the biting activity patterns of An. marajoara, partly due to difficulties in identifying the species-complex's members.In Venezuela 13 and Trinidad, 1 An.albitarsis s.l.(probably An. marajoara) showed similar activity patterns to the ones found in this study with a very pronounced crepuscular peak.
One of the most important aspects of a vector's biting activity is how it relates to human activity.Throughout rural South America, where malaria is mostly found, human activity is mostly outdoors during crepuscular hours and indoors during nocturnal hours.Therefore, to better compare human and Anopheline activities, the biting patterns in Figures 1 and 2 were converted to relative contributions of three functional periods: 1) the first three hours of the night; 2) the last three hours of dawn; and 3) the nocturnal period in between.This arrangement also facilitates comparing the same species in other regions and other species, given that different methodologies have been used in other studies regarding time-scales, study design, and collection methods.Figure 3 shows the results for both species (A and B).From these functional diagrams it can be drawn that An. marajoara has a much higher contribution to biting than An.darlingi when the human individuals are outdoors (Period 1).
Although in two out of eight observations An. marajoara showed lower densities than An.darlingi, its higher contribution to extra-domiciliar biting clearly indicates this species' importance as a potential malaria vector, especially where vector control measures are mainly house-based.
Of 34 studies examined, four also described local variation of the biting activity of An. darlingi.These studies were conducted in the state of Amazonas 6 and Rondônia, 8 Brazil, Honduras, 10  Amazonas (C).Compared to Amazonas, the observed pattern in Amapá showed considerably more early morning biting activity, although early evening activity was similar.Early morning biting activity contribution in Rondônia was similar to Amapá, but the early evening crepuscular activity in Rondônia was much greater, but also more variable.Thus, behavioral plasticity is not always observed as seen in Venezuela and Honduras.However, other data obtained in the Amazon region clearly show that this is not a peculiarity of this study area.
The likelihood of mosquito density interfering with the biting behavior was discarded in the present study, as previously reported elsewhere in Brazil 2 and French Guiana. 9However, studies conducted in Colombia* suggested a correlation between biting activity and the population density of An. darlingi.A regional study in Brazil also alluded to this possibility. 15onlight could have influenced the biting activity patterns and densities 12 of previously mentioned observations with local variation, but this was not the case for the present study since all collections were carried out during the same moon phase.A higher variability can be expected when not correcting for this influence.
The age composition of biting An. darlingi was shown to be different during the crepuscular peaks in Brazil, 2 opposing results have been found in other studies 4,9 as well as in the present one.Age-composition differences between activity periods do not necessarily indicate subgrouping, and it is more likely to be a result of specific ecological settings than of age-specific activity patterns.However, additional studies on age structure of An. darlingi and An.marajoara at the study site are intended, using more precise methods.
The observed correlation between time and the contribution to crepuscular biting for An.darlingi still requires further clarification, and indicates the existence of underlying ecological processes not yet understood.
Population variations throughout An.darlingi's wide geographical range could explain the differences in biting activity patterns. 4However, this does not provide an explanation for intra-population variationσ, which can be as extent as inter-population variations.It is beyond this study's scope to further discuss possible causes for An.darlingi being so plastic in its biting activity patterns on regional and local levels.Evidence up to date strongly indicates that ecological determinants play a role instead of genetics or local population specific characteristics.Further studies on population dynamics and individual mosquito's behavior are needed on the subject.
Plasticity in biting activity patterns can result in increased vectorial potential and control strategies may have to be adjusted to account for differences in human vector contact over time.
Knowledge of vector biting activity plasticity is also essential for specific entomological studies.Sampling assumes knowledge of the population's activity pattern.When biting pattern observations are limited to only a few nights of observation, the plasticity of this behavior can not be evidenced and could lead to misinterpretation of data.To illustrate that, a comparison between relative density estimates of whole-night and  2-hour periods (hours 2 and 3 -the most commonly used sampling period) is shown in Table 2, as well as the proportion of total biting during the 2-hour observation period for both studied species throughout eight observations.Around 6-20 % of all bites are due to An. darlingi, while 25-41% are due to An. marajoara.This difference in sampling could lead to bias due to period-specific characteristics.
The relative density estimates from partial collections are closer to all-night density estimates for An.darlingi than for An.marajoara, which are constantly overestimated.The combination of different sampling of the two species populations and overestimation of the relative density of one species could have lead to an overestimation of the importance of An. marajoara as a vector in a recent study in Amapá, 5 as it is likely to have occurred in other studies on multiple-vector settings.
In conclusion, inter-population variation in biting activity may be as great as intra-population varia-tion.Variability in biting activity should be taken into account for intervention design and specific entomological studies, especially where there is vector heterogeneity.Comparison of one species populations in different areas, or two species populations at the same site should be done with sampling methods that minimize bias for density and agecomposition estimates.It is also recommended to minimize the use of partial-night observations in situations where biting activity varies greatly.

Figure 1 -
Figure 1 -Biting activity patterns of An. darlingi during the eight observations at São Raimundo do Pirativa, State of Amapá, Brazil, from May 1999 until March 2000.

Figure 2 -
Figure 2 -Biting activity patterns of An. marajoara during the eight observations at São Raimundo do Pirativa, State of Amapá, Brazil, from May 1999 until March 2000.

Table 1 -
Number of Anopheles specimen collected per three-day observation per species, at São Raimundo do Pirativa, State of Amapá, Brazil, from May 1999 until March 2000.Collections are coded according to the week number of the year 1999, with the March 2000 collection as a continuation of the 1999 numbering.

Table 2 -
Comparison of relative density and nocturnal period contribution between all-night observations and partial (hours 2-3) observations for An.darlingi and An.marajoara at São Raimundo do Pirativa, 1999-2000.Collections are coded according to the week number of the year 1999, with the March 2000 collection as a continuation of the 1999 numbering.