Ecology of Anopheline ( Diptera , Culicidae ) , malaria vectors around the

The ecology of anopheline species (Di p t e ra , Culicidae) was studied in the vicinity of the Se r ra da Mesa Re s e rvo i r, State of Go i á s , Bra z i l . Climatic factors and frequency of anopheline populations were analyze d . Bimonthly hum a n bait and Shannon trap captures were conducted for 36 consecutive months ( Ja n u a ry 1997 t h rough December 1999). A total of 5,205 a d u l t anophelines belonging to five species were coll e c t e d . Anopheles darlingi was the most frequently collected anopheline (61.4%), f o l l ow e d by An. albitarsis s. l . ( 3 5 . 4 % ) , An. tri a n n u l a t u s. ( 2 . 5 % ) , An. oswaldoi ( 0 . 4 % ) , and An. eva n s a e ( 0 . 2 % ) . The water level and vegetation along the banks of the re s e rvoir were crucial to the f requency of the various anopheline species. Climatic factors had a secondary influence. The re s e rvo i r ’s waterl e vel stability, i n c re a s e d f requency of An. darlingi, and the arrival of gold prospectors were responsible for the inc rease in malaria cases. Ma l a r i a ; Vectors Ecology ;E c o l o gy I n t ro d u c t i o n Human intrusion in natural ecosystems and the resulting environmental changes modify established niches and create conditions for new ecological configura t i o n s. Such changes p rovide opportunities for the establishment of mosquito vector populations and tra n s m i s s i o n of the re s p e c t i ve ve c t o rb o rne diseases. A prime example of radical human intervention in the environment is the constru c t i o n of hyd ro e l e c t ric power plants, with major ecological impacts on vector population dynami c s. The risk of epidemics has been shown to i n c rease as the result of deforestation, cons t ruction of hyd ro e l e c t ric dams, and the introduction of human population groups where insect ve c t o rb o rne diseases are endemic. In the present study, we observed the p ro c e s s of human intrusion → e n v i ronmental change → anopheline population changes → and epid e m i c s, over time. To better understand the ecology of mosquito populations, we eva l u a t e d changes in their population dynamics as a result of the imposed environmental impact. Sy stematic surveillance was conducted in five t owns under the direct influence of the Se r ra da Mesa Re s e rvoir in the State of Goiás by m o ni t o ring anopheline populations from the initial dam closure and re s e rvoir flooding until the establishment of the plant’s full activities (Ja n ua ry 1997 to December 1999). Guimarães AE et al. 2 9 2 Cad. Saúde Pública, Rio de Janeiro, 20(1):291-302, jan-fev, 2004 Material and methods The Se r ra da Mesa Re s e rvoir is located in the Upper Tocantins River Basin, nort h e rn Go i á s State (13o 50’ S, 48o 18’ W ). The dam, 1,500 m long, 154 m high, and 460 m above sea level, is located on the main channel of the To c a n t i n s R i ver (Municipality of Minaçu) 1,790km fro m its mouth (Fi g u re 1). The Se r ra da Mesa Re s e rvoir is the largest in Brazil in terms of water volume flow. Du ring maximum flow the re s e rvoir contains 54.4 billion m3, measures 1,784 k m2, and extends to the limits of five main t owns: Ba r ro Alto, Ca m p i n o rt e, Minaçu, Ni q u elândia, and Uru a ç u . The re s e rvoir is located in the c e r ra d o ( s avannah) ecosystem, which in all occupies some two million km2 of Bra z i l’s terri t o ry. The Bra z i lian c e r ra d o is occupied by an extensive p l a t e a u , b e t ween 300 and 600m high, extending into the States of Amazo n a s, Rondônia, Pará, Ma ran h ã o, Piauí, To c a n t i n s, Go i á s, Mato Gro s s o, Mato Gro sso do Sul, Bahia, Minas Ge ra i s, and São Pa u l o and the Fe d e ral Di s t ri c t . This region features abundant metamorphic rock from the lower Pro t e rozoic peri o d and major natural deposits of asbestos in the municipality of Minaçu and nickel in the municipalities of Niquelândia and Barro Alto. Sm a l l g ranite outcroppings in the shape of re s i d u a l peaks are occasionally observed. The Se r ra da Mesa Dam was built on the Goiano plateau consisting of a dome called the “Granito Se r ra da Me s a” and circled by deformed metamorphic rocks (e.g., para g n e i s s, quart z i t e, and micaschist). The re g i o n’s vegetation, influenced by the Se r ra da Mesa Re s e rvo i r, consists of sparse and re l a t i vely short trees that ra rely reach a height of 10m, numerous bushes, and a lower layer of g rassy plants. The trees often have large and rough leave s, with twisted thick-barked tru n k s and bra n c h e s. These features are associated with high levels of aluminum in the soil ra t h e r than the lack of water. The twisting of tru n k s and branches is related to the frequent fore s t f i res that result in the loss of the main shoots, stimulating the sprouting of surrounding buds, thus forming irregular and twisted bra n c h e s. Even in the driest period these woody plants g u a rantee access to water through their long roots that can reach 20m below the surf a c e. Ga l l e ry forests extend non-continuously along the ri ver banks and in wet terra i n s, forming irregular patches of forest. The study are a’s climate is hot and tropical sub-humid. Ave ra g e annual rainfall ranges from 1,500 to 2,000mm, with more than 70% of precipitation occurri n g f rom October through Ap ril (Fi g u re 2). In the p e riod of heavy rainfall the re l a t i ve humidity f requently exceeds 90% and tempera t u re highs a re often above 40oC. May through Se p t e m b e r is an extended intensely dry period with re l at i ve humidity levels often below 15% and infrequent light rains (e.g., 20 mm per month in 4 or 5 days of rain and dry periods that may last for as long as 50 consecutive days). The dry peri o d is also the coolest, but tempera t u res ra re l y d rop below 10oC. Te m p e ra t u res above 25oC are not uncommon, with the mean monthly temp e ra t u res during this period above 18oC. T h i s climatic profile is chara c t e ristic in regions that a re distant from the coast and subject to stro n g continental effects. Mosquito surveillance was conducted at f i ve sites (Fi g u re 1) in the area surrounding the Se rra da Mesa Re s e rvoir for three consecutive ye a r s, f rom Ja n u a ry 1997 to December 1999. Each site has unique environmental chara c t e ri s t i c s. • Site 1: Municipality of Minaçu: the sampling point is located in the construction are a of the Se r ra da Mesa Dam, between 13o 4 9 ’ 2 6 . 9 ’’ S and 48o 18’ 56.5’’ W, next to the eastern bank of the Tocantins Rive r, near the secondary re s e rvoir (“Re s e rvoir II”) and 5km north of the p ri m a ry re s e rvoir (Fi g u re 1). This area is chara c t e ri zed by dense woods or c e r ra d ã o ( l i t e ra l l y “big sava n n a h”), with 6 meter-high tre e s, the c rowns of which limit the penetration of light to the lower underg rowth laye r s, resulting in a thick layer of organic matter. The gallery fore s t s reach the edges of the re s e rvo i r, forming shadowy creeks in the many tri b u t a ri e s. • Site 2: Municipality of Ca m p i n o rte: the s a mpling point is located at 14o 05’ 38.5’’ S and 48o 59’ 51.3’’ W, by the we s t e rn bank of the To c a ntins Rive r, in a farming pro p e rty known as Planeta Água, 18 km from the secondary access road to the town of Je rusalém, next to the BR153 highway (Fi g u re 1). Collections we re conducted along the banks of a creek that dra i n s into the re s e rvo i r. The site is chara c t e ri zed by f e rtile soil and typical gallery forest ve g e t a t i o n , f o rming dense, mainly arboreal woods, with bushes in the lower underg rowth. Hu m i d i t y and water from the drainage system sustain plant growth throughout the ye a r, providing a habitat that does not suffer the intense climatic adversities of the dry peri o d . • Site 3: Municipality of Uruaçu: the sampling point is located at 14o 32’ 04.4’’ S and 49o 01’ 07.8’’ W, on the eastern bank and under the b ridge over the Tocantins Rive r, 7km from the access cro s s roads to Uruaçu, from the GO-237 highway (Fi g u re 1). T h e re is an intense human migratory flux in this region, and there are m a n y


A b s t r a c t
The ecology of anopheline species (Di p t e ra , Culicidae) was studied in the vicinity of the Se r ra da Mesa Re s e rvo i r, State of Go i á s , Bra z i l .Climatic factors and frequency of anopheline populations were analyze d .Bimonthly hum a nbait and Shannon trap captures were conducted for 36 consecutive months ( Ja n u a ry 1997 t h rough December 1999).A total of 5,205 a d u l t anophelines belonging to five species were coll e c t e d .Anopheles darlingi was the most frequently collected anopheline (61.4%), f o l l ow e d by An. albitarsis s. l .( 3 5 .4 % ) , An. tri a n n u l a t u s.
( 2 .5 % ) , An. oswaldoi ( 0 .4 % ) , and An.eva n s a e ( 0 . 2 % ) .The water level and vegetation along the banks of the re s e rvoir were crucial to the f requency of the various anopheline species.Climatic factors had a secondary influence.The re s e rvo i r 's water-l e vel stability, i n c re a s e d f requency of An. darlingi, and the arrival of gold prospectors were responsible for the inc rease in malaria cases.

I n t ro d u c t i o n
Human intrusion in natural ecosystems and the resulting environmental changes modify established niches and create conditions for new ecological configura t i o n s.Such changes p rovide opportunities for the establishment of mosquito vector populations and tra n s m i s s i o n of the re s p e c t i ve ve c t o r-b o rne diseases.
A prime example of radical human intervention in the environment is the constru c t i o n of hyd ro e l e c t ric power plants, with major ecological impacts on vector population dynami c s.The risk of epidemics has been shown to i n c rease as the result of deforestation, cons t ruction of hyd ro e l e c t ric dams, and the introduction of human population groups where insect ve c t o r-b o rne diseases are endemic.
In the present study, we observed the p ro c e s s of human intrusion → e n v i ronmental change → anopheline population changes → and epid e m i c s, over time.To better understand the ecology of mosquito populations, we eva l u a t e d changes in their population dynamics as a result of the imposed environmental impact.Sy stematic surveillance was conducted in five t owns under the direct influence of the Se r ra da Mesa Re s e rvoir in the State of Goiás by m o ni t o ring anopheline populations from the initial dam closure and re s e rvoir flooding until the establishment of the plant's full activities (Ja n ua ry 1997 to December 1999).

Material and methods
The Se r ra da Mesa Re s e rvoir is located in the Upper Tocantins River Basin, nort h e rn Go i á s State (13 o 50' S, 48 o 18' W ). The dam, 1,500 m long, 154 m high, and 460 m above sea level, is located on the main channel of the To c a n t i n s R i ver (Municipality of Minaçu) 1,790km fro m its mouth (Fi g u re 1).The Se r ra da Mesa Re s e rvoir is the largest in Brazil in terms of water volume flow.Du ring maximum flow the re s e rvoir contains 54.4 billion m 3 , measures 1,784 k m 2 , and extends to the limits of five main t owns: Ba r ro Alto, Ca m p i n o rt e, Minaçu, Ni q u elândia, and Uru a ç u .
The re s e rvoir is located in the c e r ra d o ( s avannah) ecosystem, which in all occupies some two million km 2 of Bra z i l's terri t o ry.The Bra z i lian c e r ra d o is occupied by an extensive p l a t e a u , b e t ween 300 and 600m high, extending into the States of Amazo n a s, Rondônia, Pará, Ma ran h ã o, Piauí, To c a n t i n s, Go i á s, Mato Gro s s o, Mato Gro sso do Sul, Bahia, Minas Ge ra i s, and São Pa u l o and the Fe d e ral Di s t ri c t .
This region features abundant metamorphic rock from the lower Pro t e rozoic peri o d and major natural deposits of asbestos in the municipality of Minaçu and nickel in the municipalities of Niquelândia and Barro Alto.Sm a l l g ranite outcroppings in the shape of re s i d u a l peaks are occasionally observed.The Se r ra da Mesa Dam was built on the Goiano plateau consisting of a dome called the "Granito Se r ra da Me s a" and circled by deformed metamorphic rocks (e.g., para g n e i s s, quart z i t e, and micaschist).
The re g i o n's vegetation, influenced by the Se r ra da Mesa Re s e rvo i r, consists of sparse and re l a t i vely short trees that ra rely reach a height of 10m, numerous bushes, and a lower layer of g rassy plants.The trees often have large and rough leave s, with twisted thick-barked tru n k s and bra n c h e s.These features are associated with high levels of aluminum in the soil ra t h e r than the lack of water.The twisting of tru n k s and branches is related to the frequent fore s t f i res that result in the loss of the main shoots, stimulating the sprouting of surrounding buds, thus forming irregular and twisted bra n c h e s.Even in the driest period these woody plants g u a rantee access to water through their long roots that can reach 20m below the surf a c e.
Ga l l e ry forests extend non-continuously along the ri ver banks and in wet terra i n s, forming irregular patches of forest.The study are a's climate is hot and tropical sub-humid.Ave ra g e annual rainfall ranges from 1,500 to 2,000mm, with more than 70% of precipitation occurri n g f rom October through Ap ril (Fi g u re 2).In the p e riod of heavy rainfall the re l a t i ve humidity f requently exceeds 90% and tempera t u re highs a re often above 40 o C. May through Se p t e m b e r is an extended intensely dry period with re l at i ve humidity levels often below 15% and infrequent light rains (e.g., 20 mm per month in 4 or 5 days of rain and dry periods that may last for as long as 50 consecutive days).The dry peri o d is also the coolest, but tempera t u res ra re l y d rop below 10 o C. Te m p e ra t u res above 25 o C are not uncommon, with the mean monthly temp e ra t u res during this period above 18 o C. T h i s climatic profile is chara c t e ristic in regions that a re distant from the coast and subject to stro n g continental effects.
Mosquito surveillance was conducted at f i ve sites (Fi g u re 1) in the area surrounding the Se rra da Mesa Re s e rvoir for three consecutive ye a r s, f rom Ja n u a ry 1997 to December 1999.Each site has unique environmental chara c t e ri s t i c s.
• Site 1: Municipality of Minaçu: the sampling point is located in the construction are a of the Se r ra da Mesa Dam, between 13 o 4 9 ' 2 6 .9 '' S and 48 o 18' 56.5'' W, next to the eastern bank of the Tocantins Rive r, near the secondary re s e rvoir ("Re s e rvoir II") and 5km north of the p ri m a ry re s e rvoir (Fi g u re 1).This area is chara c t e ri zed by dense woods or c e r ra d ã o ( l i t e ra l l y "big sava n n a h"), with 6 meter-high tre e s, the c rowns of which limit the penetration of light to the lower underg rowth laye r s, resulting in a thick layer of organic matter.The gallery fore s t s reach the edges of the re s e rvo i r, forming shadowy creeks in the many tri b u t a ri e s.
• Site 2: Municipality of Ca m p i n o rte: the s a mpling point is located at 14 o 05' 38.5'' S and 48 o 59' 51.3'' W, by the we s t e rn bank of the To c a ntins Rive r, in a farming pro p e rty known as Planeta Água, 18 km from the secondary access road to the town of Je rusalém, next to the BR-153 highway (Fi g u re 1).Collections we re conducted along the banks of a creek that dra i n s into the re s e rvo i r.The site is chara c t e ri zed by f e rtile soil and typical gallery forest ve g e t a t i o n , f o rming dense, mainly arboreal woods, with bushes in the lower underg rowth.Hu m i d i t y and water from the drainage system sustain plant growth throughout the ye a r, providing a habitat that does not suffer the intense climatic adversities of the dry peri o d .• Site 3: Municipality of Uruaçu: the sampling point is located at 14 o 32' 04.4'' S and 49 o 01' 07.8'' W, on the eastern bank and under the b ridge over the Tocantins Rive r, 7km from the access cro s s roads to Uruaçu, from the GO-237 highway (Fi g u re 1).T h e re is an intense human migratory flux in this region, and there are m a n y re c reational activities associated with the re s e rvo i r.The wooded areas consist mostly of s h ru b s with few trees taller than 3m.The lower laye r consists of small bushes and leaf-cove red soil.The vegetation does not extend to the edges of the re s e rvo i r, where there are clear signs of def o restation.In the dry season, the few shadow y c reeks are exposed to the sun and the banks become ari d .
• Site 4: Municipality of Ba r ro Alto: the sampling point is located at 14 o 39' 47.9'' S and 48 o 57' 35.2''W, on the we s t e rn bank of the To c a ntins ri ver and in mining areas and is 27 km f rom the secondary access road in the distri c t of Placa, via the BR-080 highway (Fi g u re 1).This area has been heavily altered by extensive mining activity.Human settlements are conc e n t rated in crowded shacks.T h e re is heavy t raffic from tra n s p o rtation of people and mining equipment and materi a l s, resulting in erosion of the road and surrounding area.The ve getation is chara c t e ri zed as c e r ra d ã o, dominated by shrubs with occasional tre e s.The soil l a c k s o rganic matter, with frequent sandy patches exposed to intense sunlight.The sluggish waters of the Ma ranhão River form areas of sparse vegetation along the shady banks.
• Site 5: Municipality of Niquelândia: the sampling point is located at 14 o 26' 51.0'' S and 4 8 o 57' 34.6'' W, on the eastern bank, in areas of the Córrego Dantas Fa rm, 9km from the seco n d a ry access road beginning at the Pa rada da Vendinha village, via the GO-237 highway (Fi gu re 1).This region is chara c t e ristic of the Bra z i lian c e r ra d ã o.Located between tow n s, the are a consists of extensive plantations and cattle f a rms and small settlements where the local residents live off basic subsistence activities.The local population mostly lives in simple houses near the Re s e rvo i r, with eve ryday activities related to fishing.The environment has been extensively modified by deforestation and f a rming and livestock activities.Near the Re s e rvoir there are dense stretches of shrubs with soil cove red by grassy plants.
Mosquito captures we re conducted bim o n t h l y, always beginning at sunset (18:00h to 21:00h).Previous studies and random sampling around the re s e rvoir showed that sunset and early evening hours we re the most favo rable periods for presence of anophelines.Fro m October to Fe b ru a ry, during daylight savings t i m e, captures we re conducted between 19:00 and 22:00h.
Mosquitoes we re simultaneously collected using human bait and Shannon light tra p s, with 100-candle gas lamps.Two members of the re s e a rch team collected anophelines that landed on the Shannon traps or on themselve s with a mouth aspira t o r.The Shannon trap was located 150m from the human-bait capture s.
Ca p t u red anophelines we re killed with c h l oro f o rm and placed in small boxes containing p a ra f o rm a l d e h yd e.The specimens we re tra n sp o rted to the labora t o ry and mounted on paper triangles and labeled with a number and re f e rence data.
All specimens we re added to the collection of the Entomology De p a rtment of the Os w a l d o Cruz In s t i t u t e, Fi o c ruz, under the name "Se r ra da Mesa Co l l e c t i o n".Anopheles albitarsis s p e c imens we re considered sensu lato.
The relationship between the flooding s t a g e of the re s e rvoir and presence of anophelines was determined by the Williams mean test ( X w ).The level of significance between the fre-

R e s u l t s
From Ja n u a ry 1997 to November 1999, 5,205 anophelines we re collected at the five study are a s.Fi ve species we re collected: Anopheles darl i n g i , An .al b i t a r s i s , An .tr i a n n u l a t u s , An .os w a ld o i, and An .e va n s a e (Table 1).
T h roughout the thre e -year study peri o d , An .d a rl i n g i was the most frequently capture d s p e c i e s, comprising > 50% of the specimens c a p t u red at each sampling site.Ove rall, An .d a rl i n g i re p resented 61.4% of all anophelines c a p t u red.Calculation of the mean fre q u e n c y c o n f i rmed the abundance of this anopheline t h roughout the area.An .al b i t a r s i s was the second most common species in all the sampling s i t e s, with 35.4% of all the anophelines collected.An.triannulatus, An.oswaldoi, and An. e va ns a e we re captured much less fre q u e n t l y.
En v i ronmental va riations during the study resulted in va riations in the anopheline fauna d u ring filling of the re s e rvo i r.From 1997 to 1999 there was a drop in anopheline fre q u e n c y f rom X w = 169.1 in 1997 to X w = 115.2 in 1998 (a d e c rease of 32%) and X w = 20.8 in 1999 (a dec rease of 82%), or a total reduction of 88% in the anopheline population (Fi g u re 3).
The gradual reduction was common to all f i ve species over time, that is, An .d a rl i n g i w a s always the most common, followed by An .alb i t a r s i s , An .tr i a n n u l a t u s , An .os w a l d o i, and An .e vansae ( Fi g u re 3).The pro p o rtions of the two most frequently captured species, An .d a rl i n g i and An .al b i t a r s i s, remained more or less constant.In 1997 they re p resented 56% and 39% of all the anophelines, re s p e c t i ve l y, with an ave rage frequency of X w = 95.4 and X w = 66.4.T h e y re p resented 60% and 38% of all the anophelines in 1998 and 64% and 30% in 1999.
The frequency curve of the anophelines showed two peaks during the filling of the re s e rvo i r.The first peak occurred from July 1997 to Ma rch 1998, with means va rying from X w = 64.0 to X w = 155.0.The second occurred in Ma rc h 1999 (X w = 61.4)(Fi g u re 3).This pattern was observed for four of the s p e c i e s.The only exception was An .e va n s a e, which was not captured during the second peak.Small annual va riations occurred in the first larger population peak for the less commonly collected species: An .tr i a n n u l a t u s , An .o s w a l d o i, and An .e va n s a e ( Fi g u re 3).
Va riations in anopheline frequency are bel i e ved to be directly related to interventions in the environment around the Se r ra da Me s a Re s e rvo i r.Among the most important eve n t s we re the closing of the bypass tunnels and filling of the re s e rvo i r.
In October 1996 the bypass tunnels we re closed, and in Ja n u a ry 1997 the re s e rvoir began to form.The re s e rvoir was filled continuously until 1997.From the closing of the bypass tunnels until the end of the rainy season, the water leve l i n c reased from 340m to 431m above sea leve l .
Va riation in precipitation between the months of maximum and minimum ra i n f a l l was consistent with the re g i o n's climatic prof i l e. Howe ve r, the dry season was extre m e l y long, and even with the beginning of rains in November 1997, there was no significant inc rease in the re s e rvo i r's water level.Rainfall w a s a p p roximately half that observed in the pre v ious ye a r.Te m p e ra t u re and re l a t i ve humidity remained at the expected levels (Fi g u re 2).
Until Ma rch 1998, rainfall was less than expected, and the region suffered another ext remely dry period from May to Se p t e m b e r 1998.Resumption of the rainy season, fro m October 1998 to May 1999, once again fell short of the expected level.Ave rage tempera t u re and re l a t i ve humidity did not reach the levels of p revious ye a r s.

D i s c u s s i o n
Obtaining entomological information by moni t o ring populations of disease vectors has p roven to be a vital measure of disease risk assessment.T h e re f o re, detailed studies of ve c t o r population bionomics are re l e vant for deve l o ping and implementing control pro g ra m s.
It has been shown that enviro n m e n t a l changes may alter the frequency of malari a 1 , 2 , 3 , 4 , 5 , 6 Such changes are particularly re l e va n t in regions altered by human intrusion, including dam construction, mining, and ru ral settle-ment projects that provide increased opport unities for anthropophilic vector populations and where the environment favors the latter, t h e re by increasing the epidemiological ri s k s.In the current study we observed anopheline population changes over time in areas under the influence of the Se r ra da Mesa Re s e rvo i r, from the beginning of the filling of the re s e rvoir until months after the initial hyd roe l e c t ric plant operation.The five sites we re selected to provide a re p re s e n t a t i ve sample of the region.Mo n i t o ring populations over time was essential to assess environmental effects.
An .d a rl i n g i is endophilic and anthrop o p h i l i c and is captured using human bait both inside and outside re s i d e n c e s.It is also the pri m a ry human malaria vector species in much of Brazil, and even transmits malaria when the s p e c i e s' density is low. 2 , 5 , 1 3 , 1 4 The second most frequent species in our s u rve y, An .albitarsis sensu lato, also plays an i m p o rtant role in malaria epidemiology in Brazil.That is, although it is normally pre s e n t as a secondary ve c t o r, it may acquire gre a t e r significance where transmission by An .d a rl i ng i a l ready occurs. 1 3 , 1 5 , 1 6 This species is an efficient colonizer of areas altered by human act i v i t y, and the increase in its population des e rves attention due to its vector efficiency. 1 2 , 1 7 , 1 8 , 1 9 , 2 0 Wilkerson et al. 2 1 distinguished four genetically differentiated species from An .a lbitarsis sensu lato using random amplified polymorphic DNA polymerase chain re a c t i o n To t a l 1 , 0 2 5 1 0 0 .0 9 .5 1 1 9 1 0 0 .0 2 . 1 1 , 3 0 8 1 0 0 .0 1 7 .1 7 6 0 1 0 0 .0 1 3 .9 1 , 9 9 3 1 0 0 .0 2 8 .8 5 , 2 0 5 1 .0 7 4 .6   (RAPD-PCR).Wilkerson et al. 2 2 h y p o t h e s i ze d the species as An .albitarsis sensu stricto, An .marajoara, An.deaneorum, and another s p e c i e s not described.An .m a ra j o a ra and An .de a n e orum we re responsible for a malaria outbreak in Costa Ma rq u e s, Rondônia St a t e, as was An .m ara j o a ra in Ig u a p e, São Paulo St a t e, and Ma ra j ó Island, Pará St a t e. 2 2 Conn et al. 2 3 o b s e rved a significant pro p o rtion of An .m a ra j o a ra i n f e c ted with malaria parasites using ELISA and inc riminated the species as a pri m a ry vector in sites near the city of Macapá, Amapá St a t e.Other authors have implicated species fro m the An .a l b i t a r s i s complex as the pri n c i p a l m a l a ria vector in certain locations, such as the Vale do Ribeira region in the State of São Pa u l o 2 0 , around the Itaipu Hyd ro e l e c t ric Dam in the state of Pa raná 6 , and in urban areas of Bo a Vista in the State of Ro raima 2 4 .Although they consider An .albitarsis s.l.zoophilic 2 5 , Loure n ç o -d e -Ol i ve i ra et al. 2 , Fo rattini et al. 2 0 , and Guimarães et al. 6 re p o rt the species' intense a n t h ropophilic behavior, biting humans inside and outside re s i d e n c e s, especially in the absence of large animals (like cattle) in the surrounding area.

An . t r i a n n u l a t u s , An . o s w a l d o i, and
An .e vansae we re captured much less fre q u e n t l y and are typically zoophilic and exo p h i l i c.An .triannulatus a n d An .o s w a l d o i h a ve been implicated in the transmission of human malari a 2 6 .Howe ve r, Klein 2 7 , 2 8 d e m o n s t rated that An .t r i a n n u l a t u s in Costa Ma rq u e s, Ro n d o n i a St a t e, was a highly inefficient vector for both P. falciparum a n d P. v i va x.Specimens of An .osw a l d o i we re even re p o rted to infrequently have s p o rozoites in the saliva ry glands, although they presented oocysts.Since naturally infected specimens have been found, some authors i n c riminated the species as a secondary ve c t o r in certain locations 1 7 , 1 8 , 2 5 , 2 9 .The low numbers of An .e va n s a e c a p t u red and lack of re p o rts c o nc e rning its vector activity make this species' invo l vement in malaria transmission ra re.
Anopheline population dynamics are directly related to the availability of larval habitats for immature development.Some authors o b s e rved an association with increased ra i n f a l l that provides greater opportunity for accumulation of small bodies of water 8 , 3 0 , 3 1 , 3 2 .Howe ve r, Guimarães et al. 6 noted that in areas under the influence of the re s e rvo i r, anopheline population dynamics are not directly influenced by rainfall, but by the level of the re s e rvoir which, while stable, offers a constant opportunity for female anophelines to lay their eggs.
In this study we observed the re l a t i o n s h i p b e t ween va riation in anopheline populations and the water level of the re s e rvo i r.Climatic factors had a secondary influence, genera t i n g va riations in the conditions around the re s e rvoir and along the banks of tri b u t a ri e s.T h i s p rocess invo l ved three different stages.
Despite optimum tempera t u re conditions, re l a t i ve humidity, and heavy rainfall from Ja nu a ry to Ap ril 1997 (Fi g u re 2), deforestation up to 10 meters above the expected level along the banks of the re s e rvoir and tri b u t a ries was a key factor for the absence of anophelines (Fi g u re 3).Erosion, muddy water, and direct exposure of breeding sites to sunlight we re unfavo ra b l e to anopheline species in the area, especially An .d a rl i n g i ( Fi g u re 4).
With the gradual filling of the re s e rvoir and re f o restation along the banks beginning in Ma y 1997, a second stage began, with less fluctuation in the water levels and an immediate inc rease in anophelines captured.
The third stage, beginning in July 1998, was c h a ra c t e ri zed by a continuous drop in the water level (Fi g u re 4) and a reduction in the occ u r rence of anophelines.In i t i a l l y, this could be explained by the beginning of the opera t i o n stage of the Se r ra da Mesa Hyd ro e l e c t ric Pl a n t d u ring the dry season.Howe ve r, in July 1998 what would become one of the longest d ro u g h t s in history in the area hit nearly the entire count ry for the next three ye a r s.From the beginning, hyd ro e l e c t ric plants in va rious other areas of the country we re threatened by the lack of water in their re s e rvo i r s, and the Se r ra da Mesa Plant was forced to work at higher capacity to compensate for the national power d e f i c i t .The rainy period that followed was short e r than expected, failed to produce more water than was released, and lasted until the end of our study.
The impact of the increased re s e rvoir leve l on the anopheline population was not the s a m e as observed during the drop in the water leve l .From the moment the water level reached the s u r rounding vegetation at a slowly incre a s i n g p a c e, conditions favo red the development of immature anopheline forms.On the other h a n d , when the water receded the banks we re soon exposed to sunlight and erosion and became less suitable for the development of immature f o rms (Fi g u re 4).The consequences of these conditions for the anopheline population we re i m m e d i a t e, and the data show a large decre a s e in adult specimens captured, concomitant w i t h reduced water levels in the re s e rvoir in Ju l y 1998 (Fi g u re 3).
In November 1998 there was an increase in rainfall, tempera t u re, and re l a t i ve humidity, typical of summer in the study area (Fi g u re 2).Ac c o rding to Fo rattini et al. 8 and Gu i m arães et al. 3 2 , the alternation between Cu l i c i d a e population peaks and sudden decreases are o p p o rtunistic reactions to favo rable enviro nmental situations; when the latter end, the anopheline populations re t u rn to their norm a l l e vel.This phenomenon was observed part i a l l y in the area around the Se r ra da Mesa Re s e rvo i r.
The present study focused pri m a rily on entomological factors and ecological aspects affecting anopheline populations in an enviro nment altered by human activity.Howe ve r, in the epidemiology of malaria, the effects of human ecology are essential, and the flow of humans in the Brazilian countryside has always been identified as a modulating factor in the d i s t ribution and frequency of the disease 7 , 9 , 2 5 , 3 3 .
Co n s t ruction of highways, hyd ro e l e c t ric p l a n t s, and farming and mining opera t i o n s, including the best planned ones, always generate a disorg a n i zed migration of people that makes sanitation and disease control ve ry difficult.Ac c o rding to Guimarães et al. 6 , construction of the Itaipu Hyd ro e l e c t ric Dam in the State of Pa ra n a resulted in an increase in anopheline populations and the re t u rn of migrant workers who had failed to settle in the No rt h e rn re g i o n , many of whom we re infected with malaria, resulting in a serious epidemic, with nearly 3,000 cases of malaria in areas where the disease had a l ready been eliminated some 30 years before.
Human migration and malaria are common in mining areas under the impact of the re s e rvo i r, concentrating many risk factors for tra n smission of the disease.These areas lack adequate infra s t ru c t u re, and contact between humans and the vector is favo red by pre c a ri o u s housing, proximity to larval habitats, and intense exposure, since miners tend to be insufficiently clothed while working during hours when the vector is highly active.T h e re is always a multiplicity of breeding sites in these regions due to the way in which the area is modified.Besides the workers directly invo l ved in mining operations themselve s, mining are a s a t t ract individuals invo l ved in prostitution and unofficial tra d e, with high mobility and social detachment that hinder epidemiologic contro l m e a s u re s.Many of these individuals are already infected with malaria when they arri ve, having originated from or passed through regions where the disease is endemic 1 , 3 , 6 , 9 , 1 4 , 3 3 .
The arri val of gold prospectors and all the risk components mentioned above became evident in the area around the Se r ra da Me s a Re s e rvoir beginning in July 1997.As discussed, d u ring this same period we observed an in-c rease in the anopheline population, especially An .d a rl i n g i, in all the sites studied.T h e re appears to be a direct relationship between this m i g ra t o ry flow, the advent of diagnosed malaria cases, and the increased frequency of An .d a rl i n g i (and perhaps also of An .a l b i t a r s i s) ( Fi g u re 3).The important relationship betwe e n these factors is indicated by the linear corre l ation indices obtained: An .d a rl i n g i = 0.606 and An .al b i t a r s i s = 0.601 (r 0 , 1 ( 2 ) , 9 = 0.521).

A c k n o w l e d g e m e n t s
This work was accomplished with the help of Fu rn a s Ce n t rais El é t ricas S.A.The authors wish to thank the En v i ronmental De p a rtment of Fu rnas Ce n t rais El é t ricas S.A., re p resented by Dr.No rma Pinto Vilela, for p roviding logistic support during the fieldwork.We also thank our colleagues Ari ovaldo Gomes Co e l h o, Chisto So a res da Si l va, and Dion Kenede Vi e i ra for their collaboration in the field activities.

C o n t r i b u t o r s
A. E. Guimarães collaborated in the field work, s p e c i e s identification, statistical analysis, and drafting of the a rticle C. Gentile participated in the field work and statistical analysis.J. Alencar and C. M. Lopes part i c ipated in the field activities and species identification.R. P. Melo participated in the field work .
Ma l á r i a ; Ecololgia de Ve t o re s ;E c o l o g i a R e f e re n c e s

F i g u re 1
Location of the Serra da Mesa Hydroelectric Dam in Brazil and the State of Goiás, featuring anopheline collection sites 1, 2, 3, 4, and 5Pública, Rio de Janeiro, 20(1):291-302, jan-fev, 2004 quency of anophelines and diagnosed malari a cases was established using the linear corre l ation coefficient.Data on malaria cases aro u n d the re s e rvoir we re obtained from the Goiás office of the Brazilian National Health Fo u n d ation (FUNASA).

F i g u re 2
Te m p e r a t u re, relative humidity, and rainfall measurements and monthly averages, verified during sampling a round the Serra da Mesa Reservoir, State of Goiás, Brazil, from January 1997 to December 1999.Pública, Rio de Janeiro, 20(1):291-302, jan-fev, 2004 7 , 8 , 9 , 1 0 , 1 1 ,1 2   Intense and extensive deforestation, part i a l o b s t ruction of ri vers and cre e k s, highway e a rt hw o rk without proper dra i n a g e, accumulation of water, and the formation of large manmade lakes are examples of environmental changes that can result in increased vector density, favo ring ve c t o r-b o rne endemics.

F i g u re 3
(continued)   Incidence of anopheline species and malaria cases according to calculation of Williams means (X w ) a round the Serra da Mesa Reservoir, State of Goiás, Brazil, from January 1997 to December 1999.( c o n t i n u e s )

F i g u re 3
(continued)   Incidence of anopheline species and malaria cases according to calculation of Williams means (X w ) a round the Serra da Mesa Reservoir, State of Goiás, Brazil, from January 1997 to December 1999.Pública, Rio de Janeiro, 20(1):291-302, jan-fev, 2004

F i g u re 4
Serra da Mesa Dam, State of Goiás, Brazil.

F
i g u re 4a Panoramic view of the Serra da Mesa Reservoir, State of Goiás, Brazil.

F
i g u re 4bFocus on decreased water level and direct exposure to sunlight on the banks of the re s e r v o i r.S M DAlthough the numbers did not reach the normal values for the same period in pre v i o u s ye a r s, they we re sufficient to reduce the speed at which the water level decreased.Besides stabilizing the re s e rvoir for seve ral months, the rains resulted in increased vegetation, a typical c e r ra d o b e h a v i o r, and increased the potentially suitable larval habitats with a concurrent inc rease in the number of anophelines capture d up to May 1999 (Fi g u re 3).In May 1999 the drought continued, resulting in decreased re s e rvoir leve l s.Few anophelines we re capture d d u ring this peri o d .

Table 1
Number (n), percentage (%), and Williams means (X w ) of anophelines captured by sampling site, Serra da Mesa Reservoir, State of Goiás, Brazil, January 1997 to December 1999.
Incidence of anopheline species and malaria cases according to calculation of Williams means (X w ) a round the Serra da Mesa Reservoir, State of Goiás, Brazil, from January 1997 to December 1999.