Ecology of Phlebotomine sandflies ( Diptera , Psychodidae ) in Brazilian Atlantic Forest Ecologia de flebotomíneos ( Diptera , Psychodidae ) na Floresta Atlântica do Brasil

The phlebotomine sandfly fauna of a primary forest reserve at Morretes (eastern Paraná State) was studied, using CDC-like light traps, one night per month, at canopy and ground level, between April 1995 and March 1996. A total of 3,106 insects were collected, identified as belonging to nine species. Lutzomyia ayrozai and Lu. geniculata were predominant, seven other species also being present. Monthly mean temperature, rainfall and the temperature of the collection night significantly influenced the numbers of Lu. ayrozai while the two first factors influenced the numbers of Lu. geniculata, besides the collected quantities of females of the two species. The influence of the factors on Lu. ayrozai numbers was more immediate than in those of Lu. geniculata. Numbers of both species and of the females of Lu. geniculata collected in different seasons, but not at the different heights, varied significantly. Differences between the behaviour of Lu. ayrozai in Morretes and in other regions could be attributed to environmental differences and/or to regional variations in the species, which could constitute species complexes. Hourly variations of collections were different in the species and seasons. Key-words: Phlebotominae. Fauna. Seasonal variation. Psychodidae. Height influence. Lutzomyia. Resumo A fauna de flebotomíneos de uma reserva de floresta primária em Morretes (leste do Estado do Paraná) foi estudada, utilizando armadilhas luminosas tipo CDC, uma noite por mês, no nível do solo e da copa, entre abril de 1995 e março de 1996. Um total de 3.106 insetos foi coletado. Lu. ayrozai e Lu. geniculata foram predominantes, além de outras sete espécies. A temperatura média mensal, a pluviosidade e a temperatura na noite de coleta influenciaram significativamente as quantidades de Lu. ayrozai e os dois primeiros fatores influenciaram Lu. geniculata, além das quantidades de fêmeas coletadas de ambas as espécies. A influência do aumento dos fatores nas quantidades de Lu. ayrozai foi mais imediata que nas de Lu. geniculata. As quantidades coletadas de ambas as espécies e das fêmeas de Lu. geniculata nas estações, mas não nas alturas, foram significativamente diferentes. As diferenças entre os comportamentos de Lu. ayrozai em Morretes e em outras regiões poderiam ser atribuídas a diferenças ambientais e/ou a variações regionais na espécie, que poderia constituir um complexo de espécies. As variações horárias nas coletas foram diferentes nas espécies e estações. Palavras-chaves: Phlebotominae. Fauna. Variação estacional. Psychodidae. Influência da altura. Lutzomyia. 1. Departamento de Microbiologia e Parasitologia, Centro de Ciências Biológicas da Universidade Federal de Santa Catarina, Campus Trindade, Florianópolis, SC. . 2. Departamento de Zoologia da Universidade Federal do Paraná, Curitiba (UFPR), PR. . 3. Departamento de Patologia Básica/UFPR, Curitiba, PR Address to: Dr. Carlos Brisola Marcondes. Depto de Microbiologia e Parasitologia/UFSC, Campus Trindade, 88040-900 Florianópolis, SC. Fax: 55 48 331 9672 e-mail: cbrisola@mbox1.ufsc.br. Recebido para publicação em 2/8/2000. The study of the seasonal variation in populations of phlebotomine sandflies is very important to understand their biology and possible epidemiological importance. Studies of this variation have been done7 11, and some of which have provided data on the risk of transmission of parasites to man26. The numbers of insects collected at several heights above ground level can help identify to define their hosts and the transmission cycle of parasites, as shown by Williams46 47. The study of hourly variations in the activity of phlebotomine sandflies can be useful in clarifying their biology and determining their relationship to possible reservoirs of parasites and man biting. This variation has been studied in several regions23 38 43 45 46. Information on seasonal and hourly variation of sandfly activity is useful for control of transmission of parasites to man. Human16 28 40 and canine29 infections by Leishmania Ross, 1903 have been detected in western Paraná. L. braziliensis Vianna, 1911 (L. for Leishmania, to distinguish from Lu. for Lutzomyia) was isolated from sentinel animals in the Ribeira River Valley, São Paulo18 21, and L. enrietti was isolated from rodents in this region32 and near Curitiba, Paraná44. Sand fly vectorial ecology was reviewed17 in the State of São Paulo, including the important endemic region in the Ribeira River Valley. Additionally, three authoctonous

The study of the seasonal variation in populations of phlebotomine sandflies is very important to understand their biology and possible epidemiological importance.Studies of this variation have been done 7 11 , and some of which have provided data on the risk of transmission of parasites to man 26 .
The numbers of insects collected at several heights above ground level can help identify to define their hosts and the transmission cycle of parasites, as shown by Williams 46 47 .The study of hourly variations in the activity of phlebotomine sandflies can be useful in clarifying their biology and determining their relationship to possible reservoirs of parasites and man biting.This variation has been studied in several regions 23 38 43 45 46 .Information on seasonal and hourly variation of sandfly activity is useful for control of transmission of parasites to man.
Human 16 28 40 and canine 29 infections by Leishmania Ross, 1903 have been detected in western Paraná.L. braziliensis Vianna, 1911 (L. for Leishmania, to distinguish from Lu. for Lutzomyia) was isolated from sentinel animals in the Ribeira River Valley, São Paulo 18 21 , and L. enrietti was isolated from rodents in this region 32 and near Curitiba, Paraná 44 .Sand fly vectorial ecology was reviewed 17 in the State of São Paulo, including the important endemic region in the Ribeira River Valley.Additionally, three authoctonous human cases of dermal leishmaniasis, one of them due to L. braziliensis Vianna, 1911, were observed at Piçarras, in eastern Santa Catarina State 42 , about 150 km south of Morretes.A preliminary study of this focus indicated a high prevalence of dermal leishmaniasis in the human population (M.Steindel-pers.commun.).The occurrence of leishmaniasis in Piçarras indicates the possibility of natural undetected transmission of Leishmania in the forested areas between the Ribeira Valley and Piçarras, which would justify studies on the phlebotomine fauna in this region.
The phlebotomine sandfly fauna of the western region of Paraná was previously studied 4

MATERIAL AND METHODS
Two CDC-like light traps 37 were hung in a primary forest situated within IAPAR (Agronomic Institute of Paraná) reserve in the municipality of Morretes (25 o 28'37" S 48 o 50' 04" W, 10 m a. s. l.).They were separated by a horizontal distance of 5 m and were installed at heights of 1.5 and 7 m above ground level, designated respectively as ground and canopy stations.Collections were made at 200 m from the border of the forest.The vegetation is that of a rain forest, in the transition between tropical and subtropical types 31 .Collections were made from April 1995-March 1996.Due to the observations of better efficiency of light traps at new moon 6 39 , relative respectively to sandflies and Culicidae, collections were done in this phase of the lunar cycle.
Monthly temperature, precipitation and relative humidity data were obtained from the Meteorological System of Paraná (SIMEPAR), and the daily precipitation from IAPAR, at Morretes.A thermometer situated between the traps measured the temperature on each collection night.
The numbers of sandflies in each sample were transformed to the modified geometric mean (Williams mean-M w ) 9 , to prevent excessive influence of some large samples.ANOVA with two variables was used to check the possible influence of sampling height and seasonalities.The correlation between the numbers collected during each month and the climatic conditions was tested.The significance of r was tested by the t test, at the 1% and 5% significance levels 24 .
Light traps were utilised from 18.00-06.00.Collection chambers of traps were replaced at two hour intervals, the change being completed within 2-3 min.Sandflies were prepared for identification by NC method 34 .

RESULTS
The numbers of sandflies collected in the canopy and at ground level in each month are shown in Table 1, together with climatic data.Results indicate that the numbers of Lu. ayrozai (Barretto & Coutinho, 1940)  collected appeared to show a correlation with monthly mean temperatures and rainfall, and to a lesser degree to the temperature on the night of the collection.The same relationship appeared also to occur in Lu. geniculata (Mangabeira, 1941), although there was a delay between their own rise and that observed in the collected quantities.
Statistical analysis of correlation between collections of Lu. ayrozai and Lu.geniculata and climatic conditions (Table 1) showed some significant results.Lu. ayrozai numbers were correlated to monthly temperature in the month of collection, rainfall and temperature in the night of collection, and female numbers only to the first and the last of the factors above.The total numbers and those of females of Lu. geniculata were correlated to monthly temperatures two and three months before collection and the last also to monthly rainfall three months before.These data confirm the aforementioned suspicion of the influence of temperature and rainfall in the preceding months in the numbers of Lu. geniculata collected.
Collections of both sexes and of females of Lu. ayrozai and Lu.geniculata from ground level and the canopy were correlated at a significance level of 0.1%.When insects from canopy and ground were analysed together, numbers of Lu. ayrozai were correlated to all the studied climatic conditions in the month; totals of Lu. geniculata were correlated to monthly rainfall three months before and numbers of females were correlated to monthly relative humidity and temperature of night collection three months before.
Table 2 shows the seasonal M w for all species, combining samples of April, May and June for Autumn, and so on for the four seasons.No statistically significant correlation was found between the numbers of sandflies of any species and the temperature in any of the twohour periods.The mean numbers collected in each period varied according to the season.Lu. ayrozai was collected during all the periods, but occurred most frequently around midnight.In contrast, Lu. geniculata was collected more frequently before midnight, and exclusively before this hour in winter and spring.
It was difficult to establish standards of variation for the less common species.Lu. fischeri was collected almost exclusively before midnight, generally after 20.00.However, it was collected in greater numbers between 18.00-20.00,in April.Lu. shannoni was collected only in April and May, most frequently after midnight.Lu. monticola and Lu.lanei showed similar patterns, and   Lu.pascalei showed a more uniform distribution with respect to its diurnal activity and seasonal abundance.

DISCUSSION
The sandfly fauna of the area studied includes species previously found in São Paulo State, more northern areas and western Paraná.This is the first report of Lu. ayrozai and Lu.geniculata in Paraná.
The females of Lu. ayrozai and Lu.geniculata were redescribed 15 .Lu. ayrozai and Lu.hirsuta were the commonest species collected in forested areas of Rio de Janeiro, on human bait 6 and Lu.barrettoi (Mangabeira,  1942) and these in light traps 6 .Lu. ayrozai was the commonest species collected by light traps in a forested area of southern São Paulo 21 and has been collected, together with Lu. geniculata, in five of the seven American dispersion centres defined by Martins & Morales-Farias 36 .
Lu. geniculata, considered to be either a synonym of Lu. guyanensis (Floch & Abonnenc, 1941) 21 22 or a valid species 50 , is also common in the eastern region of São Paulo 21 22 and several other parts of South and Central America 27 47 48 , but it has not yet been found in collections from the Municipality of Piçarras and the Island of Santa Catarina (CB Marcondes: unpublished data).Although Lu. geniculata was not found at Serra dos Órgãos (Teresópolis, ca.850 m a. s. l.) 2 3 6 7 and Itaguaí (ca.35 m a. s. l.) 5 , it was collected at Parati (ca. 5 m a. s. l.), in an area similar to Itaguaí, although closer to the sea 1 .These Psychodopygus species therefore seem to have a patchy distribution in SE Brazil, including the possible replacement of Lu. hirsuta by Lu. geniculata at higher latitudes and lower elevations.Although some specimens of Lu. geniculata were collected biting man near houses at Parati 1 , it seems to be a typical primary forest species that occurs only near the sea.Lu. ayrozai is common in secondary forest at Florianópolis (Santa Catarina State), but Lu. geniculata was not yet found in this state (CB Marcondes: unpublished results).This could indicate that the southern limit of distribution of the last species is somewhere between Morretes and Florianópolis or can be due to a poor adaptation to secondary forests.Lu. geniculata is an infrequent man-biter in Belize 47 and it has been taken on human bait at Curiche and some other localities in Colombia 49 .No flagellates were found in 19 dissected females of Lu. geniculata from Belize 48 .
Several specimens of Lu. pascalei were collected at Morretes and this species was also very common in light trap collections made in southern São Paulo 21 .Lu. barrettoi, a very similar species, was very common in light traps 7 at Serra dos Órgãos (Rio de Janeiro) and was not collected by this method 2 6 at Morretes and at São Paulo 21 .
The numbers of Lu. fischeri collected in the canopy and at the ground level did not significantly differ at Morretes, possibly because the distance between the two stations was too small.The numbers collected at Rio de Janeiro in the canopy appear to be greater than those collected at ground level, although these were not analysed statistically 2 .This species seems to be highly anthropophilic 5 14 .
Lu. neivai (Neiva, 1926) and Lu.intermedia (Lutz & Neiva, 1912), may be predominant in forested regions, respectively, in São Paulo 13 and Rio de Janeiro 41 States.Lu. intermedia s. l. was found in forested areas of the Ribeira River Valley 18 20 21 22 in southern São Paulo, a similar environment situated close to Morretes.No specimens of this species complex 33 were collected at Morretes.This could be attributed to environmental characteristics or to the restriction of the present study to one area in the forest.Lu. neivai has been found only in houses and in the border of forested areas at Florianópolis, and some in patches of secondary forest at Piçarras (CB Marcondes: unpublished results).
The sandfly fauna in forested areas at Pariqüera Açu 20 and Cananéia 21 , about 120km to the north-east of Morretes, was much more complex than that observed in Morretes.Possibly the study of other biotopes at Morretes will reveal the presence of other species.
During the present study, Lu. ayrozai was found at ground and canopy level, without significant differences in numbers between the two stations.Sandflies of this species are common man-biters in some forest environments 2 6 21 38 and can be collected within and outside the forest in southern São Paulo, 20 occasionally even near houses 21 .Insects of this species were collected exclusively at ground level at Rio de Janeiro 6 , almost exclusively at this level at Colombia 38 and principally at this level at Southern São Paulo State 21 .Due to its occurrence in both levels, it was considered to be acrodendrophilic 21 .Lu. geniculata was collected only at ground level at Belize 47 , unlike the results of the Morretes study.The rarity of human infection by L. naiffi Lainson &  Shaw, 1989 in the Amazon basin could be due to the low degree of anthropophily shown by Lu. ayrozai 25 , although many specimens of the species were collected on human bait in the Ribeira Valley 21 , constituting additional evidence of regional variation of the species.
The regional variations in the behaviour of Lu. ayrozai and Lu.geniculata could be due to unperceived environmental differences among the localities studied, but could also indicate that species complexes, rather than single species, are involved.The wide geographical distributions of these species, like those of Lu. shannoni (Dyar, 1929), Lu. whitmani (Antunes and Coutinho, 1939) and Lu.intermedia s. l., indicates the need for studies on their taxonomy and ecology.The possible importance of such conditions in the transmission of parasites should be studied by dissection for parasites and analysis of the parous rate of the insects.The highest months of biting activity for Lu.ayrozai near Manaus were different, according to the bait animals used 8 .This illustrates the influence of the collection method on the results and the need to use several methods to study the biology of sandflies.
Lu. ayrozai was mostly collected around midnight, the hourly variation in the biting activity being similar to that observed in forest in Rio de Janeiro 7 and São Paulo 21 .The diel rhythm of some species can vary between different sites, as observed for Lu.trapidoi (Fairchild &  Hertig, 1952) in two regions of Colombia 38 45 .
There are no precise data on man biting activity at Morretes, and the seasonal fluctuation in the female numbers could give indications of this activity.
The occurrence of Brumptomyia nitzulescui in the trap hung in the canopy is curious, since this species is probably associated with armadillos.Several specimens of this species also were collected under similar conditions in Florianópolis (CB Marcondes: unpublished data).

Table 1 -
Numbers of phlebotomine sandflies collected by CDC-like light traps at Morretes, Paraná, at canopy and ground levels.

Table 2 -
Mean quantities (M w ) of sandflies collected by CDC-like light traps at canopy and ground levels in the seasons at Morretes (PR).
a: significant for seasons and not for levels, at 5%. b: significant influence of season, but not of level, at 5%, on the quantity of females.