Blood parasites in passerine birds from the Brazilian Atlantic Forest Hemoparasitos em passeriformes da Mata Atlântica Brasileira

Parasites may lead bird species to extinction, affect host temporal and spatial population dynamics, alter community structure and alter individuals’ social status. We evaluated blood parasite prevalence and intensity according to bird families and species, among 925 birds that were caught in 2000 and 2001, in the Atlantic Forest in the State of Minas Gerais, Brazil. We applied Giemsa staining to thin blood smears, to detect blood parasites. The birds (n = 15.8%) in 11  families, were infected by at least one parasite genus, especially Muscicapidae (28.3%) and Conopophagidae (25%). Among the 146 infected birds, Plasmodium was detected in all bird families and had the highest prevalence (54.8%). Trypanosoma, Haemoproteus and microfilaria had lower prevalence rates (23.3, 23.3 and 2.1%, respectively). Birds caught during the rainy season were more infected than birds caught during the dry season. The overall low prevalence of blood parasites in birds is similar to the patterns found elsewhere in the Neotropical region.


Introduction
Haematozoa, or blood parasites, may lead bird species to extinction ( VAN RIPER III et al., 1986), and alter sexual selection and the evolution of plumage color (HAMILTON;ZUK, 1982;PRUETT-JONES et al., 1990;KIRKPATRICK;RYAN, 1991).Infected individuals may be more susceptible to predators and less able to establish territories (ANDERSON; MAY, 1979).Parasite infections may determine avian death, but the cause of mortality is usually identified a posteriori (McCALLUM;DOBSON, 1995).Moreover, pathogens are usually considered to be regulators of host population size.
Several studies focusing on occurrence, prevalence or incidence of blood parasites in birds have been conducted (YOUNG et al., 1993;GILARDI et al., 1995).Within the Apicomplexa phylum, Haemoproteus, Leucocytozoon and Lankesterella are the most prevalent genera infecting birds (BENNETT; BORRERO, 1976; ATKINSON; VAN RIPER III, 1991).Plasmodium has received most attention, including in relation to its evolutionary ecology (PAUL et al., 2003), but its distribution has not been fully documented yet.Plasmodium, Haemoproteus and Leucocytozoon have been recorded in several parts of the world, with the exception of Antarctica, where the low temperatures do not allow the vectors to survive (FRIEND; FRANSON, 1999).As well as in the Apicomplexa, the blood parasite Trypanosoma has also been reported in a wide array of bird hosts (PEIRCE, 1989).However, there is still little evidence to show that this parasite genus is pathogenic in birds (BAKER, 1976).
Most studies on the prevalence of blood parasites in birds have been conducted in the Neartic region (GREINER et al., 1975;KIRKPATRICK;SUTHERS, 1988) and few have been performed in the Neotropical region (DURRANT et al., 2006;FECCHIO et al., 2007;FECCHIO et al., 2011).Some studies have found high prevalence of Haemoproteus and low prevalence of Plasmodium in the state of São Paulo, Brazil (BENNETT;SOUZA, 1980;WOODWORTH-LYNAS et al., 1989).However, very low prevalence values have been found in the Cerrado (savanna) region of central Brazil (FECCHIO et al., 2007(FECCHIO et al., , 2011)).The prevalence of blood parasites and the diversity of haematozoan haplotypes have been found to be higher in Guyana than in Uruguay (DURRANT et al., 2006).Although no relationship has been found between blood parasite diversity and latitude in Chile, a positive relationship between prevalence and latitude has been shown for Leucocytozoon, and negative relationships for Haemoproteus, Plasmodium and mixed infections (MERINO et al., 2008).Considering the poor knowledge about blood parasitism among wild birds in Brazil, this study had the aims of evaluating the prevalence of avian blood parasites in several areas of the Atlantic Forest, and estimating their occurrence and intensities in different seasons.1).Birds were caught in two forest fragments in each of these areas: one large fragment (>1000 ha) and one small fragment (up to 30 ha).The fragments were isolated from each other or from any other forested area by at least 1 km of open land, mostly used for crops and pasture.

Between
The birds were caught using up to 20 mist nets, which were left open from sunrise until 5:00 PM, with a mean total of 640 mist-net hours for each area.The birds were identified using field guides.The individuals that were caught were then banded with a metal ring from the Brazilian birding agency (CEMAVE/IBAMA), to ensure that we were not sampling the same individual twice.The birds were released after banding and after blood samples had been collected in order to prepare the thin blood smears.The species nomenclature used followed the prescriptions of the Brazilian Ornithological Records Committee (Comitê Brasileiro de Registros Ornitológicos) (CBRO, 2007), and the family nomenclature followed Comitê Brasileiro de Registros Ornitológicos (2006), so as to allow statistical analyses, because of the small samples sizes of several families.
In the field, we prepared one thin blood smear from each bird, using disposable lancets to make a small perforation in the left tarsus to collect a drop of blood.The smears were air dried and fixed with methanol.They were then stained with Giemsa solution in buffered water (pH 7.2-7.4),diluted at 1:10, in the Malaria Laboratory of the Federal University of Minas Gerais (UFMG).
We searched for blood parasites in the blood smears using an optical microscope (Olympus 3H).We inspected at least two hundred fields on each smear, and the number of parasitized cells or the number of parasites was counted at 1,000× magnification under oil immersion.
We estimated three blood parasite infection parameters: prevalence, mean intensity and relative intensity (MARGOLIS et al., 1982).We used Spearman rank correlations to evaluate the relationships between blood parasite prevalence rates and mean and relative intensities for species or families with prevalence values differing from zero.We used Kruskal-Wallis to test for differences in parasite intensities among months and between the dry season (June, July, August and September) and rainy season (October and November).This test was also used to compare the infection intensity between the dry and rainy seasons, considering each parasite separately.We used a chi-square test for pairs of independent samples to ascertain whether the prevalence rates for parasite genera were independent between the seasons.We considered statistical differences to be significant at the 0.05 level.All statistical analyses were run using the STATISTICA (version 5.1), MINITAB or EPIINFO (version 6.0) computer packages.

Results
We analyzed 925 birds belonging to 109 species and 11 families (Table 1).A total of 146 birds (15.8%) from 62 species and all 11 families (Figure 2) had at least one parasite specimen of Plasmodium, Trypanosoma, Haemoproteus or microfilaria.The total prevalence values for these four parasites were 9.2, 3.8, 3.2 and 0.03%, respectively (Table 1).Among the 146 infected individuals, we observed the following prevalence rates: Plasmodium (54.8%),Trypanosoma (23.3%),Haemoproteus (23.3%) and microfilaria (2.1%).For several birds, this was the first report showing them to be new hosts for some blood parasites.Of the 62 bird species parasitized, 31 were new host records for Plasmodium, eight for Haemoproteus and two each for Trypanosoma and microfilaria (Table 1).
The prevalence of parasitism was higher during the rainy months than during the dry months sampled.The rainy months had significantly (c 2 = 9.9; p = 0.002) higher numbers of infected birds (618 individuals analyzed, 21.2% prevalence) than the dry months (307 individuals analyzed, 13% prevalence).However, the prevalence varied greatly within the dry season, with the highest prevalence occurring in July (n = 78 birds sampled; 33.3% parasitized) (Figure 3).September was the month with most birds sampled (n = 302), but with the lowest parasite prevalence (6.9%).The prevalence of Plasmodium varied irregularly along the months (Figure 4), being highest in July (n = 78 birds sampled; 25.6% infected) and lowest in June (n = 38; 2.6%).However, no difference was detected between the dry months (n = 51; 8.4%) and rainy months (n = 34; 10.8%) in relation to prevalence (c 2 = 1.15; p = 0.28) or intensity (H = 14.90; p = 0.136).Trypanosoma had the highest prevalence values in July and August and the lowest in the other months.There was no significant difference (c 2 = 0.80; p = 0.370) between the dry months (n = 26; 4.3%) and the rainy months (n = 9; 2.8%).For Haemoproteus, there was no significant difference (H = 6.44; p = 0.27) in parasite intensity between the dry and the rainy months (Figure 4).Haemoproteus had low prevalence in all the months sampled, except in November (n = 100; 23%).Comparing the two seasons, dry (n = 4; 0.70%) and rainy (n = 26; 8.2%), there was significantly higher prevalence (c 2 = 35.63;p < 0.001) and higher parasite intensity (H = 38.16;p < 0.001) in the rainy season than in the dry season (Figure 4).

Discussion
Birds from most Neotropical regions seem to have low blood parasite prevalence rates, independent of the type of habitat or region (WHITE et al., 1978;FECCHIO et al., 2007FECCHIO et al., , 2011)).Our study showed that the overall blood parasite prevalence was 15.8% among 925 birds sampled, similar to other Neotropical areas (WHITE et al., 1978).A study on the prevalence of parasites in birds from several habitats, including open areas, in southeast Brazil, found similar prevalence values for different habitats (WOODWORTH-LYNAS et al., 1989).Similarly, the prevalence values for blood parasites in passerines from several habitats (capoeira, 'terra firme' forest and várzea) in the Brazilian Amazon region were low (10%) (LAINSON et al., 1970).It is difficult to identify nematode microfilariae at species level because of their high degree of morphological and morphometric similarities (McKEAND, 1998).Sehgal et al. (2005), for example, found eight forms of microfilaria in 12 African passerines based on morphological characteristics, but they were not able to assign them with certainty to any species.Similarly, Silveira et al. (2010) reported on two microfilarial forms from central Brazil, also with insufficient information to allow the identification of the specimens.However, one filarial species of Splendidofilariinae, Cardiofilaria pavlovskyi, has been reported in several birds, including Passeriformes (ANDERSON, 2000).

Nº of infected birds
Nº of infected birds according to parasite taxa Plasmodium was the most prevalent parasite genus, and was detected in 9.2% of the 925 individuals sampled.This pattern is the rule for most Neotropical regions (WHITE et al., 1978;FECCHIO et al., 2007).However, Haemoproteus was the most prevalent parasite, infecting 38.9% of the birds sampled in a region in southeastern Brazil, whereas in our study, this parasite was 10 times less prevalent than what they found (WOODWORTH-LYNAS et al., 1989).One explanation might be related to the fact that our study was restricted to forest birds of the order Passeriformes.Columbidae, for example, which is the most important host of Haemoproteus, was not considered in our study.Moreover, these two studies differed in relation to the time of the year when blood samples were taken in the Atlantic Forest region, and the authors did not mention the level of disturbance in their study sites, a factor that could also explain part of the differences in the samples that were collected.Nonetheless, the prevalence of both of these blood parasites varied considerably among the bird families and species that we studied.The reasons for these differences are unknown, but might involve the behavioral and ecological traits of the species.These bird-specific characteristics, such as type of nest constructed and reproductive behavior, might facilitate or prevent contact between birds and their vectors (MARINI et al., 1996;FECCHIO et al., 2011).Interspecies differences between blood parasites in terms of parasitemia might be influenced by the distribution of habitat-dependent vectors, at both micro and macro habitat scale (GARVIN;REMSEN, 1997;PIERSMA, 1997).Host specificity and host-parasite co-evolution are also factors that can affect parasite prevalence (WOODWORTH-LYNAS et al., 1989).The detection of Plasmodium in all 11 bird families suggests that the genus is not host-specific, as has also been found by Bennett et al. (1982).However, Haemoproteus is more host-specific and is restricted to some species of bird families (FALLIS et al., 1974;ATKINSON et al., 1986).Trypanosoma was found in a small percentage of infected birds in spite of having a wide variety of vectors (PEIRCE, 1989).Its detection is difficult because of its low parasitemia and high polymorphism levels, which might explain the low prevalence that we found.
Prevalence and parasite intensity were positively correlated at both family and species levels.This has also been found for bird ectoparasites (PRUETT-JONES et al., 1991;MARINI et al., 1996).Factors relating to gender, age structure, behavior, season of the year and body condition of the populations studied might be important in explaining the parasitism levels.For example, birds with high infection by Plasmodium and Haemoproteus might exhibit signs such as appetite loss, apathy, shortness of breath and weakness in one or both legs (FRIEND; FRANSON, 1999).On the other hand, some studies have shown that low levels of parasitemia in the blood or tissue of a bird stimulate some protection against infection (WILSON et al., 2002).Thus, some birds might not show signs of diseases, but keep an infection, thus allowing the parasite population to survive through the dry season when vector populations are low (FRIEND;FRANSON, 1999).Some blood parasites might change their loads or occur seasonally in the Atlantic Forest birds.In general, we found fewer parasites during the dry season than during the rainy season, even though the sample sizes were better during the rainy season.This pattern was due to the high prevalence of Haemoproteus during the rainiest months.Lower blood parasite prevalence during the dry season than during the rainy season was found in forest birds from Costa Rica and Jamaica (BENNETT et al., 1980;YOUNG et al., 1993).This pattern is in agreement with the North American model, in which it was proposed that blood parasite recurrence would be seen during the reproductive season (BEAUDOIN et al., 1971).The low prevalence during the dry months, a period when most birds are not reproducing, occurs mostly in tropical areas (BENNETT et al., 1991).During the reproductive season, nestlings still lack feathers and are more vulnerable to vectors and, thus, have high seasonal transmission.When we analyzed the results on a monthly basis, however, we detected similar peaks in both the seasons.This can be explained by the fact that even though all the data were gathered within the Atlantic Forest biome, they came from six different localities in the state of Minas Gerais.Data gathering in the same area throughout the year would enable better evaluation of this pattern.
Overall, this study revealed that Atlantic Forest birds might be infected by several blood parasites at levels similar to what has been found in other Neotropical regions.The findings of differences in prevalence rates among hosts and for different parasites, and of temporal variation in prevalence values, reveal that blood parasite-bird interactions are very complex and poorly understood.Long-term population and community-based studies are required at both local and regional scales.The high number of new host species described in this study reflects the scarcity of studies of bird parasites in Brazil.We found more new hosts in bird families with more species sampled.This constitutes evidence of the need to acquire more knowledge of blood parasite bird hosts and their ecology and behavior.

Figure 3 .
Figure 3. Percentage of individuals parasitized for each month sampled in the Atlantic Forest, state of Minas Gerais, during 2000 and 2001.

Table 1 .
Total number of birds examined, total number of infected birds and number of birds infected by each blood parasite in the Brazilian Atlantic Forest sites of the state of Minas Gerais between 2000 and 2001.