The aim of this study was to analyze the seasonal variations in habitat selection and abundance of Ciconiiformes species in four foraging habitats (rivers, channels, connected and disconnected lagoons) on the upper Paraná River floodplain, Brazil, and to conduct the surveys of wading birds in 2002 and 2003. The largest number of species and highest abundances of most species were observed in the connected lagoons. Snowy Egrets (Egretta thula), Wood Storks (Mycteria americana), Roseate Spoonbills (Platalea ajaja), and Jabirus (Jabiru mycteria) frequently used connected lagoons (habitat with the highest fish abundance) and abandoned the areas during floods, suggesting that they were able to find high quality patches for foraging. Cocoi Herons (Ardea cocoi), Great Egrets (Ardea alba), and Rufescent Tiger-Herons (Tigrisoma lineatum) used habitats with lower fish abundance and did not totally abandon these areas during floods, suggesting they were not dependent on high quality patches. Differences in foraging techniques and social behavior explained the difference between the two groups. The results showed that opportunism did not totally explain the foraging habitat selection and in low water level seasons there were higher differences in foraging behavior among the species.

Foi analisada a influência do tamanho das lagoas e da disponibilidade de presas sobre os Ciconiiformes na planície alagável do Alto Rio Paraná, Brasil. O censo dos Ciconiiformes foi realizado trimestralmente em 2002, concomitantemente às amostragens de peixes (redes de arrasto e redes de espera). Houve forte relação do tamanho das lagoas com a abundância média de aves no ano, mas modesta relação com a diversidade e o número de espécies. A densidade de peixes em redes de arrasto foi melhor prognosticadora dos Ciconiiformes do que as capturas por unidade de esforço em redes de espera. Exceto na cheia, houve forte relação da densidade de peixes com a abundância relativa de aves e mais modesta relação da densidade de peixes com a diversidade e número de espécies de aves. O tamanho das lagoas e a disponibilidade de presas foram fortes prognosticadores da abundância e abundância relativa de Ciconiiformes, respectivamente. Porém, variáveis ambientais aqui não analisadas podem ter influência significativa na determinação da diversidade e número de espécies de aves presentes.

The influence of lagoons size and prey availability on the Ciconiiformes in the upper Paraná river floodplain, Brazil, were analysed. The Ciconiiformes census was conducted quarterly in 2002, simultaneously to the fish census (gill nets and seining nets). There was strong relationship between the lagoons size and birds abundance, but modest relationship of lagoons size with the birds diversity and number of species. The presence of Ciconiiformes in the lagoons was better predicted by the fish density in seining nets than by the fish capture by unit of effort in gill nets. Except in the flood, there was strong relationship between the fishes density and birds relative abundance, and more modest relationship of the fishes density with birds diversity and number of species. The lagoon's size and prey availability were strong predictors of Ciconiiformes abundance and relative abundance, respectively. However, environmental variables not analysed probably could have significant influence in the determination of the birds diversity and number of species.

Spatial distribution of birds was studied in 10 ha forest fragment at Londrina State University, southem Brazil (23º15'S, 51º10'W). Five habitats were categorized in the forest fragment: four related to the edges and one related to the forest interior. A line transect, crossing ali these habitats, was used to identify the bird species one day each two weeks from September 1996 to September 1997 totaling 57 days of samplings. The forest was divided into three leveis: understory (bellow 3 m), mid-levels (3-7 m) and canopy (above 7 m). Occurrence frequencies in each habitat and each forest levei were calculated for each of the 62 species identified during the field work. The number of species in each levei was similar (43 species in the understory, 42 at mid-levels and 40 in the canopy). Edges had higher number of species (51, 46, 33 and 30 species) and most of them with more than 50% of occurrence frequency. By contrast, forest had the lowest number of bird species (29 species), 80% of them with less than 50% of occurrence frequency, what suggest that the edge effect is strong on this forest fragment.