Tropical Forest Bird Counts and the Effect of Sound Attenuation

Simultaneous variable circular-plot censuses were conducted by two observers at five heights above ground in lower montane rain forest in Puerto Rico. One observer began the censuses at the top of a 22-m walk-up tower, and the other observer began at the bottom. Tests for interobserver variability indicated that the observers were not significantly different in their ability to detect birds. Sufficient numbers of 6 species (Columba squamosa, Spindalis zena, Vireo altiloquus, Coereba flaveola, Nesospingus speculiferus, and Todus mexicanus) were observed to compare differences in the efficiency of canopy and ground observers. More birds that sing from the canopy were detected by the canopy observer than by the ground observer. The reverse was true for birds that sing near the ground. Species were most often detected at the station closest to their mean singing height. Population densities of canopy-singing species were underestimated by 33-46% by the ground observer. Measurements of the attenuation of pure tones of different frequency indicate that low-frequency sound attenuates less rapidly than high-frequency sound. Species with low-frequency songs are heard equally well by ground and canopy observers.