Insect abundance and colonization rate in Fontinalis neo-mexicana (Bryophyta) in an Idaho Batholith stream, U .S.A .' Abstract Insect colonization rate and abundance in the aquatic moss, Fontinalis neo-mexicana were examined in the South Fork Salmon River, Idaho, June 1978-August 1979 to determine the importance of moss as a habitat for insects . The insect communities in moss as well as in the underlying and adjacent mineral substrates were examined . Insects were sampled with a nylon organdy net in moss and with a Hess bottom sampler in mineral substrate .Insects colonized insect-free moss clumps to carrying capacity within one week . In a man-made stream channel where flow, depth and substrate were controlled, insect densities were 5 to 30 times greater in moss than the mineral substrate ; insect biomass was approximately two times greater. Insect densities in pebble substrates underlying moss and in adjacent mineral substrate were comparable . Moss cover did not appreciably alter insect densities in the underlying hyporheic zone composed of screened pebbles (1-2 cm diameter) . Insect ordinal and functional group composition was greater in moss than in the mineral substrate, however, species richness was similar . Chironomids were the most abundant insects in moss .

Acknowledgments -We thank three anonymous reviewers for their feedback to strengthen the comparative components of this manuscript. In addition, we thank Mark Richards and Maria Arendt (University of Wisconsin -Madison) for help in obtaining tissues from domestic chickens. We thank Henrik Andrén and staff at the Grimsö Wildlife Research Station for providing infrastructure and support during field collections. Portions of this study were also made possible in cooperation with the Swedish Infrastructure for Ecosystem Science (SITES).We also thank Geir Rune Rauset, Þorkell Lindberg Þórarinsson, Friðrik Jónasson, as well as Onni, the barking bird dog, for assistance with collecting birds in the field. This research was supported by the Fulbright Commission through a Fulbright-Hays Grant and National Science Foundation [DEB-1146194 and OIA-1826801 to JSF], and start-up funds from the University of Pittsburgh [to KDK]. This research fully complied with the current laws of the countries in which they were performed.

Measurements of fecal pellet size can provide important information about wild mammals, such as body size and demographic information. Previous studies have not rigorously tested whether diet can confound these measurements. Furthermore, it is unknown whether diet might alter fecal dimensions directly or through changes in animal physiology. Here, we studied three closely related rodent species that differ in natural feeding strategies. Individuals were fed diets that varied in protein and fiber content for 5 weeks. We then measured body size, fecal widths and lengths, and the radius of the large intestine. Diet composition significantly changed fecal widths in all species. High-fiber content significantly increased fecal widths and would cause overestimations of body size if applied to wild feces. Using path analysis, we found that fiber can increase fecal widths both directly and indirectly through increasing the large intestine radius. Protein affected each species differently, suggesting that protein effects vary by species feeding strategy and existing physiology. Overall, diet and large intestine morphology can alter fecal pellet measurements. Studies using fecal measurements therefore must consider these effects in their conclusions.

No abstract