Summary Vertebrate skeletons have high phosphorus (P) content relative to other tissues. Variation in skeletal investment within and among species is hypothesised to predict variation in P demand, standing stock and recycling. These relationships have been examined among fish, but not in vertebrates with more robust skeletons, such as turtles. Our objectives were to (i) describe freshwater turtle stoichiometry relative to skeletal mass, (ii) compare turtle body and excreta stoichiometry to patterns among fish and (iii) relate turtle skeletal stoichiometry to turtle nutrient storage and recycling. Skeleton constituted 82% of turtle dry mass. Total body %P increased ontogenetically with turtle mass and 93% of all P resided in the skeleton. Phosphorus storage within turtle assemblages was high (0.2–0.45 kg ha−1). Turtles excreted lower concentrations of P than fish. Excreta N:P was positively correlated with body N:P, suggesting that increased skeletal P investment was inversely related to P demand. Our results demonstrate that P stored in the bone of turtles can represent a large standing stock of P in fresh waters. Further, our work suggests skeletal investment alone is not sufficient to predict an animal's P demand and, by extension, their effects on nutrient recycling. Instead, our results indicate P demand is determined by both skeletal investment and growth rate. Consequently, taxa with high body P, extreme longevity and slow growth rates, such as adult turtles, may serve as stable standing stocks of nutrients while also contributing proportionately to nutrient remineralisation.

Our study focuses on the spatial ecology and seasonal habitat use of two aquatic turtles in order to understand the manner in which upland habitat use by humans shapes the aquatic activity, movement, and habitat selection of these species in an urban setting. We used radiotelemetry to follow 15 female Graptemys geographica (common map turtle) and each of ten male and female Trachemys scripta (red-eared slider) living in a man-made canal within a highly urbanized region of Indianapolis, IN, USA. During the active season (between May and September) of 2002, we located 33 of the 35 individuals a total of 934 times and determined the total range of activity, mean movement, and daily movement for each individuals. We also analyzed turtle locations relative to the upland habitat types (commercial, residential, river, road, woodlot, and open) surrounding the canal and determined that the turtles spent a disproportionate amount of time in woodland and commercial habitats and avoided the road-associated portions of the canal. We also located 21 of the turtles during hibernation (February 2003), and determined that an even greater proportion of individuals hibernated in woodland-bordered portions of the canal. Our results clearly indicate that turtle habitat selection is influenced by human activities; sound conservation and management of turtle populations in urban habitats will require the incorporation of spatial ecology and habitat use data.

Abstract. Animals can be important in modulating ecosystem-level nutrient cycling, although their importance varies greatly among species and ecosystems. Nutrient cycling rates of individual animals represent valuable data for testing the predictions of important frameworks such as the Metabolic Theory of Ecology (MTE) and ecological stoichiometry (ES). They also represent an important set of functional traits that may reflect both environmental and phylogenetic influences. Over the past two decades, studies of animal-mediated nutrient cycling have increased dramatically, especially in aquatic ecosystems. Here we present a global compilation of aquatic animal nutrient excretion rates. The dataset includes 10,534 observations from freshwater and marine animals of N and/or P excretion rates. These observations represent 491 species, including most aquatic phyla. Coverage varies greatly among phyla and other taxonomic levels. The dataset includes information on animal body size, ambient temperature, taxonomic affiliations, and animal body N:P. This data set was used to test predictions of MTE and ES, as described in Vanni and McIntyre (2016; Ecology

Agriculture is the dominant land use in the Lower Flint River Basin (LFRB) of south-western Georgia, USA, and is often a significant disturbance factor affecting streams and riparian habitats. Streams in the LFRB harbor a great diversity of freshwater turtles, which are among the many groups of aquatic fauna impacted by agricultural disturbance to riparian habitats. The objective of this study was to assess turtle diversity and abundance in both agriculturally impacted and unimpacted or restored reaches of streams in the LFRB. In 2007 and 2008, we used hoop traps and effort-managed snorkel surveys to sample turtles on 14 reaches of two streams (Ichawaynochaway and Spring Creeks). We recorded 823 captures of 674 individuals representing nine turtle species. There was a measurable association between the percentage of riparian undisturbed land cover and the number of turtles captured for the four most frequently captured species (Trachemys scripta, Graptemys barbouri, Pseudemys concinna and Sternotherus minor). We found a negative relationship between the total number of turtles captured and percentage of undisturbed land cover within a 287 m buffer width due to a significant increase in the number of T. scripta in less-forested sections of the creeks; however, the number of G. barbouri captures declined with reduced undisturbed land cover. Species evenness was positively correlated with percentage of undisturbed land. These results suggest that loss of riparian forest is associated with a decline in freshwater diversity (evenness) and a decline in the abundance of the endemic, state protected G. barbouri; however, overall turtle abundance may remain stable or increase with loss of riparian forest cover due to an increase in common, cosmopolitan species. Our results suggest that maintenance or restoration of riparian forests is critical to freshwater turtle conservation.