Summary 1.Habitat selection has profound ecological and evolutionary consequences. For example, there may be strong selection for nest-site choice such that oviparous females, lacking parental care, may adaptively manipulate their offspring survival or phenotype. Alternatively, nesting decisions may be a passive by-product of other processes leading to similar nonrandom nesting patterns. 2.Here we examined the nesting ecology of the turtle Podocnemis unifilis at multiple spatiotemporal scales to determine whether randomness, adaptive nest-site selection or social facilitation best explained the observed population-level patterns. We addressed these alternative strategies by exploring how environmental and geographic variation in nest sites influences embryonic survival within and across four nesting beaches in three years. 3. We found nonrandom spatial and environmental patterns of nesting within beaches and years, consistent with both the adaptive nest-site selection and social facilitation hypotheses. However, nesting patterns were unpredictable among beaches and among years. Furthermore, environmental conditions at nest sites and offspring survival were not associated, and nests from the most gregarious nesting night clustered more tightly geographically and suffered lower predation than nests laid on other nights. 4. Together, our findings provide more extensive support for social facilitation as compared to the adaptive nest-site selection hypothesis. Our results suggest that selection for female nest-site choice in reptiles may be acting more strongly via offspring survival through nest clustering derived from conspecific cueing and less strongly via environmental cueing than previously anticipated. 5. Our findings underscore the importance of examining multiple sites during multiple seasons. This approach permits testing critical predictions about the consistency of population-level patterns across space and time that enable the distinction between models. 6. Our results support a shift with regards to the trait that is usually considered the target of selection for female nest-site choice in a way that exemplifies the classic dichotomy between selection for (survival of nests and perhaps of females) and selection of (offspring phenotype, such as sex).

Mortality factors and hatching success of 422 terecay turtle (Podocnemis uni®lis) nests on seven beaches in the Nichare and Tawadu Rivers, Venezuela, were studied following the oviposition phase. Beaches (3.4±28.4 km 2 ), predominantly sandy, and bordered by riparian vegetation, were searched to record nest numbers and their distribution. Elevation and surrounding plant cover were measured for each nest. Most nests were found close to the vegetation's edge (0±11 m), at a signi®cant distance from the water's edge (21±80 m), in more open ground, and along the highest points of the beach (1±2.5 m above the water's edge). Nest density was positively correlated with beach elevation but not with beach dimensions. Most nests were concentrated in less than 10% of beach area, along the upper 20% of the beach. Average clutch size was 20.1 1.7 eggs, but larger clutches were found signi®cantly further away from vegetation. No signi®cant correlation between hatching success and clutch size was found.A large proportion of nests were subject to animal and human predation but environmental factors (especially¯ooding) affected some. The major cause of egg loss was human predation. Most clutches collected by humans were found away from vegetation, on the upper 10±20% of beaches, in exposed unvegetated sites. Animals preyed upon those nests along the vegetation's margin (1±2 m) in the upper 0±9% of beaches, in sites of high plant cover (75±100%). Nests reaching the incubation phase were mainly located in the open areas, which are more prone to human predation. About three-quarters of nests showed high hatching rates (91±100%). However, hatchability was highest furthest away from the vegetation. Our results indicate that humans are collecting eggs mostly from sites in which nests have the larger clutches and the higher potential hatching success. Because of this relationship between nest viability and location, sustainable yield programmes must consider where harvesting can take place and must avoid the application of standard harvests per nest.

We studied the population genetics of Podocnemis unifilis turtles within and among basins in the Orinoco and Amazon drainages using microsatellites. We detected high levels of genetic diversity in all sampled localities. However, 'M-ratio' tests revealed a substantial recent population decline in ten localities, in accord with current widespread exploitation. Our results reveal a consistent pattern across multiple analyses, showing a clear subdivision between the populations inhabiting the Amazon and Orinoco drainages despite a direct connection via the Casiquiare corridor, and suggesting the existence of two biogeographically independent and widely divergent lineages. Genetic differentiation followed an isolationby-distance model concordant with hypotheses about migration. It appears that migration occurs via the flooded forest in some drainages, and via river channels in those where geographic barriers preclude dispersal between basins or even among nearby tributaries of the same basin. These observations caution against making generalizations based on geographically restricted data, and indicate that geographically proximate populations may be demographically separate units requiring independent management.

No abstract

Mortality factors and hatching success of 422 terecay turtle (Podocnemis uni®lis) nests on seven beaches in the Nichare and Tawadu Rivers, Venezuela, were studied following the oviposition phase. Beaches (3.4±28.4 km 2 ), predominantly sandy, and bordered by riparian vegetation, were searched to record nest numbers and their distribution. Elevation and surrounding plant cover were measured for each nest. Most nests were found close to the vegetation's edge (0±11 m), at a signi®cant distance from the water's edge (21±80 m), in more open ground, and along the highest points of the beach (1±2.5 m above the water's edge). Nest density was positively correlated with beach elevation but not with beach dimensions. Most nests were concentrated in less than 10% of beach area, along the upper 20% of the beach. Average clutch size was 20.1 1.7 eggs, but larger clutches were found signi®cantly further away from vegetation. No signi®cant correlation between hatching success and clutch size was found.A large proportion of nests were subject to animal and human predation but environmental factors (especially¯ooding) affected some. The major cause of egg loss was human predation. Most clutches collected by humans were found away from vegetation, on the upper 10±20% of beaches, in exposed unvegetated sites. Animals preyed upon those nests along the vegetation's margin (1±2 m) in the upper 0±9% of beaches, in sites of high plant cover (75±100%). Nests reaching the incubation phase were mainly located in the open areas, which are more prone to human predation. About three-quarters of nests showed high hatching rates (91±100%). However, hatchability was highest furthest away from the vegetation. Our results indicate that humans are collecting eggs mostly from sites in which nests have the larger clutches and the higher potential hatching success. Because of this relationship between nest viability and location, sustainable yield programmes must consider where harvesting can take place and must avoid the application of standard harvests per nest.

The mitochondrial genome encodes several protein components of the oxidative phosphorylation (OXPHOS) pathway and is critical for aerobic respiration. These proteins have evolved adaptively in many taxa, but linking molecular-level patterns with higher-level attributes (e.g., morphology, physiology) remains a challenge. Turtles are a promising system for exploring mitochondrial genome evolution as different species face distinct respiratory challenges and employ multiple strategies for ensuring efficient respiration. One prominent adaptation to a highly aquatic lifestyle in turtles is the secondary loss of keratenized shell scutes (i.e., soft-shells), which is associated with enhanced swimming ability and, in some species, cutaneous respiration. We used codon models to examine patterns of selection on mitochondrial protein-coding genes along the three turtle lineages that independently evolved soft-shells. We found strong evidence for positive selection along the branches leading to the pig-nosed turtle (Carettochelys insculpta) and the softshells clade (Trionychidae), but only weak evidence for the leatherback (Dermochelys coriacea) branch. Positively selected sites were found to be particularly prevalent in OXPHOS Complex I proteins, especially subunit ND2, along both positively selected lineages, consistent with convergent adaptive evolution. Structural analysis showed that many of the identified sites are within key regions or near residues involved in proton transport, indicating that positive selection may have precipitated substantial changes in mitochondrial function. Overall, our study provides evidence that physiological challenges associated with adaptation to a highly aquatic lifestyle have shaped the evolution of the turtle mitochondrial genome in a lineage-specific manner.

BackgroundUncovering how phenotypic diversity arises and is maintained in nature has long been a major interest of evolutionary biologists. Recent advances in genome sequencing technologies have remarkably increased the efficiency to pinpoint genes involved in the adaptive evolution of phenotypes. Reliability of such findings is most often examined with statistical and computational methods using Maximum Likelihood codon-based models (i.e., site, branch, branch-site and clade models), such as those available in codeml from the Phylogenetic Analysis by Maximum Likelihood (PAML) package. While these models represent a well-defined workflow for documenting adaptive evolution, in practice they can be challenging for researchers having a vast amount of data, as multiple types of relevant codon-based datasets are generated, making the overall process hard and tedious to handle, error-prone and time-consuming.ResultsWe introduce LMAP (Lightweight Multigene Analyses in PAML), a user-friendly command-line and interactive package, designed to handle the codeml workflow, namely: directory organization, execution, results gathering and organization for Likelihood Ratio Test estimations with minimal manual user intervention. LMAP was developed for the workstation multi-core environment and provides a unique advantage for processing one, or more, if not all codeml codon-based models for multiple datasets at a time. Our software, proved efficiency throughout the codeml workflow, including, but not limited, to simultaneously handling more than 20 datasets.ConclusionsWe have developed a simple and versatile LMAP package, with outstanding performance, enabling researchers to analyze multiple different codon-based datasets in a high-throughput fashion. At minimum, two file types are required within a single input directory: one for the multiple sequence alignment and another for the phylogenetic tree. To our knowledge, no other software combines all codeml codon substitution models of adaptive evolution. LMAP has been developed as an open-source package, allowing its integration into more complex open-source bioinformatics pipelines. LMAP package is released under GPLv3 license and is freely available at http://lmapaml.sourceforge.net/.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-016-1204-5) contains supplementary material, which is available to authorized users.

In oviparous vertebrates lacking parental care, resource allocation during reproduction is a major maternal effect that may enhance female fitness. In general, resource allocation strategies are expected to follow optimality models to solve the energy trade-offs between egg size and number. Such models predict that natural selection should optimize egg size while egg number is expected to vary with female size, thus maximizing offspring fitness and consequently, maternal fitness. Deviations from optimality predictions are commonly attributed to morphological constraints imposed by female size, such as reported for smallbodied turtle species. However, whether such anatomical constraints exist in smaller-bodied females within large-bodied clades remains unstudied. Here we tested whether resource allocation of the river turtle Podocnemis unifilis (a relatively smaller member of the largebodied Podocnemididae) follows optimality theory, and found a pattern of egg elongation in smaller females that provides evidence of morphological constraints and of a reproductive trade-off with clutch size, whereas egg width supports the existence of an optimal egg size and no trade-off. Moreover, larger females laid larger clutches composed of rounder eggs, while smaller females laid fewer and relatively more elongated eggs. Elongated eggs from smaller females have larger volume relative to female size and to round eggs of equal width. We propose that elongated eggs represent a solution to a potential morphological constraint suffered by small females. Our results suggest that larger females may optimize fitness by increasing the number of eggs, while smaller females do so by producing larger eggs. Our data supports the notion that morphological constraints are likely more widespread than previously anticipated, such that they may not be exclusive of small-bodied lineages but may also exist in large-bodied lineages.