Science at OTS
Section by Elizabeth Braker. OTS CEO.
Why do large groups of harvestmen bob more frequently and longer than small groups? Aggregations of harvestmen (arachnids in order Opiolones) are often spotted on a branch or between tree buttresses in forests. When disturbed, these animals may begin to bob up and down, a behavior assumed to be a defensive response. Two research groups recently published work conducted at La Selva on the possible functional significance of this behavior as part of the defensive repertoire of harvestmen. While aggregation is thought to be a primary defense in these animals, the bobbing behavior may be a secondary defensive response that could increase the strength of the signal to potential predators. Calhoun et al. (2025)1 tested whether bobbing behavior occurred more frequently in aggregations than in isolation in Prionostemma spp. As group size increased, so did the frequency of bobbing. The authors suggested that bobbing is a context-dependent antipredator defense behavior, depending on the relative costs and benefits of its expression. Taking a different approach, Villaseñor-Amador and Escalante (2025)2 explored potential drivers of bobbing, also in Prionostemma, finding that the duration of bobbing behavior increased with group size. The authors also explored different possible triggering stimuli for bobbing behavior, finding that a touching stimulus (gentle stick touching) triggered longer bobbing than an airflow stimulus (gentle blow). Additionally, they studied whether losing one or more legs (a common response to attempted predation) affected individual engagement in bobbing, finding that there was no relationship between these variations in leg condition and duration of bobbing. These studies on Prionostemma provide new insights into a striking yet little-studied behavior, as well as suggesting pathways to explain the relative importance of different aspects of defensive behavior in animals.
Forest plots at La Selva provide a field baseline for estimates of efficiency and precision for ground-based sampling inventory. The CARBONO plots were established at La Selva for a multidisciplinary team study of forest carbon cycling. The project was designed to assess forest processes at the landscape scale by sampling with replication across the within-landscape edaphic heterogeneity typical of tropical forests. Through more than two decades, forest growth and dynamics were assessed annually (data and publications in the Dryad repository; Clark and Clark (2021)3. To validate efficiency and precision of forest sampling methods to be used for Forest Carbon Offset calculations, Tompo et al. (2024)4 used field measurements of the CARBONO plots at La Selva and, for comparison and simulations, of published tree-level data on Barro Colorado Island in Panama. Plot configurations and costs were compared using criteria commonly used by forest carbon offset project developers to meet precision of +/−10% at 90% confidence interval for above ground biomass estimation. Multiple alternative plot configurations were evaluated with one method (angle-count plots) emerging as a highly cost-efficient and versatile alternative in complex tropical forests.
Two decades of monitoring forest restoration plots show that different initial treatments persist in forest canopy composition. Schubert et al. (2024) 5 compared the results of different approaches to forest restoration in the Las Cruces region. The researchers assessed 16–18-year-old plots, initially established with four different restoration treatments, and compared them to plots in adjacent reference forests to compare tree recruitment patterns and community composition across treatments and in relation to the reference forests. Restoration treatments that incorporated active tree planting resulted in the establishment of more later-successional trees compared with natural regeneration. The authors caution that more time is needed to assess whether these differences will persist as the restored areas mature, and how quickly restored forests achieve similarity to mature later-successional canopy. The results indicate the need for long-term monitoring of restoration efforts.
Peccary wallows are biodiversity hubs. Wallows (isolated, water-filled depressions) created by the activity of collared peccary (Pecari (Dicotyles) tajacu) at La Selva functioned as attractors for many vertebrate species6. Activity and diversity of these other species higher at wallows compared to the adjacent forest. A total of 42 vertebrate species used the wallows in a range of behaviors including reproductive activity, bathing, and drinking. Wallos persisted for as long as six years and were used consistently over time.
Emerald glass frog females do not provide effective parental care to egg clutches. Although many amphibians care for their eggs and young, protecting them from predation and desiccation, the emerald glass frog (Espadarana prosoblepon) attends eggs for only a short time (1.5 hours) after oviposition. Goyes Vallejos et al. (2024)7 found that female emerald glass frogs are easily disturbed, causing them to abandon eggs even before this short interval. Similar levels of mortality were recorded in clutches with and without female presence, and female presence did not reduce mortality due to dehydration or predation. The authors challenge the assumption that simple clutch attendance is not enough to demonstrate effective parental care in these amphibians, and perhaps more broadly in other animal species.
Sloth species do not share parasite species. The interactions of gastrointestinal parasites and their vertebrate hosts are both difficult and messy to study, and are especially poorly known for free-ranging, arboreal mammals living in tropical forests. Vanderhoeven et al. (2025)8 compared diversity and host specificity of gastrointestinal parasites infecting free-ranging sloths at La Selva, asking whether host-parasite interactions were structured by host identity, the habitats in which hosts occurred, or both.
The researchers collected fecal matter from both Hoffmann’s two-toed sloth (Choloepus hoffmanni) and the brown-throated three-toed sloths (Bradypus variegatus). Both host species harbored protozoa and nematode eggs, but only C. hoffmanni had cestode eggs. The diversity of parasitic morphotypes found in this study did not differ from what has previously been described. Parasite richness did not differ between species or between habitats, despite an interesting trend of higher parasite richness for forest-living C. hoffmanni compared to that of urban individuals living in Puerto Viejo de Talamanca. However, sloth species did not share parasite species, resulting in strong differences in parasite community composition between species. Rare and possibly undescribed parasite taxa were recorded only in samples from primary forest. The study raises interesting directions for future study, including the importance of characterizing host-parasite transmission networks and the potential relevance of intermediate hosts.
Hummingbird flower mites detect and respond to electrical fields. Tropical phoretic mites that feed on nectar and pollen are transported from flower to flower by hummingbirds. In this study, Garcia-Robledo et al. (2025) asked whether electric fields facilitate transportation and detection of hummingbirds by a guild of tropical phoretic mites. Using choice experiments to study mite attraction to electrical fields and electrostatic forces, the authors found that mites use electroreception detected via sensory structures on their front legs to detect hummingbirds and use electrostatics to facilitate transportation onto their hosts.9