The Moon as Semaphore: Occupancy Modeling Reveals Contrasting Lunar Responses in Elusive Terraranas of a Tropical Montane Forest

Monitoring elusive amphibians requires understanding how environmental covariates influence detection and occupancy. We used occupancy modeling to assess the habitat use and detectability of three sympatric direct-developing frogs—the Chiriquí Robber Frog ( Pristimantis cruentus ), the La Loma Robber Frog ( Pristimantis caryophyllaceus ), and the Common Tink Frog (Diasporus diastema)—in a Costa Rican lower montane forest. Data was collected via nocturnal visual and acoustic surveys (quadratic sampling) across 14 quadrats (10x10m) over 11 monthly sessions from January 2012 to January 2013. We evaluated site covariates (tree diameter, richness, and abundance) and detection covariates (moon illumination and absolute humidity). The two Pristimantis species were old-growth forest specialists, with occupancy for P. cruentus strongly linked to larger tree diameters, increasing from ~ 0.50 to ~ 1.00 as diameter increased from ~ 30 to 35 cm. In contrast, D. diastema was a habitat generalist. Model selection revealed that detection for all three species was best explained by moon illumination and absolute humidity. Crucially, we found a clear case of temporal niche partitioning driven by the lunar cycle: P. caryophyllaceus exhibited moon phobia (peak detection: ~0.35 on dark nights), while both P. cruentus and D. diastema exhibited positive lunar tropism (peak detection: ~0.50 and 1.00, respectively, on moonlit nights). We propose that this specialized antithetical timing acts as a "semaphore," potentially minimizing acoustic competition between these morphologically similar species and with the broader nocturnal soundscape dominated by D. diastema . This finding demonstrates that the lunar cycle is a key and often overlooked environmental filter that orchestrates anuran behavior in complex tropical forests. Our results underscore that amphibian monitoring protocols must incorporate habitat structure and multiple environmental covariates, especially moon phase, to avoid biased estimates and accurately inform conservation.