Songs distinguish the cryptic giant hummingbird species and clarify range limits

Abstract

Vocal traits are often essential for distinguishing phenotypically cryptic taxa. The hummingbird genus Patagona comprises two species, near identical in plumage and morphology, that differ in almost every aspect of their ecology and evolution: The Northern Giant Hummingbird (Patagona peruviana) and Southern Giant Hummingbird (Patagona gigas). Here, we characterized the songs of both giant hummingbird species and assessed whether song can be used to distinguish the two in the field. We recorded both species in Peru, Bolivia, and Chile in 2023 and 2025 and used public data to analyze song variation of 217 individuals recorded across the Andes. Sampling spanned 49 years, >36{degrees} of latitude, and >4,300 meters in elevation. We first quantified species-level song differences in allopatric breeding populations and trained a linear discriminant model to identify individuals to species. The trained model had 100% classification accuracy. We then used our trained model to identify individuals recorded during co-occurring, non- breeding periods of overlap, and subsequently analyzed range-wide song variation; this model had 98.72% classification accuracy (1.28% error rate; one individual misidentified). We found striking song divergence between the two species, uncovering that Northern and Southern Giant Hummingbirds can be reliably and easily identified by song across their ranges and during any month of the year. We provide new data on the range limits of both species in northern Bolivia, highlighting a previously unknown zone of overlap around Lake Titicaca. Unlike other phenotypic traits, song provides a robust method for identifying the giant hummingbird species, opening doors to future research on ecology, trait evolution, hybrid zone dynamics, and conservation of the worlds largest hummingbirds.