Wildlife Society Bulletin 1–8; 2020; DOI: 10.1002/wsb.1123 Original Article Reliability of External Characteristics to Age Barrow’s Goldeneye TYLER L. LEWIS , 1,2 U.S. Geological Survey, Alaska Science Center, 4210 University Drive, Anchorage, AK 99508, USA DANIEL ESLER, U.S. Geological Survey, Alaska Science Center, 4210 University Drive, Anchorage, AK 99508, USA DANICA H. HOGAN, Environment and Climate Change Canada, 5019 52nd Street, Yellowknife, NWT X1A 2P7, Canada W. SEAN BOYD, Environment and Climate Change Canada, 5421 Robertson Road, Delta, BC V4K 3N2, Canada TIMOTHY D. BOWMAN, U.S. Fish and Wildlife Service, Migratory Bird Management, 1011 East Tudor Road, Anchorage, AK 99503, USA JONATHAN THOMPSON, Box 14 Site 21 Rural Route 2, Carvel, AB T0E 0H0, Canada ABSTRACT Accurate assignment of age class is critical for understanding most demographic processes. For waterfowl, most techniques for determining age class require birds in hand, reducing utility for quickly and efficiently sampling a large portion of the population. As an alternative, we sought to establish an observation‐based methodology, achievable in the field with standard optics, for determining age class of Barrow’s goldeneyes (Bucephala islandica). We photographed heads, wings, and bellies of 232 Barrow’s goldeneyes captured during late winter (February–April) of 2007–2015 along the north Pacific Coast. From these photographs, we focused on 5 external characteristics for both males and females, with binary states that putatively corresponded to 2 age classes—first‐year birds (<1 yr) and adults (>1 yr). For males, all 5 external traits (belly color, head color, eye color, facial crescent, median secondary coverts color) had binary states that were reliably distinguishable by observers. Moreover, all 5 external traits were highly predictive of age class (≥96% concordance between external vs. bursal‐derived ages), and novice observers, after receiving training, were able to accurately age 96% of first‐year and 99% of adult males. In contrast, patterns were weaker for females; putative external characteristics of female age class (belly color, bill radiance, bill blackness, eye color, median secondary coverts color) had 77–91% concordance with bursal‐ derived age, compared with 96–100% for males, and observers misidentified age classes of 15% of females, compared with only 2% of males. Overall, age classes of male Barrow’s goldeneyes were accurately and reliably distinguishable during winter based on several external characteristics, whereas those of females were not. Our technique may be used to estimate age composition of male Barrow’s goldeneyes during winter, providing a useful metric for monitoring annual changes in adult‐to‐juvenile ratios and other important demographic parameters. © 2020 The Wildlife Society. KEY WORDS age class, age ratio, Barrow’s goldeneye, Bucephala islandica, bursa, plumage, sea duck, sex determination. The ability to accurately assign age classes to individuals is critical for understanding demographic processes such as recruitment, survival, and age structure of populations (Saunders et al. 2014, Zhang et al. 2015). For waterfowl, external appearance of plumage and bare parts (e.g., bill, eyes, legs) has proven extremely useful for determining age, especially for differentiating first‐year birds from adults (Iverson et al. 2003, Rodway et al. 2015). Most techniques that use external characteristics for determining age classes of waterfowl were established from hunter‐killed carcasses; consequently, many of the techniques require birds in the hand to accurately identify detailed or concealed traits in- dicative of age class (Raftovich et al. 2017). For example, the United States Fish and Wildlife Service has annually collected duck wings from harvested individuals since 1961, using plumage traits of the wings to estimate annual age‐ class composition and productivity of harvested waterfowl (Pearse et al. 2014). Using carcasses or captured birds for estimating age ratios and annual productivity, however, has a number of drawbacks; carcasses require destructive sam- pling and may not be readily available for waterfowl species with low harvest, and capturing live waterfowl is often dif- ficult and expensive. Moreover, because juvenile waterfowl are more likely to be harvested or captured than are adults, Received: 1 October 2019; Accepted: 11 April 2020 Published: 1 E‐mail: tyler.lewis@alaska.gov 2 Current affiliation: Alaska Department of Fish and Game, Division of Wildlife Conservation, 525 W 67th Avenue, Anchorage, AK 99518, USA Lewis et al. • Determining Age Class of Barrow’s Goldeneye 1