Highly Variable Autumn Calling Rates of Northern Bobwhite Following Translocation

Fall covey counts are a popular index for monitoring population trends of northern bobwhite (Colinus virginianus; hereafter, bobwhite), but their utility is tenuous under different scenarios. Detecting an individual covey is the product of the probability that the covey’s activity center is located within the sampling frame, the probability the covey is located within the sampling frame during the sampling periods, the probability of the covey vocalizing, and the probability an observer will detect a calling covey. Researchers attempt to maximize detection or account for these potential sources of error using standardized protocol of limiting counts to certain weather conditions, replication, and distance sampling. Variation in calling rates across a range of bobwhite densities could lead to tenuous inference of population abundance from fall covey counts, particularly at low densities. Our objectives were to assess fall calling rates at 2 sites with low bobwhite density during population restoration. Our study sites were located in Erath County, Texas, USA and Leon County, Florida, USA and received translocated bobwhite during 2019 and 2020. We hypothesized calling rates would be influenced by the number of adjacent coveys that called, and thus, would be low for our sites. Although we did not estimate bobwhite density on our study sites, we surmised that their respective populations were /3 ha. Calling rate at the Erath County site was 0 in 2019 (n = 10 counts) and increased to 0.79 (standard error [SE] = 0.07, n = 34 counts) in 2020. Calling rate was assessed only in 2020 at the Leon County site and averaged 0.13 (SE = 0.07, n = 23 counts). Detection rate at count stations was 0 in 2019 and 0.78 (SE = 0.08, n = 27 calling coveys) in 2020 at the Erath County site. In 2020, detection rate at count stations was 0 (n = 3 calling coveys) at the Leon County site. We documented high annual variation in calling rates among low-abundance sites, suggesting researchers should seek to empirically estimate this parameter rather than applying arbitrary correction factors based on previous literature. Low and variable calling rates limit detection and can bias inference