We used 3,670 recoveries from 32,647 bandings of mallards (Anas platyrhynchos) in southwestern Saskatchewan during 1956-59 to estimate the probability of surviving from the local, flightless (classes II and III) stage to the flighted, immature stage. The probability of surviving from the local to the immature stage was 0.84 ? 0.05 (SE) for males and females. The geographic distribution of direct recoveries was similar for the birds banded as local and immature. Probabilities of survival for banded mallards can only be estimated from late summer to late summer. The estimate of survival from local to immature stage fills a gap in our knowledge of mallard mortality from female-brood breakup to the time of banding in late summer. J. WILDL. MANAGE. 53(2):428-431 Very little is known about survival of young mallards prior to the initiation of banding in late summer (Anderson 1975:25). Young mallards may experience substantial mortality, most of which occurs before 3 weeks of age (Dzubin and Gollop 1972, Ball et al. 1975, Talent et al. 1983). For example, the probability of mallard ducklings surviving from hatch to female-brood breakup was 44% in north-central Minnesota (Ball et al. 1975) and 35% in south-central North Dakota (Talent et al. 1983). The female-brood breakup generally occurs at about 7 weeks of age in early to midsummer (Ball et al. 1975, Talent et al. 1983). Probabilities of survival for banded birds can only be estimated from banding date to banding date (Brownie et al. 1985). Because most mallard bandings occur in late summer, survival estimates apply from late summer to late summer. Estimates of the probability of surviving between the later stages of brood-rearing and he time of banding flighted young do not exist. We present estimates of this probability for 1 population. We acknowledge the efforts of all individuals who banded mallards from 1956-59 in southwestern Saskatchewan and acknowledge the assistance of the U.S. Fish and Wildlife Service (USFWS) Bird Banding Laboratory in maintaining the banding and recovery records used in the analysis. Programming assistance was provided by J. E. Hines. Comments on earlier drafts This content downloaded from 157.55.39.192 on Wed, 24 Aug 2016 05:48:34 UTC All use subject to http://about.jstor.org/terms J. Wildl. Manage. 53(2):1989 MALLARD SURVIVAL * Hestbeck et al. 429 Table 1. Expected number of recoveries of mallards banded as locals under the most general model.a Expected values are expressed in terms of annual numbers of bandings of locals (N3), local to immature stage survival rate for the summer of year i (S,), annual survival rates of immatures (S) and adults (S), and annual recovery rates of immatures (f',) and adults (f,). Year Number banded 1 2 3 4 1 N, N,S,f', N,S,S'f, Nf1,SS',IS2f3 NS,SS'jS2S3f4 2 N2 N2S' fS2 N2S2S'2f3 N,S2S'2Saf, 3 N, N,3S13f'3 N3S,S'Sf4 4 N4 N4S4f'4 a Expected recoveries for immatures and ad are as specified under model H1 of Brownie et al. (1985:59-64). of the manuscript were provided by P. H. Geissler, G. M. Haramis, D.H. Johnson, R. Koford, and G. C. White.