The timing of migratory passage for sexes and ages of birds is difficult to document because migration is spread out over time and space. We compared sex and age ratios from 10 species of birds that were banded in Wisconsin during fall migration to a sample of autumn window-killed birds from Chicago, Illinois. We assumed that these specimens were a random sample of the migrating population because dates of passage in Wisconsin either predate or were coincident with specimens recovered in Chicago. More juveniles than adults collided with windows, which is expected because of the higher number of juveniles in the fall migratory population. For Connecticut Warbler (Oporornis agilis), Dark-eyed Junco (Junco hyemalis), Ovenbird (Seiurus aurocapilla), and White-throated Sparrow (Zonotrichia albicollis), statistically fewer adults collided with windows than expected. It is possible that adults of these species, who have made at least 2 prior successful migratory journeys, might have experienced some degree of learning to avoid windows. For specimen data, ratios of males to females were approximately equal, except for Swamp Sparrow (Melospiza georgiana; 1:0.82). However, banders were unable to determine sex for a high percentage of each species, ranging from 98.5% undetermined sex for Hermit Thrush (Catharus guttatus) to 48% for Common Yellowthroat (Geothlypis trichas). Taken at face value, banding records for Common Yellowthroat, Dark-eyed Junco, and White-throated Sparrow indicated a much higher percentage of males, whereas for Song Sparrow (Melospiza melodia), females were much more common; 6 species showed no apparent difference in sex ratio between the banding and window samples. In species for which banders declined to identify an individual's sex for a high percentage of birds, we questioned whether the identification of sex in the remaining small percentage was accurate. Thus, we considered the alternative possibility that the specimen data were in fact the most appropriate null hypothesis, and if so, it would suggest that banders have misidentified sexes consistently in 4 species, and ages in 7 species.