A survey was conducted at two wildlife management areas of Pennsylvania (USA) to evaluate an antigen capture enzyme-linked immunosorbent assay (AC-ELISA) for the detection of avian influenza viruses (AIV) in cloacal swabs from waterfowl and to determine the influenza A virus subtypes and the distribution of these viruses among waterfowl. We collected 330 cloacal swabs from hunter-killed waterfowl in the fall of 1990 and from cage-captured waterfowl in the summer of 1991. Thirty-one hemagglutinating agents were isolated by chicken embryo inoculation (CEI) of which 27 were influenza A viruses and four Newcastle disease viruses (NDV). The prevalence of AIV infection was 8.2%. Compared to CEI, AC-ELISA was only 15% sensitive and 61% specific. Based on the distribution of AIV by species of waterfowl, mallards (Anas platyrhynchos) and American wigeons (Anas americana) were at equal risk of AIV infection even though most of the AIV isolates came from mallards. Although significant crude effects of sampling site and season on AIV recovery could be established, juvenile age was identified as the primary risk factor of AIV recovery. Twelve AIV subtypes were identified by hemagglutination inhibition (HI) and neuraminidase inhibition (NI) tests. The most prevalent subytpes were H4N8 and H6N8. We concluded that AC-ELISA was not useful for the detection of AIV in cloacal swabs from waterfowl and that CEI, HI, and NI tests remain as the method of choice for AIV screening in waterfowl. Based on the results AIV infected preferentially the young which represent the high risk group in waterfowl populations. The results from the AIV subtyping in our waterfowl survey are consistent with the results from numerous longitudinal studies of waterfowl in North America.