The present report quantifies relations between weather and several aspects of the breeding biology of four duck species: mallard (Anas platyrhynchos), gadwall (A. strepera), blue-winged teal (A. discors), and redhead (Aythya americana). Data were obtained from two locations in North Dakota,--the J. Clark Salyer National Wildlife Refuge, intermittently during 1936-68, and the Northern Prairie Wildlife Research Center's Woodworth Station during 1965-77. Arrival dates varied with mean temperature before and during the usual arrival period; early-arriving species were affected by early-season temperatures, and later-arriving species by temperatures later in spring. After temperature effects were accounted for, first arrivals were seen at the more southern Woodworth Station a few days earlier than at Salyer. High spring temperatures also seemed to induce early nesting. The mallard, which nested earliest, was most affected by temperatures during April whereas the other species were most affected by temperatures during late April and May. Peak hatching dates were also earlier in years with higher temperatures in May. Earlier peaks were associated with early first nests, so it was difficult to separate the effects of weather and date of initial nesting. Peaks at Woodworth occurred earlier than at Salyer, after temperature differences were taken into account. The period of most active nesting was longest for the early-nesting mallard, shortest for the late-nesting gadwall, and intermediate for the blue-winged teal and redhead. For two species, precipitation during the breeding season may have prolonged nesting activities. We also found that late nesting seasons tend to be compressed. Productivity at Salyer, measured by the brood to pair ratio, was generally greater in years with higher temperatures during 23 April-3 June. Effects were more pronounced among early-nesting species. Average brood size for Classes I and II tended to decline during the 1947-62 period at Salyer; pair populations generally increased. Consequently, it was nearly impossible to distinguish the effects of pair density on brood size from those of yearly trend. In addition, Class II broods were smaller in years when temperatures were higher during late May and June.