Least terns (Sterna antillarum) were studied on the lower Platte River, Nebraska, where this endangered population nests on natural sandbar habitat and on sandpit sites created by gravel dredging adjacent to the river. Theoretically terns should select habitats according to habitat suitability. However, the introduction of sandpits and conversion of tallgrass prairies along the river banks to agriculture, residential, and wooded areas may have affected terns' abilities to distinguish suitable habitat or the suitability of nesting habitats in general. I examined habitat selection and productivity of least terns to determine if terns selected habitat according to suitability (as indicated by productivity), what factors affected habitat selection and productivity, and if estimated productivity could support this population.

Available habitats of both types were characterized and quantified using aerial videography (1989-90), and habitat use was assessed from census data (1987-90). Productivity of adults and causes and correlates of egg and chick mortality were estimated (1987-90). Population trend was assessed with a deterministic model using my estimates of productivity and a range of survival estimates for Laridae reported in the literature.

Terns tended to use river sites with large midstream sandbars and a wide channel, and large sandpit sites with large surface areas of water relative to unused sites on both habitats. Number of sites and area of sand available were estimated using discriminant function analysis of variables quantified from video scenes of both habitats. Terns apparently did not use all potentially available sandbar and sandpit sites because discriminant function factor scores for used and unused sites overlapped broadly for both habitats. Terns did not prefer 1 habitat over the other. Although proportions of available sites used were greater on sandpits than on the river, proportions of available sand used did not differ between habitats. Proportion of terns using each habitat was similar to proportion of available sand on each habitat. The distribution of nest initiation dates and rates of colony-site turnover also were similar on both habitats.

Productivity did not differ between habitats but varied significantly among sites. Nest success, fledging success, and fledglings per pair averaged 0.54, 0.28, and 0.47, respectively. Key factor analysis revealed that chick survival had a greater influence on production of fledglings (on both sandbars and sandpits) than did failure to produce a maximum clutch size or egg mortality. Most egg mortality was caused by predation on sandpits and by flooding on sandbars. Predation was suspected as the major cause of loss for chicks on both habitats. Path analysis revealed no strong or consistent correlations among mortality, numbers of nests and chicks, track trails of intruders into colonies, and habitat variables at colonies on either habitat.

Theoretically, terns should not prefer a habitat when habitats are equally suitable if terns have had time to respond to habitat changes. Although sandbars and sandpits appeared equally suitable and terns did not prefer either habitat, local productivity will not support this population unless annual postfledging survival is higher than current estimates for the species. Population trend estimated with fledglings per pair = 0.50 was negative for all but the highest (ca 0.90) rates of annual postfledging survival. Furthermore, deterministic models like the one used in this study overstimate trend.

Productivity insufficient to support the local population, in spite of habitat use that reflects habitat suitability, could be due to increased predation caused by habitat alteration adjacent to the river that may have changed the predator community. Alternatively, terns in this area could persist in spite of prevailing low productivity because they are relatively long-lived birds, if highly productive years occasionally occur or if this population is augmented by immigrants from elsewhere.