Northern Prairie Wildlife Research Center
We obtained data on the frequency of nest parasitism and cowbird/host ratios in seeded grassland, natural grassland, and cropland west of the Agassiz Lake Plain in south central North Dakota. This entire area is in the glaciated prairie pothole region of North Dakota (Stewart and Kantrud 1974). The terrain is gently to moderately rolling and contains numerous temporary, seasonal, and semipermanent wetlands. About one sixth of the upland habitat is seeded grassland, one sixth is natural grassland (mostly pastureland), and two thirds is cropland.
Our seeded grassland fields were either former cropland that had been enrolled in the Conservation Reserve Program (CRP) of the U.S. Department of Agriculture or upland areas on Waterfowl Production Areas (WPA) of the U.S. Fish and Wildlife Service. We examined six CRP fields (10-16 ha) and two fields (11-17 ha) on WPAs that covered 65-105 ha of wetland and upland habitat. The CRP fields were seeded in the late 1980s with mixtures of legumes (sweet clover, Melilotus spp., or alfalfa, Medicago sativa) and cool-season grasses. They were left ungrazed and unhayed during the study (1991-1993). In the 1970s, WPAs were seeded with similar grass-legume mixtures to attract nesting ducks.
Natural grassland in WPAs included the native grasses needle-and-thread (Stipa comata) and green needle grass (S. viridula), the exotic grasses Kentucky bluegrass (Poa pratensis) and smooth brome (Bromus inermis), forbs, and shrubs (mainly western snowberry, Symphoricarpos occidentalis). Natural-grassland fields were examined in two studies, which we refer to as the northern study and the southern study. In the northern study, we examined all species of nesting birds on 4 fields in 1991-1993. In the southern study, conducted in 1981-1987, we examined Western Meadowlarks (Sturnella neglecta) on 15 fenced fields (25-47 ha). These fields were used in a study of the effects of cattle grazing on ground-nesting birds (Bowen and Kruse 1993).
The cropland we examined in 1991-1993 consisted of three types of crop and three farming systems. The crops were wheat (the principal small grain in the region), sunflower (the principal row crop), and fallow. The three farming systems were organic (synthetic chemicals were not used), minimum till (tillage was minimal), and conventional (tillage and synthetic chemicals were used). On fallow fields, organic farmers grew sweet clover that was plowed under in June, minimum-till farmers left stubble and controlled weeds with herbicide; and conventional farmers left little or no vegetation on fields that were repeatedly tilled. We studied ten 40-ha fields of each combination of crop type and farming system for a total of about 90 fields on ten different farms.
In spring, crop fields had little or no residual cover. Cover varied from small-grain stubble to bare soil. Residual cover declined in late April and May when most organic and conventional fields were cultivated and seeded. Crops typically begin to grow rapidly in late May and June and were tall and lush by early July.
Nests in seeded grassland and northern natural grassland were located mainly by flushing birds from nests with a 25-m, hand-pulled rope. We conducted three searches on each field. In the southern study, we used a 53-m cable-chain towed between two vehicles (Higgins et al. 1977). In cropland, we conducted three searches between late April and mid-July. Two searches were conducted with a 50-m chain pulled by all-terrain vehicles (ATV) or, in rowcrop fields, a specially fitted ATV. The third search was conducted in July by walking through each field looking for nests of Red-winged Blackbirds (Agelaius phoeniceus).
We plotted the locations of all nests on maps and placed marker flags 4 m from each nest bowl. We recorded the number of host eggs, number of cowbird eggs, and the incubation stage or nestling stage during each nest visit. We monitored nests to determine their fates.
Any active nest that contained a cowbird egg or nestling on any visit was considered parasitized. The frequency of parasitism for a given species in a given habitat was the number of parasitized nests (x 100) divided by the total number of nests located for that species.
To report an average frequency of parasitism in each habitat, we calculated an unweighted mean of the frequencies for each species. Data from species with fewer than 10 nests were pooled. Data on Western Meadowlark nests in natural grassland in the northern study were treated separately from those in the southern study. There were fewer than 10 nests of Western Meadowlarks in the northern study; these nests were pooled with other species. Differences among habitats were not tested statistically because of possible confounding effects of year, species, and land use in different areas.
The sample sizes of nests do not necessarily reflect species abundance in the three habitats. Studies varied by searched area and number of study years. In particular, studies of natural grassland in the northern study area took place on fewer fields and were of shorter duration than in the southern study. Also, our search methods may have been more likely to locate nests of some species (e.g., Red-winged Blackbirds) because of their high nest visibility and adult vigilance.
Populations were censused with strip transects 100 m wide. Transects covered most or all of each field. We censused birds well after sunrise in the morning and recorded their numbers; we ignored flyovers. Single birds of unknown sex or groups of host species were assigned equally to the two sexes. We assigned a single female to each group of cowbirds because we reasoned that most groups of cowbirds probably consisted of one female and several courting males. We conducted two or three breeding-season censuses on each study field between 15 May and 15 July (the earliest cowbird eggs were found in mid-May).
We calculated the ratio of female cowbirds to male hosts in each habitat by using total counts from all censuses. Robinson et al. (1993) suggested that this ratio could be used with data derived from point counts to obtain a crude index of parasitism frequency. Data from the northern study of natural grassland were analyzed separately from the southern study. We used host species that Ehrlich et al. (1988) rated as frequent, common, or uncommon hosts. Because these hosts were the most abundant birds, the results would have been similar if we had used all possible hosts.