Northern Prairie Wildlife Research Center
Annual field investigations usually included notation of bird arrival dates, surveys of early spring residual vegetation, wetland water censuses, duck pair and brood counts, nest searches, land-use inventories, wetland classification surveys, and late summer vegetation surveys. Wetlands were classified as wet when a basin was at least 5% inundated and the water was 2.5 cm or more deep. Five water surveys were usually made annually, one soon after all snow and ice had melted from the uplands and wetlands, and four during duck pair and brood counts. For purposes of analyses, water surveys were grouped into five periods: 1-15 May, 1-15 June, 16-30 June, 1-15 July, and 1-15 August.
Counts of breeding pairs (Hammond 1969) of adult ducks (hereinafter called pairs) were made to provide indexes to the number of pairs nesting on the study area. Between 10 April and 10 July, eight pair counts were made in 1964; six in 1965-67 and 1969; five in 1968; three in 1970-71; and two in 1972-81. An evaluation of the pair counts from 1964 to 1969 indicated that pairs could be estimated as well based on data from two counts as from three or more counts. The time period with relatively stable duck numbers was usually between 25 April and 15 June.
Counts were made while walking, driving a motorbike, or driving a jeep around the periphery of wetlands. On larger wetlands and lakes or on wetlands choked with emergent vegetation, two people started counting ducks from a point on the shoreline and walked or waded in opposite directions until they met. On such counts, notes were compared to eliminate duplication. One person could count ducks on small wetlands. If ducks were flushed from wetlands during the count, their flight pattern or descent to another wetland was noted to avoid duplicate counting of the same ducks on wetlands scheduled to be counted. Ducks passing over in flight but not flushed by a counter were not included in the census.
During pair counts, all ducks were identified by species and sex within each group and recorded on field maps. The decision to segregate or aggregate duck populations on wetlands was made as follows: several ducks close to each other were considered one group; several ducks, either in singles or aggregates, spaced distantly from each other were considered several groups; and a single duck was considered a group. Final tabulation of pairs was made according to Hammond (1969): all waterfowl and other groups and flocks were recorded except those flying over the sample or transect and those flying into the sample area and landing. Indicated pairs were partitioned for final tabulation: segregated pairs and lone drakes were tabulated for both dabbling and diving ducks, lone hens were tabulated only for diving ducks and only when males were not nearby, and male groups and mixed groups of males and females were tabulated for groups up to five with the exception of northern shovelers and American wigeon for which only lone males and pairs are counted (Hammond 1969).
The time required to complete a pair count on the study area depended on the number and size of wetlands to be counted, number of people available, amount of emergent vegetation in wetlands, weather conditions (especially wind, heat, and rain), and number of ducks. During wet years, when many ducks were present, 3-4 days were required per count, and during dry years, when smaller numbers occurred, counts were completed in 1 day. Counts were usually made between 0700 and 1700 h.
Only pair data from counts made nearest 1 May and 1 June were used in the final breeding pair estimates and pairs per wetland analyses. Pair estimates for mallards, northern pintails, northern shovelers, and canvasbarks were based on early May counts, whereas estimates for gadwalls, blue-winged teal, redheads, lesser scaup, ring-necked ducks, and ruddy ducks were based on June counts. We used the higher of May or June counts for estimates of the pairs of American wigeon and green-winged teal.
Before 1967, nest searching methods included flailing the vegetation with switches or laths while walking through the cover in zigzag patterns, watching for hens to return to their nests, flushing hens from nests with dogs, towing ropes with tin cans, and finding nests during incidental field investigations. Because of the large variety and size of nesting habitats, those methods proved impractical. In late summer 1966 and early spring 1967, a cable-chain drag (Fig. 8) for finding duck nests was developed and tested (Higgins et al. 1969, 1971). The cable-chain drag has proven to be a practical method of finding a relatively large sample of duck nests in most types of upland prairie habitats; most nests in this study were found with this device.
Nest searches usually began the first week in May and were completed by mid-July each year. Three to four searches were made on the study plots annually. Most were conducted between 0700 and 1600 h.
A nest was defined as a hollowed scrape containing one or more eggs. Nest locations were usually marked with a fluorescent red surveyor's flag trimmed to 3 cm wide, but in fields with tall, dense vegetation, nests were often marked with colored ribbon or cloth flags on willow sticks. Nest markers were usually placed 3 m north of each nest. Standardization of marker direction from nests and flag color reduced the amount of time spent searching for nests during subsequent visits. A simple field map showing the general location and distance of a nest from obvious topographic features also reduced time spent finding these nests during revisits.
Data for each nest were recorded, in the field, on a standardized card (Fig. 9). Data recorded during the initial visit included date, time, location, number of eggs, stage of incubation, percent of dead vegetation at nest site, dominant plant species at the nest site, nest site physiognomy, overhead and side concealment of the nest, height of vegetation surrounding the nest, upland cover quality at and immediately surrounding the nest, and land use surrounding the nest.
The stage of embryo development of eggs in each nest was determined by candling in the field with a piece of flexible, 38 mm (1.5 inch) inside diameter black radiator hose (Weller 1956; Fig. 10), or by egg flotation in water (Westerskov 1950); the former method was preferred. Number of incubation days plus number of eggs in a clutch were then used to calculate the date of nest initiation. Mean values for species' clutch sizes, incubation terms, and ages of clutches at hatching follow Klett et al. (1986).
Data recorded during subsequent nest visits included fate of clutch (destroyed, abandoned, or unknown), number of eggs hatched, amount of down, and cause of destruction if known. A successful nest was defined as one in which at least one duckling hatched and left the nest. Rates of nest success were estimated according to the Mayfield 40% method (Miller and Johnson 1978; Johnson 1979). Duck nest densities were calculated based on the number of nests per area searched. The sample of nests in all instances was those that survived long enough to be found because nearly all nests were found by flushing hens from their nests.
Brood counts were made annually to provide an index to production of ducklings in the study area. Every pothole with water was inventoried for broods or for hens doing distraction displays, which were considered as representing broods. Brood counts were made while wading through each wetland in undulating or zigzag patterns (Evans and Black 1956) to purposely flush hens and broods into an observation field for counting. The number and age class of ducklings were recorded according to Gollop and Marshall (1954). Records of broods by species, age, number of ducklings, and location were used to screen and separate duplicate records of broods counted earlier from ones counted later.
Two brood counts were made each year. The first was in early July when some ducklings on the area were known to be Age Class II-A or II-B, and one in early August. In 1965 and 1967 three brood counts were made. Incidental brood sightings also were recorded each summer. The annual total of broods for all species was the sum of flightless broods plus distraction-displaying hens minus duplicate observations among surveys. Occasionally, incidental brood observations were added to the total when a brood was known to have hatched after the last survey, as was the case in several years for ruddy ducks.
Only data from the regular brood counts were used in analyses of brood count and wetland data. Incidental brood observations were not used in these analyses and no adjustment was necessary for duplicate observations because of the time span between counts. An index of the annual rate of brood production (hen success) was calculated by dividing the estimated number of broods by the estimated number of pairs for each species.