Northern Prairie Wildlife Research Center
Radiotelemetry. During March and April 1978 and 1979, cranes were live-trapped with rocket and cannon-projected nets positioned near groups of mounted crane decoys in cropland and pastures. To minimize disturbance, nets were fired from blinds located several hundred m from decoys. Each captured bird was weighed, aged as juvenile or adult by head plumage (Lewis 1979), and banded. Battery-operated transmitters weighing approximately 40 g (< 2% of body weight) were attached to 13 cranes in 1978 and 23 in 1979, using a neck and body loop and backpack harness. Ten birds in 1978 (8 adult, 2 juvenile) and 14 in 1979 (13 adult, one juvenile) were located frequently enough to permit statistically valid analyses of activity ranges and movements.
When feasible, each radio-equipped crane was located at hourly intervals during daylight and once each just after dusk and before dawn. Locations were determined by triangulation from ground vehicles to the nearest 100 m. Fixed-wing aircraft were used when birds could not be located from the ground. No individual was followed in more than one year.
Distances and angles between locations were calculated along straight lines. These distances represent minimal distances because cranes meandered between radio fixes. Birds that were located in the same habitat on two sequential observations were assumed to have remained in that habitat for the interval, and those located in different sites were assumed to have visited only those in which they were recorded.
Cranes center their activities within definable, undefended areas that may change in size and location through time. These "activity ranges" differ from conventional home ranges in that they are transient and occupied only for roosting and feeding. Areas of activity ranges were estimated with a harmonic means method (Dixon and Chapman 1980) including the 95% closest points. When cranes used two or more discrete (as determined by non-overlapping clusters of locations separated by at least 2 km) activity ranges, separate areas were calculated for each.
Time budgets. Diurnal time budgets were developed within each of the habitats by recording activities of individual cranes at 12-sec intervals throughout the staging period. Time spent observing in each habitat was proportional to the percent of study area covered by that habitat. Individuals were selected arbitrarily by locating a group of cranes in a spotting scope's field of view and, after looking away and slightly moving the scope horizontally and vertically, observing the individual nearest to the intersection of the cross hairs. Observations on an individual lasted from a few sec to 5 min. Observations occurred from 06:00 to approximately 18:00 h CST. Behavioral categories included "resting" (sitting or standing still), "feeding", "alert", "calling", "courting", "aggression" (fighting or being attacked), "locomotion" (walking), and "comfort" (preening). An "unknown" category contained observations that could not be classified as one of the other behaviors.
Statistical analyses. Unless otherwise stated, all values are means ± SD. Statistical treatment of distances flown, either among habitats or within a day, were complicated both by missing cells in some cases and repeated observations from marked individuals. We used repeated-measures analysis of variance (Milliken and Johnson 1984) on these data. Angular statistics and correlations were conducted following Batschelet (1981). Behavioral frequencies based on counts were analyzed with G-tests (Sokal and Rohlf 1969). Sequential behaviors may not have been truly independent, but we believe that observations were adequately dispersed over birds, groups, and contexts to warrant this type of statistical treatment.