3 9999 06317 797 4 LESSER AND CANADIAN SANDHILL CRANE POPULATIONS, AGE STRUCTURE, AND HARVEST Special Scientific Report— Wildlife No. 221 MODELING SANDHILL CRANE POPULATION DYNAMICS Special Scientific Report-Wildlife No. 222 DEPOSITORY UNITED STATES DEPARTMENT OF THE INTERIOR FISH AND WILDLIFE SERVICE Library of Congress Cataloging in Publication Data Buller, Raymond J. Lesser and Canadian sandhill crane populations, age structure, and harvest. (Special scientific report— wildlife ; no. 221-222) 1. Sandhill crane. 2. Sandhill crane shooting. 3. Bird populations- North America. 4. Birds— North America. I. Johnson, Douglas H. Modeling sandhill crane population dynamics. 1979. II. Title. III. Series. SK361.A256 no. 221-222 [QL696.G84] 639'.9'08s [598.31] 79-17226 NOTE: Use of trade names does not imply U.S. Government endorsement of commercial products. LESSER AND CANADIAN SANDHILL CRANE POPULATIONS, AGE STRUCTURE, AND HARVEST Special Scientific Report— Wildlife No. 221 by Raymond J. Buller MODELING SANDHILL CRANE POPULATION DYNAMICS Special Scientific Report — Wildlife No. 222 by Douglas H. Johnson UNITED STATES DEPARTMENT OF THE INTERIOR OT0f>. FISH AND WILDLIFE SERVICE VS A J 12k Washington, D.C. • 1979 Preface The impact that hunting may have on populations of sandhill cranes {Grus cana- densis) has been a subject of some controversy in recent years (Sherwood 1971; Miller et al. 1972; Miller 1974; Miller and Botkin 1974). Much of the debate stems from a lack of information about the population dynamics of sandhill cranes and the numbers taken by hunters. This prompted the U.S. Fish and Wildlife Service to initiate new studies aimed at a better understanding of sandhill cranes and their management needs, particularly in the Central Flyway where most of the crane hunting occurs in the United States. The results of one of these studies, conducted from 1974-77 (R. J. Buller, in coopera- tion with biologists from the Canadian Wildlife Service and several States and Cana- dian Provinces) constitute Special Scientific Report— Wildlife 221. Special Scientific Report— Wildlife 222 contains an analysis of data emanating from Buller's study, and other recent work on sandhill cranes, in the form of a series of mathematical models of sandhill crane populations. These studies and analyses provide new and useful information about sandhill cranes in the Central Flyway and should be of interest to those concerned about the welfare of these birds. It is apparent, however, that additional efforts are needed to develop reliable information on population size. Until this is done and until satis- factory measurements of other key population characteristics are obtained, mathe- matical models for sandhill crane populations must be regarded as preliminary and incomplete. For this reason, the Fish and Wildlife Service intends to continue with its studies and investigations of sandhill cranes. In addition to more intensive population sur- veys, a number of other studies are currently planned or under way. These are focused on breeding biology and behavior, breeding age, population age structure, social organization, distribution and movement patterns, sex ratios, feeding behavior, and habitat requirements. John P. Rogers Chief, Office of Migratory Bird Management LESSER AND CANADIAN SANDHILL CRANE POPULATIONS, AGE STRUCTURE, AND HARVEST Lesser and Canadian Sandhill Crane Populations, Age Structure, and Harvest1 by Raymond J. Buller Central Flyway Representative U.S. Fish and Wildlife Service P. O. Box 1306 Albuquerque, New Mexico 87103 Abstract Lesser (Grus canadensis canadensis) and Canadian (G. c. roivani) sandhill cranes were studied from 1974 through 1977 in portions of the Central Flyway and Saskatchewan, Canada. The primary purposes of the study were to obtain estimates of (1) the lesser and Canadian sandhill crane populations during the fall and spring, (2) percent of juveniles in the population, and (3) hunting pressure, harvest, and crippling loss. Visual estimates of the numbers of cranes occur- ring within the main fall staging areas and on the principal wintering grounds were made during periods of peak migration each October, and on the Platte River in Nebraska each spring. Feeding, roosting, and flying sandhill cranes were aged throughout the study area during fall and winter with the aid of binoculars and spotting scopes. Hunters were contacted in the field in 1974 to obtain information on harvest. After the three subsequent hunting seasons, 1975-76, 1976-77, and 1977-78, questionnaires were mailed to holders of crane hunting permits in an effort to deter- mine hunting pressure, harvest, and crippling loss. Results of fall, flyway-wide surveys varied from about 94,900 to 219,700 cranes and continuation of the surveys is not recommended because it is practically impossible to consistently select a time period that will coincide with the peak of migration and to avoid logistical problems. The spring inventories indicated 153,800 to 227,500 cranes in the various years and this spring inventory holds the greatest potential for assessing the total size of the lesser and Canadian sandhill crane population. The proportion of juveniles averaged 11.6%, and 20% of the successful breeders were accompanied by "twin" young. The total kill of sandhill cranes in the conterminous United States reached 14,170 in the 1977-78 hunting season. This is a report on the results of a special study, con- infrared, and white light photography) project to ob- ducted during the autumn of 1974 through the spring tain population estimates of cranes along the Platte of 1977, of lesser (Grus canadensis canadensis) and River in Nebraska during spring are also included. Canadian (G. c. rowani) sandhill cranes in relation to Sandhill cranes have been hunted in one or more hunting in the Central Flyway. The primary purposes designated areas of the Central Flyway in the United of the study were to obtain estimates of (1) the lesser States since 1 January 1961, when a 30-day season and Canadian sandhill crane populations during the was permitted in eastern New Mexico and western fall and spring, (2) percent of juveniles in the popula- Texas. Before that time hunting had not been per- tion, and (3) hunting pressure, harvest, and crippling mitted in the United States since 1916. In 1967 hunt- loss. Incidental information was also obtained con- ing was permitted in the Central Flyway portion of cerning the application of the family group technique, Colorado, exclusive of the San Luis Valley and, in the already used for winter appraisals of productivity following year, in western Oklahoma, the eastern por- among arctic-nesting geese (Lynch and Singleton tion of the Texas panhandle, and prescribed areas of 1964), to evaluate annual production among cranes. North and South Dakota. In 1972 hunting was per- Results of a remote sensing (thermal imagery, near mitted in prescribed areas of Montana and Wyoming. The birds have been legally hunted in Mexico at least 'Final report in compliance with Contract 14-16-0008-885. since 1940, and in portions of Canada since 1959. Information about the 1961 seasons in New Mexico and Texas was reported by Boeker et al. (1961, 1962). Information developed during a 4-year cooperative study, 1962-65, which preceded hunting seasons in other parts of the Central Flyway was reported by Buller (1967). Information from the present study sup- plements that provided by these earlier investigations. Study Area The present study was conducted in those portions of Saskatchewan and the States of the Central Flyway that represent important fall and spring staging areas, and wintering areas for sandhill cranes. Specifically, these included the Last Mountain-Kutawagan-Quill Lakes complex of south-central Saskatchewan and the Outlook-Kyle, Kindersley district of west-central Sas- katchewan; McLean, Pierce, Kidder, and Stutsman counties, North Dakota; the Pollock area, South Dakota; the Platte River Valley, Nebraska; southwest- ern Oklahoma; Bitter Lake and Grulla National Wild- life refuges, New Mexico; and the South and West Plains regions of Texas, including Muleshoe National Wildlife Refuge (Fig. 1). The primary study areas were harvested grain fields, pastures, unharvested sorghum and cotton fields, and roost sites (lakes, sloughs, impoundments, and playas) from Saskatchewan to Texas. Methods Population Size Visual estimates of the numbers of cranes occurring within the main fall staging areas and on the principal wintering grounds were made during periods of peak migration each October, and on the Platte River in Ne- braska each spring. Most of the estimates were de- rived from counts by observers on the ground; how- ever, aerial surveys were made in Nebraska, North Dakota, Saskatchewan, and South Dakota. The Platte River in Nebraska is a traditional spring staging area for sandhill cranes enroute to their nest- ing ground. Moreover, all the lesser and Canadian sandhill cranes that winter in Texas, Oklahoma, New Mexico, and the Republic of Mexico appear to stop along the Platte River (Wheeler and Lewis 1973). The spring inventory, started in 1959, was originally de- signed to monitor population trends rather than pro- vide a total population count. Modifications in the cen- sus techniques in 1974 were aimed at making the cen- sus a total population count; the survey was expanded to include major crane use areas in other cooperating States of the Central Flyway. The spring survey is frequently delayed or inter- Fig. 1. Congregation sites, used by sandhill cranes during the fall, winter, and spring, that were portions of the study area. rupted by logistical problems, and must be accom- plished within a very short time span. Also, the popu- lation in the Platte River Valley is concentrated in the smallest area when the cranes are roosting in the River at night. Consequently, the U.S. Fish and Wildlife Ser- vice entered into a contract with the U.S. Army to evaluate remote sensing equipment as a nighttime cen- sus technique. A prototype mission was flown at Bitter Lake National Wildlife Refuge, New Mexico, in January 1975. Three types of sensors were utilized: a classified thermal infrared (IR) line scanner, near IR photography, and white light flash photography. The same equipment was used in conjunction with the 1975 spring survey along the Platte River. Age Determination Juvenile sandhill cranes are distinguishable from adults by size, voice, head and body plumage, eye color, and the color of the bill (Miller and Hatfield 1974; Lewis 1979a), but head plumage was the most commonly used characteristic for field determination of young birds during this study. Postjuvenile cranes have papillose skin on the crown, forehead, and loral regions which is sparsely covered with short, black, hair-like bristles and is usually red to red-orange. The forehead and crown of juvenile cranes are fully feathered and are light gray to brown (Fig. 2). Fig. 2. The head of a postjuvenile sandhill crane is reddish papillose skin sparsely covered with short, black bristles (right); the juvenile's head is fully feathered and light gray or brown (left). The postjuvenile molt is under way on the specimen in the center. (Photo by Tom Smiley, U.S. Fish and Wildlife Service) The head of a juvenile presents a rounded profile, whereas that of the adult appears flattened. A juvenile crane's head profile resembles the head profile of a red- head duck [Aythya americana); a postjuvenile's head profile resembles the head profile of a canvasback duck (A. valisineria). Moreover, the red area of the crown and forehead of the postjuvenile is distinguishable under most light conditions. Observers generally posi- tioned themselves between the sun and flocks of cranes that were to be inspected for adult/juvenile ratios, and most counts were limited to morning hours before afternoon heat waves began to distort images. Neither heat waves nor position with respect to the sun were important on cloudy, overcast days. Feeding, roosting, and flying sandhill cranes were aged during fall and winter with the aid of binoculars (7 x 34, 8 x 40) and 20x to 60x spotting scopes. Cranes feeding in fields or on their roost were aged from a vehicle or other form of concealment. Feeding and roosting cranes can usually be approached within 400 m or less in a vehicle, but will flush from an ob- server in the open at this or a greater distance. As a feeding flock is approached by a vehicle most of the birds will cease feeding, but after the vehicle is stopped they will usually move away, feeding as they go. Cranes at rest on roost sites were also aged when possible, but were sometimes too densely congregated. The ideal roost situation for making age composition counts was encountered at midday and early afternoon when cranes were spaced along the shoreline. Cranes in flight were aged as they moved between their roost and primary feeding area. Others were aged as they came to watering areas following the morning feeding period or alighted at a feeding area as singles, pairs, family groups, or small flocks. In Saskatchewan, age composition data were irregu- larly recorded in 1974; however, in 1975 these data were recorded weekly at Last Mountain Lake and Quill Lakes. In 1976, age data were also recorded weekly at Last Mountain Lake and Kutawagan Lake. Elsewhere in Saskatchewan, age counts were irregularly recorded from mid- August through early November. In the United States, age structure was irregularly recorded from early September through the 3rd week of January. Since Lewis (1974, 1979a) noted that adult/juvenile ratios collected in December and January probably underestimate the true percentage of juveniles in the population, only those birds aged through November were used to determine adult/juve- nile ratios. Furthermore, R. Drewien (personal com- munication) advises that some color-marked greater sandhill cranes have acquired the unfeathered heads of postjuveniles as early as November. Therefore, adult/juvenile numbers tallied in December and January were eliminated from the estimates of annual recruitment. Limiting age counts to the August- November period reduced the total number (207,902) of sandhill cranes aged by 11,668 birds. The family group technique, used for winter ap- praisals of productivity in arctic-nesting lesser snow geese {Anser c. caerulescens) and white-fronted geese (A. albifrons frontalis), was evaluated during the col- lection of age data in eastern New Mexico and west Texas in October 1974. Family group counts were ob- tained in those situations where birds were alighting near water or moving to feeding sites. Hunting Pressure, Success, and Crippling Loss Hunters were contacted in the field during 1974 to obtain information on their numbers, success, and numbers of birds knocked down but not retrieved. This method proved to be unsatisfactory and a special sandhill crane hunting permit system was initiated during the 1975-76 hunting season. The permits were supplied to the States by the U.S. Fish and Wildlife Service and were issued free of charge to hunters upon request. Each permit holder was mailed a questionnaire at the close of the hunting season. The questionnaire included inquiries about the number of days hunted, numbers of cranes harvested, numbers crippled, and counties hunted. One followup questionnaire was mailed to nonrespondents about 3 weeks after the first mailing. Nonrespondents to the followup were assumed to have the same average hunt- ing activity and harvests as respondents, and reported harvests were expanded accordingly. Estimates of the number of successful crane hunt- ers, the mean seasonal bag per successful hunter, and the total crane bag were also derived from responses to the annual Federal Waterfowl Hunter Questionnaire Survey (Sorensen and Reeves 1976; Sorensen 1977, 1978).' Results and Discussion Fall Populations Population surveys were conducted at all important congregation sites 30 October- 1 November 1974, 20-24 October 1975, and 27-29 October 1976. These dates generally coincided with the peak of the fall migration but each survey was hampered by weather or logistical problems. Inclement weather caused a delay in the 1974 survey in Saskatchewan and North Dakota until 5-7 November, and logistical problems caused a delay in the 1975 survey in these areas until 6 November. Each year, large numbers of sandhills were noted in migration in Saskatchewan, Montana, Wyoming, Colorado, and Nebraska, and it is not known if these birds were tallied in New Mexico and Texas before the survey was concluded. During the 1974, 1975, and 1976 fall population sur- veys, observers tallied 201,100, 94,886, and 219,707 sandhill cranes, respectively (Table 1). These represent minimum estimates of the fall population because (1) it is doubtful that those birds in migration during the survey period were recorded (as previously noted counts were made during periods of peak migration); (2) generally only the main congregation areas were surveyed, e.g., the Platte River (Nebraska) was not in- cluded in the fall surveys, and smaller groups could have been overlooked; (3) many counts were made dur- ing the feeding period and feeding flocks are frequently overlooked; and (4) an unknown number arrived on wintering grounds in the Republic of Mexico before the surveys were conducted. Attempts to enumerate the fall population of sand- hill cranes in the Central Flyway were only partially successful. Moreover, it is likely that efforts to survey sandhill cranes for total population purposes during the fall never will prove satisfactory. It is difficult to preselect a survey period that will coincide with ideal weather conditions and a time when most of the cranes Table 1. Results of coordinated fall population surveys of sandhill cranes in the Central Flyway, 1974-76. Numbers of cranes Province or State 1974 1975 1976 Alberta 800 Saskatchewan 32,600 58,000 45,952 Montana — 3,000 — North Dakota 11,120 10,000 33,075 South Dakota 7,515 12.000 20.000 Colorado — 5,435 3,008 Kansas — — 5 Oklahoma 4,767 587 4,850 New Mexico 12,700 2,175 51.863 Texas 132,398 2,889 60,954 Total 201.100 94,886 219,707 are available to be counted. Furthermore, it is difficult to assemble the same survey effort from year to year. Two of the three coordinated fall population surveys resulted in population estimates of about 201,000 and 220,000. The 1975 survey was 7 to 10 days earlier than the 1974 and 1976 surveys, and more than twice as many cranes were counted in the latter 2 years. Several factors may have caused the low count in 1975. The survey period may have been too early, weather conditions may not have prompted the major migra- tion, or the bulk of the population may have been in migration, and as a consequence, not included in the tally. The low numbers of sandhill cranes tallied in New Mexico and Texas during the 1975 fall survey are further evidence that the census was too early. Pre- hunting-season population surveys have been con- ducted in these two States since 1960 (exclusive of 1973) and these surveys have yielded population esti- mates ranging from about 135,000 to 339,000 birds (Table 2). Thus, the 1975 fall population estimate was less than any of the previous preseason populations in New Mexico and Texas. The magnitude of that segment of the sandhill crane population which winters in the Republic of Mexico is also unknown. A field officer of the New Mexico De- partment of Game and Fish reported that while on patrol in the Guadalupe Mountains southwest of Ros- well in the latter part of October 1975, he daily ob- served large flocks of sandhills moving in a southerly direction. These birds were not recorded at Bitter Lake National Wildlife Refuge, nor were they tallied in the Dell City area, Trans Pecos region, Texas. Thus, it is likely these flocks were enroute to wintering grounds in Mexico. The sighting of these migrating flocks over the Guadalupe Mountains might indicate that the 1975 fall population survey was too late; however, the low Table 2. Results of preseason population surveys of lesser and Canadian sandhill cranes in eastern New Mexico and west Texas, 1960-72 and 1974-76. Table 3. Results of spring sandhill crane population surveys in Nebraska, 1959-73, and expanded spring surveys in 1974-78. No. of No. of No. of No. of Date cranes Date cranes Date cranes Date cranes 14-15 December 20 March 1959 147,496 2 April 1969 154,978 1960 134,673 30 October 1968 339,185 4 April 1960 125,870 26 March 1970 193,600 3 November 1961 147.416 31 October 1969 258,500 21 March 1961 136,276 28-29 March 1971 207,500 2 November 1962 180,901 30 October 1970 210,200a 21 March 1962 142,830 27-28 March 1972 183,600 25 November 1963 207,405 29 October 1971 147,416" 21 March 1963 101,925 26-29 March 1973 195,350 3-6 November 1964 213,896 27 October 1972 184.9013 30 March 1964 156,028 25-31 March 1974 177,015" 29 October 1965 198,027 25 October 1974 136.340 30 March 1965 80,315 25-30 March 1975 227,527 28 October 1966 139.199 22-23 October 1975 4,710 25 March 1966 123,087 25-26 March 1976 153,784 3 November 1967 210,074 27-29 October 1976 110,964 23 March 1967 123,043 17-19 March 1977 219,154 Mnromnlete survev 22 March 1968 169,194 22-24 March 1978 159,898 count (4,710) recorded in eastern New Mexico and west Texas probably resulted from a count that was made too early rather than too late. Spring Inventory The spring population surveys of sandhill cranes associated with the Platte River in Nebraska, which have been conducted since 1959, were expanded to in- clude other States in 1974. The expanded surveys were continued during the current study; however, the coverage has been highly variable (Table 3). Since the spring survey was expanded in 1974, 79-99% of the cranes tallied (171,570; 225,945; 150,119; 174,575; and 152,021 in 1974, 1975, 1976, 1977, and 1978, respec- tively) have been in the Platte River Valley. The 1975 survey was scheduled for 23 March; how- ever, spring snowstorms delayed the count along the Platte River until 29 and 30 March. The 1976 survey was conducted on 22-26 March, the 1977 survey was made on 17-19 March, and the 1978 survey on 22-24 March. During this study, observers recorded 227,527, 153,784, and 219,954 sandhill cranes during the spring survey. On the Platte River the 225,945 birds recorded in 1975 were the highest population tallied since the spring surveys were initiated. Observers recorded 159,898 sandhill cranes during the 22-24 March 1978 survey. During the prototype mission in New Mexico to test nighttime census techniques, cranes were separable as individuals under both test pen and natural conditions by all three sensor techniques. However, the semi- operational mission was delayed by equipment mal- functions, a series of blizzards accompanied by low temperatures, and slush ice in the River (Munro and Lewis 1976), and only two flights were possible in the study period 23 March to 3 April 1975. The first was flown on the evening of 31 March aIn addition to the Platte River Valley, surveys were made simultaneously at major crane use areas in other States be- ginning in 1974. These cooperating States were: 1974, the Dakotas, Kansas, Oklahoma, and Texas; 1975, as 1974 plus Colorado, Montana, New Mexico, and Wyoming; 1976, Kansas, Oklahoma, and Texas; 1977, Oklahoma, Texas, South Dakota, and partial surveys in Kansas, Nebraska, and New Mexico; 1978, Kansas, New Mexico, South Dakota, Texas, and incomplete participation data received from Colorado, Nebraska, and Oklahoma. under marginal flying conditions and examination of the processed films revealed three problems. The film magazine of the thermal IR scanner had frozen, the en- tire width of the river channel was not being photo- graphed in a single pass because the plane was too low, and cranes were flushing as the aircraft flew over. The flushing behavior of roosting cranes was unantici- pated. During the test mission at Bitter Lake National Wildlife Refuge, one roost was overflown 10 times in 2 hours without any significant movement by the cranes. On the Platte River, cranes flushed from the roost as the plane passed over, even when the white light photography, with its associated light flash, was not used. The scanner magazine was replaced on the 2 April flight and the altitude was increased to permit cover- age of the entire width of the river channel. The in- creased altitude proved to be critical. Certain features of the scanner were influenced by the increased cover- age of snow and ice, causing a loss in contrast. The effi- ciency of both the near IR and white light photog- raphy was affected by less intense light flash at the target surface which caused a loss of contrast. All images acquired on the 2 April flight were uniformly unsuitable. All sensors provided acceptable results under test conditions but the semioperational mission was not successful. Munro and Lewis (1976) concluded that prospects for nighttime remote sensing of sandhill cranes on the Platte River are not good for the follow- ing reasons: (1) restricted availability of equipment, (2) associated high costs, (3) limited width of coverage, (4) likelihood of bad weather during the survey period, and (5) the aircraft frightening the cranes from a sur- veyed strip to a parallel unsurveyed strip with the con- sequence that some birds might be counted twice. Failure of the nighttime remote sensing technique to yield an estimate of the total spring sandhill crane population on the Platte River in 1975 prompted Lewis (19796) and two assistants to make ground counts of those cranes that departed from specific roosts during the 1976 survey before the aircraft passed over. These ground counts were then used to correct the results of the aerial count and provide an estimate of the total population of sandhill cranes on the Platte River. Based on ground counts, which indicated that 3.1 to 99.0% of the cranes had left various sections of the River before the aerial crew arrived, Lewis (19796) con- sidered the minimum population on the Platte to have been 300,000 and said the total population may be close to 400,000. Attempts to replicate this procedure in 1977 were unsuccessful. The ground counts were made but a series of storms, a dead aircraft battery, and finally a mass exodus of the cranes prevented the late March combined aerial-ground inventory (Lewis 19796). Although the annual spring inventory was found to be inadequate, e.g., failure to (1) count when the popu- lation peaks in Nebraska, (2) count all cranes in the Platte River Valley, and (3) count all cranes north and south of Nebraska (Lewis 19796), it holds greater potential for enumerating the total population of lesser and Canadian sandhill cranes than does the fall population survey. Age Composition The proportion of juvenile sandhill cranes recorded during the study (Table 4) represents the combined counts of cranes at feeding fields, on their roosts, or flying between roosts and fields. The first migrants to reach south-central Saskatche- wan in August generally are accompanied by fewer young than those arriving in September and October. For example, August arrivals included 2.9% young in 1975 whereas juveniles made up 1 1.5% of the migrants in September and 13.2% in October. During the 3-year study, juveniles constituted 13.9, 10.3, and 12.1%, respectively, of the cranes in south- central Saskatchewan. In west-central Saskatchewan, young made up 10.7 to 11.4% of the population. Collec- tively, juveniles averaged 11.4% in Saskatchewan. In the Last Mountain-Kutawagan-Quill Lakes region, juveniles were almost 3 times as common (11.7 versus 4.3%) as recorded by Miller and Hatfield (1974) in 1966, 1967, 1972, and 1973. There is an inherent unknown degree of error asso- ciated with estimating annual recruitment from age structure. The frequency with which the same flocks are sampled is unknown because the rate at which birds are replaced by new arrivals is unknown. More- over, each flock may not be adequately sampled as it utilizes staging or wintering areas, nor is it possible to weight the age structure data because estimates of the numbers of cranes were not made when adult/juvenile counts were made. The age data indicate that collectively lesser and Canadian sandhill cranes experienced better produc- tion during 1974-76 than that recorded by Miller and Hatfield (1974) in 1966, 1967, 1972, and 1973. Why was the average annual recruitment rate during the present study nearly 3 times that reported by Miller and Hatfield? One possibility is that production was exceptionally good during 1974-76 whereas 1966, 1967, 1972, and 1973 were poor years. Sandhill cranes nest over a broad area including Siberia, Alaska, arctic Canada, and Baffin Island, as well as the subarctic where weather is a less important limiting factor. Broods are small, breeding is delayed until age 3 or older, and the subadult population component is large. Therefore, it seems unlikely that the age ratio in the population changes much from year to year. Miller and Hatfield (1974) noted their data were based on samples from about 10% of the cranes that migrate through the Central Flyway. They suggested that early migrants are mostly nonbreeders and that early age ratios are therefore unreliable. For example, the weekly proportion of juveniles ranged from 0.1 to 3.2% from 13 August to 16 September 1972, and 4.2 to 6.5% from 26 August to 8 September 1973. Analysis of age data of the present study by weeks (1975 and 1976) also indicates that the proportion of young increases as the fall migration progresses (Fig. 3). The 1974 age composition data were irregularly recorded by 10-day periods; thus, these data cannot be analyzed by weekly periods. Miller and Hatfield (1974) limited their observations to the east-central portion of Saskatchewan, whereas observations during the present study were much more extensive. Furthermore, they counted only flying flocks in 1972 and 1973 because they considered these gave the least biased estimate of age structure. In 1975, weekly age tallies recorded during the present study at Last Mountain Lake ranged from 0.8 to 14.2% young between 17 August and 27 September, and averaged 9.2%. In 1976, weekly age compositions collected in the Last Mountain-Kutawagan-Quill Lakes region ranged from 6.8 to 13.1% young between 22 August and 25 September, and averaged 12.1%. Possibly the low percentages of young recorded by Miller and Hatfield (1974) are related, in part, to their decision to use age data collected only from flying Table 4. Percentage of young sandhill cranes in the Central Flyway, 1974-76. (Sample size in parentheses.) August September October November Total Area 1974 1975 1976 1974 1976 1975 1976 1974 1975 1976 1975 3-year 1976 average Saskatchewan Eastern Western Total North Dakota McLean Co. Pierce Co. Kidder Co. Total South Dakota Pollock Montana Bowdoin NWR Colorado Two Buttes Reservoir Oklahoma Washita NWR New Mexico Bitter Lake NWR GruUa NWR Columbus Total Texas West Plains South Plains Lower Coast Total United States Total Grand Total 9.2 2.9 12.0 14.1 11.5 11.9 (54) (3,017) 11,665) (895) (5.759) (26.177) 5.9 20.3 9.9 11.5 7.4 10.5 (3561 (350) 1956) (2.792) (4,244) (11,846) 6.3 4.7 11.2 12.1 9.8 11.5 (410) (3,367) (2,621) (3,687) (10.003) (38,023) 7.0 12,892) 13.6 (1031 17.7 (203) 7.9 (3.198) 10.9 (3,2291 13.5 (325) 12.9 (1.331) 11.6 (4.885) 15.7 (989) 12.1 (5,580) 13.9 (677) 17.7 (2,695) 13.5 (8,952) 7.9 12.3 (3,198) (5.874) 6.3 4.7 11.2 10.1 (410) 13,367) (2.621) (6,8851 10.7 (15.877) 13.9 (8.952) 11.9 146.975) 14.6 13.2 12.2 (157) (5,474) (25.838) 10.7 13.3 12.2 (9,506) (3,954) (17.938) 10.7 13.2 12.2 (9,6631 (9.428) 143.776) 15.5 (5731 15.5 (5731 14.3 (927) 22.6 (208) 17.4 (2581 7.2 (3,476) 10.0 (3,417) 8.6 (6,893) 8.0 (3,408) 8.6 (1,553) 18.3 (525) 9.2 (5.486) 13.9 10.3 (1.106) (14.250) 11.2 10.7 10.7 10.7 (1,288) (3,774) (12,654) (9,836) 11.2 10.7 11.0 10.5 (1,288) (3,774) (13,760) (24.086) 15.5 (8501 18.8 (2.859) 7.4 13,000) 28.8 (205) 7.5 (3.391) 13.9 (1.892) 9.8 (5,283) 15.1 (284) 15.1 (284) 10.4 (1.343) 18.8 (570) 15.5 10.4 18.8 (850) (1.343) (570) 23.0 14.2 16.8 (1,313) (1,190) (113) 10.0 (2,226) 10.0 (2.226) 12.8 (2,431) 10.5 (3.0591 26.7 (319) 6.6 (761 6.6 (76) 18.8 10.6 (175) (1.112) 13.6 11.8 (110) (2.365) 12.4 (608) 12.7 11.8 (718) (2.365) 4.9 10.7 (1,404) (1.270) 11.3 14,205) 24.3 12.6 12.6 (1.028) (182) (245) 11.5 13.1 11.2 12.6 (5,490) (2.432) (5,657) (245) 7.0 (2,8921 13.6 (103) 17.2 (1.346) 10.4 (4,341) 14.3 (2.117) 22.6 (208) 22.5 (577) 7.2 (3,552) 10.0 (3,417) 8.6 (6.969) (4 7.2 12) 8.6 (1,5531 22.3 (1.553) 10.4 (7.918) 10.7 (1,270) 11.3 (4.2051 14.2 (466) 11.4 (5.941) 9.9 12.1 12.0 14.9 12.0 11.5 10.6 12.0 (14,345) (12,481) (10,372) (4,587) (6,6631 (3,722) (22,1301 (25,018) 12.1 11.7 (53,680) 11.4 11.1 (34.5141 11.8 11.4 (88,194) 11.8 10.9 (6.9231 13.9 13.8 (677) 17.7 15.9 (2.695) 13.5 12.5 (10.295) 23.0 17.8 (1.313) 11.8 (4.079) 13.5 (325) 12.9 (1.331) 12.2 (5,735) 17.9 (3.961) 7.4 (3.0001 24.2 10.6 16.4 (380) (1.112) 10.9 14.591) 7.7 (3,501) 13.9 (1.892) 12.4 (608) 10.7 10.9 (6,001) (4,591) 10.2 (24.008) 12.6 (21.909) 12.3 14.9 11.6 11.1 10.7 (54,148) (4,5871 (7,9511 (7.4961 (35,890) 11.3 (49,1041 12.8 (2.431) 10.6 (3,059) 12.6 (245) 11.6 (5,735) 12.9 (23,046) 12.0 (111,240) 9.3 19.6 11.1 11.6 33 34 35 36 3? 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 WEEK OF THE YEAR AUG. SEPT OCT. NOV. DEC. Fig. 3. Average weekly sandhill crane age ratios in the Cen- tral Flyway, 1975 and 1976. flocks in 1972 and 1973. During the present study, I aged flying birds at every opportunity. It was essen- tial that I position myself at right angles to the flight path, but not directly under it. Juvenile birds are sometimes difficult to distinguish when viewed from directly below, especially in flocks of 15-20 birds, and when the flocks are passing the observation point in rapid succession. There also is some evidence that the sandhill cranes which stage in the Last Mountain-Kutawagan-Quill Lakes region may be from a different section of the breeding range than those that stage in west-central Saskatchewan (Gollop 1976). Furthermore, Miller and Hatfield (1974) sampled during August and September only, whereas I sampled from the time the first mi- grants arrived, through the fall, and into the winter months. During the present study the proportion of juveniles averaged 11.6% (Table 4). Age composition varied greatly within the same population segment during the present study. For example, the proportion of young among the cranes that remained during the day on Bitter Lake and Muleshoe National Wildlife refuges was always less than the incidence of young among the birds that departed the refuges to feed. Thus, combin- ing all counts rather than utilizing only those counts of feeding flocks, flying flocks, or cranes on roost sites provides a more realistic estimate of the annual re- cruitment among lesser and Canadian sandhill cranes in the Central Flyway. Moreover, the proportion of young was consistently greater at some staging areas; e.g., Kidder County, North Dakota; Pollock area, South Dakota; Washita National Wildlife Refuge, Oklahoma; Aransas and Laguna Atascosa National Wildlife refuges, Texas; and Grulla National Wildlife Refuge versus Bitter Lake National Wildlife Refuge, New Mexico. Age counts can also be used to derive separate esti- mates of production in the Canadian and lesser sub- species. Cranes collected in Kidder and Pierce coun- ties. North Dakota, during the 1962-65 study (Buller 1967) and by Johnson and Stewart (1973) in 1970 and 1971 primarily represent the Canadian race. The Cana- dian race is also prevalent among sandhill cranes win- tering along the Texas Coast (Guthery 1972; Lewis 1974; Guthery and Lewis 1979). Sandhill cranes col- lected in McLean County, North Dakota, and the Pollock, South Dakota, area were predominately lessers (Buller 1967; Johnson and Stewart 1973). If the same pattern of distribution persisted during the present study, then annual recruitment averaged 16.4% in the Canadian race and 14.3% in the lesser race. A detailed examination of nearly 2,000 cranes in 756 small groups (1 to 4 birds), using the family group technique, indicated 222 families averaging 1.16 young per family (Andrews 1975). Although family group counts can be obtained simultaneously with sandhill crane age composition data, the technique has some- what less utility than when applied to goose popula- tions because most families contain only one young, and subadults cannot be distinguished from adults. Although Miller (1973) indicated that successful breeders rarely raise more than one young, noting in 623 family groups only one instance of a family with two young, the present study indicated that 16.9% of the successful breeders were accompanied by "twin" young in 1974. In 1975, twin families were recorded at the rate of 23.5% at Lake Williams, McLean County, North Dakota. At Last Mountain Lake, Saskatche- wan, 25% of the successful breeders had two young. In 1976, 17% of 703 successful pairs were tallied with twin young at Last Mountain Lake through 30 Sep- tember; in the United States, 24.8% of the successful pairs were accompanied by twin young. Thus, an aver- age of 20.0% of the successful pairs were accompanied by twin young during the present study. Hunting Pressure, Success, and Crippling Loss Sandhill crane hunters were contacted during the 1974 hunting season to obtain an estimate of their numbers and success. This proved to be an unsatisfac- tory method of obtaining hunter numbers because the survey effort could not be proportional to the hunting effort throughout the season, sample sizes were small, and the only information collected in Texas came from a "day-shoot" facility adjacent to Muleshoe National Wildlife Refuge. Although this method of deriving hunter numbers was unsuccessful, some data on suc- cess, vulnerability of juveniles, and crippling loss were obtained (Table 5). Table 5. Incomplete sandhill crane hunting statistics. Central Fly way, 1974. No. birds bagged No. birds unretrieved State No. hunters Knocked contacted Adult Juvenile down Hit Colorado Montana New Mexico North Dakota South Dakota Texas Wyoming Totals 12 18 95 12 21 293 5 456 37a 27 6 5 347 5 393 37a 6 1 163 1 181 10 6 2 0 19 aAge composition unavailable. The limited sandhill crane hunting statistics ob- tained during the 1974 season indicate that hunters averaged 1.3 birds each, bagged 2.2 more adults than juveniles, and reported a crippling loss of about 1%. The mail survey of the holders of sandhill crane hunting permits indicated that the total bag ranged from approximately 7,400 to 12,150 during the study (Table 6). Adding to the bag the estimated crippling loss indicated the total kill ranged from about 8,800 to 14,200 cranes. The estimated average crippling loss during the three hunting seasons was 15% of the total kill (18% of the bag). Fifty nine, 45, and 58% of the per- mit holders indicated that they hunted at least once in 1975-76, 1976-77, and 1977-78, respectively. Only 40, 34, and 43% of those who hunted bagged one or more cranes. Total days of hunting varied from about 18,800 to 25,400 in the three hunting seasons (Table 6). The 1975-76 waterfowl hunter survey provided a 25% larger estimate of the harvest than did the survey of sandhill crane hunters. The 1976-77 and 1977-78 waterfowl hunter surveys gave estimates of the har- vest that were 13% and 38% larger. The waterfowl hunter surveys also overestimated the number of suc- cessful hunters; e.g., 8.6, 62.9, and 78.3% more than surveys of sandhill crane hunters following the 1975, 1976, and 1977 hunting seasons, respectively. Estimates based on the special sandhill crane hunt- ing permit system during the 1975, 1976, and 1977 sea- sons in the Central Flyway, the Saskatchewan De- partment of Tourism and Renewable Resources mail surveys in 1964-75, and responses to the waterfowl hunter survey in Alaska, indicate a minimum average annual sandhill crane harvest of about 13,000 birds. Adding the harvest in Mexico, and by Indians and Eskimos in Canada and Alaska, the annual harvest, in- cluding crippling losses, may now be close to 18,000 cranes annually (see Johnson 1979). Table 6. Sandhill crane hunting statistics. Central Fly- way, 1975-76, 1976-77, and 1977-78. (Sorensen, per- sonal communication; Sorensen and Reeves 1976; Sorensen 1977). Year Hunting statistics 1975-76 1976-77 1977-78 Number of permits issued Number of mail question- naire respondents Response rate Number of active hunters3 Respondents Expanded Active hunter means Days hunted Crane bag Crane loss Expanded estimates Hunter days Crane bag Crane loss Crane kill 11,863 11.352 13,800 9,320 78.56 5,476 6,949 3.4 1.4 0.27 8,909 78.4 3,978 5,092 3.7 1.45 0.27 10,267 74.3 6,018 8,048 3.2 1.5 0.25 23,646 18.780 25,412 9,497 7,393 12,151 1,885 1,384 2,013 11,382 8,778 14,167 aHunted at least once. Management Recommendations The flyway-wide fall population survey should be discontinued because it is practically impossible to consistently select a time period that will coincide with the peak of migration and avoid logistical problems. However, the U.S. Fish and Wildlife Service and the conservation agencies of New Mexico and Texas should consider continuing the preseason population survey which was initiated in 1960. The expanded spring sandhill crane population sur- vey should be continued annually by the Office of Migratory Bird Management, U.S. Fish and Wildlife Service. Efforts such as those initiated in the springs of 1976 and 1977 by Lewis (19796) to obtain data to correct aerial estimates of sandhill cranes on the Platte River in Nebraska should be continued. The spring population survey holds the greatest potential for assessing the size of the total lesser and Canadian sandhill crane population as it stages on the Platte River enroute to the breeding grounds. There is little justification to collect adult/juvenile ratios annually because an adequate sample was col- lected during the present study. However, age data should be collected every 3-5 years for comparative purposes and to evaluate the premise that the average annual recruitment rate in the lesser and Canadian sandhill crane population is about 13%. The special sandhill crane hunting permit system was adopted for 3 years (1975-77), and the U.S. Fish and Wildlife Service and the Central Flyway Water- 10 fowl Council have agreed to continue the permit sys- tem for the 1978-79 season. These hunting permits should be required until biologists are confident that the annual waterfowl hunter survey is adequate to de- velop estimates of the number of successful crane hunters, the mean seasonal bag per succesful hunter, and the number of cranes bagged. Acknowledgments Many individuals participated in this sandhill crane study and the present report is possible because of the unselfish cooperation of the Central Flyway Water- fowl Council, the Central Flyway Technical Com- mittee, the Canadian Wildlife Service, the Saskatche- wan Fisheries and Wildlife Branch, and the U.S. Fish and Wildlife Service. I am especially grateful to J. B. Gollop, M. Gollop, B. Johns, and K. Lumbis, Canadian Wildlife Service; D. Dobson, D. Gray, J. Kinnear, and D. Mevel, Saskatchewan Fisheries and Wildlife Branch; C. Braun and H. Funk, Colorado Division of Wildlife; M. Schwilling, Kansas Forestry, Fish, and Game Commission; D. Witt, Montana Fish and Game Department; N. Dey, J. Hyland, and N. Lyman, Ne- braska Game and Parks Commission; J. Sands, New Mexico Department of Game and Fish; C. Schroeder, North Dakota Game and Fish Department; L. Due, Oklahoma Department of Wildlife Conservation; T. Kuck and W. Larsen, South Dakota Department of Wildlife, Parks and Forestry; H. Irby, M. Traweek, and R. West, Texas Parks and Wildlife; G. Wrakestraw, Wyoming Game and Fish Department; and R. Andrews, B. Blair, Jr., E. L. Boeker, D. Boggs, L. L. Brazell, C. Elliott, P. Ferguson, J. R. Foster, B. L. Gastineau, J. B. Giezentanner, R. Greenwood, W. Hawthorne, D. D. Heflebower, E. F. Johnson, H. Kantrud, E. F. Klett, B. D. Long, M. McEnroe, R. E. Munro, J. Nelson, T. W. Planz, W. Pratt, C. L. Ryan, J. B. Rodriquez, Jr., A. Sapa, S. Schleibe, B. Schranck, G. Unland, and H. Wittmier, U.S. Fish and Wildlife Service, for assistance in collection of population infor- mation and age ratio data. S. Carney and M. Sorensen, Waterfowl Harvest Surveys Section, U.S. Fish and Wildlife Service, made sandhill crane hunter question- naires available for preliminary estimates. M. Sorensen prepared the final estimates of successful crane hunters, average seasonal bag, total harvest, and crippling loss. References Andrews, Ft. 1975. A test of the family group technique for appraising productivity of sandhill cranes wintering in Texas-New Mexico. U.S. Fish Wildl. Serv. 14 pp. (Mimeo) Boeker, E. L., J. W. Aldrich, and W. S. Huey. 1961. Study of experimental sandhill crane hunting season in New Mexico during January 1961. U.S. Fish Wildl. Serv., Spec. Sci. Rep. -Wildl. 63. 24 pp. Boeker, E. L., W. S. Huey, and P. B. Uzzell. 1962. Study of Texas-New Mexico lesser sandhill crane hunting season- November 4-December 3, 1961. U.S. Bur. Sport Fish. Wildl. 11 pp. (Mimeo) Buller, R. J. 1967. Sandhill crane study in the Central Fly- way. U.S. Fish Wildl. Serv.. Spec. Sci. Rep.-Wildl. 113. 17 pp. Gollop, J. B. 1976. The sandhill cranes of Last Mountain Lake. Canadian Wildl. Serv. 13 pp. (Mimeo) Guthery, F. S. 1972. Food habits, habitat, distribution, num- bers, and subspecies of sandhill cranes wintering in south- ern Texas. M.S. Thesis. Texas A&M Univ., College Station. 81pp. Guthery, F. S., and J. C. Lewis. 1979. Sandhill cranes in coastal counties of Texas: taxonomy, distribution, and populations. Pages 121-128 in J. C. Lewis, ed. Proceedings of the 1978 Crane Workshop. Colorado State University Printing Service, Fort Collins. Johnson, D. H. 1979. Modeling sandhill crane population dy- namics. U.S. Fish Wildl. Serv., Spec. Sci. Rep.-Wildl. 222. 10 pp. Johnson, D. H., and R. E. Stewart. 1973. Racial composition of migrant populations of sandhill cranes in the northern plains states. Wilson Bull. 85(21:148-162. Lewis, J. C. 1974. Ecology of the sandhill crane in the south- eastern Central Flyway. Ph.D. Thesis. Oklahoma State University, Stillwater. 204 pp. Lewis, J. C. 1979a. Field identification of juvenile sandhill cranes. J. Wildl. Manage. 43(1):211-214. Lewis, J. C. 19796. Factors affecting the spring inventory of sandhill cranes. Pages 33-39 in J. C. Lewis, ed. Proceedings of the 1978 Crane Workshop. Colorado State University Printing Service, Fort Collins. Lynch, J., and J. R. Singleton. 1964. Winter appraisals of annual productivity in geese and other water birds. Pages 1 1 4-1 26 in 1 5th Annual Report, Wildfowl Trust 1962-63. Miller, R. S. 1973. The brood size of cranes. Wilson Bull. 85(4):436-440. Miller, R. S. 1974. The programmed extinction of the sandhill crane. Nat. Hist. 83(2):62-69. Miller, R. S., and D. B. Botkin. 1974. Endangered species: models and predictions. Am. Sci. 62:172-181. Miller, R. S., and J. P. Hatfield. 1974. Age ratios of sandhill cranes. J. Wildl. Manage. 38(21:234-242. Miller. R. S., G. S. Hochbaum. and D. B. Botkin. 1972. A simulation model for the management of sandhill cranes. Yale Univ. Sch. For. Environ. Stud. Bull. 80. 49 pp. Munro, R. E., and J. C. Lewis. 1976. Nighttime tests of re- mote sensors for census of sandhill cranes. Proc. Int. Crane Workshop 1:304-308. Sherwood, G. A. 1971. If it's big and it flies— shoot it. Audubon Mag. 73:72-99. Sorensen, M. F. 1977. Sandhill crane harvest and hunter ac- tivity in the Central Flyway during the 1976-77 hunting season. U.S. Fish Wildl. Serv. Administrative Report— 7 July 1977.9 pp. Sorensen, M. F. 1978. Sandhill crane harvest and hunter ac- tivity in the Central Flyway during the 1977-78 hunting season. U.S. Fish Wildl. Serv. Administrative Report— 14 June 1978.9 pp. Sorensen, M. F., and H. M. Reeves. 1976. Sandhill crane har- vest and hunter activity in the Central Flyway during the 1975-76 hunting season. U.S. Fish Wildl. Serv. Administra- tive Report— 9 July 1976. 9 pp. Wheeler, R. H., and J. C. Lewis. 1973. Trapping techniques for sandhill crane studies in the Platte River Valley. U.S. Fish Wildl. Serv., Resour. Publ. 107. 19 pp. MODELING SANDHILL CRANE POPULATION DYNAMICS Modeling Sandhill Crane Population Dynamics by Douglas H. Johnson U.S. Fish and Wildlife Service Northern Prairie Wildlife Research Center Jamestown, North Dakota 58401 Abstract The impact of sport hunting on the Central Flyway population of sandhill cranes [Grus cana- densis) has been a subject of controversy for several years. A recent study (Buller 1979) pre- sented new and important information on sandhill crane population dynamics. The present report is intended to incorporate that and other information into a mathematical model for the purpose of assessing the long-range impact of hunting on the population of sandhill cranes. The model is a simple deterministic system that embodies density-dependent rates of survival and recruitment. The model employs four kinds of data: (1) spring population size of sandhill cranes, estimated from aerial surveys to be between 250.000 and 400,000 birds; (2) age composi- tion in fall, estimated for 1974-76 to be 11.3% young; (3) annual harvest of cranes, estimated from a variety of sources to be about 5 to 7% of the spring population; and (4) age composition of har- vested cranes, which was difficult to estimate but suggests that immatures were 2 to 4 times as vulnerable to hunting as adults. Because the true nature of sandhill crane population dynamics remains so poorly understood, it was necessary to try numerous (768 in all) combinations of survival and recruitment functions, and focus on the relatively few (37) that yielded population sizes and age structures comparable to those extant in the real population. Hunting was then applied to those simulated populations. In all combinations, hunting resulted in a lower asymptotic crane population, the decline ranging from 5 to 54%. The median decline was 22%, which suggests that a hunted sandhill crane popu- lation might be about three-fourths as large as it would be if left unhunted. Results apply to the aggregate of the three subspecies in the Central Flyway; individual subspecies or populations could be affected to a greater or lesser degree. This report contains an analysis of data from values, reflecting a variety of plausible assumptions Buller s (1979) study of sandhill cranes (Grus canaden- about the real crane population. The resulting sis) in the Central Flyway and from other recent work scenarios can be viewed as potential behaviors to be on the species. The report describes a mathematical expected of the crane population under a variety of cir- model of sandhill crane populations, summarizes re- cumstances. No single result is offered as the "best" cent data that serve as input to the model, and pre- representation of the true population, sents the results of computer simulations of the model. The discussion section treats shortcomings of the „, . . , , model and data, and briefly recommends further re- search, operational data-gathering, and management . . . . The model reported is a simple one, incorporating age-dependent and density-dependent rates of survival The utility of a mathematical model as a predictive and recruitment. I assumed the existence of 25 age tool depends on the validity of the assumptions incor- classes, with N(l), N(2) N(25) representing the porated in the model and the accuracy of the data that number of cranes in each class (e.g., there are N(l) the model uses. Little is known about many key fea- cranes less than 1 year old). Cranes are assumed to tures of sandhill crane population dynamics. As a re- initiate breeding as 4-year-olds (age class 5) and to die suit, the model described herein, and any other that after their 25th year. The simulated population is could be created now, will be unable to predict crane otherwise homogeneous; no differences in survival or populations with any but fortuitous accuracy. With recruitment are associated with subspecies, area of that thought in mind, I present a model that can be breeding, or other features. Figure 1 displays the logi- simulated with many different sets of parameter cal flow of the model. YEARS = 0 Enter survival, recruitment, and hunting parameters Initialize count of yeere N(25) - 10.000 N - N(1HN(2H — +N<25) YEARS * YEARS+1 1+ea(N-n%) s-/s (+eb(N-HpP Initialize age class** Add to count of y*ar* Calculate • urvlval rat* Calculate half-year survival rat* Calculat* recruitment r*t* NO)- R«[N(5H ■■■♦N(2S)] Determine recruitment N(28) - 3*N(26) N - N(1HN(2K +N(25) AOE RATIO - N<1)/N VOUNQ - H1»N(1) ADULT - H2« N-NO)! TOTAL - YOUNG ♦ ADULT AGE HATIO - YOUNG/TOTAL Determine fall population Calculate tall aga composition Datermlna hunting loaa 'I NO) - (1-H1) N(l) N(2) ■ (1-H2)«N(2) N<25) - (1-H2) >N(25) N - N( HN<2H- ♦N(25) S1-S0 t«'"N-Ns> s - /5 N(1) - 0 7VS»NO) N(2) - S«N(2) NO) - S«N(3) NC25) - S>N<25) N - N(1HN<2>» *«<25> N(2) - NO) NO) • N(2) N(25) - N(24) Determine eurvlvors Irom hunting Calculate survival rat* tor r**t of year Determine population at end of year Update ag** YES Teat tor 50 year* PRINT N. AOE RATIO. HARVEST STATISTICS Print results Fig. 1. Flow chart of sandhill crane population dynamics model 0.5 100 200 300 400 POPULATION (IN 1000s) 0.4 0.3 0.2 ■ HI < or. UJ 2 (1) N is the total number of cranes in the population. SO, SI, Ns, and a are parameters of the logistic function. At very low density (N — 0) the survival rate ap- proaches the upper asymptote S = S0 + S1-S0 1 + e-"Ns = SI, approximately. As the population becomes large (N — oo) the survival rate declines, most rapidly at N = Ns, and ultimately approaches a lower limit of SO. The exponential param- eter a governs the rate of decline. Survival rates are calculated separately for the half-year beginning with the spring breeding season and ending in fall, and the half-year from fall (after hunting) to the beginning of the next breeding season. Each half-year survival rate is the square root of the value obtained from Equa- tion 1. Survival rates for spring and summer are calcu- lated at the beginning of the breeding season, and de- pend on the spring population size. Survival rates for the fall and winter period are calculated after the hunt- ing season, and depend on the number of cranes that survive to that point. The survival rate of young birds is taken to be 75% of the adult rate for the half-year from fall to spring. Recruitment rates are also density-dependent, highest at low population densities and declining (to zero) as the population becomes extremely large (Fig. 3). Recruitment functions follow the equation R0 + Rl -R0 l+e6(N-NR|' (2) where RO ( = 0) and Rl are the extremes, Nr the inflec- tion point, and b the rate parameter. The calculated re- cruitment rate is applied to all birds in the spring population 4 years of age or older, and it results in birds added to the fall population. To start each simulation, every age class at least 1 year old in the simulated population was initially set at 10,000 birds. The annual cycle was repeated 50 times, or more if the population had not converged by then. The model is deterministic, as opposed to stochastic. It will thus exhibit far less variation than does a natural population. In a sense, it will yield conserva- tive results; for example, a simulated population with random components might become extinct while its deterministic counterpart exhibits steady, or even in- creasing, numbers. The Data 250 m O O o «- 200 CO LU z < or o u. o f£ LU CO Z 150 100 50 1960 1965 1970 1975 YEAR Fig. 4. Results of spring sandhill crane population surveys in Nebraska, 1959-78. Four quantities pertinent to the population dy- namics of sandhill cranes were used in the model: (1) the size of the spring population, (2) the age composi- tion of the population during the hunting season, (3) the number of cranes killed during the hunting season, and (4) the age composition of cranes in the harvest. mates of 230,000-260,000 lessers in the Central Flyway and 15,000-20,000 Canadians. Aldrich (1979) indicated that the population size of Canadians is poorly known, but could number more than 60,000 birds. Greaters probably make up a smaller proportion of the total than does either of the other two subspecies. Spring Population Size Sandhill cranes have been counted each spring since 1959 in an aerial survey conducted along the Platte River in Nebraska (Buller 1979). The survey was origi- nally devised to monitor population trends. Counts have ranged from 80,315 to 225,945 birds, those in re- cent years averaging larger than in the earlier years of the survey (Fig. 4). It is not known whether the ap- parent increase is real or a reflection of improved sur- vey skills. Aerial surveys include only cranes remain- ing on the river roosts at the time the aircraft flies over. Because most cranes leave their roost within 30 min after sunrise (Lewis 1976), these counts are biased low. Lewis (1979) determined the proportion of cranes that had left the roost before passage of the aircraft by ground checks. Using this information Lewis (1979) considered 300,000 cranes to be a minimum population estimate and believed that the number in the Central Flyway was close to 400,000 in 1976. It would seem that the current population of Central Flyway sandhill cranes contains between 250,000 and possibly 400,000 birds. This population is a mixture of three subspecies (Johnson and Stewart 1973): the greater sandhill crane (Grus canadensis tabida), the Canadian sandhill crane (G. c. rowani), and the lesser sandhill crane (G. c. canadensis). The composition of the population according to subspecies is unknown, al- though most are lessers. Braun (1975) provided esti- Fall Age Composition The age composition of sandhill crane populations was assessed annually during 1974-76 by Buller (1979), who conducted ground counts of flocks throughout the Central Flyway. Buller found the following percent- ages of immatures in the fall populations: 10.7 of 36,000 cranes aged in 1974, 11.3 of 49,000 birds in 1975, and 12.0 of 111,000 in 1976. The unweighted average for the 3 years was 11.3%. Losses to Hunting Harvest data were not systematically collected in the United States until the 1975-76 hunting season, when crane hunters in the Central Flyway were re- quired to obtain (at no cost) a permit, and question- naires were sent to applicants after the season termi- nated. Projections from the questionnaire response in- dicated that, in 1975-76, 9,497 cranes were harvested, and an additional 1,885 downed but lost, for a total kill of 11,382 (Sorensen and Reeves 1976). The correspond- ing figures for the 1976-77 season were 7,393 cranes harvested, 1,384 crippled, and a total kill of 8,777 (Sorensen 1977). In 1977-78 the estimated harvest was 12,151 cranes, the crippling loss 2,013, and the total kill 14,164 cranes (Sorensen 1978). The 3-year average total kill was 11,441 cranes. These values are esti- mates for the Central Flyway only. Additional birds are taken in Alaska, Canada, Mexico, and possibly Siberia (Lewis 1977). Lewis (1977:27) provided estimates of harvest by hunters in Alaska as 502 cranes in 1971 and 765 in 1972. The kill in Alaska during the 1975 season, estimated by Soren- sen and Reeves (1976) from the Waterfowl Hunter Questionnaire Survey, was 288 cranes. In addition to reported harvest by hunters, cranes are taken by Alaskan natives in the spring. D. R. Klein (in Lewis 1977:26) stated that the spring harvest by natives probably does not exceed 2,000 birds, although 1,000 were taken in 1 year on the Yukon-Kuskokwim Delta. Because of increased human populations in native communities and use of modern hunting methods, cur- rent harvest could be much greater (C. P. Dau, per- sonal communication). Surveys in Canada suggest that the kill there aver- aged 2,959 cranes per year in 1972-76 (Table 1). The Canadian data do not include birds crippled but lost (Gollop 1976), nor birds taken by natives (a number not believed to be large— Lewis 1977:25). Table 1. Estimated harvest of sandhill cranes by license-holders in Canada. 1972-76.a Year Saskatchewan Manitoba Total 1972 2,030 113 2,143 1973 3,592 683 4.275 1974 3,142 58 3,200 1975 3.048 164 3.212 1976 1,757 210 1,967 Average 2,959 "Sources: Cooch and Raible (1975) for all 1972 and 1973 data and 1974 Manitoba data; Cooch, Newell, and Wendt (1978) for 1975 and 1976 Manitoba data; Smith and Cooch (1978) for 1974-76 Saskatchewan data. The kill in Mexico is unknown but, of 28 recoveries of cranes banded along the Platte River in Nebraska, 5 came from Mexico. This rate would suggest that the Mexican harvest is 22% of that in Canada and the United States. K. Baer (in Lewis 1977:28) estimated that the annual harvest in Mexico is 500-1,000 cranes. Interest in hunting cranes is reported to be increasing in Mexico. The total annual hunting kill of sandhill cranes in re- cent years can be estimated by adding together esti- mates from the various sources: Alaskan native Alaskan hunter Canadian native Canadian hunter Central Flyway Mexico 2,000 + 288-765 small 2,959 + crippling loss 11,441 500-1,000 It appears that the harvest totals about 18,000 cranes per year and could be larger. If the population esti- mates presented earlier are realistic, this harvest would represent about 5 to 7% or more of the spring population. Harvest Age Composition The age composition of harvested sandhill cranes has been reported by Buller (1979), who found from bag checks that 31.5% of 574 cranes taken in 1974 were immatures. This rate compares to a percentage of 10.7 immatures in the population that year, and sug- gests that immatures are about 3.8 times as vulner- able to hunting as are adults, the relative risk being 0.315 x (1-0.107)/[(1-0.315) x 0.107], (The relative risk to hunting is the ratio of young to adult in harvest divided by the ratio of young to adult in the exposed population.) An earlier estimate of the relative risk was given by Miller et al. (1972:23), who reported that, in Saskatche- wan, immatures made up 4% (in 1966) and 6% (in 1967) of the total population, yet they made up 19% and 20%, respectively, of the harvest. The relative risk, as estimated from their pooled data, would be 4.6. The latter figure may be biased high because the authors included ground counts made before as well as during the hunting season. It is now recognized (Buller 1979) that early-migrating flocks contain a lower percentage of immatures than do later-migrating flocks. If we as- sume the total population in 1966-67 contained 11.3% young (the 1974-76 average), then the relative risk of juveniles would be 1.9. Another estimate of the age ratio among harvested cranes was given by Boeker et al. (1961), who found 30 immatures and 107 adults in hunter bags checked in January 1961 in New Mexico. These authors indicated that the composition of 21.9% young closely paralleled age ratios of cranes trapped for banding before the hunting season. The authors did not draw any conclu- sions about differential susceptibility of young cranes to hunting, possibly because trapping itself may be age-selective. If, however, we assume that the percent- age of young available to hunters was 11.3% (the aver- age age ratio in 1974-76), then the relative vulner- ability of juveniles to hunting is 2.2 times that of adults. The Approach Although much has been learned in recent years about sandhill crane population dynamics, the lack of certain critical information precludes the construction of a single model that could somehow be deemed "valid." At best, one can incorporate assumptions thought to be reasonable, and determine the implica- tions of those assumptions as reflected by the behavior of the model. The series of models proffered by Miller and his col- leagues (Miller et al. 1972; Miller 1974; Miller and Botkin 1974) represents certain sets of assumptions about crane population dynamics. In my view, the crucial relationship is the density-dependence of natural mortality; less likely to be consequential is the density-dependence of recruitment. The dependence of natural mortality on density would be manifested in the impact that hunting has on the population level. If natural mortality rates are largely independent of the number of cranes, then reducing the population by hunting will not reduce other mortality rates, and a population otherwise at equilibrium will decline as a re- sult of hunting. If natural mortality rates are sub- stantially density-dependent, then it is possible that they will be lower in a population reduced by hunting, which will thereby "make up" at least some of the loss from hunting. Miller's models apparently were suffi- ciently density-independent to cause the simulated population to plummet with the advent of hunting. The approach I followed was to try various survival functions (Fig. 2) and recruitment functions (Fig. 3) in combination and select those combinations that yielded realistic values of known population param- eters under the situation of no hunting. Hunting was then applied to the simulated population to determine its impact. Unhappily there is little knowledge of the population dynamics of sandhill cranes that are not hunted. Virtually all data described in the previous section were collected while cranes were legally hunted. For this reason I cannot portray what the population would do without hunting. I can only as- sume that current levels of hunting have not yet seriously affected survival and recruitment rates. Moreover, no one knows which combination of parameters is "correct." It is only known that the selected ones yield results that are consistent with what is known about the sandhill crane population. Specifically, one seeks combinations of parameters that yield (1) an equilibrium spring population between 250,000 and 400,000 cranes, and (2) an age composition in the fall of 10.3-12.3% young (within 1% of the 1974- 76 mean). When hunting is applied to the population, it will remove about 5% of the entire population an- nually. Results populations between 250,000 and 400,000 cranes and fall age compositions between 10.3 and 12.3% young. Each combination of survival and recruitment func- tions described in Table 2 yielded an asymptotic popu- lation and age composition, under the situation of no hunting, consistent with what is known of the actual crane population. Thus, without additional informa- tion, one cannot discriminate among them. One cannot select from these, or from the multitude of other plausible combinations, the one that most closely parallels the true situation. Hunting was then applied to the simulated popula- tion under each of the 37 combinations. Two sets of hunting rates were used: 10.7% on young birds (H, = 0.107) and 4.3% on adults (H2 = 0.043); and 14.9% on young (H, = 0.149) and 3.7% on adults (H2 = 0.037). Under the first set of rates, young birds are about 2.5 times as vulnerable as adults; under the second, about 4.0 times. Because this differential vulnerability ratio is so poorly known, it is fortunate that very similar re- sults were obtained from both sets of hunting rates (Table 2). For the sake of brevity, I will discuss only the results under the second set of hunting rates. The first set produced similar, but slightly lower, asymp- totic populations. Under all 37 combinations the population declined with the advent of hunting. The extent of the decline varied markedly, however, ranging from 5.1 to 54.4%. The median decline was 22.3%. Figure 5 shows the re- sults of a typical simulation, this one combining sur- vival curve 8 and recruitment curve B. Without hunt- ing, the population ultimately increased from the initial 240,000 to 272,000 cranes. When hunting was applied, the population eventually decreased to 201,000 cranes, and yielded an annual harvest of about 12,000 birds. 8 250 O £ 200 CO LU 2 150 < EC o ioo CO 2 Z NO HUNTING H1 = 0.149 H2=0.037 10 20 30 YEAR OF SIMULATION A total of 768 combinations of 32 survival functions and 24 recruitment functions were executed without hunting to determine if they yielded population sizes and age ratios comparable to the data described above. Thirty-seven combinations (Table 2) gave asymptotic Fig. 5. Results of model under combination of survival curve 8 and recruitment curve B. Shown are simulated results under situation of no hunting and under hunting rates HI = 0.149 and H2 = 0.037. Also shown is the annual harvest. Table 2. Combinations of survival and recruitment curves that yielded realistic asymptotic populations and age ratios under situation of no hunting. Also shown are harvest (in 1,000's), asymptotic population (in 1,000's) and percent decrease for two hunting situations. Harvest values are numbers taken after population stabilizes. Survival curve Recruitment curve Unhunted population H, = 0.107, H2 = = 0.043 H, = = 0.149, H2 = 0.037 Harvest Population Decrease (%) Harvest Population Decrease (%) 1 F 250.8 12.2 212.3 15.4 12.3 213.6 14.8 2 A 267.3 13.4 238.1 10.9 13.5 239.3 10.5 2 B 272.9 14.6 253.5 7.1 14.8 253.9 7.0 2 C 273.8 14.0 246.0 10.2 14.1 247.0 9.8 2 F 270.8 13.7 241.6 10.8 13.7 242.8 10.3 2 G 273.9 14.5 252.4 7.8 14.7 253.0 7.6 3 C 250.0 9.1 156.9 37.2 9.4 159.7 36.1 4 A 278.5 12.6 220.1 21.0 12.7 222.2 20.2 4 B 285.9 14.4 247.4 13.5 14.6 248.6 13.0 4 C 298.4 13.0 226.9 24.0 13.2 229.6 23.1 4 F 289.2 12.7 222.4 23.1 12.9 224.7 22.3 4 G 291.8 14.2 244.9 16.1 14.4 246.2 15.6 4 H 294.7 15.5 264.3 10.3 15.9 265.0 10.1 5 C 341.9 16.2 287.2 16.0 16.2 290.0 15.2 5 D 349.9 17.9 313.5 10.4 18.1 314.6 10.1 5 I 352.8 18.0 314.3 10.9 18.2 315.5 10.6 6 D 368.9 19.1 339.4 8.0 19.1 340.7 7.6 6 E 374.6 20.4 355.0 5.2 20.6 355.4 5.1 6 I 372.8 19.4 343.1 8.0 19.5 344.3 7.6 6 J 375.8 20.3 354.0 5.8 20.5 354.5 5.7 7 A 256.5 10.1 172.5 32.7 10.5 177.4 30.8 7 C 264.4 8.7 148.3 43.9 9.4 159.7 39.6 7 F 260.5 9.9 169.9 34.8 10.4 175.6 32.6 8 A 255.1 8.6 146.0 42.8 9.1 151.2 40.7 8 B 271.7 11.8 197.5 27.3 12.3 200.7 26.1 8 C 273.4 6.8 115.8 57.6 7.4 124.7 54.4 8 F 263.6 8.6 145.5 44.8 9.0 150.2 43.0 8 G 273.8 12.0 • 200.9 26.6 12.4 203.1 25.8 9 C 322.2 8.7 148.2 54.0 9.5 160.0 50.3 9 D 336.7 14.9 252.7 24.9 15.4 257.5 23.5 9 F 308.1 10.0 170.4 44.7 10.4 176.5 42.7 9 G 306.0 13.5 230.3 24.7 13.9 233.2 23.8 9 I 339.3 14.8 251.7 25.8 15.3 256.1 24.5 10 C 348.6 8.7 149.6 57.1 9.6 161.9 53.6 10 D 357.7 15.7 268.4 25.0 16.3 274.3 23.3 10 E 367.2 19.2 325.8 11.3 19.7 328.0 10.7 10 I 362.1 15.6 265.7 26.6 16.1 271.7 25.0 Discussion Limitations of the Model and Data In their original report, Miller et al. (1972) voiced a concern that the information needed to manage sand- hill crane populations properly was lacking or inade- quate. Research in recent years has provided better data on certain characteristics, but has also demon- strated how little is known about other aspects of sandhill crane population dynamics. In this section I discuss some of the limitations of the model developed here and the data used in it. I also make recommenda- tions for further research, for data-gathering on an operational basis, and for strategies to manage sand- hill crane populations. In the model presented here, several assumptions were made about the population dynamics of sandhill cranes. It was assumed that cranes first breed as 4- year-olds. Although age at first breeding is an impor- tant species characteristic, its value is of little direct consequence to the population as modeled here. Be- cause estimates are available of the percentage of young in the fall population (e.g., 11%), it is not neces- sary to know which age segment of the population pro- duced them (e.g., 11 young produced by 50 birds age 4 or greater, or 1 1 young produced by 60 birds age 3 or greater). The age of initial breeding is an implicit parameter in the present model. Another assumption concerns the longevity of sand- hill cranes in the wild. The model assumes a 25-year life-span, after which cranes die. This limit is artificial, of course, but serves as a computational convenience. An indeterminate life-span would yield similar results. Analogously, instead of constant survival and recruit- ment rates for all adults in the population, senescence could be invoked with lowered survival and reproduc- tion among old cranes. Again, this modification would make but small changes in the results. The degree to which recruitment and survival are density-dependent is of prime importance to the modeling. A small change in the form of the recruit- ment curve or survival curve can inspire major changes in the resulting asymptotic population and its response to hunting. It is very doubtful, however, that the near future will see either recruitment or survival measured with sufficient accuracy to detect and de- lineate density dependence. Thus the information most needed to improve the model will probably be long in coming. The model is deterministic, in contrast to the sto- chastic nature of the real world. Too little is known to develop a realistic model of randomness in the param- eters. The averages of most parameters are at best poorly known; their variations and inter-correlations could scarcely be guessed. A deterministic model should suffice, however, if it is viewed conservatively. The real world will be more variable than its deter- ministic counterpart in a model; decisions based on the model should be made with that caveat in mind. As stated earlier, it is possible for a simulated population incorporating randomness to become extinct while its deterministic analog held steady, or even increased in number. An oversimplification in the model that is most troublesome to me is the aggregation of three sub- species. Birds in the Central Flyway are referred to as the sandhill crane population, but the birds involved include greaters, Canadians, and lessers, and originate in nesting areas extending from Siberia, through vast areas of Alaska and western Canada, and possibly into northern Minnesota. The Central Flyway population of lessers is large, the number of Canadians is modest, and the number of greaters is small. During fall migra- tion, different subspecies migrate along somewhat dif- ferent routes, and at somewhat different times (John- son and Stewart 1973). Hunting cannot be expected to affect each of them equally, and more attention must be given to evaluating the effects of hunting on the dif- ferent subspecies. It was further assumed that parameters measured recently, while cranes were subjected to legal hunting, are not too different from the values appropriate if cranes were not hunted. The impact of this assumption is less than one might imagine. The crucial result is that the level of a stable population lies in a balance maintained by mortality rates and recruitment rates. Any increase in mortality rates, by hunting for example, will result in a lower level of the population, unless (1) the total mortality rate stays constant, through a decrease in natural mortality, or (2) recruit- ment is increased, presumably because of lowered den- sities. The model offered here includes the possibility of compensatory effects in both natural mortality and recruitment. In the multitude of combinations examined, however, their combined effect was insuffi- cient to compensate completely for the increased mor- tality due to hunting. This result is not surprising in view of the nature of sandhill cranes. Recommendations Further Research The key concern in analyzing the impact of hunting on sandhill crane numbers is the extent to which that form of mortality is compensated for by other forms of (natural) mortality. If only a certain number of cranes are destined to survive the winter, for example, it matters little that the "doomed excess" is taken by hunters. At the other extreme, if hunting mortality is completely additive to natural mortality, then the loss of a single crane to a hunter will be reflected propor- tionately in the next year's spring population. Most likely, hunting mortality is neither completely compensatory nor completely additive. It seems prob- able that each crane shot by a hunter in the fall will be reflected in a fractional decrease the next spring. The size of the fraction is in doubt. There seems little hope of ascertaining the value of that fraction by the usual methods based on banding. Even with the intensively studied mallard {Anas platy- rhynchos), with more than two-thirds of a million pre- season bandings available for analysis, the question of compensation in hunting mortality has only recently been addressed (Anderson and Burnham 1976). I doubt that information based on sandhill crane band- ing will ever be adequate to determine the degree to which hunting mortality is compensatory. Other meth- ods are called for. A method that may prove fruitful involves the fol- lowing line of inquiry: To what degree is one mortality agent compensatory to another within a K-selected species? The question need not be restricted to only hunting mortality or to only sandhill cranes. Research on other mortality factors and on ether species of birds may provide parallels for hunting and sandhill cranes. Among birds, sandhill cranes are a relatively K- selected species (e.g., Pianka 1974; Southwood 1976). They are large, long-lived, and have low fecundity (de- ferred breeding, small clutch size). Young cranes are invested with considerable parental care and develop rather slowly. Other attributes associated with K- selection may bear directly on the question of the com- pensatory nature of mortality. Populations of K- selected species tend to be stable when at equilibrium, mortality is normally noncatastrophic and density-de- pendent, and birth rate tends to be density-dependent, but populations are relatively easily extinguished. The nature of a K-selected species is such that it cannot recover rapidly from a severe population reduc- tion (Miller 1978). Although birth rate and survival may increase as the population becomes sparse, neither rate can rise very much, and extinction looms as a real possibility. Analogies with other K-selected species, both endangered and common, may prove in- structive for making decisions about sandhill cranes. Operational Data Gathering There is little question that improvements in meth- ods of collecting sandhill crane data should be sought. Most sandhill cranes in the Central Flyway stage their spring migration along the Platte River in Nebraska. This concentration of birds should be relatively easy to count, at least compared with most populations of wild birds. Yet, counts in past years have been erratic, sug- gesting high variability from censusing procedures. Other evidence (e.g., Lewis 1979) points to biases as well as variability in the spring survey data. Modifica- tions of the survey have been made and evaluated (Fer- guson et al. 1979), and further work along this line is under way. Improved surveys may more accurately indicate the size of the sandhill crane population. It is doubtful that they will permit us to detect changes in sandhill crane numbers caused by hunting, however, except possibly over extended periods of time. The annual de- cline due to hunting (e.g., about 4,000 cranes per year in the example displayed in Fig. 5) is rather small in comparison to the variability in the counts, and would additionally be masked by natural variation in num- bers of cranes. The age composition of fall-migrating sandhill cranes has been well documented for 3 consecutive years (Buller 1979). Little annual variation was de- tected. Nonetheless, it is possible for the age composi- tion to undergo an important change, which could seriously alter the results of the model presented here. It would seem prudent to develop a periodic survey for estimating the fall age composition for a number of years. The size of the sandhill crane harvest, together with the population size, are the most critical parameters now being estimated. The requirement that hunters in the Central Flyway obtain permits to hunt cranes has imparted much more precision to harvest estimates in that area, but harvest in other important areas (e.g., Alaska and Mexico) is poorly known. The permit sys- tem should probably be continued for a time, particu- larly as changes are made in regulations. For example, in North Dakota the switch from a November season to one in September resulted in an increase in harvest from an average of about 380 cranes per year in 1968- 76 (Johnson 1977) to 4,078 in 1977 (Sorensen 1978), and about 2,800 in 1978 (M. F. Sorensen, personal communication). Better information on harvest out- side the Central Flyway and Canada is sorely needed. Only once has the age composition of harvested cranes been estimated concurrently with the age com- position in the hunted population. Fortunately, the simulation model provided very similar results when two rather different values of differential vulnerability to hunting (2.5 and 4.0) were used. Thus, this param- eter appears to be less consequential to the population dynamics than the others discussed. It is also important to delineate the various popula- tions of sandhill cranes that are subjected to hunting, and to evaluate their status. Ideally, measures of pro- ductivity as well as mortality from both natural causes and hunting should be obtained for each population. At a minimum, various marking programs could be undertaken to ascertain the derivation of harvest of cranes, and measurements should be taken on hunter- shot cranes to determine their racial composition. Management Goals In addition to implementing some operational data- gathering procedures, prudent management seems to require some difficult decisions. Primary among them are population goals for each identifiable component of the sandhill crane population. How many cranes do we want in each breeding, migrating, and wintering area? Conservationists may desire more, some grain farm- ers, fewer. Once population goals are identified, the question becomes how to attain these goals. What magnitude of harvest is consistent with the goal? And then, how shall the harvest be allocated? Indians and other natives take cranes in subsistence hunting; Canadian, American, and Mexican hunters take them for sport. Deciding who gets how many will be a chal- lenging task, particularly as demands grow and if crane populations dwindle. Acknowledgments I am grateful to R. J. Buller and J. W. Aldrich for access to unpublished data; to C. L. Nustad and D. A. Davenport for assistance with computer program- ming; and to K. P. Burnham, J. C. Lewis, H. W. Miller, J. D. Nichols, K. J. Reinecke, and D. L. Trauger for comments on an earlier version of the report. 10 References Aldrich, J. W. 1979. Status of the Canadian sandhill crane [Grus canadensis roivani). Pages 139-148 in J. C. Lewis, ed. Proceedings of the 1978 Crane Workshop. Colorado State University Printing Service, Fort Collins. Anderson, D. R., and K. P. Burnham. 1976. Population ecol- ogy of the mallard: VI. The effect of exploitation on sur- vival. U.S. Fish Wildl. Serv., Resour. Publ. 128. 66 pp. Boeker, E. L.. J. W. Aldrich, and W. S. Huey. 1961. Study of experimental sandhill crane hunting season in New Mexico during January 1961. U.S. Fish Wildl. Serv., Spec. Sci. Rep. -Wildl. 63. 24 pp. Braun, C. E., chairman. 1975. Conservation committee report on status of sandhill cranes. Wilson Bull. 87(21:297-302. Buller, R. J. 1979. Lesser and Canadian sandhill crane popula- tions, age structure, and harvest. U.S. Fish Wildl. Serv., Spec. Sci. Rep. -Wildl. 221. 10 pp. Cooch, F. G., K. Newell, and S. Wendt. 1978. The 1976 kill of migratory game birds other than waterfowl by hunters in Canada. Can. Wildl. Serv. Prog. Notes 83. 6 pp. Cooch, F. G., and H. A. Raible. 1975. Harvest of migratory game birds other than waterfowl in Canada 1974. Can. Wildl. Serv. Prog. Notes 52. 7 pp. Ferguson, E. L., D. S. Gilmer, D. H. Johnson, N. Lyman, and D. S. Benning. 1979. Experimental surveys of sandhill cranes in Nebraska. Pages 41-49;>* J. C. Lewis, ed. Proceed- ings of the 1978 Crane Workshop. Colorado State Univer- sity Printing Service, Fort Collins. Gollop, J. B. 1976. The sandhill cranes of Last Mountain Lake. Report presented at Annual Meeting of the Whoop- ing Crane Conservation Association, Regina, Saskatche- wan, 1 October 1976. Johnson, D. H. 1977. The status of the sandhill crane in North Dakota. North Dakota Outdoors 39(91:14-18. Johnson, D. H., and R. E. Stewart. 1973. Racial composition of migrant populations of sandhill cranes in the northern plains states. Wilson Bull. 85(2):148-162. Lewis, J. C. 1976. Roost habitat and roosting behavior of sandhill cranes in the southern Central Flyway. Pages 93- 104 in J. C. Lewis, ed. Proceedings of the International Crane Workshop. Oklahoma State University, Stillwater. Lewis, J. C. 1979. Factors affecting the spring inventory of sandhill cranes. Pages 33-39 in J. C. Lewis, ed. Proceedings of the 1978 Crane Workshop. Colorado State University Printing Service, Fort Collins. Lewis, J. C, chairman. 1977. Sandhill crane {Grus canaden- sis). Pages 4-43 in G. C. Sanderson, ed. Management of migratory shore and upland game birds in North America. International Association of Fish and Wildlife Agencies, Washington, D.C. Miller, R. S. 1974. The programmed extinction of the sandhill crane. Nat. Hist. 83(2):62-69. Miller, R. S. 1978. Population modeling as an aid to designing management programs. Pages 413-417 in S. A. Temple, ed. Endangered species: management techniques for preserv- ing threatened species. University of Wisconsin Press, Madison. Miller, R. S., and D. B. Botkin. 1974. Endangered species: models and predictions. Am. Sci. 62:173-181. Miller. R. S., G. S. Hochbaum, and D. B. Botkin. 1972. A simulation model for the management of sandhill cranes. Yale Univ. Sch. For. Environ. Stud. Bull. 80. 49 pp. Pianka, E. R. 1974. Evolutionary ecology. Harper and Row, New York. 356 pp. Smith, G. E. J., and F. G. Cooch. 1978. Experimental sandhill crane survey 1974-76. Pages 110-120 in H. Boyd and G. H. Finney, eds. Migratory game bird hunters and hunting in Canada. Can. Wildl. Serv. Rep. 43. Sorensen, M. F. 1977. Sandhill crane harvest and hunter ac- tivity in the Central Flyway during the 1976-77 hunting season. U.S. Fish Wildl. Serv. Administrative Report— 7 July 1977. 9 pp. Sorensen, M. F. 1978. Sandhill crane harvest and hunter ac- tivity in the Central Flyway during the 1977-78 hunting season. U.S. Fish Wildl. Serv. Administrative Report— 14 June 1978.9 pp. Sorensen, M. F., and H. M. Reeves. 1976. Sandhill crane har- vest and hunter activity in the Central Flyway during the 1975-76 hunting season. U.S. Fish Wildl. Serv. Administra- tive Report-9 July 1976. 9 pp. Southwood, T. R. E. 1976. Bionomic strategies and popula- tion parameters. Pages 26-48 in R. M. May, ed. Theoretical ecology. W. R. Saunders Co., Philadelphia. GPO 851 - 85S As the Nation's principal conservation agency, the Department of the Interior has responsibility for most of our nationally owned public lands and natural res mrces. This includes fostering the wisest use of our land and water resources, protecting our fish and wildlife, preserv- ing the environmental and cultural values of our national parks and historical places, and providing for the enjoyment of life through out- door recreation. The Department assesses our energy and mineral resources and works to assure that their development is in the best interests of all our people. The Department also has a major responsi- bility for American Indian reservation communities and for people who live in island territories under U.S. administration. UNITED STATES DEPARTMENT OF THE INTERIOR FISH AND WILDLIFE SERVICE EDITORIAL OFFICE AYLE3WORTH HALL. qSU FORT COLLINS, COLORADO 80523 POSTAGE AND FEES PAID DEPARTMENT OF THE INTERIOR INT A 23 NOTE: Mailing lists are computerized. Please return address label with change of address.