<<

;1 ~:..._ ~

WHITE-TAILED POPULATION MANAGEMENT IN THE NORTH CENTRAL STATES

North Central S~ The Wildlife Soc 1980 A\~V' Gc~s~

\\l_~~c\~O /:" J\t) ~~~= ! I' / ,_ _ ~ J !U ~ / ,1\:1 C . ., I ) ",??r, ~~'/ ;- ~~~"-I;~:_~~ ..f _ -: "-.,,~A' ~ ,J----' '19':/ ,. .' ,.,>~. '1-­ /

WHITE-TAILED DEER POPULATION MANAGEMENT IN THE NORTH CENTRAL STATES

Proceedings of a Symposium held at the 41st Midwest Fish and Wildlife Conference, Urbana, Illinois, 10 December 1979

Ruth L. Hine and Susan Nehls Editors

Sponsored by North Central Section of The Wildlife Society 1980

.... CONTENTS

3 PREFACE David A. Arnold and Oliver Torgerson 5 DEER POPULATION ESTIMATORS IN THE MIDWEST FARMLAND Lee Gladfelter 13 MONITORING DEER POPULATIONS IN THE NORTHERN FORESTED AREAS OF THE MIDWEST Jack J. Mooty 23 REPRODUCTION OF WHITE-TAILED DEER IN THE NORTH CENTRAL UNITED STATES John D. Harder 37 THE LEGAL DEER KILL - HOW IT'S MEASURED Lawrence A. Ryel 47 WINTER - THE GRIM REAPER Patrick D. Karns 55 FAWN MORTALITY ESTIMATES IN FARMLAND DEER RANGE Wayne R. Porath 65 ESTIMATING ILLEGAL KILL OF DEER Kirk H. Beattie, J. Cowles, and Robert H. Giles, Jr. 73 AGE DETERMINATION OF WHITE-TAILED DEER IN THE MIDWEST - METHODS AND PROBLEMS John L. Roseberry 83 DEER HERD MANAGEMENT - PUTTING IT ALL TOGETHER William A. Creed and Frank P. Haberland 89 POPULATION PHENOMENA - THEIR MANAGEMENT IMPLICATIONS George E. Burgoyne, Jr. 95 CONVINCING THE DECISION MAKERS Merrill L. Petoskey 99 SELLING MANAGEMENT TO THE DEER HUNTER John Madson 103 DEER MANAGEMENT FOR WHATl Edward E. Langenau, Jr. 107 WHITE-TAILED DEER MANAGEMENT IN THE MIDWEST - SYMPOSIUM SUMMARY Robert S. Cook and Rebecca Field 111 APPENDIX: A WISCONSIN DEER MANAGEMENT CHRONOLOGY (1836 - 1980) Walter E. Scott

2 PREFACE

The white-tailed deer is the "bread and but­ definite limitation on both time and space for ter" game animal for nearly all the states in the symposium. the north central part of the United States. Working with leaders of the North Central The welfare of this deer has more influence Section of The Wildlife Society, they set on the policies, the politics, the finances, and about to recruit the most competent deer the sociology of the public in this workers in the region to contribute to this region than any other game species, except book. perhaps the pheasant in South Dakota. This collection of papers attempts to deal The whitetail has generated vast numbers with the major phases of species manage­ at scientific studies, precipitated many ment for the white-tailed deer. It looks at the acrimonious debates, caused more highs and basic tools and techniques, such as aging lows in organizational morale, has triggered deer, determining populations, and measur­ departmental reorganizations, and even has ing the harvest. It also considers various been the political downfall of state legisla­ causes of mortality, factors influencing pro­ tors. The scientific literature could keep a ductivity, and goes all the way to the human reviewer busy month after month. Volumes dimensions of dealing with the hunters, the have been written on deer biology and general public, and the administrators. In the countless sporting journals tell how to hunt Appendix are highlights of an extensive the animal. In spite of the literally millions of chronology of white-tailed deer in Wisconsin, written words about the whitetail, we are prepared by a noted conservation historian. unaware of any single volume which deals We are more than grateful to the authors with the pertinent population management for their contributions to this symposium, concepts that must be dealt with annually by which will bring under one cover the pieces the managers and administrators responsible of the management puzzle and sort them out for today's deer herds. The collection of in a logical and useful manner. papers in this symposium has been designed to fill this void. A loosely organized group of biologists David A. Arnold known as the Midwest Deer and Michigan Department of Group served as an informal committee to Natural Resources P.O. Box 30028 carefully consider the major elements in the Lansing, MI 48909 deer management formula. It was decided, after due consideration, to restrict the sym­ Oliver Torgerson posium subject to the species management Missouri Department of problems and not include habitat manage­ Conservation ment. Two major reasons for this decision P.O. Box 180 Jefferson City, MO 65102 were: many deer herds occur on private lands ,1\ where little or no opportunities for range manipulation are available, and there was a Symposium Co-Chairmen

3 IVlichlgan DNR DEER POPULATION ESTIMATORS IN THE MIDWEST FARMLAND

Lee Gladfelter Iowa Conservation Commission Wildlife Research Station Boone, IA 50036

Abstract: Estimating deer population South Dakota, Nebraska, and Kansas; north­ crop damage and hunting seasons were once numbers or trends is useful in formulating harvest ern Missouri; southern Minnesota, Wisconsin, again instituted to harvest the surplus. Since regulations, long-range management goals, and and Michigan; all but the southern tip of Il­ then, scientific management of deer has public relations programs, and in evaluating linois; and Iowa, Indiana, and (Fig. 1). become increasingly difficult because of harvest goals and habitat management practices. Generally, white-tailed deer (Odocoi/eus vir­ unique biological, social, economic, and Techniques vary between states and-depend upon ginianus) density is low because of lack of ex­ agency management goals, manpower, tradition, political problems. Resource managers have habitat, deer density, and weather conditions. tensive timbered habitat. However, high attempted to establish a balance between Population ilStimators currently in. use .in the reproductive rates, minimal overwinter toss, landowner tolerance for crop damage and Midwest Farmland include: deer-vehicle accident and good body condition are benefits pro­ deer populations that are extremely reports, sex-age-kill estimates, roadside observa­ vided by extensive use of agricultural crops vulnerable to harvest. tions, aerial surveys, registered buck kill, hunter for food (Mustard and Wright 1964, Watt et kill-effort, track counts, conservation officer al. 1967, Nixon et al. 1970). Farming is the estimates, crop depredation reports, population major land use and because of problems with THE NEED FOR POPULATION models,and landownersurveys. Most statesrely on landowner intolerance for crop damage, deer more than one estimator in an attempt to obtain ESTIMATORS populations must be carefully controlled. comparable results. Numerous sources of bias are involved and somequestionableassumptions must Because of milder winters and the wide The need for proper management deci­ be made. It is doubtful that these techniques can distribution of their major food source, deer sions, based on sound biological data, is dic­ measure changes of less than 20% in population do not concentrate in winter yards as in the tated by well-informed and organized sports­ size. Finding new population estimators that can more northern parts of their range. man and nonhunter groups, competition for be statistically evaluated and can accurately History indicates that significant changes available habitat from other land uses, and predict population change should be a major goal in deer management have taken place in the competition for conservation dollars. The of future research. Midwest Farmland. Deer were common when ability of resource managers to accurately settlers first arrived and were used extensive­ predict deer population size or trends is the INTRODUCTION ly for food, clothing, shelter, and tools. But, a cornerstone to formulating management The Midwest Farmland is generally defined growing demand for food, by an expanding goals, hunting regulations, and public rela­ as the bluestem prairie region which consists human population, brought about market tions programs. of agricultural land interspersed with oak­ hunting and habitat destruction which Management goals should be developed in hickory, cottonwood-elm, or beech-maple substantially reduced deer numbers during response to deer density, available habitat, timber types. Portions of 12 states are in­ the late 1800's. In the early 19OO's, state crop depredation, landowner attitudes, and cluded in this region: eastern North Dakota, legislatures began closing or restricting hunt­ hunting and nonhunting recreational uses. A ing seasons in an attempt to reduce harvest. classic statement at this point would be that During the 1930's and 1940's, populations in­ population levels should be maintained near Gladfelter, Lee. 1980. Deer population estimators in the Midwest farmland. Pages 5-11 in Ruth L. creased rapidly in response to restricted hunt­ carrying capacity. Carrying capacity has Hine and Susan Nehls, eds. White-tailed deer ing, establishment of refuges, effective law generally been defined as the maximum population management in the north central enforcement, restocking, and habitat im­ number of deer that an area can support states. Proc. 1979 Symp. North Cent. Sect. Wild!. provement. By the 1940's and 1950's, in­ without damage to the habitat (Hosley Soc. 116 pp. creased deer densities produced widespread 1956:224). But, determining the carrying s FACTORS AFFECTING human comfort, and equipment operation. POPULATION ESTIMATORS TECHNIQUES CURRENTLY IN USE Population estimators are designed to measure some degree of change in deer IN THE MIDWEST FARMLAND population levels. A knowledge of why Population estimators have been populations change is important in develop­ developed from information collected in the ing good estimators. Population changes may following ways: take place because deer move to more (1) Field observations that rely on visual con­ suitable habitat. This movement may be in tact during the survey period, such as response to heavy hunting pressure, over­ roadside observations, aerial surveys, population, habitat changes, and food short­ population estimates by field personnel or ages. Population changes may reflect dif­ landowners, and sex and age ratio counts. ferential mortality and natality rates. Mortali­ (2) Observation of deer sign or damage, such ty rates are influenced by legal and illegal as track counts and crop depredation hunting, disease, predator density, food sup­ reports. ply, and accidents. Natality is generally FIGURE 1. The Midwest Farmland. (3) Examination of dead deer for use in related to nutritional level, weather, disease, calculating sex-age-kill estimates, buck and population density. kill, and deer-vehicle accident reports. Animal behavior should be considered in (4) Information obtained from hunters to pro­ capacity in agricultural areas where timber selecting proper population estimators. The vide estimates of deer harvest and hunter provides cover and crops provide an abun­ type of habitat selected by deer affects their effort. dant food source is difficult. Statewide deer observability and accessibility. A knowledge Conservation agencies have developed densities in the Midwest Farmland range from of movement is essential because immigra­ many different population estimators in 2 0.2 to 3.0 deer per km (Nixon 1970) but in tion or emigration may drastically affect response to their individual management some refuges where hunting is restricted and survey results on small study areas. Study goals (Table 1). Estimates are made on areas agricultural crops are available, winter con­ areas should be selected that are large ranging in size from statewide to large hunt­ centrations exceed 80 deer per krn", enough to contain daily home ranges and ing zones, counties, small management Therefore, potential carrying capacity is seasonal movements of deer. Other areas, state forests, and military reservations. great but realistic population levels depend behavioral characteristics that must be con­ Management goals may require a population upon the management goals of the conserva­ sidered are social aspects (family groups, estimator that indicates trends or one that tion agency. breeding, herding), ability to avoid detection, provides a total estimate. Most states employ Hunting seasons are the most effective and daily and seasonal activities (feeding, several different techniques for comparison management tool available to the resource watering, breeding, fawning). purposes. This paper will not deal with manager. Once management goals have Other factors involved in developing specific techniques for each population been established, hunting regulations should population estimators include habitat com­ estimator since references are given that pro­ be formulated to obtain the desired deer position and configuration, land use patterns, vide this information. Emphasis will be harvest. Harvest can be applied annually and manpower, money, and tradition. Population placed on comparing techniques currently in hunting regulations are usually flexible estimators for hardwood timber and open use and exploring advantages, disadvantages, enough to provide for full resource utiliza­ agricultural habitat differ from those used in bias, and precision of each. tion while equitably distributing benefits to coniferous forest habitat. With current the hunter. Successful accomplishment of budgetary restrictions in most conservation Deer-Vehicle Accident Reports management goals through hunting regula­ agencies, the cost of conducting population tions can only be assured when the public is estimates is an important consideration, Many states are using the number of deer­ convinced that sound biological data have especially for large survey areas. Often the vehicle accidents per billion miles traveled as been collected and presented. Agencies that cost of various techniques is directly related a population trend indicator (Table 1). l ahn have good public support for their programs to achieved accuracy. A knowledge of local (1959) evaluated the use of deer-vehicle ac­ have much more latitude in their ability to weather is also important because it affects cidents as a trend indicator in Wisconsin but manage the resource. visibility, animal behavior and movement, found that it was not a precise index of an­

~6 hunter effort, non hunting mortal ity, and natality (Eberhardt 1960, Lang and Wood TABLE 1. Population estimators used by the states in the Midwest Farmland. 1976). Wisconsin utilizes a sex-age-kill method to Population Estimator States estimate the buck population which is then Deer-vehicle accident reports lA, IL, KS, MI, MN, MO, NE, OH, WI expanded to a total population estimate Sex-age-kill estimates lA, IL, KS, MI, MO, OH, SD, WI (Creed et al. 1978). The adult buck population Roadside observations IN, KS, MI, NE, WI is estimated for some year in the past by Aerial surveys MN, MO, ND, WI adding up the subsequent legal harvest of Registered buck kill NE, OH, WI adults alive in that particular year, until the Hunter kill-effort IN, KS youngest age class (shown by age structure) Landowner surveys KS, WI of the year of interest, passes out of existence Track counts OH Conservation officer estimates IA or becomes extremely small. This total is then Crop depredation reports MO divided by 0.8 which assumes that 80% of the Population models MN bucks are recovered as a result of hunting. The buck estimate is then multiplied by an expansion factor (E.F.) that is compiled an­ nually for each hunting zone according to the following formu.a: E.F. 1.00 (B/O) (B/O)F nual change when compared to other popula­ minimum figure since many accidents are not = + + tion indices. He cited poor record keeping as reported or animals escape only to die later. where B = corrected percentage of yearling the major problem. However, McCaffery One advantage to this technique is that infor­ bucks (percentage yearling bucks divided by (1973) found a sign ificant relationship be­ mation can readily be broken down on a the number of male fawns per 100 female tween number of deer-vehicle accidents and regional basis if detailed records of fawns), 0 = percentage yearling does, and F number of antlered bucks harvested in automobile mileage are available. Cost of = number of fawns per doe. Wisconsin. He concluded that this technique reporting is minimal since field personnel are A comparative analysis indicated that sex­ could provide a reliable index to population already required to be at the scene of an acci­ age-kill estimates averaged only 70% of the change if an accurate record of loss is kept dent to remove or tag dead animals. pellet survey results in northern Wisconsin and an accurate estimate of the number of Biological data such as sex, age, reproductive while in southern Wisconsin, they were automobile miles driven is available. This status, and body condition can also be ob­ similar to pre-season population predictions demonstrates the relationship between deer­ tained. This is probably the best and most (Creed et al. 1978). An explanation for this vehicle accidents and other indexes but com­ economical technique available for may be increased vulnerability of deer in parisons with known population trends are statewide estimates of population trends. southern Wisconsin, lower nonhunting mor­ needed to increase reliability. tality, and heavier hunting pressure which The major assumption with deer-vehicle Sex-Age-Kill Estimates results in the harvest of 80% of the adult accidents is that the number of deer killed, bucks. when related to vehicle mileage, accurately Many states use biological data collected Another method of change-in-ratio analysis reflects population fluctuations. Some prob­ during hunting seasons to estimate deer requires sex and age ratios before and after lems, which may be minimized by assuming a population size (Table 1). This technique in­ the hunting season (Gilbert 1978). This can be constant bias each year, are poor reporting volves measuring changes in sex and age used when one age class is more susceptible by field personnel and disappearance of car­ ratios in the harvest (Petrides 1949, Lauckhart to hunting pressure, or when one sex and age casses from the scene of the accident. Local 1950, Davis 1963, Hanson 1963). Some resear­ class is hunted more heavily than others. To meat prices may be responsible for changes chers have developed formulas for change-in­ determine fall or winter sex and age ratios in rate of carcass loss. Other factors that may ratio estimates (Rupp 1966, Overton and from field counts, it is assumed that all affect this index are construction of new Davis 1969) and others have worked with animals have an equal chance of being highways, changes in traffic patterns or determining precision (Paulik and Robson counted. Downing et al. (1977) found that speed, and degree of cooperation from other 1969, Eberhardt 1969). The basic input factors there was rarely a period of time when this state agencies. Reported highway kill is a required are: sex and age ratios, harvest, assumption was correct for every sex and age J class. Care must be taken in determining sex times to enable statistical evaluation (Lueth found that animals with good contrast be­ ratios since small errors in this figure will be 1970). Eberhardt (1968) suggests that both tween themselves and their background were magnified many times when computing total sighting distance and flushing distance be seen more often than animals with low con­ estimates (Dasmann 1952). recorded and recommends the use of radio trast. In Colorado, Gilbert and Grieb (1957) The primary limitation to the sex-age-kill telemetry to study animal responses to found that aerial crews counted 34% of the method is obtaining accurate input data. observers. total (0. hemionus) present with Harvest estimates from check stations are The basic assumption with this method is poor to fair snow conditions but 45-49% with subject to error due to nonregistration (rang­ that deer are randomly distributed and that good to excellent snow conditions. McKenzie ing from 1% in Nebraska to 22% in Min­ they are not attracted to, or avoid, roadside (1972) reported the number of deer sighted in­ nesota). Harvest estimates from postcard areas. Weather greatly influences deer creased as snow cover increased but good surveys are influenced by hunter bias and behavior and observability so a standardized snow conditions were present only once dur­ nonreporting error. Biological data collected set of weather factors is required before con­ ing the 5-year study in North Dakota. by professional field personnel at check sta­ ducting this survey. Height above ground, speed of aircraft, tions are preferable to data obtained by gas Michigan, Nebraska, Kansas, and Wiscon­ transect width, and time of day are important station operators, grocery store owners, sin use field observations to determine sex factors in standardizing aerial surveys sportsmen, and other nonprofessionals. But and age ratios or number of fawns per doe. (Graham and Bell 1969, Caughley et al. 1976, even with professionals, inconsistencies in Population estimates are then developed Norton-Griffiths 1976). It is also important to aging deer can occur regardless of whether from harvest data utilizing observed ratios. document differences between observers, the tooth wear and replacement or incisor Ratios can be determined before (Michigan) pilots, aircraft, habitat, terrain, and deer sectioning method is used (Ryel et al. 1961, or after (Kansas) the deer season and can be densities (Jolly 1969a, b). Using a helicopter Gwynn 1978). Determining of nonhunting used to develop productivity indices instead of a fixed-wing aircraft provides mortality is also questionable because of the (Nebraska). Problems with this survey include slower speeds and increased maneuverability difficulty in accurately measuring this loss. inaccurate record keeping by field personnel (Buechner 1950, Aldous 1956). Good statis­ The major assumption with this technique and mistakes in identifying sex and age of tical design is necessary if bias is to be is that vulnerability is equal and therefore sex observed animals. estimated or if expanded population and age ratios in the harvest truly represent estimates are calculated (Caughley and those in the wild. The danger with this Aerial Surveys Goddard 1972, Caughley 1974). assumption is that differences in vulnerabili­ One important aspect of aerial surveys Aerial surveys of big game species have ty are caused by behavioral traits of animals should be a comparison of aerial counts to been studied extensively in North America as well as hunter selectivity for larger ground counts to obtain the percentage of (Saugstad 1942, Buechner et al. 1951, Petrides animals. There is evidence that 1 'h-year-olds animals seen. This will enable the resource 1953, Edwards 1954, Erickson and Siniff 1963, are slightly more vulnerable to hunting than manager to expand survey results to a Siniff and Skoog 1964, Lovaas et al. 1966, older deer (Eberhardt 1960) while fawns are population estimate so that a rate of harvest LeResche and Rausch 1974). Most studies the most vulnerable age class (Van Etten et can be determined. Gilbert and Grieb (1957) report factors that bias their results and the al. 1965). used drive counts to estimate total attempts made to standardize techniques. population size for comparison with aerial Aerial surveys are used by Missouri, Min­ Roadside Observations counts. nesota, and Wisconsin to provide total Some of the major disadvantages of the Several states use roadside counts to pro­ population estimates in winter concentration aerial survey are: (1) difficulty in observing vide population trends and estimates of sex areas or on small game management areas. animals from the air and measuring and age ratios. Indiana estimates deer density Minnesota uses aerial surveys to estimate proportion seen, (2) variation between pilots, on some military reservations with roadside total deer populations in county-sized areas observers, and aircraft, (3) high cost, and (4) counts (Stormer et al. 1974). This technique is by expanding density estimates obtained obtaining suitable weather conditions. an adaptation of the line transect method of from randomly selected sections. North However, the aerial survey provides a estimating abundance described by Gates et Dakota uses the aerial survey of permanent valuable technique for future use with al. (1968). Prescribed routes are driven, deer study areas to provide population trends for continued refinement and the possible incor­ are counted, and an estimate of right angle large management areas. poration of photography or remote sensing. distance between the deer and observer is Adequate snow cover is probably the most made to determine sample area. Counts over important factor in counting deer in hard­ the same routes must be repeated several wood timber habitat. Watson et al. (1969) 8 Registered Buck Kill (4) Conservation officers in Iowa estimate the statistically should be discontinued so that Nebraska and Ohio use the adult buck number of deer in their assigned territories resources can be used in other areas. It is harvest as an indicator of population trends. each winter (Gladfelter 1972). These easy to criticize the accuracy of population Harvest figures are compiled from man­ estimates are influenced by the number of estimators but the ability of resource datory check stations and compared to buck deer seen during field activities, number managers to utilize this information is ap­ harvest in previous years. One major assump­ of roadkills processed, and hunter success parent by many successful hunting seasons tion is that changes in the harvest represent during the previous season. It is difficult and a healthy, growing deer herd. With the fluctuations in the population. Bias may be to accurately estimate deer numbers in a development of complex systems for consistent from year to year or may vary due county-sized area so this technique is manipulation of the harvest in small to changes in hunting pressure, length of utilized as a post-season indicator of geographic areas and good law enforcement, season, weather conditions, crop harvest, population trends. deer management can become more precise. hunter selectivity, hunting regulations, (5) Crop depredation reports received from Therefore, if improved management capabili­ hunter success, and nonhunting mortality. If landowners are used as a deer trend in­ ties are to be fully utilized, more accurate possible, these factors should be evaluated dicator in Missouri. Damage to crops, or­ population estimators will be needed. The when comparing harvest figures. chards, and young trees may be in­ easy techniques have already been con­ Wisconsin uses the buck harvest indepen­ fluenced by deer population size, amount sidered. dently to indicate trends and also to estimate of food available, weather, and other fac­ New techniques should be designed to total population size. Population size tors. Korschgen (1962) found that land­ facilitate accurate data collection and re­ calculations assume that the registered buck owner complaints about deer damage in cording and statistical analysis. Assumptions harvest represents removal of from 12% to Missouri became more numerous when should be definedand then measured to iden­ 14% of the total population (McCaffery more than 50% of the deer's diet con­ tify variability. All facets of deer behavior, 1973). The removal percentage is determined sisted of agricultural crops. mortality, natality, sex and age structure, from the age structure of the harvest. (6) Population modelling is being used in Min­ vulnerability, and hunter harvest should be nesota (and tested in many other states) to explored for possible clues to the accuracy of Other Estimators estimate population trends and predict ef­ estimator techniques. New techniques should fects of various season regulation be tested in areas of known deer population Several other methods are used to obtain changes. Computer simulations of popula­ size. They should be justified in terms of ex­ deer population estimates or trends: tion changes are provided following input penditure of money and manpower and ac­ (1) Kansas and Indiana use the change in of sex and age data, reproductive rates, complishment of management goals. In­ hunter effort per deer harvested as a harvest rates, and sex and age specific creased research in this area by conservation population trend indicator. Information mortality factors. Good input data are agencies and educational institutions will be on hunter harvest and effort are obtained essential. At the present time, the most important to the future of deer management from questionnaires sent to hunters. Ac­ valuable asset of population modelling is in the Midwest Farmland. curacy is subject to truthful reporting by indicating areas of weakness in input data. hunters, sampling error, and differences in vulnerability of animals due to behavior, CONCLUSIONS Acknowledgments. I would like to sex, age, or season regulations. acknowledge the following wildlife biologists who (2) Kansas and Wisconsin periodically survey In my oprruon, most deer population provided information on deer population a random sample of landowners who are estimator techniques probably cannot estimators used in their respective states: Dave asked to estimate the number of deer on measure changes of less than 20% in deer Arnold, Michigan Department of Natural their land, and in Kansas opinions on hunt­ populations. Utilizing more than one popula­ Resources; AI Berner, Minnesota Department of .ing and deer damage are also obtained tion estimator allows comparisons that Natural Resources; Bill Creed, Wisconsin Depart­ (Peabody 1976). Survey results are used as enhance the probability of detecting more ment of Natural Resources; Forrest loomis, Illinois Department of Conservation; Jim McKenzie, North a trend indicator but probably are more subtle changes. Resource managers have a Dakota Game and Fish Department; Karl Menzel, important in measuring landowner at­ good knowledge of assumptions and limita­ Nebraska Game and Parks Commission; John titudes. tions for each technique, but tend to Olson, Indiana Department of Natural Resources; (3) Track counts are used to estimate popula­ overlook them because of the difficulty in ac­ Wayne Porath, Missouri Department of Conserva­ tion trends in Ohio. This technique is curately measuring their effects. Techniques tion; les Rice, South Dakota Department of described in the paper by Mooty. which are not reliable or cannot be measured Wildlife, Parks, and Forestry; Keith Sexson, Kansas 9 j Forestry, Fish, and Game Commission; and Bob servo Game Div. Rep. No. 2282. 192 pp. Taylor, ed. The deer of North America. Stoll, Ohio Department of Natural Resources. 1968. A preliminary appraisal of line transects. The Stackpole Co., Harrisburg, Pa. 668 Funded in part by the Federal Aid to Wildlife J. Wildl. Manage. 32(1):82-88. pp. Restoration Act under Pittman-Robertson Project 1969. Population analysis. Pages457·95 in R.H. JAHN, L.R. W-115-R. Giles, [r., ed. Wildlife management 1959. Highway mortality as an index of deer­ techniques. The Wildl. Soc., population change. J. Wildl. Manage. LITERATURE CITED , D.C. 623 pp. 23(2):187-97. EDWARDS, R.Y. JOLLY, G.M. ALDOUS, S.E. 1954. Comparison of an aerial and ground cen­ 1969a. The treatment of errors in aerial counts 1956. Conducting deer studies with the use of sus of . J. Wildl. Manage. of wildlife populations. East Afr. Agric. a helicopter. J. Wildl. Manage. 18(3):403-4. and For. J. 34(spec. issue):5G-55. 20(3):327-28. 1969b. Sampling methods for aerial censuses of ERICKSON, A.W. and D.B. SINIFF wildlife populations. East Afr. Agric. and BUECHNER. H.K. 1963. A statistical evaluation of factors in­ For. J. 34(spec. issue):46-49. 1950. Use of the helicopter in wildlife work. J. fluencing aerial survey results on brown Wildl. Manage. 14(4):472-73. bears. Trans. North Am. Wildl. and Nat. KORSCHGEN, L.J. Resour. Conf. 28:391-409. 1962. Foods of Missouri deer, with some BUECHNER, H.K., 1.0. Buss, and H.F. BRYAN management implications. J. Wildl. 1951. Censusing by airplane in the Blue GATES, C.E., W.H. MARSHALL, and D.P. OLSON Manage. 26(2):164-72. Mountains of Washington. J. Wildl. 1968. Line transect method of estimating Manage. 15(1):81-87. grouse densities. Biometrics 24:135-45. LANG. L.M. and G.W. WOOD 1976. Manipulation of the deer CAUGHLEY, G. GILBERT, J.R. herd. Wildl. Soc. Bull. 4(4):159-66. 1974. Bias in aerial survey. J. Wildl. Manage. 1978. Estimating population characteristics. 38(4):921-33. Pages 297-304 in J.L. Schmidt and D.L. LAUCKHART. J.B. Gilbert, eds. Big game of North America, 1950. Determining the big-game population CAUGHLEY, G. and J. GODDARD ecology and management. Wildl. from the kill. Trans. North Am. Wildl. 1972. Improving the estimates from inaccurate Manage. Inst., Washington, D.C. 494 pp. Conf. 15:644-50. censuses. J. Wildl. Manage. 36(1):135-40. GILBERT. P.F. and J.R. GRIEB LERESCHE, R.E. and R.A. RAUSCH CAUGHLEY. G., R. SINCLAIR, and D. SCOTT·KEMMIS 1957. Comparison of air and ground deer 1974. Accuracy and precision of aerial moose 1976. Experiments in aerial survey. J. Wildl. counts in Colorado. J. Wildl. Manage. censusing. J. Wildl. Manage. Manage. 40(2):290-300. 21(1 ):33-37. 38(2):175-82. CREED, W.A., B.E. KOHN, and K.R. MCCAFFERY GLADFELTER, H.L. LOVAAS, A.L., J.L. EGAN. and R.R. KNIGHT 1978. Deer population measurements in 1972. Deer in Iowa. Iowa Conserv. Comm. 1966. Aerial counting of two Montana elk management units. Wis. Dep. Nat. Wildl. Res. Bull. NO.1. 19 pp. herds. J. Wildl. Manage. 30(2):364-69. Resour. Perf. Rep. Pittman-Robertson Proj. W-141-R-13. 12 pp. GRAHAM, A. and R. BElL LUETH, F.X. 1969. Factors influencing the countability of 1970. Drive, roadside and track counts for deer DASMANN, R.F. animals. East Afr. Agric. and For. J. census or indices: a review. Pages 27-28 1952. Methods for estimating deer populations 34(spec. issue):38-43. in Deer population dynamics and census from kill data. Calif. Fish and Game methods: a review. Proc. Conf. 38(2):225-33. GWYNN. J.Y. 1978. Tooth wear and replacement vs. cemen­ Southeast. Assoc. Game and Fish Comm. DAVIS, D.E. tum analysis age assignments of yearly 59 pp. 1963. Estimating the numbers of game popula­ Virginia white-tailed deer. Pap. MCCAFFERY, K.R. tions. Pages 89-118 in H.S. Mosby, ed. presented at 14th Northeast. Deer Study 1973. Road-kills show trends in Wisconsin deer Wildlife investigational techniques. The Group meet. 5 pp. populations. J. Wildl. Manage. Wildl. Soc., Washington, D.C. 419 pp. 37(2):212-16. HANSON, W.R. DOWNING, R.L., E.D. MICHAEL. and R.J. POUX. JR. 1963. Calculation of productivity, survival, McKENZJE. LV. 1977. Accuracy of sex and age ratio counts of and abundance of selected vertebrates 1972. Evaluation of the winter aerial white­ white-tailed deer. J. Wildl. Manage. from sex and age ratios. Wildl. Monogr. tailed deer census technique with special 41(4):709-14. No.9. 60 pp. reference to snow cover, five-year sum­ EBERHARDT, L.L. HOSLEY, N.W. mary. N.D. Game and Fish Dep. Prog. 1960. Estimation of vital characteristics of 1956. Management of the white-tailed deer in Rep. Pittman-Robertson Proj. W-f>7-R-11. Michigan deer herds. Mich. Dep. Con- its environment. Pages 187-259 in W.P. 19 pp.

10 MUSTARD. E.W. and V. WRIGHT PAULIK. G.J. and D.S. ROBSON SAUGST AD. S: 1964. Food habits of Iowa deer. Iowa Conserv. 1969. Statistical calculations for change-in­ 1942. Aerial census of big game in North Comm. Prog. Rep. Pittman-Robertson ratio estimators of population Dakota. Trans. North Am. Wildl. Conf. Proj. W-99-R-3. 35 pp. parameters. J. Wildl. Manage. 33(1):1-27. 7:343-56. NIXON. e.M. PEABODY. W.e. SIN IFF. D.e. and R.O. SKOOG 1970. Deer populations in the Midwest. Pages 1976. Results of winter1975-76 landowner deer 1964. Aerial census of caribou using stratified 11-18 in White-tailed deer in the survey. Kans. For., Fish, and Game random sampling. J. Wildl. Manage. Midwest. U.S. For. Servo North Cent. For. Comm. Prog. Rep. Pittman-Robertson 28(2):391-401. Exp. Stn. 34 pp. Proj. W-23-R-14. 26 pp. STORMER. F.A., T.W. HOEKSTRA. e.M. WHITE. and e.M. PETRIDES. G.A. KIRKPATRICK NIXON. e.M., M.W. MCCLAIN. and K.R. RUSSElL 1970. Deer food habits and range 1949. Viewpoints on the analysis of open 1974. Assessment of population levels of characteristics in Ohio. J. Wildl. season sex and age ratios. Trans. North white-tailed deer on NAD Crane. Purdue Manage. 34{4):87G-86. Am. Wildl. Conf. 14:391-410. Univ. Res. Bull. No. 910. 11 pp. 1953. Aerial deer counts. [. Wildl. Manage. VAN ETTEN. R.e., D.F. SWITZENBERG. and L. EBERHARDT NORTON-GRIFFITHS. M. 17(1):97-98. 1965. Controlled deer hunting in a square-mile 1976. Further aspects of bias in aerial census enclosure. ]. Wildl. Manage. 29(1):59-73. of large mammals. J. Wildl. Manage. Rupp. R.S.. WATSON. R.M., G.H. FREEMAN. and G.M. JOLLY 1966. Generalized equation for the ratio 40(2):368-71. 1969. Some indoor experiments to simulate method of estimating population abun­ problems in aerial censusing. East Afr. OVERTON. W.S. and D.E. DAVIS dance. J. Wildl. Manage. 30(3):523-26. Agric. and For. 34{spec. issue):56-59. 1969. Estimating the numbers of animals in I. wtldltf.e populatt6M,.pag5403-ss in RvEt.l.A., Ie;!); FAY. and R.C. VAN fHEN WATT, P.G., G.L. MILLER, and R.I. RoBn R.H. Giles, Ir., ed. Wildlife management 1961. Validity of age determination in 1967. Food habits of white-tailed deer in north­ techniques. The Wildl. Soc., Michigan deer. Mich. Acad. Sci., Arts, eastern Kansas. Trans. Kans. Acad. Sci. Washington, D.e. 623 pp. and Lett. 46:289-316. 70(2):223-40.

0: Z o c ·Vi c o u '" ~

11 Wisconsin DNR MONITORING DEER POPULATIONS IN THE NORTHERN FORESTED AREAS OF THE MIDWEST

Jack J. Mooty Minnesota Department of Natural Resources 201 S.W. Golf Course Road Grand Rapids, MN 55744

Abstract: Population management of white­ compromise between manpower,. money, needs, to apply at any time of the year. Bergerud tailed deer (Odocoileus virginianus) requires and efficiency. (1968:22) stated that the Iiterature abounds reliable census data. Over the past 50 years with methods and unverified results of big numerous techniques have been developed and INTRODUCTION game counts (Hazzard 1958). Further, he pro­ evaluated. Some methods which have not proved posed the following minimum standards: (1) workable are string counts, bed counts, roadside It is difficult to obtain accurate and effi­ that the accuracy of the method be verified counts, and highway mortality. Methods of limited cient counts of free-ranging ungulates. usefulness Include deer drives,strip counts, various with a known population, or (2) that 2 com­ However, such knowledge is essential for mark-recapture systems, and aerial counts. Some pletely independent census methods give techniques in current use in northern forests of the population management. The white-tailed similar results. midwestern United States are counts of trails, deer in the northern forested areas of the Census techniques for white-tailed deer tracks, sex-age-kill analysis, pellet group counts, midwestern United States (Fig. 1) is no excep­ have shown a degree of imagination on the and population modelling. Numbers of deer trails tion. Heavy forest cover and the secretive part of the investigators, ranging from field (paths) have been correlated with other indexes of nature of deer make them very difficult to counts to complex multivariate models. Most deer abundance in Wisconsin, and showpromise as census. Direct counts are not reliable techniques must meet one overwhelming an efficient and accurate way to census deer in because the observer never knows what pro­ criterion - to fit within the limited budgets of forested areas. Roadsidetrack counts will estimate portion of the population has been sighted. deer populations, but are limited by the non­ most wildlife agencies. For this reason wildlife managers usually rely random distribution of suitable roads and traffic volume. Use of the sex-age-kill technique for on counts of sign left by deer. This paper em­ METHODS IN GENERAL USE monitoring deer populations requires data on age phasizes the review and application of these ratios, legal kill, hunting effort, and the rate of indirect counts. Trail Counts natural mortality. At present, the pellet group Most of the methods used to census deer in count is the preferred deer census method for the northern forest are census indexes, i.e, Bartlett and Stephenson (1929) attempted northern Midwest forests. This method lends itself counts or ratios which are relative in some to use trails to census deer in wintering areas to unbiased random sampling, and appropriate sense to the total number of animals in a in Michigan. However, the use of trails to statistical analysis. However, large amounts of specified population (Overton and Davis census deer during the snow-free period is a manpower are required to complete the counts. 1969:404). As such, these methods are merely recent innovation. McCaffery (1976) used 50 Estimating deer numbers by population modelling 2 is becoming increasingly popular. Models require a refinement and systematic application of O.4-km transects/1,000 km to estimate deer baseline data on natality and mortality. The tech­ the hunter's technique of "reading sign" to populations in Wisconsin. Completion of niques employed by various agencies are often a determine animal numbers (Bennett et al. these counts required 3-4 man-days/1,000 1940). Be humbled by the fact that, as krn". The number of deer trails was positively wildlife resource managers, we are merely related to other indexes of deer abundance Mooty, Jack J. 1980. Monitoring deer populations defining the limits of a relationship which has such as the adult buck kill (r = +0.91), pellet in the northern forested areas of the Midwest. been known to exist since man became a group counts (r = + 0.89), and sex-age-kill Pages 13-22 in Ruth L. Hine and Susan Nehls, eds. ratios (r 0.94). Best survey results (closest White-tailed deer population management in the hunter. = north central states. Proc. 1979 Symp. North Cent. The ideal census method for deer would be agreement with "best" population estimate Sect. Wildl. Soc. 116 pp. precise, accurate, inexpensive, fast, and easy available for the particular area) were ob­

13 ~ are made the next day. I found while doing track counts in Minnesota that dragging, especially after a rain, was most convenient and produced a surface suitable for counting. In addition we took advantage of road grading when it was done, but could not rely on this as it occurred at irregular intervals and did not always fit our work schedule. I never felt comfortable with relying on rain alone to prepare the road because light rains might not eliminate some old tracks and heavy rains influence deer activity. Variability of track count data is usually HURON high because of the smaller sample size (the route is the sampling unit). For example, on a MINNESOTA 842-km l study area in northeastern Min­ nesota, 10 years of track and pellet group count data had average 95% confidence in­ tervals ± 30% and ± 20% of the means, respectively. One of the critical limitations of the track ~ Forest Area WISCONSIN count technique is that in most cases routes MICHIGAN cannot be selected at random, introducing a possible source of bias. This bias should be reduced in direct porportion to the extent and distribution of countable roads in relation to FIGURE 1. Northern forest area of the midwestern United States. deer habitat. During the spring and fall, phenological tained when there was a -2C temperature correlate well with other indexes of deer changes and green vegetation attract deer to prior to 20 October. That is, counts con­ abundance. McCaffery (pers. comm.) cor­ roadsides, and counts made at these times ducted in the fall or spring following frosts in related track counts with the kill of bucks (r would probably be highly variable. Our early autumn produced the best results. = +0.85) and Mooty (in prep.) and Moran in counts in Minnesota were made between 1 McCaffery (1976) noted very little change Michigan (pers. comm.) found significant cor­ July and 15 August. Earlier work in Minnesota in trail abundance from fall to spring. relations between populations estimated by (Kohn and Mooty 1971) indicated that deer However, 65% of the trails marked in the track counts and populations estimated by .do not avoid or prefer roadsides at this time. spring were not visible in July. New and re­ pellet group counts (r = + 0.83 and + 0.85, Other problems associated with deer track opened trails began to appear in August. respectively). Howe (1954), in comparing counts include varying observer ability to see Trail abundance was also positively related several census techniques (i.e. strip census, tracks, observer fatigue, erasure of tracks by to the percentage of shade-intolerant forest Delury method [Delury 1947], deer drives, vehicle traffic, and differences in judgment types (r = 0.79, P < 0.01). Based on the ex­ mileage counts, sex and age ratios, shining as to how many deer are represented when perience in Wisconsin, the trail count seems counts, time area counts [Chiavetta 1952], several sets of tracks occur together. Addi­ to hold promise and further trials are war­ and track counts) on the Rifle River Area of tional problems include the trend towards ranted. Michigan concluded that track counts gave a blacktopping of rural roads, and the attrac­ conservative estimate of the deer population. tion of deer to logged areas which are usually Track Counts Preparation of the surface for counting and concentrated along roads. elimination of old tracks is done the day Given proper distribution of roads, ac­ Work in Wisconsin, Michigan, and Min­ before the count through rainfall, or by curate counts, and correct timing, I believe nesota indicates that deer tracks along roads grading or dragging. Counts of fresh tracks this technique will accurately reflect deer ~ 14 numbers. The following papers provide more assumptions implicit in the use of pellet were 17.8% lower than re-counts. More re­ information on track counts: Schrader (1949), group counts (Ryel 1971 :16) are: cent data from Minnesota suggest a trend Tyson (1959), Brunett and Lambou (1962), (1) Deer defecate at a rather constant fre­ towards undercounting equal to 15-20% (Mooty, unpubl.). Ryel (1971) stressed the im­ Downing et al. (1965), Harlow and Downing quency. portance of training and monitoring field (1967), and Daniel and Frels (1971). (2) Pellet groups persist long enough to be crews and Van Etten and Bennett (1965) Track counts may also lend themselves to counted. believed that errors would be reduced by measuring habitat selection by recording the (3) Pellet groups can be found and counted in using a plot shape that can be readily vegetation adjacent to tracks, and produc­ the field. searched and by using 2 experienced tion and survival by recording fawn tracks (4) An explicit deposition period can be observers to check each other's findings. As a separately from adults (Moran 1978). In Min­ delineated. result of this study Michigan went to a rec­ nesota we found a significant correlation (r (5) The age of pellet groups, which are pre­ = +0.91) between fawn tracks and the sent, can be established relative to the tangular plot (21.8 X 3.6 m, 0.008 hal. Their second recommendation for reducing error is previous winter's severity, as measured by deposition period. Verme's (1964) Winter Severity Index. As applied in Minnesota the procedure is that the position of pellet groups upon the Iit­ ter be used as the age criterion - i.e., all roughly as follows: groups on top of the leaf litter should be Pellet Group Counts (1) Define the area to be censused. counted as new. In areas without leaf litter, (2) Determine required sample size. such as meadows and pine stands, criteria for Any deer hunter "worth his salt" would (3) Stratify the census area by expected deer aging of pellet groups need to be developed. leave an area where he consistently sees very density. little deer sign. When he does this he is in ef­ (4) Determine which areas are accessible fectffiakingOa crude census, The pe1fef group such as those within 1.0 km of a driveable Population ModeIHng count technique refines this evaluation road. These form the population from While models are not an actual census through the use of an equation with which the sample is drawn. technique, they can be used to estimate deer "knowns" and an "unknown". The knowns (5) Allocate sample among strata using op­ numbers provided some baseline data are (pellet groups, defecation rate, deposition timum allocation. available (Gross 1973). For most models in period, and size of area searched) are used to (6) Select a random sample for each strata. use the basic data needed are a population determine the unknown deer density. Of Assuming a stratified sample, analysis con­ estimate, and estimates of age-specific course, the knowns must be accurately sists of calculating a weighted mean by strata natality and mortality. Models are measured. of pellet groups per course and then using the "powerful" in that we can look at the effects The first reported use of deer pellet group following formula: of various rates of mortality or natality on the counts occurred around 1940 (Bennett et al. total population or a given segment (Walters 1940, Interstate Deer Herd Committee 1946). x pellet and Gross 1972). However, the use of models At present it is the most accepted and used groups per x 1 x size Deer course plot size of area will only be as good as the data that goes into deer census method in the northern and = population deposition x defecation X no. them. One of the important benefits of using western states. Surprisingly though, only 2 period rate of plots modelling is that it stimulates interest in get­ studies (Eberhardt and Van Etten 1956, Ryel ting good data. An update on the current 1971) have tested the technique on an area For more detail on design and calculations, status of big game population modelling is with a known population of white-tailed deer. see Neff (1968), Smith et al. (1969), and Ryel given by Pojar and Strikland (1979). They concluded that while the technique is (1971). subject to serious errors resulting from The most serious problem associated with Sex-Age-Kill Estimates missed and incorrectly aged pellet groups, pellet group counts in the forested area of that these can be controlled sufficiently to the Midwest is that of making accurate This method is similar to those known in give reliable results. counts of the groups present on plots and fisheries work as estimates of virtual popula­ The pellet group count is at present the distinguishing between old and new groups. tions (Ricker 1958) in which the annual preferred method for censusing white-tailed Rechecks of plots indicate that large errors catches of a given year class are summed un­ deer in the northern forests of the Midwest. are possible. Van Etten and Bennett (1965) til the class disappears from the catch. The This is because the technique lends itself to reported an undercount of 22% of the new method, when used for deer, involves 2 steps: unbiased random sampling, and appropriate groups actually present. Petraborg and (1) estimating the buck population, and (2) ex­ statistical analysis. Several important Idstrom (1972) found that original counts panding the estimate to the total deer

15 population. The necessary data are, ac­ sample. The other major problem with drive all the 16-ha lines and one-half of the boun­ cord ing to Eberhardt (1960), age ratios, counts is the requirement for 100 or more dary lines on a 10.4-km 2 area. All of the number of deer legally killed, hunting effort people for the census. A technique described animals flushed are counted and the flushing records, and an assumed specific rate of by Morse (1943) greatly reduces the man­ distance is measured. The average flushing natural mortality. If there are large losses to power required in the drive by using deer distance from the observer determines the sources other than legal harvest, then there tracks in the snow instead of standers to tally width of the sample strip. Intuitively, the strip are problems with this techique. However, in the number of deer leaving the drive area. In census has much appeal because it is easy to postseason dead deer surveys in Michigan a report on deer drives in Minnesota, Olson do and large areas can be covered in a short (Eberhardt 1960, and D.A. Arnold, pers. (1938), stated that the average drive area was time with limited manpower. comm.) and Wisconsin (Creed 1966) very few 224 ha. Drives on areas larger than this were This type of census was first used to census adult male deer were found. So, the assump­ less successful and more difficult to manage. (Bonasa umbel/us) by King tion that legal harvest accounts for most of According to Adams (1938), the main disad­ (Leopold 1933), and to census deer by the hunting kill of adult males is apparently vantages of the drive method are disor­ Erickson (1940) in Minnesota. Krefting and valid. Large, unaccounted-for poaching ganized drive lines, exaggeration of the Fletcher (1941) used the strip census on deer losses will also affect sex-age-kill calcula­ number of deer seen, heavy cover, and in Oklahoma. Hahn (1949) applied a variation tions. Recent studies (Vilkitis 1968, 1971) in­ adverse weather. of the strip method in censusing deer in the dicate that such losses may be common. These counts were used extensively during Edwards Plateau region of Texas. Wisconsin presently uses the sex-age-kill the 1930's when the Civilian Conservation However, there are some important method to estimate deer populations, and Corps (CCC) provided large amounts of man­ unanswered questions in extrapolating data assumes that 80% of the adult bucks are power. The longest continuous record of from the strips to the census area. For exam­ taken annually by legal hunting. drive counts (1933-61) is from the George ple, one does not know what proportion of Additional references for this technique Reserve in Michigan (O'Roke and the animals are seen at a given distance from are Kelker (1940), Allen (1942), Lauckhart Hamerstrom 1948, Chase and Jenkins 1962). the observer (visibility bias, Eberhardt 1968). (1950), Dasmann (1952), Selleck (1957), and In 1971 there was renewed interest in the Eberhardt (1968:22) stated: "For line Hanson (1963). deer drive census in Minnesota (Carter 1971). transects, the major unknowns pertain to the Twenty-three areas totalling 31.5 km 2 were behavior of individual animals. Very likely METHODS OF LIMITED USE driven by private citizens. On these fall the most serious problem has to do with the counts an average of 4.7 deer/krn" were prospect that some individuals will move Drive Counts counted. Pellet group counts the previous before 'flushing'. Most of the methods for spring on 26,000 km 2 of northern deer range dealing with transect data assume some kind Briefly, drive counts are made on areas estimated an overwinter population of 5.3 of frequency distribution model and are sub­ with a distinguishable boundary - usually deer/km" (Karns 1971). ject to bias if the assumed distribution does roads. Observers (standers) are placed on the Drive counts are of only limited value for not apply." Eberhardt (1978) recommended roads on 3 sides, and drivers begin on the censusing deer over large areas in heavily using Hayne's method (Hayne 1949) when fourth side and move through the area. Deer forested parts of the Midwest. They will prob­ working with animals that flush, as it does not which leave are counted by the standers and ably work best on small areas where there is a assume a specific model. those animals which cut back through the good road system and a readily available Recent improvements in correcting for the drive line are counted by the drivers. manpower supply. The U.S. Forest Service visibility bias (Anderson and Pospahala 1970) Theoretically, this results in a total or near (1935) and Trippensee (1948) provide more of this method make its use for censusing of total count of the deer on the area; and with a detail on setting up and conducting deer immotile objects, such as dead deer, more well-trained crew working on an area with drives. feasible. Robinette et al. (1974) found that the good visibility on the boundaries and good in­ method of Anderson and Pospahala (1970) terval marking of the drive area, this is prob­ Strip Census and that of Kelker (1945) gave close estimates ably the case. of a known population of inanimate objects. In areas with a good road system The strip census is dependent upon the King's method, while biased, was less so distributed throughout the deer range, drive number of animals flushed in a sample strip than other methods based on sighting counts will work well. However, in most and an expansion of these numbers to the re­ distances. Anderson et al. (1979:72) stated northern areas, forest roads are poorly maining unsampled area. As devised by King that the following assumptions are critical for distributed, leading to problems of a biased (Leopold 1933) the census consists of walking valid results: 16 (1) Points directly on the Iine are seen with Petraborg and Idstrom (1972) found that spot­ areas can be covered in a relatively short probability 1. lighting counts of marked and unmarked deer time. However, in reality the technique is no (2) Points are fixed at the initial sighting posi­ underestimated the population as more problem-free than any other. Because tion (they do not move before being determined by pellet group counts. of heavy forest cover, their more accessible sighted) and none are counted twice. To overcome some of the difficulties with range, and the availability of other census (3) There are no measurement errors and no the mark-recapture technique, Martin (1970) methods, aerial counts have not been used rounding errors (distances are measured presented a regression method of correction extensively for deer in northern forests of the exactly). for unequal catchability, and Eberhardt Midwest. The problem of visibility bias is (4) Sightings are independent events. (1969) recommended shifting traps at least summed up by Caughley (1974:92) when he In addition, a minimum of 40 objects should once as a check on bias. Another way to stated: "Aerial censuses of large mammals be located. overcome this problem is to "recapture" are inaccurate because the observer misses a While conducting strip counts, it is also marked animals visually from aircraft. This significant number of animals on the possible to record data on habitat conditions, method has been used successfully by Rice transect. The accuracy deteriorates pro­ habitat selection, and herd composition. and Harder (1977) and Floyd et al. (1979). At gressively with increasing width of transect. Other wildlife flushed may also be tallied. best, the mark-recapture technique using cruising speed and altitude". He suggested With these improvements in the strip census, aircraft seems suited for use on areas such as that the best strategy may be to measure the this method should work well for estimating military reservations, wildlife management bias and correct the estimates accordingly. dead deer numbers. The most serious prob­ units, and research areas. One recent study which attempts to measure lems are finding the required minimum of 40 A considerable volume of literature is this bias when counting deer is that of Floyd objects and the necessary manpower. For a available regarding this technique. Several et aL(1979).lnmaking counts of deer on sam­ comprehensive, current discussion of this methods for analysis of mark-recapture data pie areas in northeastern Minnesota where a technique, see Burnham et al. (1980). are presented by Parr et al. (1968), Paulik and known number of radio-marked deer were Robson (1969), and Manly (1970). A located, 42% of the radio-marked animals Mark-Recapture bibliography of statistical methods was were visible from the aircraft. This obser­ prepared by Tepper (1967). The following vability factor was then used to estimate the The mark-recapture method is one of a papers give more information on this deer population in a similar census of the group of population estimators based on technique: Petersen (1896), Lincoln (1930), larger area. Woolf and Harder (1979) saw an some change in the ratio of two identifiable Ricker (1958), DeLury (1958), Seber (1965), average of 32% of the marked deer in an characteristics of a population (Paulik and Jolly (1965), Edwards and Eberhardt (1967), aerial helicopter census of white-tailed deer Robson 1969). Other approaches utilize Nixon et al. (1967), and Anderson (1975). in Pennsylvania. For mule deer (Odocoileus changes in observed sex ratios and/or age hemionusJ on their winter range in Colorado, composition (Kelker 1940, Rasmussen and Aerial Counts Gilbert and Grieb (1957) determined air Doman 1943). observability to be 43% of the animals ob­ This technique is not, to my knowledge, Direct Counts. Since the pioneering work served on the ground by deer drives. Jacob­ currently used as a regular census method for of Saugstad (1942), aerial censusing of big son and Weatherill (1975), using multiple estimating deer numbers in the northern game animals has increased and been re­ regression analysis, determined a visibility forests of the Midwest - the major problem fined. From the general survey approach bias of at least 46% for an aerial census of being that when standard trapping methods (Morse 1946), efforts have progressed to white-tailed deer in Manitoba. In Texas (Hahn are used, the assumption of equal probability transects of a determined width (Banfield et 1949) reported that only 50% of the deer of recapture is violated (Eberhardt 1969). For al. 1955, Bergerud 1963) to stratified sampling observed on cruise counts were seen from the example, Strandgaard (1967), in a Petersen and optimum allocation (Siniff and Skogg air. LeResche and Rausch (1974) determined method (Petersen 1896) census of 1964, Evans et al. 1966). Aerial counts have observability for moose to be 68% and 43% (Capreo/us cspreolus}, found this procedure been most often used in the north to census for experienced and inexperienced observers, unsatisfactory because deer observed caribou (Rangifer tarandusJ and moose (Alces respectively. These data suggest that about (recaptured) did not constitute a truly elces) in remote areas (Edwards 1954, one-half of the animals are seen in an aerial random sample from the total population. Banfield et al. 1955, Bergerud 1963, Bergerud census in forested areas. Andersen (1962) found that 75% or more of and Manuel 1969, and Timmermann 1973). In an attempt to improve the aerial census the population of roe deer had to be trapped Aerial census of wildlife has much appeal. of deer in Manitoba's Interlake Region, to get a reliable estimate of percent fawns. Flying is exciting, animals are seen, and large Jacobson and Weatherill (1975) compared

17 2.6-km 2 quadrats and variable-sized strip i.e., counts of breaks in a string strung concentrate on these "verified census" plots flown at various air speeds and altitudes through the woods (Morse and Burcalow studies. While this will be repetitious I think and recorded the deer observed by individual 1942), and bed counts (Bartlett and it's important to bring these studies together crew members. Based on costs, they recom­ Stephenson 1929). Neither of these methods in one part of this paper. This discussion will 2 mended using strip plots (S0.50/km ), pro­ seemed to work very well and received little be limited to studies of white-tailed deer in vided adjustments for visibility bias are possi­ attention beyond these initial efforts. the Midwest. ble, although the accuracy of the quadrat Deer seen while driving roads during a Ryel (1971) compared 10 years of pellet 2 census (S1.oo/km ) was improved by using 2 specified period were used as an index to group counts with known populations based observers and overlapping strips. Quadrats deer numbers in Minnesota from 1942 to 1953 on the aging method (Jenkins 1964) for the are recommended in hilly or river valley ter­ (Erickson et al. 1961). This type of census is George Reserve of southern Michigan. The rain, and when a greater degree of accuracy subject to the vagaries of weather, results indicated poor agreement between and precision is required. As with strip plots, phenological conditions, and varying work the 2 methods. The 95% confidence limits a correction for visibility bias, based on crew schedules, and was discontinued after 1953. for pellet group counts included the known performance, is necessary. Caughley (1977) Such counts are used in Wisconsin to deter­ population in only 2 of the 10 years. The also concluded that transect sampling was mine sex and doe:fawn ratios during summer linear correlation coefficient (r = -0.62) in­ more efficient than quadrat sampling, and months. Michigan also uses deer seen by dicates an inverse relationship, but was not anyone planning an aerial census should DNR employees per 100 hours of driving by quite significant at the 0.05 level (0.6319 for 8 refer to this publication and that of Jolly county as a population trend indicator. df). Ryel (1971) also reported on a similar (1969). To overcome wartime restrictions on comparison for the Cusino enclosure in the Although there are problems with an aerial travel, and limited manpower, a variation on Upper Peninsula of Michigan. Here popula­ census, this technique may be the only this observation technique was the Railroad tion estimates based on pellet group counts answer when large areas must be censused in Engineers' Deer Tally (Erickson et al. 1961). were compared with the known population. a short time. Further improvements in aerial Engineers and firemen of 4 major railroads in The known population was determined by census will, no doubt, increase its use for Minnesota recorded all of the deer seen removal of all animals each year. The Cusino deer. The "state of the art" will improve in along the railroad rights-of-way between 15 results were somewhat better than at the direct proportion to experience and rigorous September and 15 December. Because of George Reserve. Calculated confidence and disciplined approach to study design as wartime restrictions on the railroads, the cen­ limits of 2 standard errors of the mean in­ suggested by Jacobson and Weatherill (1975). sus was discontinued after 1 year (1942). cluded the known population in 4 of the 5 Remote Sensing. Published trials on infra­ Two published studies, both from Wiscon­ years. However, the correlation coefficient red or remote sensing for the census of wild sin, using highway mortality as an index to for the data was not significant (r = +0.45, P ungulates are scarce. However, those that deer populations were reviewed. Jahn (1959) > 0.05). Rechecks of plots on both areas in­ have been reported (Croon et al. 1968, Mc­ concluded that because of inconsistencies in dicated serious undercounting of the pellet Cullough et al. 1968, Graves et al. 1972, and recording, reporting, and traffic pressure, groups actually present. Ryel (1959:18) Wride and Baker 1977), mention 2 serious these data could not be used in an index to stated: "Missing groups are by far the com­ problems with this technique: annual changes in the deer population. monest source of error we can demonstrate (1) Inability of the equipment to detect However, McCaffery (1973) found a highly with pellet group surveys. Other possible ex­ animals under conifer or deciduous leaf significant correlation (r = 0.97, P < 0.01) planations for the differences include use of cover; and between numbers of roadkills, adjusted for incorrect defecation rates and incorrect (2) Inability to distinguish between animals of changes in traffic volume, and the antlered aging of pellet groups. The better perfor­ similar size. buck harvest. mance of the pellet group count technique at Until these problems are overcome, this Cusino may be related to the better compac­ method of census will be of limited VERIFIED CENSUS STUDIES tion of the leaf litter and herbaceous vegeta­ usefulness in the northern forests of the tion by snow, and the absence of large grassy Midwest. let me return here to the census standards openings and abundant oak trees which are suggested by Bergerud (1968). Very few commoo in the George Reserve." MISCELLANEOUS METHODS studies comparing a census with a known Petraborg and Idstrom (1972) estimated population are available. Most verified deer populations by track and pellet group Two of the less well-known methods used studies are based on 2 independent estimates counts and by the lincoln Index in the Camp in attempts to census deer are string counts, of the population. In this section I want to Ripley Military Reservation in Minnesota.

18 Pellet group and track count population correlation coefficient for these 2 estimates (r timber (Canis lupus) sign, dead deer, and estimate differences ranged from 11% to = +0.43) was not significant. The pellet habitat conditions. An additional benefit is 30%. The 95% confidence limits for track group counts were completed using a simple that on the ground field surveys, such as counts included the mean deer per square random sample design. Kubisiak (1976) pellet group counts, biologists are out in the mile estimate by pellet group counts for 3 of believed that stratification would improve woods on an annual basis, thereby keeping the 4 years. Lincoln Index population the accuracy of the estimate, because deer them in touch with what is happening in the estimates based on recapture of collared distribution changes with winter conditions. woods. This is especially important for local deer by spotlighting averaged only 38% of Lincoln Index population estimates were wildlife managers and biologists. the population estimates by pellet group calculated for 5 years. These estimates ex­ As wildlife resource managers we have a counts for the 5 years. Roadside counts of ceeded the mrpe in 4 of the 5 years by responsibility to monitor deer populations deer tracks were significantly correlated with 17-66%. The correlation coefficient for the 2 using the best method that time, money, and other population indexes in Minnesota estimates (r = +0.78) was not significant. manpower will permit. Our various publics (Mooty and Karns in prep.) (tracks and pellet Trail count population estimates were close are much too sophisticated today to settle for group counts, r = + 0.83), Michigan (R}. to the mrpe for the 4 years they were com­ "seat of the pants" estimates. Intuition is no Moran pers. comm.) (tracks and pellet group pleted. In a later study, Kubisiak (1979) found substitute for hard data and often leads to er­ counts, r = +0.85), and Wisconsin (K.R. a highly significant correlation (r = + 0.97, P roneous conclusions (Gross 1972). McCaffery pers. comm.) (tracks and buck kill, < 0.01) between 11 years of trail count and r = +0.85). In addition, Mooty and Karns (in mrpe data. Aerial helicopter counts were LITERATURE CITED prep.) found that the 95% confidence limits completed in March of 1971 and 1972. These for the pellet grcupcountslncluded the track counts exceeded the mrpe by 37% and 24%, ADAMS; H.E. count average population estimate in 7 of the respectively. Sex-age-kill analyses were made 1938. Deer census and kill records of the Lake 10 years. McCaffery (1976) found a highly for 2 years and resulted in close agreement States. Trans. North Am. Wild/. Conf. significant correlation between deer trails with the mrpe. 3:287-95. and 3 other population indexes (buck harvest, ALLEN. D.L. pellet group counts, and sex-age-kill calcula­ CONCLUSION 1942. A pheasant inventory method based tions). upon kill records and sex ratio. Trans. Howe (1954) attempted to census deer on It has been said before but bears repeating: North Am. Wild/. Conf. 7:329-32. the Rifle River Area of Michigan using 8 dif­ select a census method which is valid, design ANDERSEN. J. ferent methods. The strip census (King varia­ the survey carefully for the best possible 1962. Roe-deer census and population analysis tion by cover type) gave good results. results and execute the same with a high by means of modified marking release DeLury's regression equation (DeLury 1947) degree of precision. Population data provide technique. Pages 72-82 in E.D. LeCren of kill per unit of effort and total kill pro­ a base or starting point for many other ap­ and M.W. Holdgate, eds. The exploita­ duced a good estimate of the buck popula­ proaches such as population modelling. tion of natural animal populations. Br. tion. Deer drives and mileage counts pro­ Eco/. Soc. Symp. No.2. John Wiley, New These subsequent calculations will only be as York. 399 pp. duced reasonably accurate population good as the input data. estimates. Shining and track counts tended to Generally, the greater the desired precision ANDERSON, D.R. underestimate the population. of a census, the more expensive it will be. So, 1975. Population ecology of the : V. Temporal and geographic estimates of Kubisiak (1976) censused deer on the Sand­ take time beforehand to decide how precise survival, recovery, and harvest rates. hill Wildlife Area of Wisconsin using pellet the result must be to meet your needs. Also, U.S. Fish and Wild/. Servo Resour. Pub/. group counts, mark-recapture, trail counts, consider the benefits of other information No. 125. 110 pp. aerial census, and sex-age-kill analysis. which can be recorded in conjunction with a ANDERSON, D.R., J.L. LAAKE, B.R. CRAIN. and K.P. Population estimates from these methods deer census. While the cost of a deer census BURNHAM were compared with a minimum recon­ alone might be high, by getting information 1979. Guidelines for line transect sampling of structed population estimate (mrpe) com­ on other species at the same time, the cost biological populations. J. Wild/. piled by backdating harvest, mortality, and per unit of information will be reduced. For Manage. 43(1 ):70-78. trapping records for the years 1963-72. The example, on pellet group counts in Min­ ANDERSON, D.R. and R.S. POSPAHALA 95% confidence limits for the pellet group nesota, record was also made of ruffed 1970. Correction of bias in belt transect studies count included the mrpe estimate for 5 of the grouse flushes and roosts, of immotile objects. J. Wildl. Manage. 7 years they were completed. However, the (Lepus americanus) pellets, moose pellets, 34(1):141-46. 19 Iivetrapping data. J. Wildl. Manage. BANFiElD. A.W.F., D.R. FLOOK. J.P. KElSALL. and A.G. CREED. W. 31(1):87-96. LOUGHREY 1966. Calculating deer populations by the sex­ 1955. An aerial survey technique for northern age-kill method. Wis. Conserv. Dep. 3 ERICKSON. A.B. big game. Trans. North Am. Wild!. Conf. pp. 1940. Notes on a method for censusing white­ 20:519-32. CROON. G.W., D.R. MCCULLOUGH. C.E. OLSON. JR..and tailed deer in the spring and summer. J. Wild!. Manage. 4(1):15-18. BARTLETT, I.H. and J.H. STEPHENSON l.M. QUEAL 1929. A preliminary survey of deer yards in the 1968. Infrared scanning techniques for big ERICKSON, A.B., V.E. GUNVALSON. M.H. STENLUND. Upper Peninsula of Michigan. Mich. game censusing. J. Wild!. Manage. D.W. BURCALOW. and l.H. BLANKENSHIP Acad. Sci., Arts, and Lett. 10:1-4. 32(4):751-59. 1961. The white-tailed deer of Minnesota. Minn. Dep. Conserv. Tech. Bul!. NO.5. 64 BENNETT. Ll., P.F. ENGLISH. and R. MCCAIN DANiEl. W.S. and D.B. FRELS pp. 1940. A study of deer populations by use of 1971. A track count method for censusing pellet group counts. J. Wild!. Manage. white-tailed deer. Tex. Parks and Wildl. EVANS. C.D., W.A. TROYER. and C}. LENSINK 4(4):398-403. Dep. Tech. Ser. NO.7. 18 pp. 1966. Aerial census of moose by quadrat sampling units. J. Wild!. Manage. BERGERUD. A.T. DASMANN. R.F. 30(4):767-76. 1963. Aeriat winter census of caribou. J.Wild!. 1952. Methods for estimating deer populations Manage. 27(3):438-49. from kill data. Calif. Fish and Game FLOYD. T.J., l.D. MECH. and M.E. NElSON 1968. Numbers and densities. Pages 21-42 in 38(2):225:33. 1979. An improved method of censusing deer F.B. Golley and H.K. Buechner, eds. A DELuRY.D.B. iii deciduous-coniferous forests. J. Wildl. practical guide to the study of the pro­ 1947. On the estimation of biological popula­ Manage. 43(1):258-61. ductivity of large herbivores. Int. Bioi. tions. Biometrics 3(4):145-67. GILBERT. P.F. and J.R. GRIEB Prog, Handb. NO.7. Blackwell Sci. Publ., 1958. The estimation of population size by a 1957. Comparison of air and ground deer Oxford. 308 pp. marking and recapture procedure. J. counts in Colorado. J. Wild!. Manage. BERGERUD, A.T. and F. MANUEL Fish. Res. Board Can. 15(1):19-25. 21(1):33-37. 1969. Aerial census of moose in central New­ DOWNING, R.l., W.H. MOORE. and J. KIGHT GRAVES. H.B., E.D. BElLIS. and W.M. KNUTH foundland. J. Wildl. Manage. 1965. Comparison of deer census techniques 1972. Censusing white-tailed deer by airborne 33(4):910-16. applied to a known population in a thermal infrared imagery. J. Wild!. BRUNETT. l.E. and V.W. LAMBOU Georgia enclosure. Proc. Cont. Manage. 36(3):875-84. 1%2. A test of track counts as a measurement Southeast. Assoc. Game and Fish Comm. GROSS. I.E. of deer population size. Pap. presented 19:26-30. 1972. Criteria for big game planning: perfor­ at Southeast Sect. The Wild!. Soc. meet. EBERHARDT. l. mance measures vs. intuition. Trans. 7 pp. 1960. Estimation of vital characteristics of North Am. Wild!. and Nat. Resour. Conf. Michigan deer herds. Mich. Dep. Con­ BURNHAM K.P., D.R. ANDERSON. and Ll., LAAKE 37:247-59. servo Game Div. Rep. No. 2282. 192 pp. 1980. Estimation of density from line transect 1973. Push-button deer management: boon or sampling of biological populations. 1968. A preliminary appraisal of line transects. boondoggle? Trans. Conf. West. Assoc. Wild!. Monogr. No. 72. 202 pp. J. Wild!. Manage. 32(1):82-88. Game and Fish Comm. 53:157-73. 1969. Population estimates from recapture fre­ CARTER. J. HAHN. H.C., JR. quencies. J. Wild!. Manage. 33(1):28-39. 1971. A look at the public's role. Pages58-63in 1949. A method of censusing deer and its ap­ M.M. Nelson, ed. The white-tailed deer 1978. Transect methods for popu lation plication in the Edwards Plateau of in Minnesota. Minn. Dep. Nat. Resour. 88 studies. J. Wild!. Manage. 42(1):1-31. Texas. Tex. Game, Fish, and Oyster pp. EBERHARDT. l. and R.C. VAN ETTEN Comm. Final Rep. Pittman-Robertson Proj. W-25-R. 24 pp. CAUGHLEY. G. 1956. Evaluation of the pellet group count as a 1974. Bias in aerial survey. J. Wild!. Manage. deer census method. J. Wild!. Manage. HANSON. W.R. 38(4):921-33. 20(1 ):70-74. 1963. Calculation of productivity, survival, 1977. Sampling in aerial survey. J. Wild!. EDWARDS. R.Y. and abundance of selected vertebrates Manage. 41(4):605-15. 1954. Comparison of an aerial and ground cen­ from sex and age ratios. Wild!. Monogr. sus of moose. J. Wildl. Manage. No.9. 60 pp. CHASE. W.W. and D.H. JENKINS 18(3):403-4. 1962. Productivity of the George Reserve deer HARLOW. R.F. and R.l. DOWNING herd. Proc. 1st Nat!. White-tailed Deer EDWARDS. W.R. and l. EBERHARDT 1967. Evaluating the deer track census method Dis. Symp. 1:78-88. 1967. Estimating cottontail abundance from used in the southeast. Proc. Conf.

20 Southeast. Assoc. Game and Fish Comm. ferential hunting 1055 in the sexes. Proc. MCCAFFERY, K.R. 21:39-41. Utah Acad. Sci., Arts, and Lett. 17:65-69. 1973. Road-kills show trends in Wisconsin deer HAYNE. D.W. 1945. Measurement and interpretation of populations. J. Wildl. Manage. 1949. An examination of the strip census forces that determine populations of 37(2):212-16. method for estimating animal popula­ managed deer. Univ. Mich., Ann Arbor. 1976. Deer trail counts as an index to popula­ tions. J. Wildl. Manage. 13(2):145-57. PhD Thesis. 422 pp. tions and habitat use. J. Wildl. Manage. 40(2):308-16. HAZZARD, l.K. KOHN, B.E. and J.J. MOOTY 1958. A review of literature on big game cen­ 1971. Summer habitat of white-tailed deer in MCCULLOUGH, D.F., e.E. OLSON, JR., and l.M. QUEAL sus methods. Colo. Game, Fish, and north central Minnesota. J. Wildl. 1968. Progress in large animal census by ther­ Parks Dep. Compl. Rep. Pittman­ Manage. 35(3):476-87. mal mapping. Pages 138-47 in P.l. Johnson, ed. Remote sensing in ecology. Robertson Proj. W-38-R-11. 76 pp. KREFTlNG, loW. and J.B. FLETCHER HOWE, D.l. 1941. Notes on the cruising method of census­ Univ. Ga. Press, Athens. 244 pp. 1954. A comparison of deer census methods in ing white-tailed deer in Oklahoma. J. MORAN. R.J. the Rifle River Area. Mich. Dep. Conserv. Wildl. Manage. 5(4):412-15. 1978. Monitoring of deer populations. Mich. Game Div. Rep. No. 2020. 33 pp. Dep. Nat. Resour. Perf. Rep. Pittman­ KUBISIAK. J.F. Robertson Proj. W-117-R-11. 11 pp. INTERSTATE DEER HERD COMMITTEE 1976. Experimental reduction of the Sandhill 1964. Progress report on the cooperative study deer herd through controlled hunting. MORSE, M.A. of the interstate deer herd and its range. Wis. Dep. Nat. Resour. Final Rep. Study 1943. Techniques for reducing manpower in U.S. For. Servo Reg. 5. 11 pp. No. 211. Pittman-Robertson Proj. the deer drive census. J. Wildl. Manage. W-141-R-11. 27 pp. 7(2):217-20. JACOBSON. 1.0. and R.G. WEATHERILL 1975. Application of quadrat and strip plot 1979. Deer population and range changes 1946. Censusing big game from the air. Minn. sampling techniques to an aerial census following herd removal at the Sandhill Conserv. Volunteer 9(52):26-33. Wildlife Area. Wis. Dep. Nat. Resour. of white-tailed deer in Manitoba's In­ MORSE, M.A. and D.W. BURCALOW Perf. Rep. Study No. 212. Pittman­ terlake. Manit. Dep. Mines, Resour., and 1942. Wildlife restoration and management Environ. Manage. 53 pp. Robertson Proj. W-141-R-14. 8 pp. planning project. Minn. Dep. Conserv. JAHN, loR. LAUKHART, I.B. Wildl. Res. Q. Prog. Rep. 1(4):2-19. 1959. Highway mortality as an index of deer 1950. Determining the big game population NEFF. D.J. from the kill. Trans. North Am. Wildl. population change. I. Wildl. Manage. 1968. The pellet-group count technique for big 23:187-97. Conf. 15:644-49. game trend, census, and distribution: 11 JENKINS. D.H. LEOPOLD, A. review. J. Wildl. Manage. 32(3):597-614. 1933. Game management. Charles Scribner's 1964. The productivity and management of NIXON, e.M., W.R. EDWARDS, and l. EBERHARDT deer on the Edwin S. George Reserve, Sons, New York. 481 pp. 1967. Estimating squirrel abundance from live­ Michigan. Univ. Mich., Ann Arbor. PhD LERESCHE, R.E. and R.A. RAUSCH trapping data. J. Wildl. Manage. Thesis. 193 pp. 1974. Accuracy and precision of aerial moose 31(1):96-101. JOLLY, G.M. censusing. J. Wildl. Manage. 1965. Explicit estimates from capture­ 38(2):175-82. OLSON, H.F. recapture data with both death and im­ 1938. Deer tagging and population studies in LINCOlN, F.e. Minnesota. Trans. North Am. Wildl. migration; stochastic model. Biometrika 1930. Calculating waterfowl abundance on the Conf. 3:2~6. 52(1 &2):225-47. basis of banding returns. U.S. Dep. Agric. 1969. Sampling methods for aerial census of Circ. No. 118. 4 pp. O'ROKE. E.e. and F.N. HAMERSTROM, JR. wildlife populations. East Afr. Agric. and 1948. Productivity and yield of the George For. J. 34(spec. issue):46-49. MANLY, B.F.J. 1970. A simulation study of animal population Reserve deer herd. J. Wildl. Manage. KARNS, P.D. estimation using the capture-recapture 12(1):78-86. 1971. 1971 deer pellet survey - Itasca Deer method. I. Appl. Ecol. 7(1):13-39. Management Unit. Minn. Dep. Nat. OVERTON. W.S. and D.E. DAVIS Resour. Game Res. Q. Prog. Rep. MARTIN, G.G. 1969. Estimating the numbers of animals in 1970. A regression method for mark-recapture 31(2):75-83. wildlife populations. Pages 403-55 in estimation of population size with R.H. Giles, [r., ed. Wildlife management KELKER, G.H. unequal catchability. Ecology techniques. The Wildl. Soc., 1940. Estimating deer populations by a dif­ 51(2):291-95. Washington, D.e. 623 pp.

21 PARR, M.J., T.J. GASKELL, and B.J. GEORGE known herd size. Mich. Dep. Conserv. Nat. Can. 101:615-29. 1968. Capture-recapture methods of Game Div. Rep. No. 2252. 26 pp. TRIPPENSEf, R.E. estimating animal numbers. J. BioI. 1971. Evaluation of pellet group survey for Educ. 2(2):95-117. 1948. Wildlife management. McGraw Hill, estimating deer populations in Michigan. New York. 479 pp. PAULIK, G.J. and D.S. ROBSON Mich, State Univ., East Lansing, PhD 1969. Statistical calculations for change-in­ Thesis. 237 pp. TYSON, E.l. ratio estimators of population 1959. A deer drive VS. track counts. Trans. SAUGSTAD, S. North Am. Wildl. Conf. 34:456-64. parameters. J. Wildl. Manage. 33(1 ):1-27. 1942. Aerial census of big game in North PETERSEN, e.G.J. Dakota. Trans. North Am. Wildl. Conf. U.S. FOREST SERVICE 1896. The yearly immigration of young plaice 7:343-56. 1935. Wildlife handbook North Central Region. 51 pp. into the limfjord from the German Sea. SCHRADER, T.A. Rep. Dan. BioI. Stn. for 1895. 6:1-77. 1944. Roadside deer counts as an emergency VAN ETTEN, R.e. and CL BENNETT. JR. PETRA BORG, W. and J. IDSTROM census method. Trans. North Am. Wildl. 1965. Some sources of error in using pellet­ 1972. Population dynamics of deer in Camp Conf. 9:150-55. group counts for censusing deer. J. Wildl. Manage. 29(4):723-29. Ripley Military Reservation. Minn. Dep. SEBER, G.A.F. Nat. Resour. Game Res. Q. Prog. Rep. 1965. A note on the multiple recapture census. VERME, t.r. 31(4):194-218. Biometrika 52(1 &2):249-59. 1964. An index of winter weather severity for POlAR, T.M. and D. STRICKLAND, ed s. northern deer. J. Wildl. Manage. SELLECK, D.M. and e.M. HART 32(3):566}3, 1979. A workshop on the status and applica­ 1957. Calculating the percentage of kill from tion of big game population modelling. sex and age ratios. Calif. Fish and Game VILKITIS. J.R. Colo. Div. Wildl. 53 pp. 43(4):309-16. 1968. Characteristics of big game violators and RASMUSSEN, D.1. and E.R. DOMAN extent of their activity in Idaho. Univ. SINIFF, D.B. and R.O. SKOOG Idaho, Moscow. MS Thesis. 202 pp. 1943. Census methods and their application in 1964. Aerial censusing of caribou using 1971. The violation simulation formula proves the management of mule deer. Trans, stratified random sampling. J. Wildl. North Am. Wildl. Conf. 8:369-80. Manage. 28(2):391-401. as reliable as field research in estimating c1os."ed-season illegal big game kill in RICE, W.R. and J.D. HARDER SMITH, R.H., D.J. NEFF, and e.Y. MCCULLOCH Maine. Trans. Northeast Sect. The Wildl. 1977. Application of multiple aerial sampling 1969. A model for the installation and use of Soc. 28:141-44. to a mark-recapture census of white­ deer pellet group survey. Ariz. Game and tailed deer. J. Wildl. Manage. Fish Dep. Spec. Rep. No.1. 30 pp. WALTERS. e.J. and J.E. GROSS 41(2):197-206. 1972. Development of big game management STRANDGAARD, H. plans through simulation modelling. J. RICKER, W.E. 1967. Reliability of the Petersen method tested Wildl. Manage. 36(1 ):119-28. 1958. Handbook of computations for on a roe-deer population. J. Wildl. biological statistics of fish populations. Manage. 31(4):643-50. WOOLF, A. and J.D. HARDER Fish. Res. Board Can. Bull. No. 119. 300 1979. Population dynamics of a captive white­ pp. TEPPER, E.E. tailed deer herd with emphasis on 1967. Statistical methods in using mark­ reproduction and mortality. Wildl. ROBINETTE, W.l., e.M. LOVELESS, and D.A. JONES recapture data for population estima­ Monogr. No. 67. 53 pp. 1974. Field tests of strip census methods. J. tion. U.S. Dep. Inter. Bibliogr. No.4. 65 Wildl. Manage. 38(1):81-96. pp. WRIDE, M.D. and K. BAKER 1977. Thermal imagery for census of RYEL, l.A. TIMMERMANN, H.R. ungulates. Proc. Int. Symp. on Remote 1959. Deer pellet group surveys on an area of 1973. Moose inventory methods: a review. Le Sensing of the Environ. 11:1091-96.

22 REPRODUCTION OF WHITE-TAILED DEER IN THE NORTH CENTRAL UNITED STATES

John D. Harder Dept. Zoology, 1735 Neil Avenue The Ohio State University Columbus, OH 43210

Abstract: Natality is the most easily studied and pregnant fawns were observed only in popula­ population parameter that requires estimates of all the major population parameters. Valid tions in which the mean whole body weight of all of the sex and age structure of the population estimates of ovulation rates based on counts of doe fawns exceeded 38 kg. Computer simulation of as well as estimates of the reproductive per­ corpora albicantia require careful examination of deer population dynamics can be used to illustrate formance of the female. For example, data the effect of changes in fetal sex ratios, fetal rates histological preparations; this approach is not indicating that an average yearling doe (1-2 recommended for general surveys. Conversely, of fawns, and fetal rates of mature doeson natality years old) will give birth to 1.20 fawns provide gross examination of ovaries collected from deer and population growth. Use of computerized after mid-December prcvides.accurateinformation models for predicting population size and explor­ very little predictive capability regarding on current ovulation rates, and uterine examina­ ing the consequences of different hunting regula­ change in population size, if the number or tion of deer killed after early February yields valid tions and harvest strategies is encouraged. relative proportion of yearling does in the estimates of pregnancy rate and the number and population is unknown. Natality may be sexratio of fawns born per doe the following June. INTRODUCTION delineated in a number of ways, but an im­ Unfortunately, the relationship between gross portant distinction is made between gross Estimates of natality, mortality, immigra­ natality and net natality, i.e., the fall fawn "crop," and net natality. Gross natality is estimated tion, emigration, and density are essential for is often unknown because early postnatal mortali­ for the moment of birth by projection of fetal ty can be high and, in most cases, difficult to understanding the dynamics of a population. rates, whereas net natality is a measure of the measure. Still, uterine examination is, by far, the Of these, natality is the most readily number of independent young added to the most efficient and valid procedure for estimating estimated and, perhaps, most useful of all. population each fall. This distinction reproductive performance, and is currently Estimates of the other parameters often re­ recognizes the importance of postnatal mor­ employed throughout the north central region. quire ambitious mark-recapture efforts or in­ Since reproductive rates increasesignificantly dur­ tality (death rate of fawns during the first 5 volve statistical models with unrealistic ing the first 2 or 3 years of life, estimates of the months of life) and the limited utility of sim­ assumptions. Given unbiased sampling, female age structure of populations, as well as ple measurements of reproductive perfor­ natalities of deer populations can be ob­ adult sexratios, are essential for calculating natali­ mance. ty. Reported fetal rates of mature does (2'12 + tained from does killed in sport hunts or in The difference between reproductive per­ years) collected in 10 north central states ranged highway accidents. Furthermore, reproduc­ from 1.53 to 2.10, whereas the percentage of preg­ tive performance data, the basis of natality formance and natality is important. nant fawns varied from zero to over 60%. estimates, serve as convenient monitors of Reproductive performance (or reproductive Agricultural regions produced does with higher herd and range conditions. rate) is a measure, usually age specific, of fer­ reproductive performance than those from north­ tility or fecundity. Deer biologists typically ern, wooded, or arid regions. Nutritional status is Definitions and Terminology express it as the average number of ova shed the major determinant of reproductive perfor­ (ovulation rate) or fetuses (fetal rate) or fawns mance in white-tailed deer. Reported fetal rates in Natality is defined as a rate, the number of produced per year by females in a given age mature does increased with whole body weight, new individuals added to a population by class. These ratios are valuable biological Harder, John D. 1980. Reproduction of white-tailed birth (NrJ per unit of time (t) per unit of measurements and essential components of deer in the north central United States. Pages 23-35 population (N): natality estimates, but they are not in themselves parameters of population in Ruth L. Hine and Susan Nehls, eds. White-tailed D.N rIND.t deer population management in the north central ecology. states. Proc. 1979 Symp. North Cent. Sect. Wildl. Where D. change (Odum 1971). It is a Productivity is a popular term, particularly Soc. 116 pp. 23 among wildlife biologists, generally meaning large follicles (> 3 mm diameter) can be 12 months after parturition, and thus, were a young produced per 100 females or the observed in the ovaries of deer collected at reliable indicator of the ovulation rate replacement rate of the population. Unfor­ almost every month of the year. Most of associated with the previous pregnancy. tunately, the term has a host of meanings in these follicles do not ovulate but, rather However, Golley (1957), Mansell (1971), and agriculture and industry, and it has been used become atretic. During the onset of the rut in Teer et al. (1965) found that CA persisted for 2 classically in the biological sciences as the early November, the percentage of large, or 3 years and that CA counts often exceeded rate of carbon fixation, i.e., photosynthesis in atretic follicles decreased markedly, but yet the true ovulation rate, particularly in older primary production or assimilation in secon­ only 1 or 2 follicles per ovary were in precise­ does. Gibson (1957) demonstrated that a dary productivity. Therefore, I choose not to ly the right stage of development to respond variety of ovarian structures, including lu­ use the term productivity in a technical to the sequence of hormonal dynamics that teinized follicles, regressing Cl of estrus, and sense, using instead, natality. culminate in ovulation (Harder and CA will form rust-colored scars. Moorhead 1980). The endocrine-ovarian­ Objectives environmental mechanisms controlling the Uterine Examination number of ova ready for ovulation in a given This paper is concerned with reproductive estrous cycle are not well understood for any Necropsy of pregnant deer is, by far, the performance and gross natality in white­ species, but therein lie the keys to under­ most efficient and productive approach to tailed deer [Odocoileus virglnienus), and standing much of the observed variation in the measurement of reproductive perfor­ therefore the physiological events of the reproductive performance of white-tailed mance. In utero mortality of fetuses is ap­ female reproductive cycle from ovulation to deer. parently very low (Teer et al. 1965), and, birth. Spermatogenesis and associated After ovulation, the granulosa cells of the unless does are severely undernourished, physiological events in the male are, of ruptured follicle begin to luteinize under the counts of fetuses can be projected to live course, important but are not considered influence of luteinizing hormone (lH), and births per doe. Uterine examinations also pro­ here. Estimates of population sex and age the corpus luteum (Cl) begins to form. vide valid estimates of fetal sex ratios. structure are essential for calculation of Thomas and Cowan (1975) reported that the Reliable procedures for estimating early natality. However, such estimates depend on first ovulation of the breeding season in postnatal losses are needed; without them, valid sampling schemes rather than an black-tailed deer [D. hemionus colutnbienus) estimates of gross natality are of diminished understanding of reproduction, and the seldom led to conception or a lasting value, particularly if early postnatal mortali­ technical aspects of sampling will not be pregnancy. The resulting Cl regressed quick­ ty is potentially high. Verme (1962) observed treated here. Net natality is primarily a func­ ly and a second ovulation occurred about 8 that most fawns of undernourished does were tion of early fawn mortality, the subject of days later. Harder and Moorhead (1980) born dead or died within 48 hours of birth. the paper by Porath in this symposium. estimated the interval between the first and langenau and lerg (1976) described a mater­ second ovulations of the season in white­ nal rejection syndrome in undernourished REPRODUCTIVE BIOLOGY tailed deer to be no more than 15 days and does that included neglect of the newborn AND NATALITY otherwise confirmed the observations of fawn and refusal to nurse despite the Thomas and Cowan (1975). Perhaps due to an presence of milk in the udder. Teer et al. Ovarian Analysis incomplete hormonal regime, the first ovula­ (1965) estimated that about 11% of the fawns tion of the breeding season is not accom­ on the Llano Basin of Texas were lost shortly Reproduction in the female begins with panied by estrus; the ovulation is "silent". after birth. oogenesis and the associated events of The second ovulation of the season, ap­ In that a large portion of postnatal mortali­ follicular development, ovulation, and cor­ parently associated with estrus, results in fer­ ty apparently occurs within 1 or 2 days after pus luteum formation. These processes may tilization and the resulting Cl persist as Cl of birth, the ability to predict such early mortali­ be viewed histologically in specimens col­ pregnancy. ty on the basis of a fetal or maternal condi­ lected during the breeding season. The scars After parturition, in early June, Cl of tion index would be most valuable. Verme of corpora lutea of pregnancy regress slowly, pregnancy begin to regress. The resulting (1977) developed fetal growth curves, which, after parturition, and can be used to assess scars, the corpora albicantia (CA) or corpora when projected to term, indicated mean birth ovarian activity many months afterward. For rubra, accumulate a rust or reddish pigment weights ranging from 3.24 to 2.27 kg. Based these reasons, examination of ovaries col­ that is grossly visible on razor-sectioned on weight-death relationships of newborn lected during the hunting season has been slices of Formalin-fixed ovary. Cheatum fawns in pen studies, projected fetal weights practiced for over 30 years. (1949a) estimated that CA persisted for 8 to indicated probable mortality rates in fawns 24 ranging from 10% to 70%. The relationship postnatal mortality, calculation of gross 60% in the agricultural regions of some between projected natal mortality and the natality is a relatively simple task. Fetal rates states. fawn:doe ratio in the fall harvest was signifi­ must be expressed as the average number of Recent reproductive data contributed by cant, indicating that the technique was a fetuses per all does examined in a given age biologists in Michigan and Minnesota re­ potentially reliable indicator of early fawn class, rather than fetuses per pregnant doe. In vealed significant(P < 0.05) variation in fetal losses. many cases, small sample sizes and low rates between years and between regions variation between older age classes will within those states (Table 2). Some variation Population Structure and Natality justify the use of average fetal rates for does in fetal rates might be related to yearly dif­ older than 3'11 years or even 2 Y2 years. The ferences in sampling locations within regions Reproductive performance data have no actual number of does used in calculation of or annual variation in weather. Higher fetal direct application to population ecology natality should be obtained from the propor­ rates in the southern sections of both states unless they are used with estimates of tion of the total number in the population in probably reflected the higher nutritional population size and structure to calculate each female age class. Then, the total plane of deer associated with agricultural gross natality. The prefawning population number of fawns produced can be obtained land use patterns. size might be known from a census, or it by summing the products of doe number and could be set arbitrarily, e.g., as a multiple of fetal rate in each age class. These and more DETERMINANTS OF an abundance index for a given management complex calculations involving early REPRODUCTIVE PERFORMANCE unit. Accurate estimates of sex and age struc­ postnatal losses, net natality, and population ture are essential but difficult to obtain, par­ growth over a series of simulated years can Aside from variation in the percentage of ticularlv those the of the conceming size be done readily with relatively simple, com­ does breeding; ovulation rate dearlyac­ fawn component of a population. puterized projection models. counts for most of the variation in reproduc­ Information on population sex and age tive performance of female white-tailed deer. structure is available from 3 sources: sport That is, loss of ova and fetuses from a fertile hunts, highway accidents, and field observa­ REPRODUCTIVE estrous cycle normally totals less than 15%. tions. Sex ratios obtai-ned from hunter returns Roseberry and Klimstra (1970) observed a are usually biased in favor of bucks, but short PERFORMANCE OF DEER IN 12.6% loss of ova in yearling and older does seasons (2-3 days) might reduce hunter selec­ THE NORTH CENTRAL UNITED from Illinois, and estimates of ova loss in tivity (Nixon 1971). Large sampIes from STATES other areas ranged down to less than 5% hunters probably provide reasonable (Ransom 1967, Haugen 1975). However, estimates of the female age structure, ex­ Reproductive performance of mature does Ransom (1967) also found ova loss exceeded cluding fawns. Large samples of deer killed (2Y2 + years) collected throughout the north 20% when more than 2 ova were shed. In on highways might provide useful data on central United States ranged from 1.53 to utero mortality, as evidenced in resorbing population structure. Numbers of deer killed 2.10 fetuses per all does examined (Table 1). fetuses, has been low in most studies, typical­ on Wisconsin highways correlated well with Variation in reproductive performance within ly involving less than 5% of the fetuses, often trends in buck harvests (McCaffery 1973). some states was as large as variation between less than 1% (Teer et al. 1965, Ransom 1967, Numbers of fawns and yearlings killed on a states. This was especially true of states with Woolf and Harder 1979). Thus, following section of Interstate 80 in Pennsylvania did a wide range of habitat types and climatic ovulation, the next important adjustment of not differ significantly between sexes, but conditions. The southern agricultural regions productivity apparently comes in the form of among adults many more does than bucks of Michigan, Minnesota, and Wisconsin pro­ stillbirths or early postnatal mortality (Verme were killed (Bellis and Graves 1971). Field duced does with higher performance than 1962). observations could potentially yield useful those of the northern, wooded regions. Does data on adult sex ratios and fawn:doe ratios, from the western, more arid sections (Black Age of the Doe but standardized, unbiased procedures are Hills) of South Dakota showed lower fetal needed. Clues to the solutions of these and rates than the East River Area of the same Age and nutritional status are the 2 major other sampling problems are the subject of state. Fawns exhibited the most striking varia­ factors affecting reproductive performance other papers in this symposium (Ryel, Beattie tion in reproductive performance of any age in white-tailed deer. The 2 are related, but et al., and Burgoyne). group. The percentage of fawns breeding regardless of nutrition, ovulation rate tends Given accurate age-specific fetal rates, sex ranged from 6% or less in the northern to increase with age, at least through the first and age structure data, and negligible early reaches of Michigan and Wisconsin to over 3 or 4 years of life. The ovulation rate and

25 TABLE 1. Comparison of reproductive performance of white-tailed deer in the north cen­ tral United Stares". fetal rate of fawns are always significantly Mature Does Year of lower (P < 0.05) than those of older females. State Fawn, Yearlings (2'/, + years) Deer In some populations, does 2112 + years and Percent Percent Fetal Percent Fetal Collec­ showed significantly higher ovulation and Regi""-. Pregnant Pregnant Rate Pregnant Rate tion Reference fetal rates than yearlings (Ransom 1967, ILLINOIS Nixon 1971, Kirkpatrick et al. 1976), while in Crab Orchard 41 96 168 98 188 1966 Roseberry and Ref uge (160) (110) (278) Klirnstra other areas, differences between yearlings (1970) and older deer were not significant (Ransom INDIANA 1967). In general, differences between fetal NAD Crane 20 100 153 100 194 1964-71 Kirkpatrick rates of yearlings and older does were small Depot (45) (34) (93) et al (1976) in populations containing a large proportion (>40%) of pregnant fawns (Table 1). In­ IOWA Statewide and 74 85 1.66 100 210 1957-il6 Haugen creases in reproductive performance after Desoto Range (240) (472) (655) (1975) the third year of life are generally small and Statewide 68 96 1.82 90 1.63 1978 Gladfelter nonsignificant. (83) (22) (30) (1978) MICHIGAN Region I 6 91 125 96 175 1951-79 Friedrich Nutritional Status of the Doe (UP) (489) (626) (2.350) (1979) Cheatum and Severinghaus (1950) were Region III 52 96 1.83 96 1.94 1952-79 J Vogt (Southern) (966) (412) (665) (1979, pc) among the first to clearly demonstrate the MINNESOTA relationship between range conditions and North Forest 11 88 1.34 85 1.53 1971-79 P Karns fertility of white-tailed deer. During the last (822) (110) (421) (1979, pc) 30 years overwhelming evidence has ac­ South Agri­ 31 92 149 94 185 1979 J Ludwig cumulated indicating that nutrition is the ma­ cultural (179) 47) (161) (1979, pc) jor factor controlling reproductive perfor­ MISSOURI mance in white-tailed deer. Poor nutrition NE, Agri­ 56 89 189 100 1.92 1978 W Porath cultural (9) (9) (12) (1979, pc) can stem from a variety of sources ranging E. Central Hills 48 91 173 100 193 1978 W. Porath from a negative energy balance during severe (23) (11) (15) (1979, pc) winter weather in Manitoba (Ransom 1967) to NEBRASKA drought-induced range deterioration in Texas Statewide 63 94 174 99 195 1961-73 K Menzel (Teer et al. 1965). Sileo (1973) divided the (295) (143) (182) (1979, pc) eastern United States into 6 major deer "fer­ OHIO Statewide 77·· 99 1.87 100 2.04 1962-67 Nixon tility regions" based on a forest type by soil (335) (180) (215) (1971) type classification. These factors were SOUTH DAKOTA related to the percentage of land farmed East River 57 97 185 1978 L Rice which was, in itself, a reliable indicator of (21) (39] (1978, pc) reproductive performance in deer. Black Hills 96 1.54 L Rice Conclusions from field studies regarding (28) (1978, pc) nutrition and reproduction have been con­ WISCONSIN Northern 3 Hale firmed in pen studies where short-term reduc­ (203) (1959) tion in rations caused reduced fawn produc­ Statewide 74 120 86 156 195&-58 Sileo (1973, tion (Verme 1969) and delayed maturity in (230) (538) from Hale female fawns (Abler et al. 1976). Conversely, 1959) Robinette et al. (1973) found that litter size in South and 29 Sileo (1973). penned mule deer (0. hemionus) increased East Central (147) Hale (1959) when their nutritional level was raised. • States with a 1978 or 1979 collection date estimated reproductive rates On a yearly basis. Fetal Ovulation rate is probably most strongly rate indicates the number of fetuses from all does examined in a given age class Sample sizes are given in parentheses. Reference date with pc indicates personal communication influenced by the nutritional status of the ··Percentage that ovulated. doe in late summer and early fall. Although 1Yearlings included with does 2'1, + years. 26 forage availability in late summer is un­ doubtedly important to the pre-ovulatory TABLE 2. Examples of geographical and chronological variation in fetal rates in does from condition of the doe, maternal obligations Michigan and Minnesota, 1972-79*. during the preceding 6 months might also be a factor, particularly in northern regions. Minnesota2 That is, if spring is delayed or if the doe is in Michigan1 Fetuses per Pregnant Doe poor condition at the onset of lactation, low Fetuses per Mature Doe (2Y, + years) Itasca Mille lacs nutritional status might be maintained Food Shortage Rainy River Big Woods Area of (Northern (South & Central throughout much of the summer. Verme Region II Region III Region) Region) (1969) suggested that the undernourished Year N x ± SO** N x ± SO N x ± SO N x ± SO does cannot recover from the stress of lacta­ tion in time to be in prime condition for 1972 63 1.88 ± 0.54 (No Data) 45 1.64 ± 0.47 51 1.63 ± 0.51g 1973 42 1.83 ± 0.49 58 1.95 ± 0.63 18 1.56 ± 0.33 38 1.68 ± 060 ovulation during the rut and are, therefore, 1974 26 1.96 ± 0.20a 12 2.00 ± 0.43 27 1.70 ± 0.50 47 1.57 ± 0.44h,i,j less productive. 1975 24 1.83 ± 0.48 19 2.10 ± 0.57b 49 1.55 ± 0.4ge,f 53 1.81 ± 0.50f,h Numerous measurements can be made of 1976 41 1.88 ± 0.51 21 1.67 ± 0.66b 22 1.77 ± 0.61 63 1.81 ± 0.43i the bodies of female white-tailed deer that 1977 14 1.79 ± 0.58 15 2.06 ± 0.46 16 1.69 ± 0.48 64 1.70 ± 0,40 reflect nutritional status and history, but they 1978 51 1.78 ± 0.54c 40 2.10 ± 0.50c 14 1.64 ± 0.63 28 1.89 ± 0.57g,j vary considerably in reliability, sensitivity, 1979 44 1.54 ± 0.62a,d 47 1.94 ± 0.53d 63 1.81 ± 0.50e 46 1.78 ± 0.51 and convenience. Body weight reflects * Data from Region II (northern one-half of the lower Peninsula) and Region III (southern growth as well as seasonal accumulation and agricultural one-half of the Lower Peninsula) represent only the 8 most recent years of a con­ depletion of fat reserves. As such it is an old, tinuous series of annual data sets dating back to 1951. reliable index of nutritional status and, when * * Data are presented as the mean ± standard deviation (X ± SO). Column means bearing the measured during the fall hunting season, it is same alpha superscript are significantly different (P < 0.05). particularly relevant to puberty in fawns and Means and t-tests calculated from age-specific fetal counts given in Ryel and Youatt (1972), the breeding condition of mature does. Body Burgoyne and Youatt (1973), Youatt et al. (1974, 1975). Purol et al. (1976). Friedrich et al. (1977), weight is also well suited for use in com­ Friedrich and Fay (1978), and Friedrich (1979). Means include nonpregnant does; 95-100% of the parisons of deer from large geographic areas does were pregnant. or animals taken over long time spans. Unfor­ tunately, body weights are difficult to 2 Data were presented as fetuses per pregnant doe and notation of significant differences (P < 0.05) was provided by P. Karns (pers. comm.). measure in the field and are normally only recorded at selected check stations. Indexes of fat reserves are valuable because they reflect the energy balance of an animal, the most popular being those of kidney (Riney During the last 15 years, blood chemistry ficult than analysis of physical 1955) and bone marrow fat (Cheatum 1949b, and hematology have attracted interest as in­ measurements. LeResche et al. (1974) Verme and Holland 1973). In recent years, dicators of nutritional condition of deer. evaluated a number of blood parameters Michigan biologists have collected man­ Blood parameters reflect vital processes and relative to nutritional assessment. dibles of all deer killed on the highways in should be sensitive to subtle or recent winter for age determination and fat analysis changes in nutrition. Measurements of red Female Body Weight and (Friedrich and Fay 1978). Ozoga and Verme blood cells, hemoglobin, cholesterol, serum Reproductive Performance (1978) found that the growth and involution urea, lactic dehydrogenase, thyroxine, and of the thymus gland in fawns varied with nonesterified fatty acids reflected dif­ Populations with high female body weights dietary plane, and they encouraged use of ferences between populations reievant to tended to have high reproductive rates (Table thymus gland weights as an index of nutri­ habitat and nutritional status (Seal et al. 3). Crab Orchard deer, the apparent excep­ tional status. An informative review of 1978). Unfortunately, biochemical tion, had a relatively high fetal rate despite measurements commonly used to estimate parameters can be affected by factors low body weight. However, these deer from physical condition was provided by unrelated to ongoing nutritional status, and the southern tip of Illinois were not under­ Hesselton and Sauer (1973). thus their interpretation is generally more dif­ nourished, but were smaller because they

27 probably belonged to the inherently smaller subspecies, O. v. virginianus, while other populations in Table 3 belonged to the larger TABLE 3. Relationship between whole body weights and reproductive performance of O. v. borealis O. Roseberry, pers. comm.). Ex­ female white-tailed deer from northern United States. cluding Crab Orchard data, fetal rates of mature does in Table 3 were closely cor­ Fawns Mature Does (2V, + years) Location Mean Mean Fetuses per related with body weight (r 0.982, Y = = and Weight Percent Weight Percent All Does -1.23 0.049X). Gill (1956) found does in + Reference (kg) Pregnant (kg) Prl~gnant Examined the West Region of West Virginia to be larger (mean dressed weight = 41.3 kg) and to have IOWA 41.2 65-74 68.9 100 2.10 (288)"" (240) (244) (655) a higher ovulation rate (1.90) than does in the Kline (1967)" East Region which averaged 32.7 kg and 1.51 Haugen (1975) corpora lutea per doe. SENECA ARMY DEPOT1 42.4 36 62.7 100 1.93 The Age of Puberty (New York) (23) (22) (29) (29) Hesselton (1967) Reproductive performance of yearling and CRANE NAD DEPOT1 38.0 36 63.2 100 1.94 mature does was higher in populations that (Indiana) (1,146) (45) (1,145) (93) showed a high percentage (>30%) of preg­ Kirkpatrick et al. nant fawns (Table 1). This might simply (1976) reflect a higher overall nutritional plane for these populations, or it might indicate that CRAB ORCHARD NATIONAL 29.5 41.6 53.2 98 1.88 WILDLIFE REFUGE (178) (160) (266) (278) rapid growth and sexual development as a (Illinois) fawn enhance reproductive performance Roseberry and Klimstra throughout life. Some of the physiological (1970,1975) mechanisms involved in puberty might also be active during the transition from seasonal PLUM BROOK STATION 32.7 0 59.3 95 1.67 anestrus to the rut and thus explain the (Ohio) (35) (35) (197) (197) Harder and Peterle (1974) absence of breeding in some older animals. Bell and Peterle (1975) Attainment of puberty is closely related to nutritional and physical development. The RACHELWOOD WILDLlFE1 30.9 0 52.3 94 1.32 amount of growth achieved by fawns before RESEARCH PRESERVE (187) (35) (314) (46) the onset of winter and decline of natural (Pennsylvania) forage in December probably determines Woolf and Harder (1979) whether or not a fawn participates in the rut. " Kline (1967) and Roseberry and Klimstra (1975) were cited for weight data only. All data for Plum Robinette et aJ. (1973) concluded that the Brook Station represent weighted averages of values presented in Harder and Peterle (1974) and minimum whole body weight for first estrus Bell and Peterle (1975). in mule deer fawns was 41 kg; Hesselton and "" Sample sizes given in parentheses. Sauer (1973) thought that this threshold weight for white-tailed deer was 38-40 kg 1 Whole body weight calculated: dressed weight X 1.28. (30-32 kg, dressed). Data from the several northern locations seem to support the critical weight hypothesis (Table 3). However, the threshold weight might be lower in Fawns normally comprise the single largest large changes in gross natality. Unfortunate­ smaller subspecies. Crab Orchard fawns were age class in a population. Thus, even though ly, despite the fact that reproductive perfor­ well nourished and productive but inherently fawns seldom, if ever, carry more than 1 fetus mance of fawns is highly variable (Tables 1 small, being approximately 21% lighter than and even if 1-year-olds are less successful in and 3) and sensitive to environmental fawns from northern Illinois 0. Roseberry raising their young, small changes in the changes, this age class has received relatively pers. comm.). percentage of pregnant fawns could lead to little attention in reproductive studies.

28 sampling effort are frequently as important Population Density and COMPUTER SIMULATION IN as statistical determinations. In areas of low Reproductive Performance THE STUDY OF NATALITY AND deer density, it might be necessary to take POPULATION GROWTH whatever samples are available and then Low fecundity in deer is usually attributed pool data over several years. Shortages of to a low nutritional plane. However, this rela­ A computerized version of the Leslie-Lewis funds, trained personnel, or opinions of ad­ tionship is often associated with high popula­ matrix model (Pennyquick et al. 1968) ministrators can present obstacles. tion density (Teer et al. 1965), and the poten­ developed by Rice (1976) was used to Therefore, biologists should be able to clear­ tial effect of density of social stress alone simulate the dynamics of the deer population ly justify collections, define the level of (Christian and Davis 1964) should be con­ on Plum Brook Station in northern Ohio (Fig. precision' necessary to meet objectives, and sidered. Data are available from 2 popula­ 1). A 2.44 m fence enclosed the Station and then determine appropriate sample size for tions that apparently had good-to-adequate eliminated significant immigration or emigra­ each management unit. nutrition during the time of high population 2 tion. How often should reproductive rates be density, i.e., more than 20/km • Deer numbers The parameters required by the model measured? This is an important question in­ on the Crab Orchard National Wildlife were: (1) initial population size, (2) sex and volving the time and efforts of many people. Refuge had reached an estimated density of age structure of the initial population, (3) About one-half of the 10 states surveyed 2 about 30/km (Autry 1967) in 1965, just prior fetal sex ratios, (4) age-specific reproductive (Table 1) took statewide observations of fetal to the data collection of Roseberry and rates (fetuses/doe), (5) number of animals rates on a yearly basis. The remainder con­ Klimstra (1970) (Table 1). Rice and Harder removed each year, e.g., from hunting and ducted special studies of deer natatity on (1977) documented the presence of 2,499 ± trapping, and (6) natural mortality, or the selected localities or statewide on an 94 (±2 SE) deer on the 2,176-ha Plum Brook 2 reciprocal, survival. The first 5 parameters unscheduled basis, perhaps every 5 or 10 Station in 1975 for a density of 115 per km , were estimated directly from samples taken years. the highest density ever reported for white­ from the Station. Natural mortality or sur­ Objective answers to questions regarding tailed deer. Computer simulations indicated vival of deer on the Station, independent of sampling frequency must be based on some that the deer density on Plum Brook Station removals, was unknown, so estimates from knowledge of annual variation in reproduc­ was at least 80/km2 during the period that similar herds were assigned to each age class, tive rates for the management units in ques­ reproductive data were collected (Table 3). beginning with fetuses to the first 6 months tion. Significant differences in fetal rates oc­ Although Plum Brook fawns did not breed, after birth (0.775 survival). High annual sur­ curred between units and between years reproductive performance for adults taken vival after the first 6 months (0.9(){).95) was (Table 2). Annual differences in fetal rates of from these high density herds was well within assigned to females. Male survival ranged 0.2 to 0.3 were common, and they occasional­ the range of values reported for deer from from 0.90 to 0.70 for the first 3 years of life ly approached 0.5. Differences of the latter low density herds (Table 1). and then declined steadily to 0.40 at 7 years magnitude could lead to marked variation in High density has had no discernible effect of age. natality and population growth (Fig. 5). on reproductive performance of penned States are typically divided into manage­ It is recommended that annual statewide deer. Mature does held in pens at a high nutri­ ment units based on topography, climate, surveys of reproductive rates be conducted tional plane produced an average of 1.80 agricultural land use, and other factors rele­ over a sufficient number of years to cover the fawns (Verme 1969). Mule deer held in 1.2-ha vant to deer biology. Sampling effort within a range of annual variation in winter severity, enclosures at densities ranging from 20-30/ha, given management unit, ideally, should be 2 precipitation, or other factors known to in­ equivalent to densities of over 2,OOO/km , pro­ commensurate with temporal and spacial fluence reproductive rates. Data from at duced an average of 1.97 fawns per doe when variation in reproductive rates. That is, units least 5 years would be required in most cases, maintained at a high dietary plane (Robinette that show high performance with little varia­ and sample sizes would have to be large et al. 1973). Thus, I believe it may be safely tion within or between years could be enough (probably >40) to support statistical stated that population density per se has little sampled less intensely than units exhibiting comparisons between years and management or no influence on the reproductive perfor­ depressed or variable rates. Once a desired units. mance of mature does. Whether density in­ level of discrimination or precision has been fluences the initiation of breeding in fawns, selected, the variance of previous estimates Collection and Analysis of postnatal mortality, or other determinants of can be used to determine the required sam­ of Biological Material net natality is another question. ple size (Ostle 1963). Practical considerations regarding Examination of ovaries collected during 29 the hunting season is recommended for states or management units with low numbers of deer killed on the highways in winter or early spring. Hunters can be en­ couraged, through instructions and adver­ tisements, to save the reproductive tract when dressing the carcass; biologists have used this approach with varying degrees of success for several decades (Cheatum and Severinghaus 1950, Haugen 1975). For reasons outlined previously, CA counts from gross ovarian examination (Cheatum 1949a) should not be used to estimate the ovulation rate associated with the previous pregnancy in mature does. Careful histological examination of ovaries by trained technicians could produce useful in­ dexes of ovulation rate in older does (Mansell 1971), but this approach probably would not be practical in most cases. More emphasis should be placed on gross ovarian analysis of yearling does and carefully developed criteria of CA indicating pregnancy rates in fawns. If estimates of pregnancy rates in fawns are not forthcoming from winter necropsies, then they should be obtained from the occurrence of CA in yearling does or from the proportion of yearlings in the fall harvest that show enlarged teats, a reliable indicator of lactation and prior pregnancy (Sauer and Severinghaus 1977). NO Assigned annual survival values used in the simulation model were confirmed with a combination of simulation and field observa­ tions. Rice and Harder (1977) estimated the population at 2,499 ± 94 (± 2 SE) deer in January 1975; approximately 90 more were YES NO present in October 1974. Sport hunting and STOP finished year = year + 1 removal by trapping reduced the population to low levels by the fall of 1977 (Fig. 2), and removals thereafter, mostly fawns and does (67-84%), maintained the low density. Fifty­ one deer were counted on the Station in Oc­ tober 1979 (Edward Butler pers. comrn.). Not all deer were seen, and the total number pre­ sent was probably at least 2-4 times the number counted. Thus, the computer simula­ FIGURE 1. Flow diagram of the computerized projec­ tion using high female survival (0.90-0.95) tion model of the Plum Brook deer population (Rice agreed with field observations in that at least 1976]. 30 100-200 were known to be on the Station in the 3 populations, and during the first 3 healthy fetuses (Armstrong 1950) or possibly the fall of 1979. However, when a slightly simulated years, the percentage removed to predict early postnatal mortality (Verme lower female survival (0.85-{).90) was used in from each population was approximately 1977). Interested personnel can be easily the data set, the simulation terminated in equal (22-28%). Nevertheless, population A trained to collect data, but dead deer can be 1978, when the simulated population was grew steadily, so that the numbers removed widely distributed over time and space. reduced to zero (Fig. 2), indicating that the during the last 2 simulated years represented Biologists often must depend on a large estimate of 0.9Q-{).95 annual female survival only about 21% of the fall population levels. number of cooperators willing to work over a used in the model was not too high. Population C declined gradually at first and period of 2-4 months, and inconsistency in In addition to confirming assumptions then, steeply during the last 2 years when the procedures and effort could be a source of regarding natural mortality, the projection or numbers removed represented 32-42% of the biased data. Thus, ovaries should be col­ simulation model has been applied in 3 fall populations. Population B maintained lected from a subsample of does killed on general ways: (1) year-by-year projection of the initial size while supporting annual highways in order to confirm reported fetal population size (Fig. 2) based on known removals of 23-29%. rates and to determine the relationship be­ removals and reproductive parameters; (2) ex­ tween ovulation rate and fetal rate of does in ploration of management strategies, e.g., RECOMMENDATIONS FOR a given area. hunting, trap-removal, and the application of STUDIES OF NATALITY antifertility agents (Rice 1976); and (3) simula­ APPLICATION OF tion of population dynamics with varied Sampling Procedures REPRODUCTIVE DATA TO reproductive rates and sex ratios (Figs. 3, 4, DEER MANAGEMENT and 5). In most populations, ovaries of mature Fetal sex ratios normally range between does collected after late November will con­ Estimates of natality and reproductive 45% and 50% female, but extremes of 60% tain active CL of pregnancy, which provide a data are applicable to 2 major areas of deer female fetuses (Woolf and Harder 1979) and reliable indication of ovulation rate and have management: (1) prediction of population 30% female fawns (Verme 1969) have been a predictable relationship to fetal rate. CL of size and setting harvest quotas, and (2) recorded. The causes and theoretical or pregnancy can be easily identified and evaluation of the general vitality of a deer evolutionary implications of shifts in fetal sex measured in razor-sectioned ovaries. population and the condition of the habitat ratios are controversial and have been However, care must be given to the collec­ or range. State game agencies spend a large discussed elsewhere (Verrne 1969, Trivers and tion dates relative to timing of the rut, and amount of time in setting and enforcing hunt­ Willard 1973, Woolf and Harder 1979). Of im­ size criteria for CL of pregnancy should be ing regulations and in evaluating the harvest. portance to this discussion, however, are the established (Trauger and Haugen 1965). Clearly, estimates of abundance are essential results of simulations indicating that a wide All states that made annual surveys of to this process, but natality estimates can be range of fetal sex ratios might have a relative­ reproduction in deer (Table 1) did so by particularly useful in calculating antlerless ly small impact on natality as evidenced in uterine analysis, i.e., by counting fetuses in permit quotas for management units (J. population growth (Fig. 3). does killed on highways. This is, by far, the Ludwig pers. comm. 1979). Setting regula­ Does that produce a viable fawn on their most efficient and meaningful measurement tions is, in many cases, an unavoidable pro­ first birthday could have a tremendous im­ of reproductive performance. Michigan and cess of reacting to prevailing conditions on a pact on natality (Haugen 1975). Computer other states concentrate their efforts in year-to-year basis. Land acquisition or habitat simulations showed that populations differ­ March, April, and May, but collections made alterations may be planned years in advance, ing only in the proportion of breeding fawns any time after the first of February would but populations are typically viewed in would respond quite differently to exploita­ give reliable estimates of the percentage of retrospect. This need not be true. tion (Fig. 4). breeding does in all age classes, the number Computerized projections or simulation Different fetal rates, covering the range of fetuses per doe, and fetal sex ratios. models offer predictive capability, and observed in mature does from across the Uterine analysis is advantageous because biologists in several states are currently using north central United States (Tables 1, 2, and data can be collected during field necropsies this tool. Population modelling can be used 3) led to substantial differences during a without the need to collect or preserve to predict population size (Fig. 2), explore the 5-year projection of population size (Fig. 5). tissues. At the very least, age of the doe and impact of various hunting regulations on sex The high fetal-rate population (A) grew to be the sex and number of fetuses should be and age classes, and determine optimum 3 times larger than the low fetal-rate popula­ recorded. Measurements of forehead-rump yield (Walters and Bandy 1972). Realism of tion (C). Equal numbers were removed from length and body weight can be used to age the projections will depend on the nature of 31 3,000 FIG. 2 FIG. 3 3,800

2,600 -- HIGH ANNUAL FEMALE SURVIVAL (090-095J

-- - MOOERATE ANNUAL FEMALE 3,600 --55% FEMALE FETUSES SURVIVAL (085- 090J 2,200 --- 50% FEMALE FETUSES ----45% FEMALE FETUSES 0:: 0:: 3,400 / w w w w ·· ..·····40% FEMALE FETUSES a 1,800 a / u, u, o O ,/ 3,200 ,//'/ 0:: 0:: / w 1,400 w co lD "/' -: :::!: // :::!: ::> ::> z z 3,000 /'>/ .., ~ -~:-~ 1,000 . ~-:::::;..-_/ .. " 2,800 #P~«"'~'~'-;:""'" 600 ~-:::;:;.-~~ 2,600 200 1974 1975 1976 1977 1978 1979 0 YEAR 1974 1975 1976 1977 1978 1979 YEAR

FIG. 4 POPULATION A FIG. 5 3,200 11- High Fetal Rate 4,000 POPULATION B Moderate Fetal Rote 3,000 POPULATION C 3,600 Low Fetal Rate

2,800

0:: 0:: 3,200 W w W ------­ w a 2,600 -----­ a u, u, o o 2,800 0:: ------, 0:: , w ------­ ~ 21'100 lD :::!: ~ ::> " z ~ 21'100 ------­ " ...... , ...... 2,200 -- 06 FETUS per FAWN ' ...... ' ...... --- 04 FETUS per FAWN 2,000, <, 2,000 --- 02 FETUS per FAWN ...... 00 FETUS per FAWN -. 1,600 -,

1,800 I i -"­ 1974 1975 1976 1977 1978 1979 YEAR 1,200 1974 1975 1976 1977 1978 1979 YEAR ~ the model and the accuracy of the data used LITERATURE CITED in the simulations. Nevertheless, modelling FIGURE 2. Computer projection of the fall can provide reasonable answers to the "what ABLER. W.A., D.E. BUCKLAND. R.l. KIRKPATRICK. and deer population on Plum Brook Station, based if" questions and could contribute to P.F. SCANLON on the 1975 census (Rice and Harder 1977) enlightened management decisions (Walters 1976. Plasma progestins and puberty in fawns as influenced by energy and protein. J. and known removals of 384, 1,603, 1,292, 171, and Gross 1972). Wild!. Manage. 40(3):442-46. and 90 deer from 1975 through 1979, respec­The second major use of reproductive tively. With the high female survival, the com­data, that of evaluating the general vitality of ARMSTRONG, R.A. puter simulation projected the population to deer or range condition in a given area, is 1950. Fetal development of the northern white­ tailed deer (Odocoileus virginianus be about 300 deer in the fall of 1979, a well established, dating from the classic reasonable estimate relative to minimum borealis Miller). Am. MidI. Nat. publication of Cheatum and Severinghaus 43(3):650-66. numbers known to be on the Station at that (1950). Fetal rate, yearling antler develop­ time. ment, and body weight were the criteria for AUTRY,D.C. 1967. Movements of white-tailed deer in assignment of Pennsylvania counties into 1 FIGURE 3. Computer simulations of fall deer response to hunting on Crab Orchard Na­ of 5 deer-quality classes (Lang and Wood tional Wildlife Refuge. South. 1/,. Univ., populations differing only in fetal sex ratios 1976). Indices of population size, and recruit­ Carbondale. MS Thesis. 44 pp. (percentage of female fetuses). Simulated an­ment and harvest quotas were then nual removals were proportional to the sex calculated for each county class. Reproduc­ BEll. R.l. and T.J. PETERLE 1975. Hormone implants control reproduction and age structure of the population and tive parameters, e.g., percentage of fawns totaled SOD, 500, 600, 600, and 700 deer from in white-tailed deer. Wild!. Soc._Bull. breeding, might be more reliable or even 3(4):152-56. 1975 through 1979, respectively. more sensitive indicators of nutritional plane Bn.us, E.D. and H.B. GRAVES or habitat change than body weight. 1971. Deer mortality on a Pennsylvania in­ FIGURE 4. Computer simulations of fall deer Reproductive parameters are of greatest populations differing only in the proportions terstate highway. J. Wild!. Manage. value, however, because they integrate the 35(2):232-37. of fawns breeding (equal to fetal rates of effects of nutrition over the entire year and 0.0-0.6 because pregnant fawns carry one probably over the life of the individual. All BURGOYNE, G.E., JR. and W.G. YOUATT fetus only). Simulated annual removals were 1973. 1973 spring deer survey.Mich. Dep. Nat. things considered, the study of reproduction Resour. Rep. No. 128. 29 pp. (mimeo) proportional to the sex and age structure of and natality is one of the most efficient, prac­ the population and totaled 600, 600, 700, 700, tical, and meaningful ways to monitor a deer CHEATUM, E.l. and 800 deer from 1975 through 1979, respec­population. 1949a. The useof corpora lutea for determining tively. ovulation incidence and variations in fertility of white-tailed deer. Cornell Vet. FIGURE 5. Computer simulations of fall deer 39(3):282-91. populations differing only in adult fetal rates. 1949b. Bone marrow asan index of malnutrition Population A produced 0.20 fetuses/fawn, in deer. N.Y. State Conserv. 3(1):29-32. 1.50 fetuses/yearling, and 2.00 fetuses/doe 2 Vz years or older. Population B produced 0.2 CHEATUM, E.l. and C.W. SEVERINGHAUS Acknowledgments. Information in Tables1 1950. Variations in fertility of white-tailed deer fetuses/fawn, 1.20 fetuses/yearling, and 1.65 and 2 came as generous contributions from deer related to rangeconditions. Trans. North fetuses/doe 21;2 years or older. Population C biologists from throughout the north central states Am. Wild!. Conf. 15:170-90. produced 0.20 fetuses/fawn, 1.00 and their time and effort is sincerely appreciated. fetuses/yearling, and 1.30 fetuses/doe 21;2 Ted Bookhout, Tony Peterle, John Roseberry, and CHRISTIAN, J.J. and D.E. DAVIS years or older. Simulated removals were the lou Verme made valuable suggestions for improv­ 19&4. Endocrines, behavior and population. same as in Figure 4. ing the manuscript. Science 140(3651 ):1550-00

FRIEDRICH, P.O. 1979. Doe productivity and physical condition: 1979 spring survey results. Mich. Dep. Nat. Resour. Rep. No. 2843. 12 pp. (mimeo)

33 FRIEDRICH. P.D. and L.D. FAY HESSElTON. W.T. and P.R. SAUER PENNYQUICK. CJ., R.M. COMPTON. and L. BECKINGHAM 1978. Doe productivity and physical condition: 1973. Comparative physical condition of four 1968. A computer model for simulating growth 1978 spring survey results. Mich. Dep. deer herds in New York according to of a population, or two interacting Nat. Resour. Rep. No. 2816. 13 pp. several indices. N.Y. Fish and Game I. populations. I. Theor. BioI. 18(3):316-29. (mimeo) 20(2):77-107. PUROl. D.A., L.D. FAY.and H.D. HARTE FRIEDRICH. P.O., L.D. FAY. and D.A. PUROl KIRKPATRICK.CM., CM. WHITE. T.W. HOEKSTRA. F.A. 1976. Doe productivity and physical condition: 1977. Doe productivity and physical condition: STORMER. and H.P. WEEKS spring survey results. Mich. Dep. Nat. 1977 spring survey results. Mich. Dep. 1976. White-tailed deer of U.S. Naval Ammuni­ Resour. Rep. No. 2793. 24 pp. (mimeo) Nat. Resour. Rep. No. 2793. 13 pp. tion Depot Crane. Purdue Univ. Agric. (mimeo) Exp, Stn. Res. Bull. No. 932. 42 pp. RANSOM. A.B. GIBSON. D. KLINE. P.D. 1967. Reproductive biology of white-tailed 1957. The ovary as an indicator of reproduc­ 1967. Satus and management of white-tailed deer in Manitoba. I. Wildl. Manage. tive history in the white-tailed deer deer in Iowa, 1954-1962. Proc. Iowa 31(1):114-23. (Odocoileus virginianus borealis Miller). Acad. Sci. 72:207-17. Univ. Toronto. MA Thesis. 61 pp. RICE. W.R. LANG. L.M. and G.W. WOOD 1976. A census and population model of the Gill. I. 1976. Manipulation of the Pennsylvania deer Plum Brook deer herd. The Ohio State 1956. Regional differences in size and produc­ herd. Wildl. Soc. Bull. 4(4):159-66. Univ., Columbus. MS Thesis. 115 pp. tivity of deer in West Virginia. I. Wildl. LANGENAU. E.E., JR. and J.M. LERG Manage. 20(3):286-92. 1976. The effects of winter nutritional stresson RICE. W.R. and J.D. HARDER 1977. Application of multiple aerial sampling GLADFElTER. H.L. maternal and neonatal behavior in 1978. Productivity of white-tailed deer in Iowa. penned white-tailed deer. Appl. Anim. to a mark-recapture census of white­ Iowa Conserv. Comm. 8 pp. (typed) Ethol. 2:207-23. tailed deer. I. Wildl. Manage. 41(2):197-206. GOllEY. F. B. LERESCHE. R.E., U.S. SEAL. P.D. KARNS. and A.W. 1957. An appraisal of ovarian analysis in deter­ FRANZMAN RINEY. T. mining reproductive performance in 1974. A review of blood chemistry of moose 1955. Evaluating condition of free-ranging red black-tailed deer. J. Wildl. Manage. and other Cervidae with emphasis on deer (Cervus elaphusJ, with special 21(1):62-65. nutrition assessment. Le Nat. Can. reference to New Zealand. N.Z. J. Sci. 101(1-2):263-90. and Technol. 36(5):429-63. HAlE.I.B. 1959. Final report on deer reproduction study. MANSEll. W.D. ROBINETTE. W.L., C.H. BAER. R.E. PlllMORE. and CE. Wis. Dep. Conserv. 8 pp (typed) 1971. Accessory corpora lutea in ovaries of KNITTLE white-tailed deer. I. Wildl. Manage. HARDER.I.D. and D.l. MOORHEAD 1973. Effects of nutritional change on captive 35(2):369-74. 1980. Development of corpora lutea and mule deer. I. Wildl. Manage. plasma progesterone levels associated MCCAFFERY. K.R. 37(3):312-26. with the onset of the breeding season in 1973. Road-kills show trends in Wisconsin deer white-tailed deer (Odocoiieus vir­ populations. J. Wildl. Manage. ROSEBERRY. J.L. and W.D. KliMSTRA ginianus). Bioi. Reprod. 22(2):185-91. 37(2):201-16. 1970. Productivity of white-tailed deer on Crab Orchard National Wildlife Refuge. J­ HARDER. J.D. and T.I. PETERlE NIXON. CM. Wildl. Manage. 34(1):23-28. 1974. Effect of diethylstilbestrol on reproduc­ 1971. Productivity of white-tailed deer in Ohio. tive performance of white-tailed deer. l. Ohio I. Sci. 71(4):217-25. 1975. Some morphological characteristics of Wildl. Manage. 38(2):183-96. the Crab Orchard deer herd. J. Wildl. ODUM. E.P. Manage. 39(1 ):48-58. HAUGEN. A.O. 1971. Fundamentals of ecology. W.B. 1975. Reproductive performance of white­ Saunders Co., Philadelphia. 574 pp. RYEl. L.A. and W.G. YOUATT tailed deer in Iowa. J. Mammal. 1972. 1972 spring deer survey. Mich. Dep. Nat. OSTlE. B. 56(1 ):151-59. 1963. Statistics in research. Iowa State Univ. Resour. Rep. No. 271. 24 pp. (mimeo) HESSElTON. W.T. Press, Ames. 585 pp. SAUER. P.R. and CW. SEVERINGHAUS 1967. Evaluation of deer life equation data OZOGA. 1.1. and L.I. VERME 1977. Determination and application of fawn from Seneca Army Depot. N.Y. Conserv. 1978. The thymus gland as a nutritional status reproductive rates from yearling teat Dep. Job CampI. Pittman-Robertson indicator in deer. I. Wildl. Manage. length. Trans. Northeast Sect. The Wildl. Proj. W-89-R-ll. 10 pp. (mimeo) 42(4):791-98. Soc. 33:133-44. 34 SEAL, U.S., M.E. NELSON, L.D. MECH, and R.L. TRAUGER, D.L. and A.D. HAUGEN WALTERS, C}, and P.J. BANDY HOSKINSON 1965. Corpora lutea variation of white-tailed 1972. Periodic harvest as a method of increas­ 1978. Metabolic indicators of habitat dif­ deer. J. Wildl. Manage. 29(3):487-92. ing big game yields. J. Wild/. Manage. ferences in four Minnesota deer popula­ 36(1):128-34. tions. J. Wildl. Manage. 42(4):746-54. TRIVERS, R.L. and D.E. WILLARD 1973. Natural selection of parental ability to WALTERS, C}, and J.E, GROSS vary the sex ratio of offspring. Science 1972. Development of big game management SILEO, L. JR. 179(4068):90-92. 1973. Fertility analysis of the ranges of white­ plans through simulation modeling. 1. tailed deer in the eastern United States. VERME, L.J. Wild/. Manage. 36(1):119-28. Agric. Exp. Stn. Storrs, Conn. Rep. Proj. 1962. Mortality of white-tailed deer fawns in No. 340. 43 pp. (mimeo) relation to nutrition. Proc. 1st Nat/. WOOLF. A. and J.D. HARDER White-tailed Deer Dis. Svmp. 1:15-28. 1979. Population dynamics of a captive white­ 1969. Reproductive patterns of white-tailed tailed deer herd with emphasis on ITEER. J.G., JW. THOMAS, and E.A. WALKER deer related to nutritional plane. J. reproduction and mortality. Wild/. 1965. Ecology and management of white­ Monogr. No. 67. 53 pp. tailed deer in the Llano Basin of Texas. Wildl. Manage. 33(4):881-87. Wild/. Monogr. No. 15. 62 pp. 1977. Assessment of natal mortality in upper YOUATT, W.G., L.D. FAY, and H.D. HARTE Michigan deer. J. Wild/. Manage. 1974. 1974 spring deer survey, doe productiVity 41(4):70

"80

35

THE LEGAL DEER KILL - HOW IT'S MEASURED

Lawrence A. Ryel Michigan Department of Natural Resources Box 30028 Lansing, MI 48909

Abstract: The size of the legal deer harvest in hunters afield in 1978 with gun and bow. however, much related information is also a state or province is not only of great interest to Many Michiganians believe that the opening desirable and useful in developing deer sportsmen, but it also provides necessary informa­ of the firearm deer season is the second most management strategies. For example: tion for wildlife managers. Measuring the legal significant day of the year next to Christmas. (1) Determ ine the sex and age of the deer harvest is much easier to do than estimating values Many northern schools close, hundreds of for other mortality factors. Basically only 3 tech­ killed. thousands of people begin their vacations, niques are used by midwestern states and pro­ (2) Determine the distribution of the harvest and factories and offices report high vinces: compulsory deer registration, compulsory over time and by geographic areas. absenteeism. In 1968, over one-quarter of the hunterreport cards.. and mail out-mail back ques­ (3) Determine the number of hunters, their resident males in Michigan between the ages tionnaires. Some states use 2 systems. Registration distribution, hunting effort, and relative provides the most detailed, the most timely, and of 25 and 44 hunted deer (Ryel et al. 1970). success over time and by geographic the most acceptable harvest figures to sportsmen. The deer season is news. If the Mi'chigan areas. Mail surveys can produce much additional infor­ Department of Natural Resources does not (4) Determine the harvest by traditional mation such as hunter success, hunting pressure, provide deer harvest information, the news firearms as well as by primitive weapons origin-destination, and opinions on regulations. media will fabricate it. Report cards combine advantages and disadvan­ such as long bow, cross bow, and muzzle The legal harvest is potentially the most tages of both of these systems. Mail surveys are loaders. accurate, precise and easily obtainable life generally less expensive than registering deer or (5) Obtain information on hunter attitudes history component for deer. Estimates of using report cards; however, care must be taken to concerning management practices and other mortality factors such as predation, insure precision and accuracy in the estimates. regu lations. Deer registration generally underestimates the true malnutrition, disease, crippling loss, or (6) Obtain economic data on deer hunting. legal harvest since accuracy depends entirely on poaching are not likely to be anywhere near (7) Obtain demographic information on the degree of compliance. Report cards are more as precise or accurate. Obtaining good hunters. difficult to assess, depending on return rates and population estimates of most wild animal how estimates are produced. All methods can be species, including deer, is notoriously dif­ The information given in this paper is made to perform satisfactorily. ficult. Still, great care must be taken in ob­ based in large part on the experiences of INTRODUCTION taining deer harvest information because all workers in Michigan where deer harvest sorts of errors can be present which will pro­ estimates have been calculated by various The white-tailed deer (Odocoileus v;r­ duce misleading results. methods since 1880. In addition, information ginisnus) is the most heavily hunted big game Over the years in the United States and was suppl ied by the other 13 states and pro­ species in the Midwest. In Michigan it has Canada, a number of approaches have been vinces making up the Midwest Fish and been the most hunted game species since employed to obtain deer harvest data - some Wildlife Conference. Finally, much material 1963, with some 760,000 individual deer good and some not so good. Trial and error was also obtained from literature searches. has weeded out most of these until today In the Midwest, 3 main deer harvest Ryel, Lawrence A. 1980. The legal deer kill - how only a few are in use. Bennett et al. (1966) estimate methods are currently in use: it's measured. Pages 37-45 in Ruth L. Hine and have chronicled the various systems used in registration, mail out-mail back surveys, and Susan Nehls, eds. White-tailed deer population Michigan. hunter report cards. These will be described management in the north central states. Proc. 1979 In all of these systems the primary objec­ and evaluated in some detail, along with Symp. North Cent. Sect. Wildi. Soc. 116 pp. tive is to estimate the total legal harvest; brief presentations of certain other methods. 37 NOTES ON SAMPLING of the sampling method used, the essential must present it to an authorized registration feature is randomness which insures the or deer checking station. They may be re­ Before discoursing on harvest surveys, absence of subjectiveness in the selection quired to answer certain questions concern­ some general comments on sampling seem in process. Virtually all statistical methods ing where and when the animal was taken, order. Kill surveys are really sampling prob­ presuppose randomly selected samples. and perhaps have a confirming seal attached lems and anyone involved in producing deer For various reasons, probability sampling is to the deer. In some states and provinces, harvest estimates by whatever means should not employed in all deer harvest surveys. In registration stations are operated exclusively be at least somewhat knowledgable about many cases there is no way of assigning a by the natural resource agency's personnel. proper sampling methods. Fortunately there probabil ity of selection to any population In others, private business establishments are several excellent sampling texts which element. Statistics such as means and such as gasoline service stations, motels, can be consulted: for example, Cochran variances can be calculated for nonprobabili­ restaurants, sporting goods stores, and (1977), Kish (1965), and Jessen (1978). ty samples, but these can really apply only to grocery stores are authorized to register deer The population is the central concept in the sample itself. There is no theoretical basis for a fee. sampling. It can be defined as a set of for arguing from such a sample to some In a nationwide survey, Guynn et al. (1977) elements having at least one characteristic in population. The sort of nonprobability reported that some 25 states used a statewide common. A sample, then, is some subset of a samples dealt with in deer harvest surveys are compulsory deer registration system in the population. Ideally, a sample will be a commonly called "handy" samples. 1976-77 hunting seasons, although in 4 of representative portion which will allow an ac­ Estimates derived from surveys can be these the hunter had to stop only if he en­ curate characterization of the entire popula­ evaluated in 2 ways, accuracy and precision. countered a checking station on his travel tion for the items of interest. One cannot say To the layman these 2 terms are synonymous. route. Some states registered deer only on whether a given sample is "good" or "bad" To the statistician they are quite different. certain management areas. In the Midwest, 6 without knowing population values. Samples Precision is a measure of the inherent states and provinces now use compulsory can only be rated on the basis of the selec­ variability of the data collected during a registration: Illinois, Minnesota, Missouri, tion process used and on long-term probabili­ survey and can be expressed in terms of Nebraska, Ohio, and Wisconsin. ties. That is, representativeness is inherent in range, standard deviation, variance or con­ In a deer registration system, the target the sampling plan and not necessarily in the fidence intervals. Accuracy refers to the population is made up of all deer legally sample at hand. closeness of the estimate to the true popula­ killed by hunters. The sampling frame of In most "real world" situations, one cannot tion value. Precision can be calculated from necessity consists of the successful hunters. usually sample the entire population of in­ the results of a survey, while accuracy can be In theory, a 100% sample is used. The ac­ terest, or target population, for one reason or determined only in rare instances where the curacy of a deer registration system depends another. Instead we settle for something Jess, true population value is known. Precision and entirely on relative compliance, which may called the sampled population, which can be accuracy are not necessarily related. For ex­ vary from area to area. Harvest figures will covered by something called a sampling ample, a subjective sampling procedure may always be equal to, or less than, the true kill. frame. A frame is a device, real or implied, produce an estimate with high precision but No statistical basis exists for estimating the which allows us to take hold of the popula­ low accuracy. total harvest from an incomplete count, tion piece by piece. At times one must work Bias is another term which has several which amount to a nonprobability sample. with a very incomplete frame. In a sample meanings. In the context of sample surveys, it Appraisal of the accuracy is necessarily sub­ survey one can really only estimate the commonly means systematic distortions due jective, largely based on prosecution reports characteristics of that part of the target to subjectiveness or nonrandomness in and rumors of noncompliance. Very likely population covered by the frame. sampling procedures. those choosing not to register their deer will Probability sampling is the highest ranked be different in some respects from those who form of sampling. A probability sample is one do. Accurate harvest estimates under this selected by a chance mechanism where the DEER HARVEST system depend on such things as aggressive chance of selection for each population ele­ ESTIMATING SYSTEMS law enforcement and judicial support, the ment entering the sample is known. In effect backing of hunters and biologists, an ade­ this means that population elements are Registration quate number and distribution of registration selected through the use of random numbers stations, and tradition. Without all of these (Rand Corporation 1955) at one or more steps Deer registration, as used here, means a ingredients in the recipe, the cake will likely of the selection process. Whatever the details procedure whereby hunters who kill a deer be a failure. Almost certainly the early years 38 of a registration system will produce serious Other phases of registration systems are biologists to registration stations in these underestimates of the harvest. also expensive. A large number of stations localities. must be used, several in each county or The accuracy of kill locations will depend Michigan registered antlerless deer in 2 management unit open to deer hunting. basically on the hunters' knowledge of local counties in its northern Lower Peninsula Wisconsin used 417 in 1978. Stations must be geography, although registrars may be able (Lake and Newaygo) during the 1956, 1957, adequately staffed and should be open dur­ to help identify colloquial place names. and 1958 seasons. At the same time, mail ing some portion of the evening hours. It re­ Successful hunters are a captive audience surveys were also conducted of hunters who quires a great deal of preseason effort to line while registering deer. All sorts of collateral received permits to kill an antlerless deer. In up enough registration stations in the proper information can be gleaned from them: 1956, 2,028 antlerless deer were registered places to provide adequate coverage of the animal sightings, hours of hunting, number in while the mail survey estimate was 2,076. state. It also requires a great deal of careful party, kind of lodging used, land ownership, Following the season, conservation officers planning to insure that biologists are assigned crippling loss, other hunters seen, kind of contacted 602 of the 2,463 apparently unsuc­ to stations in the most efficient manner. If gun, number of shots fired, costs of hunting cessful permit holders (not a random sample) private businesses are included in the net­ trips, miles driven, violations observed, type and found 16 (2.7%) who admitted they work, then contracts must be signed, liability of vehicle, opinion on regulations, and so on. failed to register an antlerless deer. Similarly, coverage provided, and instructional pro­ But there are 2 serious problems with this in 1957, 1,999 antlerless deer were registered, grams conducted. Normally the businesses approach. First of all, to be meaningful, per­ 2,016 were estimated by the mail survey, and receive a small fee for registering each deer, sonal interviews need to be carefully con­ officers queried 414 of the 2,055 apparently except that there is generally a minimum pay­ ducted by trained interviewers. With the unsuccessruf .tnct a random sample) and ment. Records must be audited prior to large number of peopleinvolved in a registra­ found 6 (1.4%) who did not register an issuance of checks. tion system, there is little hope that this can antlerless deer (Hawn and Ryel 1969). In Regardless of whether the stations are be accomplished. Secondly, only successful 1958, no postseason checks were conducted. manned by private citizens or natural hunters are seen. They do not constitute a By and large, deer registration is well liked resource agency personnel, they must be sup­ probability sample of all hunters nor in all by wildlife managers in those states and pro­ plied with signs, reporting forms, confirming likelihood are they a representative sample. vinces where it has been used for several seals, and maps. Stations require continual Successful hunters are not distributed years. Illinois, Missouri, and Wisconsin have servicing during the season. Following the uniformly but tend to be concentrated where all registered deer for over 20 years and Ohio season, all records must be promptly re­ the deer herd is largest. It would be unwise to has for 18 years. They all report high com­ trieved. expand such information to all hunters. pliance and satisfaction with the system Sex and age composition of the harvest Because of this, the midwestern states of (Lynch and Carr 1974). Registration provides must be based on the deer examined by com­ Minnesota, Missouri, and Wisconsin also harvest information for small areas which is petent biologists. Such data when recorded utilize mail surveys to provide information on often difficult to obtain with mail surveys. by lay workers are not reliable and their relative hunter success, hunting pressure, and The hunters and the public generally seem to records are useful only as "body counts". other information. accept the figures as valid, and registration Agencies will have little control over who Finally, it is important to realize that in a records can be easily checked by the skep­ registers deer at privately run stations. Some period of oil shortages, registration forces tical. Hand counts of kill data are available states, such as Michigan, can examine large many hunters to drive extra miles to register soon after the season. Final summaries of numbers of hunter-killed deer by using a few their deer since there are only a few stations registration data, however, may require checking stations (4 in Michigan) during peak per county. While this may not amount to a significant amounts of data processing com­ periods of hunter traffic on major highways. large distance for an individual hunter, in the pared to mail surveys. For example, in 1978, With a registration system, an overall ap­ aggregate it could amount to millions of about 145,000 deer were estimated to have praisal of the physical condition and com­ miles if appreciable numbers of deer are been harvested in Michigan's firearm season position of the harvest will require that killed. Rain and snow can also make deer based on a sample of some 33,000 hunters biologists be sent to a large number of registration an uncomfortable or even (Ryel 1979a). Registration would have re­ registration stations. Wisconsin regularly dangerous experience. quired over 4 times the data entry without mans about 70 during the early part of their yielding one iota of information about the firearm deer season. On the other hand, if Mail out-Mail back Surveys 600,000 hunters who did not kill a deer that detailed information is desired for specific year. areas, then it is a simple matter to send Mail out-mail back surveys, called mail 39 surveys hereafter, are used by about 26 states those recervrng special permits to harvest the confidence limits requires 4 times the to estimate deer harvests (Guynn et al. 1977). antlerless deer. Michigan regularly conducts sample size. In a management context, a In the Midwest, 7 states and provinces a separate survey in each of its hunter's sample is considered too small if its results employ this method for gathering harvest choice permit areas - some 113 in 1978 (Ryel are not precise enough to make appreciable data: Indiana, Iowa, Kansas, Manitoba, 1979a). Here variable sampling rates are used contributions to decisions (Kish 1965). In Michigan, North Dakota, and Ontario. In ad­ to obtain the most precise information for a practical situations, sample sizes are often dition, Minnesota, Missouri, and Wisconsin given overall sample size. Each area is in ef­ determined by costs. use mail surveys to obtain supplemental fect a stratum in a stratified sampling plan Response biases are caused by false report­ hunting information. (Cochran 1977). ing. This may involve only the kind of deer Basically this method involves sending Variable sampling rates can also be killed or whether or not a deer was killed at questionnaires to a sample of hunters and employed. A more accurate assessment of all. There is ample evidence that some asking them questions about their hunting ex­ the hunting effort and deer harvest for a par­ hunters upgrade the kind of deer killed in periences in the recent season. ticular county can be accomplished by in­ their questionnaire response. Doe fawns Mail surveys have several distinct advan­ creasing the sampling rate for the residents of become does, buck fawns become does or tages. When statistically valid sampling those counties who do the bulk of the hunt­ bucks, etc. The magnitude of this depends designs are used, the precision of the ing there. Or one could study the hunting ex­ partly on the hunting traditions of the state or estimates can be estimated from the survey periences of the hunters from Detroit by province. In a 1950's Michigan study, handy results themselves. Mail surveys are the most sampling licensees with certain ZIP codes. samples of deer were examined at locker economical of the several methods and costs Licenses can be sampled in 2 main ways: plants, seal numbers were recorded, and later and precision can be easily controlled by by a systematic sample with a random start these hunters were sent a questionnaire. manipulating sample size. Both successful or completely at random using random Compared to the deer actually observed, and unsuccessful hunters can be contacted. numbers. If the file is small or if all names mail survey results in general indicated a Questions can be asked about success, hunt­ and addresses are on the computer, a random small increase in the number of adult bucks ing locations, days hunted, and weapons selection is preferred. If a large file is reported, a relatively larger increase in adult used, as well as information on economics, sampled by hand, systematic sampling is the does, and a large decrease in fawns, especial­ attitudes, and sex and age of the hunters more practical approach. Two kinds of ly doe fawns. Most surprising, however, were themselves. The work is generally all carried systematic samples can be used: a straight a few that reported they did not kill a deer out at a central location with much of the ef­ systematic sample of every nth name or ran­ (Ryel 1960). MacDonald and Dillman (1968) fort done by clerical personnel. domly selected terminal license digits. Since reported similar results from a study in New The target population is licensed deer there should be no correlation between Mexico. hunters. The ideal sampling frame is a central license position or license number and hunt­ In Michigan, there is a long, historical file of copies of all licenses. Although even ing success, such systematic samples should tradition for members of a party to help each with a central license file the frame used may approximate true random samples and can other fill their tags, even though this is not not exactly correspond to the entire popula­ be treated as such for most analyses. An strictly legal. Hence, we have modified our tion of licensed deer hunters. For example, in alternative is to use the technique of questionnaire in recent years to find out if a Michigan not all carbon copies of hunting replicated or interpenetrating sampling sampled hunter's tag has been placed on a licenses are available for sampling. Some which allows valid computations of means deer rather than whether he or she killed a license agents go bankrupt each year and and variances from a series of systematic deer. In Michigan, data from deer checking their records are tied up in litigation. Some samples (Deming 1960). stations are used to determine the composi­ agents experience fires. Some license copies The accuracy of harvest estimates derived tion of the deer killed by hunter's choice per­ apparently become lost in transit, and some from mail surveys depends on the magnitude mittees. There is also a tendency for casual are returned too late to be included. Further­ of sampling error, response biases, and hunters, for instance some wives or youths, to more, Michigan and some other states sell nonresponse biases (Hayne and Eberhardt have the husband or father fill out any ques­ various combination of sportsman's licenses 1954). The probable magnitude of chance tionnaire sent to them. The average person which enable sportsmen to hunt or fish for sampling error (precision) can be estimated seems to have an uncomfortable feeling several species. Hence, not all people pur­ from the results of the sample itself and can about not being able to provide something chasing one of these licenses are deer be manipulated by sample size, assuming a positive. Questionnaires and cover letters hunters. random sample is used (Cochran 1977, Turner need to emphasize that the experience of the Other frames that might be considered are 1970). For a simple random sample, halving sampled person is what should be reported. 40 Finally, it is widely supposed that some believed to have been received and do not in­ rection factors based on the results of years hunters report killing a deer when in reality clude those where the addressee moved when special efforts are made to obtain high they do not. MacDonald and Dillman (1968) without leaving a forwarding address, died response rates. Correction factors can also be reported about 9% in their study did so. The after the season, gave a ficticious name developed by contacting a high proportion of difficulty here, however, is to contact a ran­ and/or address, or wrote illegibly on the car­ a random subsample of the nonrespondents dom sample of truly unsuccessful hunters bon copy of the license (phone directories through telephone or personal contact whose kill tags were not placed on a deer will often help decipher some of these). Nor­ (Hansen and Hurwitz 1946). without them being aware that such a list was mally, a 2-week period between mailings Correction factors have also been being assembled. Similar criticisms can be seems about ideal based on the usual pattern developed from comparisons of mail survey leveled at locker plant or checking station of returns. estimates with checking station data for studies which are certainly not random For a small annual survey of about 500 special management areas. This approach samples of the deer harvest. firearm deer hunters to obtain license pur­ has a high potential for trouble. First of all, It is important to conduct mail surveys as chase information, Michigan regularly hunters using such areas are not a random close to the end of the hunting season as achieves about a 98% response on delivered sample of all hunters, nor are they necessari­ possible. In South Carolina, Webb and questionnaires through a combination of 4 ly representative of all hunters. Secondly, it is Loadholt (1971) sent questionnaires to the reminder mailings followed by telephone extremely difficult to obtain a 100% kill same sample of archers 4-5 months apart and calls and finally personal contacts (Ryel estimate for a given area unless there are found both inflated kill and hunting effort 1979b). limited access points and all vehicles are estimates from the later survey. In a related Questionnaires should be attractive, short, searched as they leave. Otherwise an study Johnston and Webb (1975) found that unambiguous in wording and easily filled out . underestimate will almost certainly occur. man survey estimates of hunting effort and with the questions in a logical order. A cover Quality control checks made at Michigan's deer harvest on South Carolina game letter is very important to both explain the Mackinac Bridge suggest that the accuracy management areas were lower' than those reason for the survey and provide instruc­ of the deer count there varies with the length produced by a subsequent personal interview tions for filling out the questionnaire. Filion of the lineups to pay tolls, the individual survey of a random sample of 6% of the (1974, 1978) has described a number of booth operators, and the mood of returning respondents to the mail survey. The harvest methods for increasing return rates on Cana­ hunters. Probably less than 85% of the true estimate for the identical subsample of dian Wildlife Service surveys. Experiences in total is counted in most years. Finally, results hunters was 23.3% higher for the interview. Michigan suggest that response rates are in­ from poorly conducted mail surveys are Finally, since seemingly minor changes in versely related to time delays in sending typically used for comparison, i.e., one hav­ the wording of questions may cause dramatic questionnaires to the hunters. Ideally, forms ing a very incomplete sampling frame, or a changes in response, it is important to pretest should be mailed near the end of the hunting poor response rate, or both. As Deming (1960) questionnaires. As in all professions, wildlife season; however, waiting for license agents to stated: "Good agreement or poor agreement biologists have developed a unique jargon return carbon copies may mean a delay of 2 between the results of a sample and of a which is not necessarily understood by months before initial mailing. complete census taken at about the same hunters. Mitigating the effects of nonresponse bias time does not of itself establish or disprove Nonresponse bias refers to possible dif­ can be done in a variety of ways. In Michigan, the precision of the sample, nor the quality of ferential results by different classes of the approach is to obtain response rates near the complete census." hunters. In a mail survey those responding to 90% and assume the remaining 10% have Some workers seem to feel that an in­ the original mailing tend to be more suc­ similar experiences. To use this same ap­ creased sample size by itself will somehow cessful than those responding to reminder proach with low response rates seems quite improve the quality of mail survey results. mailings (Eberhardt and Murray 1960). Hence, risky. The effect of even relatively low Deming (1960) spoke to this point very suc­ it is important to obtain a high response nonresponse on confidence limits for the cinctly: "One cannot offset the hazard of through the use of several reminders. In mean success can be quite serious (Cochran nonprobability sampling by increasing the general, the lower the response rate, the 1977). Substituting new individuals for size of the sample." Rather than increasing higher the inflation of the estimates. nonrespondents is of little use and may make initial sample sizes in order to obtain some Michigan currently uses an original and 4 . matters worse. It has the effect of replacing stated number of returns, it is far better to reminder mailings to achieve return rates of non respondents with other people who close­ spend extra money increasing the response about 90% of delivered questionnaires (Rye) ly resemble those who have already re­ rate of a smaller sample. Similarly, a valid 1979a). Delivered questionnaires are those sponded (Kish 1965). Some states apply cor­ sampling plan should be substituted for a

41 poor one even if the former is more expensive The basic advantages of hunter report 820,000 deer licenses of various types (Ryel and the sample size must be smaller. cards are that the hunters receive them 1979a) would be a massive and expensive Various mathematical approaches have generally at the time of license purchase, undertaking even in the computer age. been used to estimate true hunter success. thus saving postage and insuring receipt. In A compulsory report card system can be For example, the relation between addition, the hunters are aware of the infor­ thought of as a hybrid between mail surveys cumulative returns and cumulative success mation which is desired prior to the season. and deer registration, but it lacks the major can be projected by regression techniques to Both the target population and associated advantages of either system. The problem of estimate the 100% point. Others have fitted a sampling frame are all of the licensed obtaining the necessary high return rates is second degree polynomial to these data and hunters. In theory, memory bias should not akin to that in mail surveys except on a much projected the tangent to this curve at the last be a problem since all hunters are aware at larger scale. The idea of obtaining data from point and to the total sample size (Legler and the outset of the season that they must all hunters, or all successful hunters, is not Hayne 1967). A very conservative approach is supply their hunting results. unlike examining all deer in a registration to assume that none of the non respondents The chief disadvantages include the high system. killed a deer. costs involved with supplying all hunters with Some states have attempted to adjust for Besides introducing biases, the problem of cards and, more importantly, the problems incomplete returns in calculating deer non response on mail surveys also affects the associated with incomplete returns. As with harvest estimates by employing a Lincoln­ time when data are available. If an agency mail surveys, successful hunters are more apt Index-type estimator. The idea here being must make 4 or 5 mailings at 2-week inter­ to respond than unsuccessful, causing in­ that in the population of hunter-killed deer, a vals, then a minimum of 2 to 2Yz months is flated estimates. The degree of inflation is in­ sample would be "marked" (m) when deer spent in just obtaining returns. To this must versely related to the response rate (Hayne were exam ined at roadside deer checking sta­ be added the time preceding contacts for and Eberhardt 1954). As with deer registra­ tions by recording seal numbers. The second sample selection and the time following con­ tion, accuracy depends on achieving near sample would be the returns of hunter report tacts required for analyses and report writing. 100% compliance. Here the logical approach cards (n) which contain j( of the "marked" When the start of the survey is delayed would be to use several reminder mailings deer. Then an estimate of the total kill can be because lists of hunters are unavailable for and/or to treat nonrespondents as a separate calculated according to Overton and Davis sampling immediately after the season, stratum and subsample them. This approach (1969) by: harvest estimates may not be available until must also be followed where only successfuls A nm several months following the season. are asked to report, since there is no way of N A mail survey of deer licensees was first knowing who is successful. x conducted by Michigan in 1925, but since Michigan used compulsory report cards, only 50% responded to the single mailing, which were attached to all hunting licenses, In practice, there are some serious problems the workers at the time felt the survey wasn't from 1937 through 1954; however, the report­ with this system. The following discussion accurate enough. Had they followed through ing regulation was never enforced. Returns follows Ryel (1965): with reminder mailings, Michigan could have declined from 66% in 1937 to 18% in 1951 (1) All tag numbers must be accurately had a 27-year head-start on their current (Bennett et al. 1972). Numerically, even the recorded and properly matched against system. low returns tended to overwhelm the the mail sample. Mistakes would mostly available staff and subsampling of returns cause X' to be smaller than actual and Hunter Report Cards was used in some years to produce harvest cause an overestimate of the kill. estimates. During this period, firearm deer (2) The proportion of hunters who have their Hunter report cards are considered here to license sales ranged from a low of 158,720 in deer checked and also report by mail must be questionnaires supplied to all hunters, 1937 to a high of 386,400 in 1951. Hence the be the same as the proportion of hunters usually at the time of license purchase, which 18% return in 1951 amounted to some 69,500 who did not have their deer checked and are to be returned by mail at the end of the cards. Because of the need for labor saving, reported by mail. In other words, the fact season. Some states require only successful the old Michigan Conservation Department that a hunter has his deer examined at a deer hunters to report. Guynn et al. (1977) in­ turned at an early date to punched cards and biological checking station must not in­ dicate 11 states used hunter report cards in performed analyses with the forerunners of fluence his decision whether or not to 1976. In the Midwest, South Dakota and Il­ today's sophisticated electronic data process­ send in his voluntary report card by mail. linois (bowhunters only) are currently using ing hardware. Attempts to use a report card Furthermore, this must be true at all this method. system today in Michigan with sales of over geographic levels in order that estimates

42 will be valid for counties or groups of southern lower Peninsula but hunt in the chase a license the following year (Ryel counties. If hunters are more likely to re­ northern two-thirds of the state. Hence, 1967). ply by mail after having their deer counts are made at the Straits of Mackinac (3) Hunters hunting 2 consecutive years were checked, the kill estimate is deflated; if and across the middle of the lower Peninsula more successful than those that didn't less likely, the kill estimate is inflated. on major north-south highways. At the (Ryel 1968), hence, expanding the results This has a high potential as a source of Mackinac Bridge, ticket booth operators over the license sales for the previous year trouble. For instance, a good share of the make a continuous tally for the entire season. produced inflated harvest figures. hunters whose deer are being examined At all other locations deer counts are made (4) license agents neglected to ask the re­ are likely to inquire if they should also during samples of time periods. These counts quired questions of many hunters and send in their deer report. By answering, the are for 30 minutes and are distributed roughly some hunters refused to answer the ques­ examiner has influenced the hunter's deci­ proportioned to expected deer flow. Observa­ tions. sion. tions from these samples are first expanded In New York, comparisons between to total southbound visible deer, then, using Michigan had also employed a similar hunters whose deer were checked only the proportion of visible deer among those system from 1932 to 1936 with the same poor once and those checked more than once examined at roadside checking stations, to an results. (i.e. at a locker and a checking station), estimate of the total southbound deer. Multi­ disclosed that a higher proportion of those ple regression prediction equations, Telephone Surveys. Telephone surveys undergoing multiple checks mailed in calculated from the relationship between ad­ have been used to estimate game harvests, their reports (Hesselton 1964). This sug­ justed deer observations and the mail survey particularly in the Southeast, with some suc­ gests that a higher proportion of con­ harvest estimates for prior years, are used to cess (Stern et al. 1962). The chief advantage tacted hunters woutd report compared. predict what the mail survey estimate will be of telephone surveys is the ability to quickly with those not contacted at all. If true, this for the just completed season. Estimates are contact a sample of people. However, the would result in a deflated kill estimate. available about a day or two after the season sampling frame is typically very incomplete. (3) The deer examined on roadside checking and provide a total kill estimate which has Not all people have telephones and for those stations must be a random sample of the been quite accurate in the past, averaging who do, some numbers are unlisted or not up­ kill, and hunters who report killing a deer within 6.9% of the mail survey for the 25-year to-date. Generally, it is not easy to trace on their report cards must be a random period 1954 to 1978. The kill is also appor­ someone who has moved. Michigan workers sample of all hunters who shot a deer. This tioned into bucks and antlerless deer by have used telephone surveys to reach chronic latter condition does not hold in Michigan region of the state using preliminary sum­ nonrespondents on certain mail surveys to since placement of deer checking stations maries of checking station data and a subjec­ deer hunters. With a nearly complete set of depends on the fact that most deer are tive appraisal by field wildlife biologists in current Michigan phone directories, one can killed in and most the southern lower Peninsula. obtain phone numbers for about 70% of the hunters live in southern Michigan. Such an deer hunters (Ryel 1971). Our feeling is that unbalanced sample would cause x to be License Stub Survey. Hawn and Ryel (1969) those with listed telephone numbers are small relative to n and would result in an reported on a system imposed on the probably different in some respects from inflated estimate. Michigan Department of Natural Resources those for whom we cannot obtain numbers; by the Michigan legislature in 1967 wherein however, no real evaluation has been made. Other Methods hunters purchasing a deer hunting license Telephone company records of residential were queried about their hunting during the phone numbers provide a more complete and Deer Traffic Survey. To provide a kill figure previous year, 1966. The advantages of this up-to-date sampling frame, including unlisted right after the season for the news media and system were similar to those of the hunter numbers, for the much smaller target popula­ the winter "hot-stove" league, Michigan report card, as were the disadvantages. In ad­ tion of hunters. Even so, hunters without developed the so-called deer "traffic survey" dition, there were some unique problems: telephones are not included in the frame at in 1952 (Hayne and Eberhardt1956). Actually, (1) Estimates were not obtainable until 15 all and some phone numbers will include traffic isn't surveyed at all. Rather, observa­ months after the season. several hunters in the same household. In tions of deer going south on hunters' vehicles (2) Not all hunters in 1967 hunted in 1966. spite of these deficiencies, where no central are used. These are conducted entirely Michigan has found that 75-80% of Iicense file exists and/or where there are large separate from roadside deer checking sta­ firearm licensees and 60% of bow and numbers of unlicensed hunters, telephone tions. In Michigan most hunters live in the arrow licensees in a given year will pur­ surveys provide virtually the only approach. 43 In Michigan a year-long telephone survey LITERATURE CITED presented at 16th Midwest WildI. Conf., of recreation participation was carried out in St. louis, Mo. 1976 with telephone interviews - some 17,781 BENNETT, CL, JR, Ll. HAWN, D.H. JENKINS, and l.A. 1956. Estimating deer kill from counts of deer in all (Kennedy 1977a, 1977b). Compared to RYEL on automobiles. Pap. presented at 18th the regular mail survey of 1976 firearm deer 1972. Background material helpful in Midwest Wildi. Conf., Lansing, Mich. hunters, estimates of hunter-days from the understanding the Michigan system for: HESSELTON. W.T. telephone survey were 26.3% higher. estimating deer numbers, estimating the 1964. Report on percentage of hunters report­ annual kill of deer by hunters, and ing the deer they take - 1963. N.Y. Con­ DISCUSSION measuring the number of deer lost by servo Dep. 5 pp. starvation, illegal in-season kill, ac­ The ideal system for determining the deer cidents, predation, disease, etc. Mich. JESSEN, R.J. 1978. Statistical survey techniques. John Wiley harvest should produce accurate, mean­ Dep. Nat. Resour. Res.and Dev. Rep. No. 52A. 18 pp. & Sons, lnc., New York. 520 pp. ingful, timely, and acceptable information on harvests, hunting and hunters at minimum BENNETT, c.l., JR., l.A. RYEL, and l.J. HAWN JOHNSTON, W.E. and l.G. WEBB costs. Unfortunately, no such system exists. 1966. A history of Michigan deer hunting. 1975. Postal and personal interview surveys of Whichever methodology is employed will be Mich. Dep. Conserv. Res. and Dev. Rep. South Carolina deer hunters. Proc. Conf. No. 85. 66 pp. a compromise in one or more respects. Southeast. Assoc. Game and Fish Comm. Regardless of the system (or systems) used, COCHRAN, W.G. 29:655-64. 1977. Sampling techniques. John Wiley & however, it should be carried out in a careful KENNEDY, J. Sons, Inc., New York. 428 pp. and thorough fashion. Too often data are col­ 1977a. Michigan 1976 recreation survey, design lected in a manner not consistent with good DEMING, W.E. and application. Mich. Dep. Nat. Resour. statistical practices. Furthermore, it is easy 1960. Sample design in business research. John Recreat. Servo Div. 110 pp. for procedures to become routine and for in­ Wiley & Sons, Inc., New York. 517 pp. 1977b. Michigan resident recreation activities herent biases to creep in and become EBERHARDT, l.l. and R.M. MURRAY and providers. Mich. Dep. Nat. Resour. perpetuated. Workers should be aware of 1960. Estimating the kill of game animals by Recreat. Servo Div. 20 pp. potential problems and take great pains to licensed hunters. Proc. Soc. Stat. Sect., KISH, l. 120th Annu. Meet. Am. Stat. Assoc. avoid them. Where possible, independent 1965. Survey sampling. john Wiley & Sons, 120:182-88. quality control checks should be built in. Inc., New York. 643 pp. In conclusion, I hesitate to make specific FILION, F.l. recommendations. The methods described in 1974. Methods for increasing returns in mail LEGLER, E., JR. and D.W. HAYNE 1967. Sampling to determine unreported deer this paper can all be made to perform hunter surveys. Can. Wildl. Servo Biom. kill in Tennessee, 1964-1966. Proc. Conf. satisfactorily. The choice depends on many Sect. Rep. NO.7. 54 pp. Southeast. Assoc. Game and Fish Comm. 1978. Increasing the effectiveness of mail factors. The important thing is for an agency 21:69-73. to adopt a harvest-estimating system with surveys. Wildi. Soc. Bull. 6(3):135-41. their eyes open - to be aware of the pluses GUYNN, D.C., JR., H.A. JACOBSON, and W.A. MITCHELL LYNCH, G.M. and H.D. CARR and minuses. Caveat emptor! 1977. Deer data collection systems used by 1974. Compulsory registration of big game state wildlife management agencies. kills in North America. Alberta Dep. Miss. Game and Fish Comm. Perf. Rep. Recreat., Parks, and Wildl. Wildl. Res. Pittman-Robertson Proj. W-48-24. 17 pp. Ser. No.1. 18 pp.

HANSEN. M.H. and W.N. HURWITZ MACDONALD, D. and E.G. DilLMAN Acknowledgments. I wish to thank my col­ 1946. The problem of non-response in sample 1968. Techniques for estimating nonstatistical league, louis Hawn, and my secretary, Marilyn surveys. J. Am. Stat. Assoc. 41:517-29. bias in big game harvest surveys. J. Moss, for their invaluable help in the preparation Wildl. Manage. 32(1):119-29. of this paper. In addition I offer my sincere thanks HAWN, l.J. and l.A. RYEL 1969. Michigan deer harvest estimates: sample to the many people in the other Midwest states and OVERTON, W.S. and D.E. DAVIS surveys versus II complete count. J. provinces who contributed information about their 1969. Estimating the numbers of animals in Wildl. Manage. 33(4):871-80. surveys. wildlife populations. Pages 403-55 in This study was financed in part by Federal Aid in HAYNE. D.W. and l. EBERHARDT R.H. Giles, Jr., ed. Wildlife management Wildlife Restoration Act, under Pittman-Robertson 1954. Nature of the bias of estimates com­ techniques. The Wildl. So c., Project FW-3-R. puted from voluntary reports. Pap. Washington, D.C. 623 pp.

44 RAND CORPORATION 1955. A million random digits with 100,000 nor­ mal deviates. The Free Press, Glencoe, III. 600 pp.

RVEL, L.A. 1960. Reporting bias in mail surveys of deer hunters. Mich. Dep. Conserv. Game Div. Rep. No. 2312. 10 pp. 1965. Comparison of the Michigan, Wisconsin and New York systems for estimating deer harvests. Mich. Dep. Conserv. Res. and Dev. Rep. No. 24. 8 pp. 1967. Hunter participation survey D, 1966. Mich. Dep. Conserv. Res. and Dev. Rep. No. 123. 7 pp. 1968. 1968 deer hunter success survey. Mich. Dep. Conserv. Res. and Dev. Rep. No. 129.5 pp. 1971. 1970 deer hunter success survey. Mich. Dep, Niit.Re-SOUf. Res. and Dev, Rep. No, 248. 10 pp. 1979a. The 1978 deer seasons. Mich. Dep. Nat. Resour. Surv. and Stat. Servo Rep. No. 184.12 pp. 1979b. Deer hunter participation survey, 1978. Mich. Dep. Nat. Resour. Surv. and Stat. Servo Rep. No. 185. 3 pp.

RVEL, L.A., G.c. jAMSEN, and L]. HAWN 1970. Some facts about Michigan hunters. Mich. Dep. Nat. Resour. Res. and Dev. Rep. No. 197. 26 pp.

STERN. H., jR, W.S. OVERTON, L. SOILEAU, and E. LEGLER, )R 1962. Results of design tests of methods of estimating doe harvest. Proc. Conf, Southeast. Assoc. Game and Fish Comm. 16:85-103. <~ TURNER, D.W. { 1970. Mail surveys of hunting - precision and sample size. Proc. Conf. Southeast. ,=<_K~'~?> ~-'> Assoc. Game and Fish Comm. 24:292-303.

WEBB, L.G. and C.B. LOADHOLT 1971. The significance of a "time lag" in con­ ducting a postal survey of archery deer hunters. Proc. Conf. Southeast. Assoc. Game and Fish Comm. 25:82-85.

4S Wisconsin DNR WINTER - THE GRIM REAPER

Patrick D. Karns Minnesota Department of Natural Resources 201 S. Golf Course Road Grand Rapids, MN 55744

Abstract: White-tailed deer are considered re­ rapidly filled by deer from existing stock, as into Minnesota, ending here about 1920, cent invaders in the northern portion of their range, was the case in Michigan (Bartlett 1938, thence proceeding into northwestern Ontario and as such their physiological adaptation is in­ 1950), Wisconsin (Dahlberg and Guettinger where deer reached their maximum north­ complete. This is manifested by overwinter mor­ 1956) and parts of Minnesota (Erickson et al. ward extension at Sioux lookout in the 1950's tality of deer when winter severity exceeds the 1961, Petraborg and Burcalow 1965). The ex­ (Ontario Deer Technical Committee 1978). deer's physiological limits. Fawns require the treme northern portion of Minnesota and Along with the new habitat came control greatest amount of nutrition through the winter, and are hence the first to feel the effects of under­ northwestern Ontario was occupied by of major deer predators, the timber wolf nourishmerrt. Various means of assessing winter moose and caribou and was devoid of deer (Canis lupus). (Felis cosicolor], and severity and effects on overwintering deer are until the late 1800's. The logging operations man (Homo sapiens); and a conservation con­ reviewed. that created this habitat began in eastern On­ science that led to a very protectionist at­ tario in the 1830's and proceeded westward titude towards deer (Flader 1974). Michigan HISTORICAL PERSPECTIVE The energy crunch is nothing new for our whitetails. Ever since they've "moved north" there have been winters too severe for them to keep the fuel tank full. As a result of these MANITOBA ONTARIO severe winters, deer have died in large numbers. Historical evidence places the major por­ tion of the presettlement white-tailed deer populations of the upper Great lakes region south of 45°N latitude (Bartlett 1938, 1950; Dahlberg and Guettinger 1956; and Erickson et al. 1961; Ontario Deer Technical Commit­ tee 1978) (Fig. 1). As whitetails are basically grazers (Cook and Hamilton 1942, Mooty 1976), the vegetational changes brought about by logging, fires, and settlement created a considerable amount of suitable habitat north of the 45°N latitude that was

Karns, Patrick D. 1980. Winter - the grim reaper. FIGURE 1. Approximate northern limits of major white­ Pages 47-53 in Ruth L. Hine and Susan Nehls, eds. White-tailed deer population management in the tailed deer populations prior to logging and settlement north central states. Proc. 1979 Symp. North Cent. (from Bartlett 1938, Dahlberg and Guettinger 1956, Wildl. Soc. 116 pp. Erickson et al. 1961).

47 implemented the buck law for deer hunting port 2-3 deer/acre/year, and less than 1 deer range north of the 45th parallel, and in 1921, and "by 1936 starvation was rampant deer/acre/year after being browsed out. The literally hundreds of thousands of deer have . ' ." in the northern lower peninsula (Bartlett carrying ability of unbrowsed cedar areas died as a result of these severe winters since 1950:11). Similar losses were also occurring in deteriorated almost as fast as the browsed the 1930's, and yet, throughout most of the Wisconsin, which had enacted the buck law areas through natural events of self-pruning area the deer population persists. in 1915 (Swift 1946:30). The buck law was not and decreased plant reproduction. The carry­ The exception to this persistence is in the adopted by Minnesota, but alternate year ing capacity of hardwood yards was deter­ most northerly expansion in northwestern seasons were adopted from 1922 to 1932, mined to be less than 1 deer/acre/year Ontario, from lake Nipigon to Sioux lookout, after which annual any-deer seasons have (Davenport 1939). where deer, as plentiful as they were in the been the rule until 1970. As a result deer A similar study in a "typical" Wisconsin 1950's and 1960's, now only occur as scat­ populations erupted (leopold et al. 1947), as deer yard indicated this type could withstand tered individuals throughout the area. The did the controversies over proper herd approximately 60 "deer browse same applies to northeastern Minnesota, and management. All of this took place north of days"/acre/year and remain productive. It is overall the combination of changing habitat, the traditional deer range, an area where interesting to note there was recovery of weather, , and hunting has been winter weather entered the arena as "the plants in the pens 2 years after all deer were responsible for such population declines Grim Reaper". removed, but in unbrowsed control areas the (Mech and Karns 1977). Wisconsin, in response to the wintering number of stems decreased 73% (Dahlberg problem, began an artificial feeding program and Guettinger 1956:197). in 1934, which lasted until 1956 (Haberland All of this, and that reported by many other PHYSIOLOGICAL pers. comm.). This practice was avoided on a workers, dwelt on the theme that because CONSIDERATIONS large scale by Michigan and Minnesota. there were overwinter losses, deer exceeded From the 1930's on, losses of deer to starva­ the carrying capacity of the range, as, by A considerable amount of research effort tion were documented as a result of severe definition, there should be no overwinter has been expended over the past 3 decades in winters throughout the area, and biologists losses if the herd was "in balance" with the attempting to understand more about the fought with almost everyone regarding deer range, This would imply then that no over­ nature of wintering deer. It wasn't until the population management. In severe winters winter losses would occur, regardless of any pioneering work of Helenette Silver's group the deer population occupied less than 10% extrinsic factors. (Silver et al. 1969, 1971) that we began to of the total range in "deer yards". Population Studies by Aldous (1952), Krefting (1941), grasp the adaptations the deer had to surviv­ problems were more apparent in these yard­ and Krefting et al. (1966) demonstrated that ing the winter months. The basic fact that the ing areas than on other areas of the range. It browsed plants such as willow, dogwood, and physiological mechanisms of deer resulted in was during this period that the concept of mountain maple actually produced more a reduced fasting metabolic rate during carrying capacity for deer was developed as foods than unbrowsed stems, and were very winter months, which in turn would allow " ... the number of deer a unit of range can tolerant to the annual removal by deer. them to exist on reduced rations, was a real support for a full year without serious Krefting concluded that what had been re­ breakthrough in our understanding of deer damage to the plants that provide deer food ferred to in the past as "overbrowsing" survival. Moen (1978) has added significantly and cover or to the deer themselves" should have been termed "heavy browsing" to this basic framework in describing the (Dahlberg and Guettinger 1956:192). This and had the beneficial effect of producing seasonal nutrient requirements as a smooth became dogma in the field of wildlife more browse over a longer period of time flowing sine wave, reaching its nadir in winter management. than unbrowsed plants. In a 25-year study of and apex in summer. Expressed as multiples Early research efforts were focused on forest development in Minnesota, Krefting of basal metabolism: deer winter concentration areas, referred to could find no effects of deer browsing (1975). as deer yards (i.e., Bartlett and Stephenson Perhaps a better definition of carrying BMR = 70wO.75 1929), food habits, feeding trials, and carrying capacity would be that density at which deer capacity of deer yards (Davenport 1939, can attain their full optimum biotic potential, where BMR = basal metabolic rate and W = Davenport et al. '1944, Davenport et al. 1953, realizing in some winters there will be mor­ weight of the deer in kilograms, these Kabat et al. 1953, Dahlberg and Guettinger tality, and in some areas deer may control multiples range from 1.5 in the winter to 3.5 1956:196). From the work of Davenport et al. vegetation to their advantage. 'in summer. The multiples during winter (1953), it was determined that cedar deer Severe winters, in terms of deer survival, months also increased with increasing snow yards, in the peak of production, could sup­ are more frequent in the recently occupied depths.

48 Quite simply, in the northern environment, The Ontario snow rating system (Passmore ceed their physiological limits?), and the in­ there are winters that extend beyond the 1953) uses basically the same criteria as the creased nutrient demands placed on preg­ deer's physiological limits, either in total "46 cm for 13 weeks", but adds to it a factor nant does in the last trimester of pregnancy, length of the winter or snow depth, that result for evaluating crust conditions by the ability beginning in mid-March. No attempt has in mortality. This is not a function of deer of the snowpack to support a man on been made under Michigan conditions to population density, as length of winter and snowshoes. Crust on the snowpack can act in relate the Winter Severity Index of Verme to snow depths are independent of deer popula­ a positive manner by elevating deer to new total overwinter mortality (Vogt pers. tions, but is a basic physiological property of food supplies, making travel beyond yard cornrn.), white-tailed deer. The deer are unable, under limits possible, and decreasing the amount of certain conditions, to meet their nutrient energy required for travel (Moen and Evans EFFECTS OF WINTER ON DEER needs, and die. Even on white cedar (Thuja 1971, Moen 1978). A crust can also be a occidentalis], the most "complete" deer negative factor by enabling predators to Minnesota has employed Verme's Index food, Ullrey et al. (1970) concluded pregnant move advantageously over the snowpack or since 1967. Despite the fact that it is only a adult does would lose approximately 23% of sealing off herbaceous food supplies. Crust crude approximation of winter severity their fall weight over a 9O-day wintering formation is variable across the area, depend­ (Moen 1978), an analysis of the data has period. This is close to the critical 30% over­ ing on climatic factors. In the more eastern revealed significant correlation between the winter weight loss, the point at which deer areas, and near the Great Lakes, crusts may Index and the dead deer found on spring deer die. The length of winter has its effects by form at almost anytime during the winter. In pellet count courses conducted since 1973 (r reducing the nutrients available to deer while Minnesota and northwestern Ontario, the = 0.98, N = 7 years) and with the rate of their metabolic demands are stitl high in the snowpack usually stays loose and fluffy all femur fat depletion in adult does and fawns. fall and spring. Snow depth increases the winter, with crust formation usually limited Although there is a good correlation be­ nutrient expenditure for obtaining food. to late winter and early spring. When crusts tween dead deer and the Winter Severity In­ Moen (1968) demonstrated that deer can form in late winter in the more northerly dex, we have been unable to arrive at a withstand the rigors of our northern winters areas, they have resulted in overkill by timber satisfactory means of going from dead deer with little or no overhead canopy if they wolves (Mech et al. 1971) and provide an ad­ found to total overwinter losses. For 4 years maintain a high plane of nutrition. Again, we vantage for , , and domestic the fat loss in the femur shows strong correla­ have an animal that, while it is adapted to dogs in preying on deer. tions with the total Winter Severity Index for seasonal changes in nutrient availability in its Kohn (1978), working in Wisconsin, has fawns (r = -0.83) and adult does (r = traditional range, is not completely adapted developed an index to winter severity using -0.94) (Fig. 2). Data for bucks are too limited to life in the northern forests. Hence, en­ Climatological Records. The number of days for similar analysis. Although these relation­ vironmental extremes have more of an effect with 46 cm or more of snow is added to the ships are apparent, it is not presently possible on deer in the more northerly areas of their number of days the minimum temperature to extrapolate them to survival, but they are range. goes below -18 C. The resulting index cor­ another index of winter severity. Based on relates well with dead deer found on inten­ findings in dead deer, 30% fat in the bone WINTER SEVERITY sive dead deer searches in northern Wiscon­ marrow has been described as the starvation ASSESSMENT sin. level. No one has performed serial marrow Verme (1968), working in Michigan, biopsies to see if surviving deer reach a Several schemes have been devised to developed a rating scheme for winters taking similar plane and recover. assess winter severity in terms of deer mor­ into account snow depth, snow support, and Stepwise multiple regression between per­ tality. The simplest is merely the rule of the cooling power of the environment. The cent of femur fat loss in fawns and monthly thumb used in Minnesota that "46 cm or cumulative index for the winter is taken as an Winter Severity Index values revealed that more of snow for 13 weeks" results in deer index of winter severity as it relates to deer February and March were the critical months. mortality. The 13 weeks means early snows and fetal development (Verme 1977). From The March Index' was critical to fat depletion that leave late, thus imposing a restriction on the Winter Severity Index, under Michigan in does. As the onset of fat depletion in the nutrient intake at times (late fall and early conditions, Verme states that weather in femur is rapid (Harris 1945), and represents spring) when metabolic demands are not December and April is the determinant of loss of the last vestiges of stored fat, the down to winter levels, and also increases the overwinter and newborn fawn survival, weather preceding February in the case of amount of energy required to obtain the respectively. This accounts for how long the fawns is what sets the stage for this depletion. foods available during the winter. deer are on a winter diet (does it meet or ex­ Similarly for adult does, while condition of

49 the fawns, those deer entering their first winter. Bartlett (1950) reported 90% of the 100 overwinter losses to be fawns in Michigan, whereas Swift (1946) reported 76% in Wisconsin. Fawns comprise about 85% of the 90 deer found dead by wildlife managers as a result of winter weather in recent years. This z disproportionate loss of fawns has its ex­ 0 80 i= planation, at least in part, in the fact that w they have physiological demands imposed on ...J 7'0 a.. them for continued growth during the winter w months (Thompson et al. 1973). Cl 60 !c{ As overwinter losses are primarily L&.: restricted to young of the year in the popula­ a:: 50 tion, the principal effect is a reduction in the :::::> next year's sport harvest, as a proportion of ~ w 40 the yearlings which make up a large share of u, the harvest, particularly in bucks-only hunt­ I­z ..+ ing situations, are absent. Prior to 1973, Min­ ., 0'" nesota seasons were for deer of either sex and w 0(" 0 u Q o· any age, and as a result approximately 1/3 of a:: / wa.. :~ the deer harvested were fawns. Thus, larger­ o·'0'" scale overwinter losses were averted because " 10 ... a large share of this cohort had already been removed by hunters. Estimating the overwinter losses is a her­ 0' iii ii'I iii I culean task. Large-scale dead deer searches, 15 30 45 60 75 90 105 120 135 150 as described by Whitlock and Eberhardt WINTER SEVERITY INDEX (1956) have been applied in Michigan follow­ ing severe winters (e.g., Burgoyne and Moss FIGURE 2. Percent femur fat depletion in fawns and 1978, Ryel1959 and others) and in Wisconsin does as related to the total Winter Severity Index in (e.g., Thompson 1979). The searches require a northern Minnesota. large amount of manpower and generally lack in the precision of their estimate. The the marrow may decline rapidly in March, which continue tissue growth throughout the winter of 1978-79 was described as "the worst depending on March weather, the winter up winter (Thompson et al. 1973), the critical winter east of the rockies". Overwinter losses to that point will largely determine if fat time in determining their survival begins in estimated for Michigan's lower peninsula stores are mobilized. Additional field February. were approximately 83,000 deer (Burgoyne measures of stress are needed to fully assess and Moss 1978) and were estimated as 30,000 the environmental effects on deer survival. ESTIMATING OVERWINTER ± 20,000 in the northern area of Wisconsin Stepwise multiple regression of monthly DEER LOSSES (Thompson 1979) based on statistically Winter Severity Index values to fetus designed dead deer searches. Haberland development indicated March as having the Most large-scale dead deer searches (pers. comm.) placed Wisconsin's total over­ most significance. generally conclude that insufficient numbers winter losses at 100,000 animals. Thus, in accord with the physiological of deer were found to arrive at any sex and Techniques currently used to estimate demands of deer, months when the age distribution of the overwinter losses. overwinter losses rely on a large number of metabolic rate is above the mid-winter nadir Despite these limitations of data, it has been people searching for dead deer in the spring. have the most bearing on fat depletion in almost universal experience of the deer Techniques to reduce the amount of man­ does and fetal development. For fawns, managers that losses are largely apparent in power, and improve the precision of the 50 estimates are needed. Recent developments and disease, increases predation, provides in­ McCullough (1978:175) described the quality in the mathematical treatment of strip count digestible artificial food, etc. All of these and quantity of the late summer fattening data may allow their use for estimating the arguments were developed under a period, which is dependent upon precipita­ losses (Eberhardt 1968,1978; Anderson et al. superabundance of deer, deer that could tion, as critical for overwinter survival in Tule 1979, and Burnham et al. 1980). have been harvested before they got into this elk (Cervus e/aphus nannodes] of California. Recent developments in treating strip predicament but weren't, so that they starved Stewart et al. (1976) presented a model for count data should permit their use in and the high populations persisted. Artificial overwinter losses of moose in Saskatchewan estimating overwinter losses (Eberhardt 1978, feeding may be, as one manager put it, the based on the growing season. In Minnesota, Anderson et al. 1979). Although in Minnesota lowest form of deer management known. Yet, as elsewhere, we have demonstrated the the number of dead deer found on spring deer can be fed successfully if provided a seasonal cycle of nutrient availability, and pellet counts has a strong correlation with complete diet. Feeding may be justified in are now into a long-term project to determine the Winter Severity Index, this number (from some situations, such as trying to maintain or what the differences may be between years. 3 to 25 deer/year) may not be sufficient for a increase a low population in face of a severe Differences in crude protein are apparent good estimate. Anderson et al. (1979) stated winter, or to avert crop damage. It is a dif­ over the past 4 years, and we expect other dif­ that a minimum of 40 "objects" must be ficult decision to make, and each situation ferences to manifest themselves as the found before estimates from strip counts are must be judged on its own merits. analysis continues. Indeed, the whole aspect valid. Perhaps with additional research, the A management scheme for cedar deer of overwinter mortality is a combination of strip count can become a valid technique for yards was outlined by Verme (1965), and for factors that are present throughout the year. estimating overwinter losses. mixed conifer swamps by Krefting and To this end, the Great lakes Deer Group Phillips (1970). Silvicultural practices to Committee stated that factors other than DEER POPULATION assure perpetration of cover in conjunction winter alone must be considered in determin­ MANAGEMENT with food-producing areas is stressed by ing carrying capacity (Krefting 1964). these investigators. Cutting practices that "Managing" the overwinter losses has pro­ assure a continuum of cover and food are to vided the battleground between wildlife be favored over a strict preservationist at­ SUMMARY managers and the populace since the deer titude towards a yarding area. Allowed to moved north. Winter feeding programs were mature, a yarding area will die in its ability to Physiological adaptation of white-tailed strongly advocated by a large segment of the winter deer. deer to the seasonal weather pattern is an an­ public to "save the deer". Wildlife managers While I have dwelt largely on what hap­ nual metabolic rhythm that assures survival denounced the program for many reasons; pens during the winter months, the rest of the through most years. Occasional severe for example, winter feeding doesn't solve the year cannot be ignored. What of the nutrient winters exceed these physiological limits, problem of overpopu lation, costs too much, content of the plants during the summer resulting in overwinter mortality. Fawns have can't get to all the deer, can't be a sustained months, and those left for the winter? higher metabolic demands during winter effort, causes concentration of deer with an Einarsen (1946) was able to predict over­ months, due to their continued demands for adverse effect on the remaining browse, pro­ winter survival based on the crude protein growth, therefore winter weather is a large motes the likelihood of spreading parasites content of key browse species in Washington. determinant in their survival.

51 LITERATURE CITED EBERHARDT, L.L. KREFTlNG, L.W., M.H. STENLUND, and R.K. SEEMEl 1968. A preliminary appraisal of line transects. 1966. Effect of simulated and natural deer ALDOUS, S.E. J. Wildl. Manage. 32(1):82-88. browsing on mountain maple. J. Wildl. 1952. Deer browse clipping study in the Lake Manage. 30(3):481-88. States Region. J. Wildl. Manage. 1978. Transect methods for population 16(4):401-9. studies. J. Wildl. Manage. 42(1):1-31. LEOPOlD, A.L., L.K. SOWLS, and D.L. SPENCER 1947. A survey of overpopulated deer ranges in ANDERSON, D.R., J.L. LAAKE, B.R. CRAIN, and K.P. EINARSEN, A.S. the United States. J. Wildl. Manage. BURNHAM 1946. Crude protein determination of deer 11(2):162-77. 1979. Guidelines for line transect sampling of food as an applied management tech­ biological populations. J. Wildl. nique. Trans. North Am. Wildl. Conf. MCCULLOUGH, D. Manage. 43(1 ):7()'78. 11:309-12. 1978. Case histories - the Tule elk (Cervus BARTLETT, I.H. ERICKSON, A.B., V.E. GUNVALSON, M.H. STENLUND, canadensis nannodes). Pages 173-84 in 1938. Whitetails. Presenting Michigan's deer D.W. BURCALOW, and L.H. BLANKENSHIP Alden Press, Oxford. Threatened deer. problem. Mich. Dep. Conserv. 64 pp. 1961. The white-tailed deer of Minnesota. Proc. IUCN Threatened Deer Progr. 434 Minn. Dep. Conserv. Tech. Bull. NO.5. 64 pp. 1950. Michigan deer. Mich. Dep. Conserv. 50 pp. pp. MECH, L.D., L.D. FRENZEl, and P.O. KARNS FLADER, S.L. 1971. The effect of snow conditions on the BARTLETT. I.H. and J.H. STEPHENSON 1974. Thinking like a mountain. Univ. Mo. vulnerability of white-tailed deer to wolf 1929. A preliminary survey of deeryards in the Press, Columbia. 284 pp. predation. Pages 51-59 in L.D. Mech and upper peninsula of Michigan. Mich. L.D. Frenzel, eds. Ecological studies of Acad. Sci., Arts, and Lett. 10:1-4. HARRIS, D. the timber wolf in northwestern Min­ 1945. Symptoms of malnutrition in deer. J. BURGOYNE, G.E., JR. and M.L. Moss nesota. U.S. For. Servo North Cent. For. Wildl. Manage. 9(4):319-22. 1978. Estimated winter deer losses in Michigan Exp. Stn. Res. Pap. No. NC-52. 62 pp. 1977-78. Mich. Dep. Nat. Resour. Surv. KABAT, C; N.E. COlllAS, and R.e. GUETTINGER MECH, L.D. and P.O. KARNS 1953. Some winter habits of white-tailed deer and Stat. Servo Rep. No. 171. 9 pp. 1977. Role of the timber wolf in a deer decline and the development of census methods BURNHAM, K.P., D.R. ANDERSON, and J.L. in the Superior National Forest. U.S. For. in the Flag Yard of northern Wisconsin. LAAKE Servo North Cent. For. Exp. Stn. Res. Pap. Wis. Conserv. Dep. Tech. Wildl. Bull. No. 1980. Estimation of density from line transect No. NC-148. 23 pp. 7. 32 pp. sampling of biological populations. MOEN, A.E. Wildl. Monogr. No. 72. 202 pp. KOHN, B. 1968. Surface temperatures and radiant heat COOK. D.v. and W. HAMILTON, JR. 1978. Winter severity index for central Wiscon­ loss from white-tailed deer. J. Wildl. 1942. Winter habits of white-tailed deer in sin. Wis. Dep. Nat. Resour. Final Rep. Manage. 32(2):338-44. north-central New York. J. Wildl. Job 210.9. Pittman-Robertson Proj. 1978. Seasonal changes in heart rates, activity, Manage. 6(4):287-91. W-141-R-14. 11 pp. metabolism and-forage intake of white­ DAHLBERG, B.L. and R.G. GUETTINGER tailed deer. J. Wildl. Manage. KREFTlNG, L.W. 42(4):715-38. 1956. The white-tailed deer in Wisconsin. Wis. 1941. Methods of increasing deer browse. J. Conserv. Dep. Tech. Wildt. Bull. No. 14. Wildl. Manage. 5(1):95-102. MOEN, A.E. and K.E. EVANS 282 pp. 1975. The effect of white-tailed deer and 1971. The distribution of energy in relation to DAVENPORT, L.A. snowshoe hare browsing on trees and snow cover in wildlife habitat. Pages 1939. Results of deer feeding experiments at shrubs in northern Minnesota. Univ. 147-62 in Snow and ice symposium. Iowa Cusino, Michigan. Trans. North Am. Minn. Agric. Exp. Stn. Tech. Bull. No. State Univ., Ames. 280 pp. Wildl. Conf. 4:268-74. 302.43 pp. MOOTY, J.J. ' DAVENPORT, L.A., W. SHAPTON, and W.e. GOWER KREFTlNG, L.W., ed. 1976. Year-round food habits of white-tailed 1944. A study of the carrying capacity of deer 1964. Research for deer management in the yards as determined by browse plots. deer in northern Minnesota. Minn. Dep. Great Lakes region. Great Lakes Deer Trans. North Am, Wildt. Conf. 9:144-49. Conserv. Wildt. Res, Q. Prog. Rep. Group. 73 pp. 36(1 ):11-36. DAVENPORT, L.A., D.F. SWITZENBERG, R.e. VAN ETTEN, and W.O. BURNETT KREFTlNG, L.W. and R.L. PHILLIPS ONTARIO DEER TECHNICAL COMMITTEE 1953. A study of deer yard carrying capacity 1970. Improving deer habitat in upper 1978. Ontario deer management - problems by controlled browsing. Trans. North Michigan by cutting mixed conifer and solutions. Ont. Minist. Nat. Resour. Am. Wildl. Conf. 18:581-96. swamps. J. For. 68(11):701-4. 260 pp. 52 PASSMORE, R.C. STEWART, R.R., R.R. MAcLENNAN, and J.D. KINNEAR Michigan white-tailed does. J. Wildl. 1953. Snow conditions in relation to big game 1976. Annual variation of plant phenological Manage. 34(4):863-69. in Ontario during winter of 1952-53. Ont. events and its theoretical relationship to VERME, L.J. Dep. Lands and For. Wildl. Res. Rep. No. energy balance in moose. North Am. 1965. Swamp conifer deeryards in northern 2.12 pp. Moose Conf. and Workshop. 12:1-30. Michigan: their ecology and manage­ PETRA BORG, W. and D.W. BURCALOW SWIFT, E. ment. J. For. 63:523-29. 1965. The white-tailed deer in Minnesota. 1946. A history of Wisconsin deer. Wis. Con­ 1968. An index of winter weather severity for Pages 13-58 in J.B. Moyle, ed. Big game servo Dep. Publ. No. 323. 96 pp. northern deer. J. Wildl. Manage. in Minnesota. Minn. Dep. Conserv. Tech. THOMPSON, C.B., J.B. HOLTER, H.H. HAYES, H. SILVER, 32:566-74. Bull. No.9. 231 pp. and W.E. URBAN 1977. Assessment of natal mortality in upper RYEL, L.A. 1973. Nutrition of white-tailed deer. I. Energy Michigan deer. J. Wildl. Manage. 1959. Technical report on the 1959 dead deer requirements of fawns. J. Wildl. Manage. 41(4):700-8. searches. Mich. Dep. Conserv. Game 37(~):301-11. WHITLOCK, S.c. and L.L. EBERHARDT Div. Rep. No. 2242. 43 pp. THOMPSON, D.R. 1956. Large scale dead deer surveys: methods, SILVER, H., N.F. ~OLOVAS, J.B. HOLTER, and H.H. HAYES 1979. Survey of dead deer, northern Wiscon­ results and management implications. 1969. Fasting metabolism of white-tailed deer. sin, spring 1979. Wis. Dep. Nat. Resour. Trans. North Am. Wildl. Conf. 21:555-66. J. Wildl. Manage. 33(3):49G-98. Bur. Res. Surv. Rep. 3 pp. SILVER, H., J,B. HOLTER, N.F. COLQVAS, and H.H. HAYES ULLREY, D.E., W.G. YOUATT, H.E. JOHNSON, L.D. For. 1971. Effect of falling temperatures in fasting B.L. SCHOEPKE, and W.T. MAGEE white-tailed deer. J. Wildl. Manage. 1970. Digestible and metabolizable energy re­ 35(1}:37~. quirements for winter maintenance of

53 Michigan DNR FAWN MORTALITY ESTIMATES IN FARMLAND DEER RANGE

Wayne R. Porath Missouri Department of Conservation 1110 College Avenue Columbia. MO 65201

Abstract: Reproduction and mortality are the plane of pregnant and lactating does and/or preda­ identification of causes and rates of mortality essential components of deer population tion, neither of which should constitute a signifi­ will at least facilitate more informed deci­ dynamics. Estimates of preseason mortality are im­ cant factor in farmland deer range. Application of sions on harvest management. This portant to deer harvest strategies. Fawn mortality principles of mortality assessment to farmland knowledge becomes increasingly important rates are most critical because they typically are deer populations is discussed. as biologists begin to use population modell­ higher than those of older age classes. Age com­ position in the harvest and preseason doe:fawn INTRODUCTION ing in their management programs. Age­ ratios may provide indexesto production and mor­ specific mortality rates are important com­ tality,but fail to identify temporal distribution, Minimum information needed to under­ ponents in most models; and although causes, or magnitude of fawn mortality. Mortality­ stand the dynamics of a population are pro­ knowledge of all age-specific rates is sensing transmitters attached to fawns allow con­ duction and mortality rates, rate of increase necessary, mortality rates for the age class tinuous observation and rapid retrieval of dead or decrease and number of animals within comprising the "weakest link" have the animals, thus are preferred over visual tagging that population. Knowledge of these greatest effect on simulations. systems. Fawn capture methods have included parameters is extremely elusive and wildlife Obtaining reliable mortality estimates has walk and horseback searches, use of gill nets, biologists usually are forced to make gross been and continues to be one of the most dif­ tracking with trained dogs, aerial surveillance with helicopter and fixed-wing aircraft, ground estimates or educated guesses about them. ficult problems facing deer managers. surveillance with spotting scope from fixed posi­ The quality of estimates may be dictated as Historically, data have been limited and have tion, monitoring radio-equipped does at parturi­ much by personnel and dollar availability, been derived largely from reproductive rates tion, and landowner contacts. During the period habitat and weather conditions as it is by and harvest sex ratios (Eberhardt 1960). 1977-79 in Missouri, 62 fawns averaging8 days old logic. Biases such as variation among age classes in were captured through landowner-cooperator con­ The parameter being addressed in this harvest vulnerability and inaccurate check tacts (34), walk searches (10), aerial surveillance paper is mortality between ages 0-6 months station sampling may render such data in­ with helicopter (7), doe observations(8), and track­ for quantifying recruitment into prehunt valid. Usually herd dynamics have been ing with trained dogs (3). Observed mortality rates populations of white-tailed deer (Odocoileus assessed and harvest recommendations made in 1978 were 35.5% and 53.3% during 0-90 and 91-180 day postpartum periods, respectively. In virginianus). Mortality patterns for most without identification of preseason fawn 1979, mortality was 28.6% within 90 days postcap­ mammals, including ungulates, typically mortality (Lang and Wood 1976). Some ture. No specific mortality sourcewas identified as follow a "U"-shaped pattern (Caughley 1966). assessment systems, however, require such most important in Missouri. White-tailed deer fawn Mortality is highest in the youngest age class, data. mortality rates have been reported between 8 and declines rapidly, then increases in older age The nutritional plane of does during 96% in other studies.High ratesin theseareashave classes. pregnancy has substantial effect on prenatal usually been associated with a low nutritional It is important to recognize how mortality and early postnatal mortality rates. Does ex­ patterns change in response to environmental perimentally fed a high or moderate quality fluctuations. The weakest' link in the life Porath, Wayne R. 1980. Fawn mortality estimates diet in winter plus a high quality spring ration in farmland deer range. Pages 55-63 in Ruth l. Hine history will be the first to be affected, either produced large, healthy fawns, and early and Susan Nehls, eds. White-tailed deer population positively or negatively, by environmental fawn mortality from nutritional deficiencies management in the north central states. Proc.1979 change. Reducing mortality rates of the was only 7%. Early postnatal mortality in­ Symp. North Cent. Sect. Wildl. Soc. 116 pp. youngest age class may not be possible, but creased to 90% when both winter and spring

55 diets of does were of low quality (Verme fetus:doe counts to fall fawn:doe counts for differential observability must be 1962). Low levels of protein in the diet of made via helicopters. Mortality estimates established for estimating net production penned, pregnant females had more effect following a hard winter were 64% and a mild through ratio counts. on early postnatal mortality of fawns than on winter were 39%. Chronology of California production of fawns in Missouri. Forty-two mule deer (0. hemionus califomicus) mortali­ HARVEST SEX-AGE STRUCTURE percent of fawns produced by females on a ty was estimated by field observation of 7% protein diet died of malnutrition within a changes in sex and age composition, and by Harvest sex and age data collected at few days after birth, while none on a 13% noting temporal pattern of fawn remains in checking stations and life tables derived from protein control ration died (Murphy and scats (Salwasser 1974, Salwasser et al. these data have commonly been used as Coats 1966). Weekly patterns of black-tailed 1978). O'Pezio (1978) collared females in descriptors of population phenomena. deer (0. hemionus columbianus) fawn re­ winter and recorded observations of marked Special emphasis has concentrated on mains in coyote scats supported the females and their subsequent families to measurement of those parameters which hypothesis that early fawn loss is a function estimate mortality of fawns. identify the success or failure of a particular of inadequate nutrition of the pregnant doe Fall fawn:doe ratios do not identify causes harvest strategy as a tool for population con­ (Salwasser 1974). Verme (1977) described a or indicate temporal distribution of fawn trol. method of assessingthe annual magnitude of mortality, important considerations in Methods of estimating total populations natal mortality in white-tailed deer due to in­ understanding herd dynamics. Also, obser­ from sex, age and kill data were described by adequate maternal nutrition. Based on birth vability of fawns was significantly less than Eberhardt (1960). These methods, or slight weights predicted from indexes of road-killed does during fall months in Virginia (Downing modifications of them, are currently used by pregnant does, early postnatal mortality et al. 1977). December was the only month biologists in several midwestern states to varied from 10% following years with low when observability was equal, but fawns had describe herd dynamics and formulate winter severity indexes to 70% following shed their spotted coats by then and ages allowable harvest figures (Table 1). Fawn winters with high severity indexes. Thus, were not discernible. Observed fall fawn:doe mortality is important to these methods only winter and spring nutritional planes of preg­ ratios, therefore, would be below the real as it affects net production indexes, while nant females may have a pronounced effect values and would bias prehunt fawn mortali­ adult buck populations, adult sex ratios, and on early postpartum mortality. This mortality ty estimates. fawn:doe ratios are the more critical factor must be considered in design of fawn Fall fawn:doe ratio counts are used to parameters of the system (Creed et aI. 1978). mortality studies. assess recruitment in several midwestern Lang and Wood (1976) indicated that equa­ Three types of data have been used to states. Daytime fawn:doe ratio counts were tions based on harvest reports, sex ratios, at­ assess fawn recruitment into prehunt popula­ not significantly different from spotlight ratio trition rates, and recruitment rates adequate­ tions - fall sex and age ratio counts, sex-age counts in Wisconsin (McCaffery pers. comm. ly described deer population dynamics in structure of the harvest, and fawn capture­ 1979) and were used to estimate an annual Pennsylvania. Weaknesses of this population radio tagging studies. Discussion of the ap­ net production index, which is incorporated estimation technique were described as plicability of each to farmland deer popula­ into the sex-age-kill method of herd assess­ underestimation of mortality rates, and tions follows. ment (Rusch 1978). Research is underway in failure to delineate specific causes of mor­ South Dakota to evaluate 2 methods of tality. FALL SEX AND AGE fawn:doe ratio counts-random daylight The usual life table unfortunately lacks RATIO COUNTS counts and spotlight counts (Rice 1978). Ran­ reproductive data and assumes births dom daylight counts appear to represent ac­ balance deaths and all rates are constant Fall sex-age ratio counts have been used tual conditions and are preferred over over the history of the population tabled widely within the range of white-tailed deer. spotlight counts. Random daytime ratio (Eberhart 1969). More importantly, harvest Typically the ratios have been based on counts are also used in the fall in Nebraska, sex and age data do not describe the spotlight counts, counts along established Michigan, Ohio, and Indiana (Table 1). magnitude, causes, or temporal distribution survey routes, or random daytime observa­ The reliability of fall fawn:doe ratio counts of prehunt fawn mortality. If knowledge of tions. There is no doubt that fawn:doe ratios is a function of ability to correctly identify those parameters is necessary in a particular decline through time and by fall are substan­ and classify animals, sufficient numbers of population analysis procedure, whether tially lower than spring production estimates. observations from a representative sample of simulation is used or not, other methods must Hall (1972) estimated mortality at Camp the population, and differential observability be used to obtain the information. While the Wainwright, Alberta, by comparing spring of sex-age classes of deer. Correction factors importance of harvest sex and age data can­ 56 TABLE 1. Status of fawn mortality assessment in midwestern states. not be discounted, their applicability in quan­ tifying prehunt fawn losses is limited. Need for Techniques Prehunt Potentially Mortality Currently Mortality Causes of Applicable State and Source> Estimates Used Rate (%) Mortality Techniques FAWN CAPTURE-RADIO Illinois Modelling Harvest sex-age Unknown Variety Fawn capture ­ TAGGING (F. Loomis) ratios radio tracking Radio telemetry may prove to be the best Indiana Modelling, Daytime Unknown Variety system for assessing fawn mortality in 0. Olson} fawn:doe farmland deer populations. Causes, rates, ratio counts and temporal distribution of mortality can be Iowa Modelling Harvest sex-age Unknown Variety, Fawn capture ­ identified by continually monitoring in­ (l. Gladfelter) ratios proportions radio tracking strumented fawns. Fawn capture-tagging unknown studies only recently have been attempted on deer in the Midwest and elsewhere. There are Michigan Modelling, Daytime 20-30 Variety. Field collection several reasons for this, but primary causes 0. Vogt and regional fawn:doe proportions of lactating are difficulties encountered in capturing an R. Aho} life tables, ratio counts, unknown does, fawn antlerless track counts capture, radio adequate sample of fawns and the absence quotas tracking of reliable tagging systems. Precision biotelemetry equipment with mortality­ Minnesota Assess turn- Harvest sex-age 10 Variety, Sex-age-kill, sensing components is now available and has (J. ludwig) over, popula- ratios proportions fawn:doe ratio been adapted for use on young ungulates tion trends, unknown counts, fawn (Cook et al. 1967, long 1977, Masters 1978). antlerless capture - radio quotas tracking Missouri Modelling Harvest sex-age 35 Variety Fawn capture ­ Fawn Capture Methods (W. Porath) ratios, fawn radio tracking capture - Several methods have been employed to radio tracking capture white-tailed and mule deer fawns (Table 2). Some considerations in selecting Daytime Unknown Variety, Production Nebraska Modelling, methodology include deer density, terrain, (K. Menzel) allowable fawn:doe assume indexes vegetative cover, availability of personnel; harvest ratio counts predation highest and land ownership. However, in any method it is important to consider abandonment­ Ohio Modelling Daytime 20 Variety Fawn:doe related mortality. Handling fawns may cause (R. Stoll) (planned) fawn:doe ratio counts abandonment, especially if it occurs before ratio counts, parental bonds are established (White et al. harvest sex-age ratios 1972). Hesitant, but eventual, acceptance of radio-collared and ear-tagged mule deer South Dakota Calculate Daytime 35 Variety, Fawn:doe newborn fawns has been reported (Goldberg (L. Rice) allowable fawn:doe predation ratio counts and Haas 1978). Mortality related to capture harvest ratio counts. may be high and tagging can be minimized by following harvest sex-age certain guidelines - allowing time for mater­ ratios nal bonds to develop (about 2 days postpar­ Wisconsin None, use Daytime Unknown Early post- Net production tum), handling fawns as little as possible, (K. McCaffery and sex-age-k ill fawn:doe natal, winter index as com- wearing sterile gloves, and avoiding con­ W. Creed) ratio counts, stress on doe ponent of spicuous markers. harvest sex-age sex-age-kill One of the first capture techniques ratios described was that by Downing and >Personsresponding to questionnaire sent July 5, 1979. Responses on file at Missouri Department McGinnes (1969) who found that searching by of Conservation, Columbia, MO 65201. vehicle followed by a noisy, fast-run ap­ 51 TABLE 2. Capture and marking techniques used in fawn mortality studies. proach to observed fawns was successful in State Deer Capture Marking pastureland with occasional wooded areas. (Reference) Species Technique Technique They suggested concentrating the search for Virginia -Bradford Army Depot w.t." Vehicle search followed ear tag and a single doe, not leaving the vehicle until the (McGinnes and Downing 1969) by tracking with trained streamer, fawn is sighted, and using a noisy, fast ap­ dogs and fast, noisy tattoo proach to encourage the fawn to drop. chase on foot Night searches aided by spotlights and dip Texas -Welder Wildlife Refuge w.t. Same as Virginia Radio nets have been used in southern (Carroll and (Cook et al, 1971) Brown 1977) and eastern Texas (Veteto and Hart 1976). Sites in southern Texas were gen­ Texas -Lavaca and Gonzales cos. w.t. Night searchwith spotlights Radio tly rolling coastal prairies with timber adja­ (Carroll and Brown 1977) and nets cent to major creeks and gently sloping Texas -eastern w.t. Night - vehicles with Radio upland with scattered trees and wooded bot­ (Veteto and Hart 1976) spotlights tom lands. Eastern Texas habitat was primari­ ly post-oak (Quercus stellata) savanna. Oklahoma -Cookson Hills w.t. Walk and horseback searches Radio (Bolte et al, 1970, Logan 1972) with throw net and V,-mile Understanding and concentrating on the gill net behavior of the doe can facilitate capture of hidden young fawns (White et al. 1972). Oklahoma -Wichita Mts. w.t. Ground surveillance with Radio and Observation platforms were used to study (Garner et al. 1976, spotting scope from fixed ear streamer mannerisms of the doe to alert observers to Bartush 1978) position. night search with possible locations of fawns. The most helpful spotlight, and aerial surveillance with helicopter followed by fast. cues of the doe to locate fawns were long noisy chase on foot solitary walk, close approach to the fawn, searching, nursing, watching, and aggression. New York -Seneca Army Depot w.t. Tranquilize Observation This observation technique has been used (O'Pezio 1978) of family successfully in Colorado (Anderson 1975), groups, radio Oklahoma (Garner et al. 1976, Bartush 1978), Arkansas -Svlarnore Forest w.t. Walk search; 100+ persons Radio Oregon (Trainer 1975), Missouri (Bryan 1980), enclosure and Montana (Dood 1978). (Cartwright and Rogers 1977) Searches on foot or horseback have been New Jersey -west central w.t. Walk search with 2-4 man Ear tag and used by several investigators. Phenomonal (Lund 1975) teams, throw net and recording streamer success with this technique was reported by of fawn bleating Lund (1975) in New Jersey. During a 5-year period, 496 fawns were captured by 2-4 man Missouri -east central w.t. Walk search, aerial Radio teams walking hay meadows. Density (Bryan 1980) surveillance with helicopter, landowner-cooperator contacts, estimates were not presented, but either-sex and tracking with trained dogs harvest was 4.74 per krn". Capture success in­ creased each year and was 3.1 man-hours per Oregon -Steens Mt. mule" Ground surveillance with Radio captured fawn during the final year of the (Trainer 1975) spotting scope from fixed position, study. A recorded bleating of a fawn was and tracking with trained dogs sometimes played before searching a field in Colorado -north central mule Ground surveillance with Radio an attempt to trick does into giving away the (Anderson 1975) spotting scope from fixed position locations of their fawns. Occasionally, a doe would emerge from cover and walk toward Montana -Missouri River Breaks mule Ground surveillance with Radio, ear (Dood 1978) spotting scope from fixed position, streamer its hidden young. Lund (1975) suggested fur­ and aerial surveillance with fixed­ ther experimentation with high quality re­ wing aircraft cording and amplifying equipment. , White-tailed deer. Others who have used walk searches in­ clude Bolte et al. (1970) and Logan (1972), " Mule deer. 58 both on Cookson Hills Wildlife Refuge in (5%) were captured with the aid of a trained Several collar designs have been described Oklahoma. Horseback riders and persons labrador retriever (Bryan 1980). for use on fawns. The "Oregon-Trainer" type afoot directed fawns into a 0.8 km long, 1.2 m (Trainer 1975) has been used extensively in high gill net. Walk searches using up to 15 western states and was used in Missouri. persons were used during the period, 1977-79, Fawn Tagging Methods Transmitters and batteries were mounted on to capture 10 of 62 fawns marked on a an oval-shaped collar with inside dimensions 131-km2 study area in Missouri (Bryan 1980). Choice of appropriate tagging methods of 8.9 by 10.2 cm formed by preheating a 42.3 Effort required for this method was high depends on research objectives. Visible by 1.3 by 0.2 cm strip of Kydex plastic (Rohm (34-176 hours/capture). Small crews of 2-4 markers would be preferred for observation and Haas, Philadelphia, Penn.*). Growth of persons have been most effective. Deer den­ of family groups to document changing fawn necks was accommodated by 1.3 cm sities should be very high for walk searches to fawn:doe ratios. Ear tags, ear streamers, ear strips of polyurethane foam glued to the in­ be effective in farmland deer range. streamers combined with tattoos, and leg side of the collar. A strip of amber latex tub­ Aerial search and surveillance with air­ streamers have been used as permanent, ing (4.76 mm inside diameter and 2.38 mm ground communications have been suc­ visual tagging systems (Queal and wall thickness) was anchored near each end cessfully used to locate and capture fawns in Hlavachick 1968, Downing and McGinnes of the collar to fasten the ends together. Ex­ Oklahoma (Bartush 1978), Montana (Dood 1969, Lund 1975, Bartush 1978, Dood 1978). posure to sunlight caused the tubing to rot 1978), and Missouri (Bryan 1980). Both fixed­ These techniques have the disadvantage of allowing the collar to drop after about 9 wing aircraft and helicopters have been used. not allowing continuous monitoring of months. Transmitters and batteries were en­ Helicopters are more versatile, allow for con­ marked animals, thus removing the possibili­ capsulated in dental acrylic or a urethane tinuous surveillance, and can be used to ty for determination of time and cause of resin to provide protection from shock and guide running fawns toward strategically mortality. Ear streamers have been shown to moisture. The completed transmitter collar placed ground personnel. Fawns up to 2 increase mortality by predation (White et al. weighed 220 g. weeks of age typically occupy grass or hay 1972) possibly because of accentuated, more This collar design with some modifications fields and field border cover types in visible ear movements (Queal and is now available commercially (Telonics, farmland range and are visible from an Hlavachick 1968), although mortality was not Mesa, Arizona*). The transmitter and lithium altitude of 25-75 m. It is important to note increased due to attachment of ear streamers batteries are hermetically sealed in a metal that some negative reaction by private land­ on fawns in Virginia (Downing and McGinnes housing. Operational life is 10 to 20 months owners occurred in Missouri because of their 1969). depending on pulse rate chosen. Typical feelings about cost (too expensive) and worry Use of radio telemetry to identify early pulse rates are 60-80 pulses per minute (ppm) of frightening livestock or flattening hay or mortality of white-tailed deer fawns on during the "alive period" and 120-160 ppm cereal grain with downdraft. All landowners Welder Wildlife Refuge in southern Texas during the "dead period". The mortality sen­ in an area should be contacted prior to was described by Cook et al. (1967). Expand­ sor is a mercury switch which responds to helicopter usage. able collars containing mortality sensing, movement. The sensor can be calibrated to Cooperating landowners were extremely lithium-battery-powered transmitters were switch to the "dead period" according to the helpful in locating and capturing fawns in successfully used by Flock et al. (1975) on user's specifications, usually 3-6 hours Missouri. All land within the study area was in mule deer fawns. Telemetry has provedto be following cessation of movement. The private ownership, and approximately 50% a very useful tool in identifying causes and transmitter collar weighs 220-30 g. was comprised of agricultural crops, with hay rates of prehunting season fawn mortality Radio-tracking of instrumented fawns can being a primary component. Cooperating because fawns can be continuously be accomplished with either portable or base­ landowners were provided nylon laundry monitored. It has been employed for white­ station receivers, the former preferred bags and were asked to capture fawns, place tailed deer in Texas (Cook et al. 1971, Veteto because of maneuverability. Available an­ the fawn in a bag, lay the bag in a shady and Hart 1976, Carroll and Brown 1977), tenna systems include variable-element place and contact study personnel. Of 62 Oklahoma (Bolte et al. 1970, Logan 1972, hand-held vagi, hand-held loop, vehicle-roof fawns averaging 8 days old captured between Garner et al. 1976, Bartush 1978), Arkansas mounted, or fixed-site designs. Fawn bearings 1977-79, 34 (55%) were caught as a direct (Cartwright and Rogers 1977), and Missouri are determined by the strongest signal and/or result of cooperating landowners. Walking (Bryan 1980). Mule deer fawns have been crews captured 10 (16%), helicopter radio-tagged in Oregon (Trainer 1975), Col­ "Mention of trade names or manufacturer names surveillance resulted in 7 (11%) captures, 8 orado (Anderson 1975, Anderson and Bowden does not constitute endorsement by the Missouri (13%) by observing postpartum does, and 3 1975), and Montana (Dood 1978). Department of Conservation. 59 the null-average methods. An excellent TABLE 3. Percentage estimates and primary causes of early mortality of deer fawns. discussion of bias and sampling error in­ volved in telemetric location fixing by Days of triangulation has been presented by Springer location Deer Percent Primary Causes Sample Observa­ (Reference) Species Mortality of Mortality Size tion (1979). Ability to obtain an accurate location Virginia -Bradford Army Depot w.l.* 8 258 60 is important in rapid retrieval of dead fawns. (McGinnes & Downing 1969) A recent innovation in mortality research has been the use of heat-sensitive, vaginally Texas -Welder Wildl. Ref w.l. 72 Coyote predation, 81 60 (Cook et al. 1971) sta rvation/disease implanted transmitters. (B. lange, New Mex­ ico Department of Game and Fish, las Texas -lavaca and Gonzales w.l. 47 Coyote preda tion, 120 60 Cruces, pers. comm. 1979). Radios are in­ cos. starva tion/disease (Carroll and Brown 1977) serted in the vagina of pregnant females and retained in position by suturing the lip of the Texas -eastern w.l. 37 Coyote predation, 65 90 vulva. When parturition occurs, the transmit­ (Veteto and Hart 1976) starva tion/disease ter is expelled with the fawn and the pulse Oklahoma -Cookson Hills w.t 34 18 frequency changes. Continuous monitoring Wildlife Refuge (1968-69) (Bolte et al. 1970) of females must occur near parturition so newborn fawns can be found before they Oklahoma -Cookson Hills w.l. 18-64 Blood loss and 33 60 move from the birth site. Approximately 200 Wildlife Refuge (1970-72) infection from (logan 19721 ticks vaginal transmitters have been implanted in mule deer and elk in New Mexico. Results of Oklahoma -Wichita Mts. w.l. 88 Coyote predation 35 365+ this work will soon be published. (1974-75) (Garner et al 1976)

Oklahoma -Wichita Mts. w.l. 90 Coyote predation 48 49 FAWN MORTALITY ESTIMATES (197&-77) (Bartush 1978) Published data have shown great variabili­ ty in fawn mortality estimates, ranging from 8 New York -Seneca Army Depot w.l. 24 Early postnatal 46" 90 to 96% (Table 3). Primary causes were preda­ (O'Pezio 1978) tion by coyotes, starvation, disease, winter Arkansas -Sylamore w.l. 100 Coyote predation 23 Until nutritional stress on pregnant females and enclosure death blood loss and subsequent infection due to (Cartwright and Rogers 1977) high infestation of ticks. less important Missouri -east central w.t 33 Variety 39 90 causes of mortality of instrumented fawns in­ (Bryan 1980) cluded dog or predation, highway Missouri -penned deer w.l. 0-42 Nutrition stress 12 1 Until kills. illegal and hunter kills and accidents (Murphy and Coats 1966) on doe death caused by fences or other factors. During the Michigan -penned deer w.t. 7-90 Winter/nutrition 215 Until summers of 1977 and 1978 in Missouri, 29 (Verme 1962) stress on doe death nontagged dead fawns were found by land­ South Dakota w.t. 35 Predation 2 2 owners and search crews, of which 20 were (Rice 1978) suspected killed by hay mowing machines. This may be 3 a significant mortality factor in areas where Oregon -Steens Ml. mule 27 Coyote predation, 106 90 (Trainer 1975) starv ation/disease hay crops are an important component of agricultural operations. Colorado -north central mule 43 Unknown 23 127 Predation by coyotes has been the primary (Anderson 1975) cause of white-tailed deer fawn deaths in Montana -Missouri River mule 34 Coyote predation 32 90 several areas, including Texas, Oklahoma, Breaks and Arkansas (Table 3). This type of mortality [Dood 1978) typically occurred within 30 days postpar­ 'white-tailed deer. 1 does. 3 mule deer. tum. The predator/prey relationships between "marked adult does 2 fall fawn:doe ratio counts.

60 deer and coyotes are complex and not easily deer on a Wichita Mountains, Oklahoma Fall fawn:doe ratio counts when coupled interpreted. Knowlton (1976) described the study area. Losses of mule deer fawns were with spring productivity data provide net pro­ relationship of weather factors and confined to about 45 days postpartum in duction indexes and the magnitude of fawn deer/coyote interactions in western ranges. Oregon (Trainer 1975). Since the majority of losses. However, biases of observability Following years of below average precipita­ deaths typically occur early postpartum, it is among various sex-age groups must be over­ tion, conception of fawns was delayed, and important to capture and instrument fawns as come for accurate estimates. Also, it is im­ parturition occurred midway between spring soon as possible following development of possible to determine causes or temporal and summer fruiting periods. In the absence maternal bonds. distribution of mortalities with ratio counts. of fruit as a food source, coyotes actively It may be important to categorize causes Legal harvest sex-age data have similar defi­ searched for fawns which were likely to be of death as predation-excluded or predation­ ciencies. weaker than in years of adequate rainfall. involved (Cook et al. 1971). White (1973) Radio telemetry has provided a method for Conversely., when rainfall was abundant, a described remains of deer fawns killed by daily observation of young fawns and rapid major portion of parturition occurred while coyotes and Garner et al. (1976) provided a detection of mortality. Lightweight, durable coyotes were feeding on fruits, and fawns detailed description of criteria used to deter­ transmitters with mortality-sensing com­ were healthier and more viable. mine predator species in predator-involved ponents and life expectancies of 1 year or Nutritional stresses placed on females in mortalities. Necropsies of carcasses should more have been developed and are available northern portions of the Midwest during be performed to aid in establishing cause of commercially. Scanning receivers are severe winters (Verme 1977) may result in death, especially when disease, parasites or available which enable simultaneous mortality of malnourished fawns, and these starvation are suspected. Salmonellosis monitoring of many individuals. losses may be accentuated by coyotes or (Salmonella sp.) was diagnosed as the cause Capture techniques which have potential other predators taking advantage of these of mortality of one fawn in Texas (Cook et al. in farmland areas include walk searches, weakened animals. This topic merits addi­ 1971). Rectal swabs made at time of capture aerial surveillance, ground surveillance with tional investigation for farmland deer, can be helpful in identifying the role of a spotting scope from a fixed position, and although it is unlikely that most farmland Salmonella or other bacteria in fawn mortali­ landowner contacts. Vaginally implanted deer are often subjected to food shortages in ty. Blood loss and infection from the feeding transmitters for monitoring activities of does winter or spring. of lone star ticks (Amblyomma americanum) nearing parturition have not been used on No particular cause of death was outstand­ were associated with the deaths of 71% of in­ white-tailed deer, but they have proved suc­ ing in Missouri. Of 32 fawns averaging 8 days strumented fawns in eastern Oklahoma cessful in mule deer and elk. Plans are under­ of age at capture in 1978, mortality rates (Logan 1972). way for use of these devices in white-tailed were 35.5 and 53.3% during 0-90 and 91·180 deer in Missouri and Illinois (Loomis pers. days postpartum, respectively. In 1979, mor­ DISCUSSION comm. 1979). tality totaled 28.6% of 7 radio-equipped deer Mortality estimates for fawns 1-90 days of monitored 90 days postpartum (Bryan 1980). Although fawn mortality research is in its age have been reported between 8 and 96%. Causes of death within 30 days postpartum infancy in the Midwest, sufficient work has High incidences of mortality have usually were primarily starvation and bobcat preda­ been done elsewhere to suggest needs of been caused by either early postnatal losses tion; those after 30 days were exclusively farmland deer population management. The associated with winter nutritional stress on man-related (highway, illegal, hunter, dog, first and most important consideration is pregnant females in northern regions or and fence kills). Although man-related mor­ whether quantification of early fawn mortali­ coyote mortality in southern and western tality has not been a primary source in other ty is critical to understanding herd dynamics areas. A study is planned in Illinois to in­ mortality studies, mortality has been highest and whether such information is needed to vestigate the contribution of deer, especially during the first 30 days postpartum. At develop harvest strategies. This information fawns, to the overall diet of coyotes (Loomis Welder Wildlife Refuge in southern Texas, is of minor importance to some analysis pers. comm. 1979). A population study of 93% of the mortality, which was primarily systems, notably the sex-age-kill method. bobcats is in progress in Missouri which may coyote predation, occurred within 30 days Energies would be better directed toward the better define the role of this predator in deer (Cook et al. 1971). Garner et al. (1976) weak links of these systems such as estima­ mortality (Hamilton et al. 1979). Fritts and reported that 83% of the mortality occurred tion of the proportion of the adult buck Sealander (1978) suggested that bobcat within 60 days postpartum in 1974-75 while cohort not recovered through legal harvest predation of fawns may be common in areas Bartush (1978) reported 91% of the total mor­ and observed net production (McCaffery of high deer and bobcat densities. The com­ tality occurred within 30 days in 1976-77 for pers. comm. 1979). plexity of predator/prey relationships as 61 described by Knowlton (1976) indicates that LITERATURE CITED DOOD, A.R. subtle factors are involved and these factors 1978. Summer movements, habitat use, and are not easily quantified. ANDERSON, A.E. mortality of mule deer fawns in the Wide variations in fawn mortality 1975. Evaluation of radio telemetry. Colo. Div. Missouri River Breaks, Montana. Mont. estimates have been reported within the Wildl. Perf. Rep. Work Plan No. 19. lob Dep. Fish and Game Perf. Rep. Study No. range of white-tailed deer. Although No.1. Pittman-Robertson Pro]. NG-47-1. Job No.9. Pittman-Robertson W-38-R-29. pp. 413-73. Proj. W-12D-R-8,9. 55 pp. preliminary work has been done in Missouri, studies should be done elsewhere within ANDERSON, A.E. and D.C. BOWDEN DOWNING, R.L. and B.S. MCGINNES farmland deer range to identify probable 1975. Mule deer-coyote interactions. Pages 1969. Capturing and marking white-tailed deer regional differences in fawn mortality. 15-16 in H.A. Swope, ed. Colo. Div. fawns. J. Wildl. Manage. 33(3):711-14. Wildl. Colo. Game Res. Rev. 73 pp. Reducing or eliminating major causes of DOWNING, R.L., E.D. MICHAEL, and R.I. Poux,IR. mortality in young fawns would not be prac­ BARTUSH. W.s. 1977. Accuracy of sex and age ratio counts of tical. However, identification of mortality 1978. Mortality of white-tailed deer fawns in white-tailed deer. I. Wildl. Manage. rates caused by those factors will fill a signifi­ the Wichita Mountains, Comanche 41(4):709-14. County, Oklahoma, Part II. Okla. State cant gap in the knowledge of deer herd EBERHARDT, L. Univ., Stillwater. Ms Thesis. 161 pp. dynamics and will improve predictive 1960. Estimation of vital characteristics of capabilities of deer herd managers. BOLTE, I.R., I.A. HAIR, and I. FLETCHER Michigan deer herds. Mich. Dep. Con­ 1970. White-tailed deer mortality following servo Game Div, Rep. No. 2282.192 pp. tissue destruction induced by lone star 1969. Population analysis. Pages 457-95in R.H. Acknowledgments. This study was ticks. I. Wildl. Manage. 34(3):546-52. Giles, Ir.. ed. Wildlife management financed in part by Federal Aid in Wildlife Restora­ BRYAN, D.A. techniques. The Wild I. Soc., tion Act, under Pittman-Robertson Project W-13-R. 1980. White-tailed deer fawn mortality, home Washington, D.C. 623 pp. The assistance of deer researchers in the Midwest range, and habitat utilization in east cen­ Ftocx.A., R. HAINES, and T. CAVIN in the compiling of data for this manuscript was tral Missouri. Univ. Mo., Columbia. MS 1975. Fawn telemetry and collar development. greatly appreciated. Thesis. 45 pp. Nev. Dep. Fish and Game Perf. Rep. CARROLL, B.K. and D.L. BROWN Study Rl-C, Job NO.1. Pittman-Robertson 1977. Factors affecting neonatal fawn survival Proj. W-48-R-6. 7 pp. in southern-central Texas. J. Wildl. FRITTS, S.H. and I.A. SEALANDER Manage. 41(1):63-69. 1978. Diets of bobcats in Arkansas with special CARTWRIGHT, M. and M.I. ROGERS reference to age and sex differences. I. 1977. Svlarnore deer mortality study. Ark. Wildl. Manage. 42(3):533-39. Game and Fish Perf. Rep. Study No.3. GARNER, G.W., J.A. MORRISON, and l.C. LEWIS Pittman-Robertson Proj. W-53. pp. 7-28. 1976. Mortality of white-tailed deer fawns in CAUGHLEY, G. the Wichita Mountains, Oklahoma. 1966. Mortality patterns in mammals. Ecology Proc. Conf. Southeast. Assoc. Fish and 47(6):906-18. Wildl. Agencies 30:493-506.

COOK, R.S., M. WHITE. D.O. TRAINER, and W.e. GOLDBERG, 1.5.and W. HAAS GLAZENER 1978. Interactions between mule deer dams 1967. Radiotelemetry for fawn mortality and their radio-collared and unmarked studies. Wildl. Dis. Assoc. Bull. fawns. I. Wildl. Manage. 42(2):422-25. 3(4):160-65. HALL, W.K. 1971. Mortality of young white-tailed deer 1972. Natality and mortality of white-tailed fawns in south Texas. J. Wildi. Manage. deer in Camp Wainright. Alberta Fish 35(1 ):47-56. andWildl. Div. Proj. No. W-R-72. lob No. A-4. 27 pp. CREED, W.A., B.E. KOHN, and K.R. MCCAffERY 1978. Deer population measurements in HAMILTON, D.A., D.W. ERICKSON, and F.B. SAMSON management units. Wis. Dep. Nat. 1979. Bobcat studies in Missouri. Mo. Conserv. Resour. Perf. Rep. Study No. 209. Dep. Perf. Rep. Endangered Species Pro], Pittman-Robertson W-141-R-13. 25 pp. sE-1. Study NO.3. 6 pp,

62 KNOWLTON, F.F. so, Game, Fish, and Parks Dep Perf. 1976. Potential influence of coyotes on mule Rep. Study No. 7516. Pittman-Robertson deer populations. Pages 111-18 in Proj W-75-R-20. 17 pp. Workman, G.W. and J.B. Low, eds. Mule deer decline in the west - a symposium. RUSCH, AJ. Utah State Univ. 134 pp. 1978, Summer observations of deer. 1978. Wis. Dep. Nat. Resour. Bur. Res. Surv. Rep 4 LANG, L.M. and G.W. WOOD pp. 1976. Manipulation of the Pennsylvania deer herd. Wildl. Soc. Bull. 4{4):15%5. SALWASSER. H. 1974. Coyote scats as an indicator of time of LOGAN, T. mortality in the North Kings deer herd, 1972. Study of white-tailed deer fawn mortali­ Calif. Fish and Game 62(2):84-87 ty on Cookson Hills deer refuge, eastern Oklahoma. Proc. Conf. Southeast. SALWASSER, H., S.A. HOLL. and G.A. ASHCRAFT Assoc. Game and Fish Comrn. 26:27-38. 1978. Fawn production and survival in the North Kings River deer herd Calif, Fish LONG, F.M., ed. and Game 64(1):38-52 1977. Proc. First Int. Conf. on Wildl. Biotelemetry. Laramie, Wyo, 159 pp. SPRINGER. JT 1979. Some sources of bias and sampling error LUND, R.C. in radio triangulation. J. Wildl. Manage 1975. The capture and marking of wild, 43(4)926-35, newborn white-tailed deer fawns. Pages :25-3-3 in D: DeCarli, ed. Trans. Northeast. TRAINER. C Sect. The Wildl. Soc. 213 pp. 1975.' Direct causes of mortality in mule deer fawns during summer and winter periods MASTERS, R. on Steens Mountain, Oregon -- a progress 1978. Deer trapping, marking and telemetry report. Proc. Conf West. Assoc. Game techniques. State Univ. N.Y., Newcomb. and Fish Comm. 12 pp. 72 pp. VERME. L.J. MCGINNES. B,S. and R.L. DOWNING 1962. Mortality of white-tailed deer fawns in 1969. Fawn mortality in a confined Virginia relation to nutrition. Proc. 1st Natl. deer herd. Proc. Conf. Southeast. Assoc. White-tailed Deer Dis. Svrnp 1:15-38. Game and Fish Comm. 23:188-91. 1977. Assessment of natal mortality in upper ~"'-'''''''''''\''\ MURPHY, D.A. and].A. COATS Michigan deer. J Wildl. Manage ~,-- ..\ 1966. Effects of dietary protein on deer. Trans. 41(4)700-8, \ - ,::' j /I(':;y, ,- o . ti\ I: \'-,~. North Am. Wildl. and Nat. Resour. ConI. VETETO, G. and R. HART ~,-~.,. ~~./~~,~'~:')')' ~ 31:129-39. ,."._S .. 1976. Factors affecting fawn survival. Tex. \ .-;; <--. , . ,,,' >r,. _~~.~,_ / O'PEZIO, J,P. Parks and Wildl. Dep Perf, Rep. Job No. 1",I~. . , tI:~ " _._-<"~<:::"'~ 1978. Mortality among white-tailed deer fawns 13. Pittman-Robertson PrOj. W-74-R-21 .»:"--. .;J,""-.... ~ '~l:~,•. t . on the Seneca Army Depot. N.Y, Fish and 19 pp. ',:.( ~ . / ...... -;;'f~.; Game J. 25(1):1-15. WHITE, M. "~, 1".,- .. s: I,.", QUEAL, L.M. and B.O, HLAVACHICK 1973, Description of remains of deer fawns --- 1968, A modified marking technique for young killed by coyotes, ). Mammal ungulates. J. Wild!. Manage. 54(1):291-92. 32(3):628-29. WHITE, M., F,F. KNOWLTON. and W.e. GLAZENER RICE,IA. 1972. Effects of dam-newborn fawn behavior 1978. Mortality rates of fawn age class in on capture and mortality, J. Wildl. South Dakota deer populations, 1977-78. Manage 36(3):897-906,

63 Wisconsin DN R ESTIMATING ILLEGAL KILL OF DEER

Kirk H. Beattie College of Natural Resources University of Wisconsin-Stevens Point Stevens Point, WI 54481 Cleveland J. Cowles and Robert H. Giles, Jr. Department of Fisheries and Wildlife Sciences Virginia Polytechnic Institute and State University Blacksburg, VA 24061

Abstract: Methods presented for estimating INTRODUCTION Public hunting in America is based on the in-season illegal deer kill involve: (1) extrapolating principle of e,g..ual opportunity. Many hunters statistics from dead deer surveys during and/or Illegal kill of North American deer do not harvest a deer in any 1 year, other following a hunting season, (2) making a simple (Odocoi/eus spp.) is of vital interest to hunters may legally harvest as many as 30 count of the number of illegally killed deer wildlife managers, deer hunters, and non­ deer/year (e.g., in Alabama). Although suc­ discovered by enforcement officers, (3) determin­ hunting observers of deer. Managers of the cess per hunter varies, each hunter has the ing an index to illegal deer kill based on the deer resource are charged with providing same legal opportunity to harvest the same number of violation reports received by enforce­ benefits to the American public via manage­ number of deer as other hunters. Kill of deer ment officers, and (4) using McCormick's com­ pliance rate technique. The latter method assumes ment of deer. Since many of the benefits out of season decreases the opportunity of in­ that total violations detected by wardens are derived from deer management accrue to season hunters to harvest deer because of a related to the total number of hunters checked in hunters, maintaining a harvestable surplus of reduced number of deer available, and the same way that total violations occurring are deer is normally a paramount objective of violates the principle of equal opportunity. related to the total number of persons hunting. The wildlife management agencies (subject to This results in hunter frustration, anger, and only known technique for estimated closed-season constraints such as big game crop damage, reduced quality of the overall hunting ex­ illegal deer kill is a violation simulation technique. deer-vehicle accidents, disease outbreaks, perience. The rationale is that the total number of violations etc.), Desired harvest levels for deer popula­ Out-of-season illegal kill of deer also ir­ for a study period is related to the total detected tions are established by estimating preseason ritates or disgusts nonhunting observers of violations in that time approximately the same way that violation simulations are related to total deer populations and subtracting from these deer through their knowledge that illegal kill simulations reported by enforcement personnel. acceptable postseason populations. By know­ is occurring, by reducing the number of deer Two major drawbacks to the use of the violation ing probable success rates for hunters, a fixed available for observation, by driving deer simulation technique are presented. An untested number of hunting licenses or permits may be away from easily observed areas near public procedure for estimating the number of individuals issued (or, as is often the case, season lengths roads, and, more importantly, by intruding on killing deer illegally during a specified period is may be regulated). Illegally killed deer may otherwise tranquil and safe experiences presented. result in the desired harvest being exceeded. afield. In 1975, 2 districts of the Shenandoah If the likely number of illegally killed deer is National Park in Virginia were closed to after­ overestimated in preseason computations of dark travel for 2 weeks to combat deer Beattie, Kirk H., Cleveland J. Cowles, and Robert H. Giles, Jr. 1980. Estimating illegal kill of deer. desired harvest, resource waste and poaching. The closure was extended to 3 Pages 65-71 in Ruth L Hine and Susan Nehls, eds. agricultural damage may result. Thus, lack of weeks in 1976. White-tailed deer population management in the good estimates of mortality from illegal Assessing the magnitude of loss due to il­ north central states. Proc. 1979 Symp. North Cent. causes may compromise scientific manage­ legal deer kill is important for at least 3 Sect. Wildl. Soc. 116 pp. ment of the deer resource. reasons: (1) to allow deer managers to incor­

65 porate loss figures into deer harvest about relative rates of violations between estimation procedure. The estimate obtained strategies; (2) to provide an indication of the regions and over time. This implies little need is obviously an underestimate of actual loss magn itude of the deer poaching "problem" for unbiased absolute estimates of illegal kill because many deer will have been removed to wildlife law enforcement divisions (and and more need for consistent application, and failure to discover all carcasses is prob­ subsequent measures of effectiveness of en­ over time and between regions, of able. l)nless the illegality of the discovered forcement effort); and (3) to provide the economical and less sophisticated indicators. carcasses is obvious (e.g., female in a bucks­ potential for transfer of an illegal deer kill Information thus obtained would be useful in only hunting season, decapitated carcass), estimate methodology to illegal kill of other enforcement decision making such as deter­ much depends on the inference of the game animals. mining optimal distributions of a usually observer. Finally, and most importantly, time, However, many of the techniques for fixed supply of manpower. manpower, and budgets are not sufficient to estimating illegal game kill require a sizable Since the techniques of estimating illegal allow the method to be employed on a commitment of time and money to obtain un­ harvest have been developed in response to statewide basis. biased statistics. Therefore, it is of primary management needs, it can be assumed that Another method of estimating illegal deer concern prior to embarking on any research the potential for use of the illegal kill kill also provides an underestimate of actual aimed at estimating illegal harvest to con­ estimate in management decisions is high. loss. This method involves simply counting sider: (1) the degree to which the accuracy However, this does not negate the need for the number of illegally killed deer discovered and precision of the illegal kill estimator is attention to the third consideration upon by enforcement officers. For example, compatible with the population estimator; (2) each planned application of an illegal kill Michigan conservation officers discovered the need to ascertain absolute or relative estimator. 330 deer killed illegally during the bow estimates of illegal kill; and (3) the potential season in Michigan in 1972 and 1,112 during for translation of research findings to ESTIMATING IN-SEASON the regular firearm season (Michigan Law En­ management decisions. forcement Division 1973). From the standpoint of the deer manager, ILLEGAL DEER KILL The third method, developed by Kaminsky the first consideration is of importance. If the One method involves systematically and Giles (1974), involved determining an in­ manager bases his estimate of total game searching a hunting area during and/or dex to illegal deer kill based on the number population on information of low accuracy following a hunting season and counting the of violation reports received by enforcement such as that obtained at public hearings, then number of deer known to be killed illegally or officers. In a survey of all Virginia game it would be wasteful to attempt an intensive thought to have been killed illegally. Hardin wardens in the fall of 1969 the number of ac­ field study to try to obtain an accurate or and Roseberry (1975) systematically searched tual and casual complaints of spotl ighting precise estimate of illegal harvest. 5,200 acres on Crab Orchard National received by each warden was recorded, The second consideration is of importance Wildlife Refuge (Illinois) following the deer although some of the spotlighting reports when contemplating estimation of illegal kill hunting season to assess unreported losses of were probably not related to deer poaching. from the standpoint of law enforcement. deer. They examined various circumstances Each of the 109 agents received an average Several of the more sophisticated approaches surrounding located dead deer and inferred of 3.94 actual spotlighting complaints per of estimating illegal kill have been employed intentional abandonment of deer (illegal) on week from 1 October to 31 December. Based by various state enforcement divisions once the basis of these circumstances. Their on these figures, the authors estimated that and never repeated. It could be asked why so estimates showed that at least 13 (20%) of 5,153 spotl ighting complaints occurred dur­ much interest is expressed by enforcement the 64 carcasses located had been inten­ ing this 12-week period. This period con­ officials in the absolute illegal kill since no tionally abandoned. One fawn had been hid­ tained an average of 86.3% of all spotlighting one knows yet how to relate empirically the den in a brush pile, 3 deer had been field­ arrests in the 3 years from 1968 to 1970 and benefits of enforcement manpower to the dressed, and 3 adult males had been the authors assumed that arrests were made costs of illegal kill. If it was known how many decapitated and their carcasses left in the in proportion to the total amount of illegal enforcement agents were required to effect a field. Hardin and Roseberry suggested that spotlighting activity occurring. During the re­ desired change in illegal kill, then an unbias­ the remaining 6 deer had been intentionally maining 40 weeks of the year, 13.7% of the ed estimate of absolute illegal kill would be abandoned because they had apparently sus­ total spotlighting arrests were made. By of value. However, the present state of tained instantly fatal wounds. assuming that the same proportion of the knowledge regarding manpower re­ The drawbacks accompanying this method 5,153 complaints would be reported during quirements in wildlife law enforcement is so of estimating illegal deer kill are probably the 40-week period, the authors estimated primitive that it would be wiser to learn more sufficient to preclude its use as a standard that 5,895 spotlighting violations occurred

66 during 1969. This figure was an overestimate An example of this compliance rate pro­ season deer violations by the compliance to the degree that duplication in reports of cedure is provided by the Michigan Law En­ rate method. It seems intuitive that those per­ spotlighting and reports of legitimate hunting forcement Division (1974). During the 1971-72 sons who know they are in violation of a deer (e.g., hunting) as spotlighting deer hunting season, Michigan conservation law or regulation will take steps to evade occurred. However, the figure is probably officers inspected 77,450 deer hunters, made detection by law enforcement agents. If severely underinflated because only a por­ 2,438 arrests, and issued 3,765 warnings. Mail violators are more evasive than the average tion of spotlighting violations were reported surveys for 1971 estimated there were person who unknowingly commits a deer (based on similar studies). Kaminsky and 605,740 big game hunters in Michigan and crime or a nonviolator, random inspection of Giles' procedure does not estimate the they spent a total of 4,520,340 days hunting. hunters will result in the violation estimate number of deer lost to illegal spotlighting ac­ The total number of deer law-related viola­ being an underestimate of the actual number tivity nor does it accurately estimate the tions was estimated by: of violations occurring. magnitude of spotlighting activity. However, Cowles et al. (1979) suggested that the under proper administration, the technique total following procedures are minimal re­ (4,520,340) (2,438 + 3,7(5) may be valuable as a relative index to the ex­ violations quirements for estimating compliance with tent of illegal spotlighting activity. 77,450 wildlife laws from arrest data: A fourth method of estimating in-season il­ (1) Comprehensive definition of the problem legal deer kill, developed by James total with sufficient justification of why com­ McCormick of the California Wildlife violations = 3&2,034 pliance should be estimated for a specific Resources Agency (McCormick 1968, 1970), resource user group. Not all of these violations were related to il­ has received the widest acclaim and, with (2) Specific definition of the resource user legal deer kill. Many were related to failure proper implementation, probably shows group whose behavior is being studied. to post a camp registration perm it. This would require a detailed analysis of more promise than other methods developed. Separating those violations related solely to The procedure assumes the proportionality: illegal deer kill would allow an uncon­ the distribution of the population of users being sampled in both time and area. taminated estimate of illegal deer kill. Cowles et al. (1979) listed 7 assumptions (3) Specific definition of behaviors which will be considered violations and the role of violations among for compliance estimates based on the ratio total violations those checked officer discretion. If warnings are issued, of detected violations to total resource users total hunters total hunters checked sampled: they should be consistently included or ex­ cluded from compliance estimation. Thus an estimate (,,) is possible: (1) Wildlife law enforcement agents contact (4) Design and implementation of sampling resource users at random. that will maintain randomness and (violations representativeness. total (total among (2) Resource users contacted are represen­ (5) Attention to changes in administrative hunters) those checked) tative of the entire user population being violations studied. policy, technology, or legislative action total hunters checked (3) Resource users commit a known average which would affect reports of arrests and number of violations per day in the field. inspections. (6) Concomitant studies of violators within ;~. (4) Contact of a specific user is not duplicated on the same day. each resource user type to establish an Since the above formula assumes that each (5) All wildlife law enforcement agents are estimate of the mean number of violations hunter hunts only one day of the year, "total consistent in defining user behaviors. expected per violator day. Self-report or hunters" is replaced by "total days of hunt­ (6) Wildlife law enforcement agent reports of spy-blind techniques may be applicable. ing" to adjust for the number of days spent in inspections or violations are accurate or, if (7) Familiarity with, and critical appraisal of, the field by all deer hunters: inaccurate, error does not change over surveys providing supporting data, such as time. hunter-use surveys. The analyst should be (violations (7) If needed for computational purposes, the aware of major changes in such surveys total (total days among estimated days-afield per user is accurate. and how they may affect his conclusions. of hunting) those checked) violations Assumption number two is probably the Use of the compliance rate procedure for total hunters checked most critical in attempts to estimate in- estimating illegal deer kill is restricted to

67 hunting seasons because the technique is The estimate is unbiased provided that the 15,825 + (0.1395 X 15,825) = 18,033. This based on random inspections of hunters. In­ product of Ma times C is 3 or 4 times as large estimate does not apply solely to deer spections made during nonhunting seasons as the estimate (I) and guarantees that the because Vilkitis also collected moose (Alces are not normally random and are instigated rediscovery sample (R) includes at least one alces) and black bear (Ursus americanus). The by probable cause (e.g., a person carrying a previously noted event. If Ma times C is not 3 low value of R suggests: (1) the error may be firearm in a hunting area). or 4 times larger than " the estimate is biased quite large, and (2) Poissonal or other downward by a factor equal to the probabili­ distributions of events and detections. ESTIMATING CLOSED-SEASON ty of getting no repeats or rediscoveries Vilkitis also simulated 49 night-hunting ILLEGAL DEER KILL (Robson and Regier 1964:216): violations from 15 September to 2 December 1970. None of the 49 simulations (C) resulted The only known technique for estimating 100 -(Ma+l)(C+l) in an arrest (R = 0). Based on 403 night­ closed-season illegal deer kill was developed e R hunting arrests made by agents during the by Vilkitis (1968) who studied closed-season period, night-hunting violations for the state big game poaching in Idaho and then In Maine, Vilkitis (1971) simulated 81 of Maine during the hunting season were retested his methods in Maine. First he de­ closed-season kills from 27 November 1970 estimated to be: fined: (1) violation as an offense against ex­ to 2 April 1971 using parts of 15 collected isting fish and game law; (2) arrest as an act of animals. Vilkitis dressed as a woods worker, 1= (403+1)(49+1)_1 = 20,199 seizing or taking into custody when enough drove a Scout, and used a .30-.30 rifle or (0+1) information is available to prosecute an 12-gauge shotgun. Shots were fired and alleged violator; (3) incident as a situation in evidence of blood, hair, hides, and drag trails The unbiased estimate of , was 27,780. which positive evidence of a violation is was left. Simulations were performed only Field detection of violations was almost iden­ available, but information is insufficient for during daylight hours in areas where big tical in both Idaho (1.1%) and Maine (1.2%). an arrest; and (4) violation simulation as an game was known to be concentrated. Vilkitis Thus, Vilkitis developed the following equa­ act performed by a specially authorized performed simulations in all enforcement tion for estimating illegal kill: agent of the state to duplicate, in every way divisions in Maine in relative proportion to possible, a violation. the mean number of convicted prosecutions Illegal kill = 122.9 X closed-season The violation simulation technique from 1962 to 1965. In Idaho, Vilkitis simply arrests employed the rationale that the total number displayed a legal collecting permit to the ar­ of violations (I) for a study period was related resting officer. In Maine the lower court trial The New Mexico Game and Fish Depart­ to the number of violations detected in that was waived and the superior court hearing ment replicated Vilkitis' closed-season viola­ time (either arrests Ma or incidents Mb) ap­ postponed until after field research was to be tion simulation technique during '1975 proximately the same way that the total completed. During the study period, 1 of the (Pursley 1977). The simulator hired by the number of violation simulations performed 81 closed-season simulations (C) was reported department was described as a "very good (C) was related to the number of violation as an incident (R). Based on the fact that en­ poacher." During the deer hunting season of simulations detected and reported by en­ forcement agents reported 136 closed-season 1975, the simulator raided dumps, trash cans, forcement personnel (R). That is, arrests and 249 incidents, illegal c1osed­ and abandoned camps to obtain deer car­ season kill for the state of Maine for the casses. Nineteen animals were killed by the _,_ =i study period was estimated as: simulator but were apparently not counted as Ma (or Mb) R simulations. In contrast to Vilkitis' studies 1= (136+249+1)(81+1) -1 = 15,825 (Vilkitis 1968, Vilkitis and Giles 1970), the and (1+1) locations and times in which simulations = Ma(C) = Mb(C) 1 orl were performed were selected randomly. R R This figure was underestimated by 13.95% since: (1) no arrests were made for simulated Simulations (discards) were staged over a An unbiased estimate is achieved by (Bailey acts; (2) recording of incidents was a new pro­ period of several months and their locations 1951, Chapman 1951): cedure and some agents did not record all il­ recorded. By checking incoming reports from legal kills; and (3) the product of M times C agents, matches could be made. One of the I = (Ma +l)(C+l) - 1 was not at least 3 times the estimate of l. The 144 simulations was reported as an incident. (R+l) unbiased estimate of , was calculated as Based on 236 violations reported by New Mexico agents during the 50-week closed by increasing the frequency of simulations with different characteristics. The simulated season in 1975, the following estimate was performed by each simulator. The difference act (hide, trails, blood) is continuous over made: between 1 and 3 detections is the difference some time (several days), not instantaneous between a reasonable and an unreasonable as the poaching act, therefore presenting ad­ A oM N =-­ estimate (i.e., 2,447-662,314 with 1 detection ditional problems in computing the probabili­ X and 834-13,993 with 3 detections in Pursley's ty of an agent encountering evidence. study). Beattie et al. (1978) presented an untested N = (144)(236) The New Mexico researchers described the procedure for estimating the number of in­ second drawback (Pursley 1977:1): "Our dividuals illegally killing deer during a 1 agent was armed with a substantial supply of specified period. The Lincoln-Petersen mark­

A dismembered deer. He was faced with the recapture method with apprehended N = 33,984 problem of deciding what the typical violators would employ the following for­ where N = total estimated number of deer poacher does with his refuse. We do not mula: lost during closed season, 0 = total number know if 'typical' is casually tossing it by the of simulated violations, M = total number of side of the road or if it is carefully hiding the N = fA +1HC +1) _ 1 reported violations, and X = total number of unwanted remains under a log. The only solu­ R+1 tion was to try to produce a spectrum, which simulated violations reported. The 95% con­ would range from easy to detect, through where N" = estimated number of individuals fidence limit was given as 2,447-662,314. The possible, through improbable, in hopes of violating, at least once, a deer-related figure of 662,314 is almost 20 times the closely reproducing the real situation," wildlife statute during the period for whkh R number of deer harvested annually in New The researchers attempted to overcome is measured; A = number of individuals Mexico. Due to a breakdown in officer report­ this drawback, not knowing the habits of a issued a verbal or written warning or citation ing, Pursley discovered that one of the agents "typical" poacher, by staging simulations for committing 1 or more violations of a deer­ had discovered a simulation but had not randomly in space and time. However, there related statute during a specified period; C = reported it. Using this figure of 2 detections, a are indications from 2 previous studies that number of individuals apprehended for revised estimate was given as: deer-related violations do not occur random­ violating a deer-related statute during the ly in time or space. period for which measurement of C occurs; N = (144) (236) = 16,992 Sawhill and Winkel (1974) interviewed 148 and R = number of A individuals appre­ 2 admitted illegal deer hunters from 9 counties hended during a period equal to but follow­ of New Jersey. According to self reports of ing the period for which A was measured. with a 95% confidence limit of 1,303-47,934. the Violators, the majority of illegal deer ac­ Roughly the same assumptions of the The method of unbiased estimation was not tivity occurred on any given day of the week Lincol n-Petersen formula application to wild employed. between 12 and 2 a.m, during the winter animal populations would have to be made There are 2 major drawbacks to the use of months. Eighty-four percent of the time, the the violation simulation technique as a valid violation was committed by 2-3 men. (Note (e.g.. emigration from violator population means of estimating closed-season illegal that past violation simulation studies equals immigration, A's keep violating and deer kill. First, in the 3 deer violation simula­ employed 1 simulator.) do not change behavior patterns). One advan­ tion studies reported previously, confidence Kaminsky and Giles (1974) reported that tage of this procedure is that previous arrest limits on estimates of illegal kill were unac­ spotlighting in Virginia was most frequent records could be employed to explore its ceptably large. Although Vilkitis did not between 10 and 11 p.m. on Saturdays in suitability. report confidence limits in his studies, his November near cornfields surrounded by a estimates must have been similar to Pursley's mixture of oaks and pines. Spotlighting viola­ RELATIVE VERSUS ABSOLUTE confidence limits of 2,447-662,314 animals. tions did not occur randomly. The upper limit seems unrealistic based on Since no typology has been constructed for ESTIMATES OF ILLEGAL biological carrying capacity and hunting ac­ the "typical" deer poacher, violation simula­ DEER KILL tivity required to harvest the animals. Clearly tion studies should probably be implemented the distribution is not distributed normally. A by several people. To improve on the All of the techniques discussed for low number of detections could be overcome representativeness of the simulation, at­ estimating illegal deer kill or number of deer­ by employing more than 1 simulator and/or tempts should be made to select simulators law violations produce absolute rather than 69 relative measures. Absolute measures are that most citizen reports of violations would the validity of the technique. useful for justifying increases in enforcement be made by telephone. Since only 20 of 50 None of the previously mentioned pro­ manpower levels to produce and document a state wildlife law enforcement divisions were cedures for estimating violations, whether reduction in absolute levels of violations. found to have standard record-keeping of providing a relative or absolute measure, will However, relative measures of the level of citizen violation reports, and, since 7 of the probably be reliable or valid when employed violation occurrence may be more desirable 50 listed agent affiliations in telephone direc­ alone. To the extent that results obtained since manpower levels are normally static, tories, the joint probability of a successful from 2 or more of the procedures are and since information necessary to compute call and record of the call may be as low as significantly correlated, convergent valida­ indexes of the level of violation activity will 0.06. tion has been achieved and greater con­ probably be less costly to acquire than ab­ The final example of a relative index would fidence can be placed in statements regard­ involve comparison of deer-law violations ing the validity of inferences about the "deer solute levels of violation activity. Indexes of reported by hunters themselves. Previous law violation problem" (Cook and Selltiz the level of deer-law violations may be self-reported violation studies of hunter il­ 1967). developed using: (1) a Harris Poll-type obser­ legal activities have been conducted by Most of the estimation techniques have vation panel for wildlife violations, (2) a Chesness and Nelson (1964), Sawhill and never or only occasionally been employed by general probability survey of observed Winkel (1974), and Smith and Roberts (1976). wildlife Jaw enforcement divisions. Because wildlife violations, (3) standardized violation Research is needed to develop a validation of this lack of consistent use, problems reporting rates by agents and the public, or (4) methodology for self-reported violations by associated with employing the techniques self-reported violations by hunters (Beattie et hunters. have not been solved and an "optimum" set al. 1978). of indexes or estimators cannot be recom­ A Harris Poll-type observation panel would CONCLUSIONS mended at this time. The validity of estima­ involve selecting randomly a sample of tion procedures must be established to im­ residents, probably in rural areas, and re­ McCormick's compliance rate procedure is prove the procedures' usefulness as an aid to questing them to supply periodic written probably the most economical and readily in­ allocation of enforcement manpower and reports of the deer-law violations they corporated method currently available for related processes. Once the validity of 1 or observe and/or hear about during the course estimating in-season game violations. It more estimation procedures has been should be employed cautiously, and only of normal activities. Panel responses would established, the procedures must be used provide an index to the change in frequency after considering the assumptions presented repeatedly and consistently to provide a by Cowles et al. (1979), and deciding whether of deer-law violations over time. basis for manpower allocation and other the assumptions c",n reasonably be met and A general probability survey would be decisions. The trend has too often been to what adjustments for bias may be made. Im­ similar to using a Harris Poll-type panel. A employ a procedure once, such as the viola­ proved systems of receiving and recording random sample of hunters and/or state tion simulation technique, obtain a number, citizen reports of violations would avoid residents would be contacted periodically and remove the procedure from considera­ some of the bias problems encountered in and respondents requested to provide infor­ tion for future use. Validated estimation McCormick's technique as well as providing mation about the type and frequency of techniques, whether providing absolute or both in-season and closed-season study of violations they observed over a period of relative measures, must be incorporated into time. relative violation rates. seasonal or annual enforcement manage­ Vilkitis' violation simulation technique is Many state wildlife law enforcement divi­ ment plans. the best method available for estimating ab­ sions collect information enabling a com­ Poaching estimation is a new and only parison of violation reporting rates over time. solute closed-season illegal deer kill. There is slightly studied area of research. Progress in no reason why the technique cannot be ex­ Attempting to standardize report recording in estimating deer mortality from poaching may conjunction with improved agency/public panded to estimate both closed-season and likely beachieved in: communication (increased reporting of viola­ in-season illegal deer kill. Performing simula­ tions) would allow such comparisons. Cowles tions randomly in time and space, drawing (1) Poacher interviews to determine mean an­ et al. (1978) found that the nationwide prob­ simulators from different social groups, per­ nual take per poacher, a shift in emphasis ability of a citizen-observed violation being forming sufficient simulations to achieve from deer to poacher estimates. recorded by a state wildlife Jaw enforcement reasonable confidence limits on estimates, (2) Changes in winter roadside spotlight division may be as low as 5.6%. This prob­ and studying appropriate distributions of counts of deer (adjusted by weather ability was computed under the assumption poaching and agent searches should improve variables). The assumption is that popula­

70 tions under poaching pressure will CHAPMAN, D.G. Polytech. Inst. and State Univ., decrease more rapidly than others. 1951. Some properties of the hypergeometric Blacksburg. MS Thesis. 10& pp. distribution with applications to (3) Regression analyses. Mechler (1970) MICHIGAN LAW ENFORCEMENT DIVISION zoological census. Univ. Calif. Publ. in showed high Ri values for simple equa­ 1973. Michigan deer kill estimates 1971-1972. Stat. 1(7):131-{)(). tions that were predictive of legal deer kill Mich. Dep. Nat. Resour. 1& pp. (mimeo) in Virginia counties. The implication is CHESNESS, R.A. and M.M. NELSON 1974. Measurement of the effectiveness of law that illegal kill is either so slight as not to 19&4. Illegal kill of hen pheasants in Min­ enforcement effort. Mich. Dep. Nat. influence the variance in annual harvest nesota. J. Wildl. Manage. 28(2):249-53. Resour. 9 pp. (mimeo) or to be a constant proportion of the legal COOK. S.W., and C. SELLTIZ harvest. Which of these options (or others) 19&7. A multiple-indicator approach to at­ PURSELY; D. 1977. Illegal harvest of big game during closed is operational needs study. titude measurement. Pages 220-35 in M. season. N.M. Dep. Game and Fish. & pp. (4) Substantial rewards for information about Fishbien, ed. Readings in attitude theory and measurement. Wiley and Sons, New (mimeo) deer poaching. York. 587 pp. (5) Mathematical analyses of time between ROBSON, D.S. and H.A. REGIER deer poaching arrests, enforcement effort, COWLES, c.j., K.H. BEATTIE, and R.H. GILES, JR 19&4. Sample size in Petersen mark-recapture and probability of repeat arrests. 1978. A survey of methods of recording reports experiments. Trans. Am. Fish. Soc. of fish and wildlife violations. Fisheries (6) Educational and enforcement efforts in­ 93(3):215-2&. 3(2):8-11. crementally applied that reduce deer 1979. Potential limitations of wildlife law com­ SAWHlll., G.S, aDd R. WINKEL poaching arrests may be beck-projected to pliance estimators. Wildl. Soc. Bull. 1974. Methodology and behavioral aspects of estimate the initial 'levels of poaching ar­ 7(3):188-91. the illegal deer hunter. Proc. Conf. rest potentials. Southeast. Assoc. Fish and Wildl. Agen­ (7) Computer simulations of population HARDIN, J.W. and J.L. ROSEBERRY cies 28:715-19. 1975. Estimates of unreported deer loss dynamics may be used to interpret a range resulting from a special deer hunt on SMITH, R.1. and R.J. ROBERTS of feasible poacher-caused mortalities Crab Orchard National Wildlife Refuge. that may allow the legal harvest or other 197&. The waterfowl hunters' perceptions of Proc. Conf. Southeast. Assoc. Fish and the waterfowl resource. Trans. North herd characteristics (e.g., sex ratios) to be Wildl. Agencies 29:4&0-&&. Am. Wildl. and Nat. Resour. Conf. achieved. KAMINSKY. M.A. and R.H. GILES. JR. 41:188-93. 1974. An analysis of deer spotlighting in VILKITIS, J.R. Virginia. Proc. Conf. Southeast. Assoc. 19&8. Characteristics of big game violators and Fish and Wildl. Agencies 28:729-40. extent of their activity in Idaho. Univ. LITERATURE CITED MCCORMICK, J.B. Idaho, Moscow. MS Thesis. 202 pp. 19&8. A procedure for evaluating the effec­ 1971. The violation simulation formula proves BAILEY, N.T.j. tiveness of wildlife law enforcement. as reliable as field research in estimating ~ 1951. On estimating the size of mobile popula­ Proc. Conf. West. Assoc. Fish and Wildl. closed season illegal big game kill in tions from recapture data. Biometrika Agencies 48:&2&-39. Maine. Trans. Northeast Sect. The Wildl. 38:293-30&. 1970. An evaluation of wildlife law enforce­ Soc. 28:141-44. ment effort. Proc. Conf. West. Assoc. BEATTIE. K.H., R.H. GILES, JR. and C}. COWLES VILKITIS, J.R. and R.H. GIlES, JR. 1978. Quasi-experiments, multiple indicators, Fish and Wildl. Agencies 50:517-40. 1970. Violation simulation as a technique for and enforcement effectiveness. Proc. MECHLER. J.L. estimating closed-season big game kill. Conf. West. Assoc. Fish and Wildl. Agen­ 1970. Factors influencing the white-tailed deer Trans. Northeast Sect. The Wildl. Soc. cies 58:7&-90. harvest in Virginia, 1947-19&7. Va. 27:83-87.

71 Wisconsin ON R AGE DETERMINATION OF WHITE-TAILED DEER IN THE MIDWEST - METHODS AND PROBLEMS

John L. Roseberry Cooperative Wildlife Research Laboratory Southern Illinois University-Carbondale Carbondale, IL 62901

Abstract: This paper reviews the various of individual animals is an acknowledged re­ deer into age classes. Hunter (1947) methods of age determination for Odocoileus and quisite for effective management. attempted to segregate male mule deer (0. examines comparative results from their applica­ Knowledge of age distribution enhances hemionus] into age classes by a combination tion. A survey of midwestern state wildlife agencies understanding of population dynamics; and, of dental and antler characteristics. Subse­ indicated that ages are assigned to harvested important parameters such as fecundity and quent work confirmed that antler dimensions animals based primarily on criteria involving survival rates are most useful when age­ tend to increase with age but are unsuitable molariform tooth eruption and wear, external in­ cisor characteristics, and cementum annuli counts. specific (Caughley 1977). for assigning animals to year classes Most agencies expressed satisfaction with their Age data present 3 major concerns to the (Severinghaus et at. 1950, Severinghaus and particular methodology although the various deer biologist: (1) are the data accurate, (2) Cheatum 1956). Nevertheless, weights of techniques have been shown to produce conflict­ are they representative, and (3) how can they shed antlers might provide clues to popula­ ing results. Nonuniformity of aging criteria be used to improve management? This paper tion age structure (Anderson and Medin hampers the interstate exchange of information considers only the first problem, especially as 1969). and the interpretation of age-specific research it relates to analysis of harvested samples. results. Needed is a well-planned, coordinated, and Various methods of age determination are definitive evaluation of the various techniques Maturation of Long Bones using adequate known-age material representing reviewed and comparative results examined. all geographic and ecological regions of the Procedures currently used in the Midwest are An index of skeletal maturity (degree of Midwest. viewed in light of these findings. epiphyseal closure) developed from known­ Technological aspects of the techniques are age black-tailed deer (0. h. columbianus] INTRODUCTION not emphasized. enabled lewall and Cowan (1963) to ac­ curately estimate age to 72 months. Wildlife management has been criticized However, the technique has obvious limita­ for failure to properly emphasize the prin­ METHODS OF AGE tions for harvested samples and has received ciples and concepts of population dynamics DETERMINATION little further attention. (Scott 1954, Caughley 1976). It is therefore Techniques for determining age of fetal appropriate that the present symposium on Molariform Tooth Eruption white-tailed deer (Odocoileus virginianus] (Armstrong 1950, Hudson and Browman 1959) consider the management implications of and early postnatal deer (Haugen and Speake and Wear 1958) will not be reviewed. The following em­ population phenomena. Many such The normal sequence of tooth develop­ phenomena are reflected by and/or con­ phasizes techniques for assigning age past 4-5 months. ment in white-tailed deer permits segregation tribute to changes in population age struc­ of autumn-killed animals into 3 age groups: ture. The ability to accurately determine ages Antler Dimensions fawns (4-7 months) possessing less than 6 molariform (cheek) teeth, yearlings (16-19 Roseberry, John L. 1980. Age determination of A once popular, though presently less com­ months) showing deciduous or newly erupted white-tailed deer in the Midwest - methods and problems. Pages 73-82 in Ruth L. Hine and Susan mon, notion among laymen was that number permanent 4th premolars, and adults (28 + Nehls, eds. White-tailed deer population manage­ of antler points was indicative of age. months) having fully erupted permanent ment in the north central states. Proc. 1979 Symp. Cahalane (1931) suggested that beam teeth. Assignment of the latter group to in­ North Cent. Sect. Wildl. Soc. 116 pp. diameter could segregate male white-tailed dividual age classes is by degree of wear on

73 the occlusal surface of molariform teeth It is generally acknowledged that tooth wear examination of jaws in situ may present some (Taber 1969). Terminology used here and may vary according to climate, soil type, difficulties. From a sample of 418 whitetails, elsewhere follows Riney (1951) who de­ food habits, and genetic makeup (Ludwig Roseberry and Hardin (1978) judged that il­ scribed the dental formula of Odocoileus as 1967). Ryel et al. (1961:310) applied linois check station operators correctly follows: Severinghaus's (1949) New York criteria to segregated 98% of all adults from younger Michigan whitetails and concluded: "Agree- . animals and classified 92% of all yearlings 1.2.-.2.-.2.­ c.2.­ P .2.-_Ll.! M..l2 3 ment is good in the 2'h- and 3'h-year age correctly. Kuehn (1970) reported that 1 2 3 1 o 2 3 4 1 2"3 groups but falls off rapidly as the age in­ previously instructed Minnesota check sta­ creases." Ludwig (1967), in contrast, reported tion personnel correctly identified 98% of all According to Severinghaus and Cheatum correctly aging 33 of 36 known-age (2'h -4 Y,) fawns and 95% of all yearlings from clean, (1956), American pioneers used tooth wear to whitetails from southern Illinois using excised jaws. differentiate among young, prime, and old Severinghaus's criteria. deer. Biologists as well have long regarded Methods of age determination must be External Incisor Characteristics tooth replacement and/or wear as potential judged for precision and applicability as well Severinghaus (1949) described the se­ criteria of age in deer (Cahalane 1932, as accuracy, i.e. are the criteria unambiguous quence of incisiform tooth replacement for McLean 1936, Cowan 1936, Park and Day and will they produce consistent results? known-age O. virginianus from New York; 1942). Unquestionably though, the signal con­ Brown (1961) found that individual examiners however, Ryel et al. (1961) found the criteria tribution was by Severinghaus (1949) who tended to consistently err in one direction or too variable to permit aging Michigan fawns described tooth replacement and wear for a the other. Ryel et al. (1961) noted significant to the nearest month. Brown and Peabody large collection of known-age whitetails from differences among individuals and several (1972) and Rice (1977) mentioned the use of New York. Various adaptations of the tech­ levels of proficiency in assigning ages based external incisor characteristics to identify nique were subsequently applied to o. he­ on tooth wear. Others reporting variable in­ fawns, yearlings, and adults. The latter 2 mionus and o. h. columbianus (Robinette and terpretation of the same material include classes are separated by crown wear whereas Jensen 1950, Leopold et al. 1951, Moreland Low and Cowan (1963), Gilbert and Stolt fawn incisors are deciduous; or, if permanent, 1952, Jones 1953). Robinette et al. (1957) later (1970), Brown and Peabody (1972), and show incomplete closure of the root canal. described these criteria for known-age mule Roseberry and Hardin (1978). Brown (1961) deer and developed a quantitative index of recorded slightly greater accuracy in assign­ Eye Lens Weight molariform tooth wear (molar ratio). This in­ ing ages to males than females among black­ dex is not generally reliable for assigning in­ tailed deer. There is some indication that The lens of the mammalian eye dividual ages (Brown 1961, Ludwig 1967, teeth of males of O. hemionus wear faster presumably grows continuously throughout Erickson et al. 1970). With respect to quan­ leading to underestimation of female ages life, although not at a constant rate. Use of titative indexes, Roseberry and Hardin (1978) (Erickson et al. 1970, Thomas and Bandy eye lens weight as a criterion of age was first suggested that the multivariate statistical 1975). However, Hill (1973) and Roseberry described by Lord (1959) for cottontail procedure, discriminant analysis, might be and Hardin (1978) reported no significant dif­ (Sylvilagus floridanus] and subsequently ap­ used to distinguish among age groups and ference in rates of agreement between age plied to a variety of wildlife including deer classify unknown subjects based on a com­ assignments based on tooth wear and dental (Friend 1967). Lord (1962) concluded that the bination of dental and mandibular cementum annuli for male and female o. eye lens growth curve in deer was influenced measurements and characteristics. virginianus from Pennsylvania and Illinois, by nutritional status thus limiting its Subjective (visual) evaluation of tooth respectively. Likewise, Kuehn (1970) noted no usefulness as an indicator of age. Longhurst eruption and wear has been tested on known­ difference in quantitative measurements of (1964:773), however, developed regression age specimens with mixed results. Ryel et al. tooth wear for male and female whitetails equations describing eye lens weight-age (1961) and Brown (1961) reported the follow­ from Minnesota. relationships for known-age black-tailed deer ing accuracy by age class for white-tailed and Despite evidence that tooth eruption and and pronounced the technique: "satisfactory black-tailed deer, respectively: 1'h (97 and wear is not a precise indicator of age past 1 Y, for aging deer at least through 5 years ....t r 90%), 2'h (69 and 60%), 3'h (57 and 42%), years, it is generally assumed that the tech­ Longhurst's data, along with additional 4'h (41 and 26%), and 5 1.1, (23 and 22%). Ex­ nique is adequate to definitively segregate known-age specimens, were reanalyzed by perienced personnel in both studies tended to fawns, yearlings, and adults (Brown 1961, Connolly et al. (1969:703) who concluded: "a overestimate ages of jaws in younger classes Ryel et al. 1961, Connolly et al. 1969, considerable fraction of adult deer aged by and underestimate those in older age groups. Erickson et al. 1970). However, check station the lens method may be assigned to the

74 wrong age class .... " Downing and normal restriction of food intake technique. Specifically, northern deer may Whittington (1964) recommended the tech­ characteristic of a deer's annual cycle. Sauer have more clearly defined cementum struc­ nique for separating fawns, yearl ings, and (1973) further investigated and described the ture than do animals from southern regions adults; however, ludwig (1967) considered physiology of cementum annulus formation. (Matson 1978). Regardless of geographic Jens weight no better than body weight or Ransom (1966) examined known-age origin, all areas of the cementum do not pro­ beam diameter for assigning age to male molars from Michigan and Minnesota white­ duce uniformly distinct growth layers; and, whitetails from Illinois. lueth (1963) earlier tailed deer and claimed that growth layers false, split, and compound annuli can occur reached a similar conclusion for deer in were visible without histological preparation. (Gilbert 1966, Campbell 1967, Erickson and Alabama. Friend (1968:285) concluded from a Campbell (1967) was unable to duplicate Seliger 1969, Lockard 1972, Thomas and review of published and unpublished studies Ransom's findings with Illinois material but Bandy 1973, Rice 1980). Gasaway et al. (1978) that: "fawns and yearlings could be reported that cementum annulations from reported that 2 experienced workers agreed separated with a high degree of accuracy, sectioned and stained molars were in agree­ on only about 60% of the moose incisor sec­ yearlings and 2V, year olds with less certain­ ment with known age for 75 of 76 animals. tions they examined. Gwynn (1978) also ty, and thereafter year class assignment was Subsequent cementum analyses have presented evidence of variable interpretation unreliable for individual deer." Other studies generally involved microscopic examination of cementum growth layers from white-tailed have also shown that eye lens weights are un­ of histologically prepared incisor sections. deer in Virginia. In addition, paired incisors satisfactory for assigning individual ages, Complete agreement between incisor cemen­ do not always show identical patterns of an­ especially past 1V, years (Hoffman and tum annuli counts and known age of nulation (Rice 1977). In a test of precision Robinson 1966, Urbston 1968, Erickson et al. Odocoileus has been reported from several achieved by a commercial laboratory, 30 1970, Keller and landry 1976). studies (Low and Cowan 1963, Gilbert 1966, pairs of incisor teeth (I,) from southern Il­ In a related effort, ludwig and Dapson Erickson and Seliger 1969, Lockard 1972, linois whitetails were labeled and submitted (1977) attempted to correlate the accumula­ Thomas and Bandy 1973); however, except separately for sectioning and age determina­ tion of insoluble lens proteins with age in for the latter study, it is unclear whether blind tion. Twelve (40%) of the 30 deer were assign­ white-tailed deer. They found, however, that tests were conducted, i.e. final annuli deter­ ed a different age from their 2 incisors individual ages could be estimated only minations made prior to knowledge of true although the discrepancies never exceeded 1 within 2 or 3 years. age. Sauer (1971) achieved 84% agreement year (Table 2). This particular laboratory between incisor cementum annulations and assigns 1 of 3 certainty codes to each age Dental Annulations known age for white-tailed deer from New estimate: "results nearly certain", "some er­ Morphological characters that change con­ York while Hill et al. (1975) recorded 86% ac­ ror possible", or "error likely" due to confus­ tinuously with age are inherently less ac­ curacy for a small sample of known-age ing or indistinct annulus patterns. Such curate for age determination than those Pennsylvania deer. classifications assigned to a sample of deer which change by annual quanta (Caughley The above cited studies dealt primarily from III inois varied significantly by region 1977). Annular growth patterns of mam­ with cementum analysis of limited and age class but not sex (Table 3). Therefore, malian teeth are an example of the latter geographic material at University it would not be possible to simply ignore the phenomenon (Scheffer 1950, laws 1952). The laboratories. Increasingly though, wildlife "error likely" samples without biasing results. first attempt to so age cervids was by agencies are considering commercial Sergeant and Pimlott (1959) who related laboratories for processing large samples. COMPARATIVE RESULTS FROM growth layers in the cementum of moose One of the more popular of these, Matson's CRITERIA OF AGE (Alces alces) incisors with age. Based on early Commercial Microtechnique, has been the work with Odocoileus, Low and Cowan subject of limited evaluation with known-age The literature on deer aging methods con­ (1963:468) wrote: "The passage of a year in a material. Citing personal communication tains several examples of ambiguities in deer's life usually results in a distinctive pat­ with R.L. Cook, Texas Parks and Wildlife reporting results of accuracy tests. One in­ tern of increment to the cementum. A pale­ Department, Hackett et al. (1979) reported an volves the base from which percentages of staining zone of rapid increment is added 83% error for 29 known-age white-tailed deer agreement are computed and reported. For during the spring and summer. During the from Texas. In contrast, 13 (72%) of 18 age example, suppose 140 animals were iden­ winter, there is a reduction in increment rate assignments based on cementum annuli for tified as 2V, years old from a large group that shows as a dark-staining zone." Low and Illinois deer were correct and none erred by which actually contained 100 animals of that Cowan associated the latter, called "rest more than 1 year (Table 1). These results sug­ known age. Suppose also that 70 of the lines" by Thomas and Bandy (1973), with the gest possible differential applicability of the animals termed 2'11 were actually that age.

75 TABLE 1. Dental cementum annuli and tooth wear age TABLE 3. Certainty of annuli counts for I1linois white­ assignments for known-age white-tailed deer from Crab tailed deer based on region, age, and sex, 1977 and Orchard National Wildlife Refuge, southern 1/linois. 1978.

Age According to 2 Different Techniques Percent Assigned to No. Certainty Levels" Cementum Annuli' Known Age Incisor NO.1 Incisor NO.2 Tooth Wear" Variable' Deer A 8 C Region 2'12(8) 2'12 2'1, North (42°) 159 19 50 31 3'12 2'12(8) 3'12(8) 3'12 Centra I (40°) 275 23 61 16 3'12 3'12(8) 3'12(8) 3'12 South (37°7') 335 11 50 39 4'12 4'h(A) 4'12 (A) 3'12 4'12 4'12 Ag,e Class1 6'12 5'1,(8) 4'/,+ '12 360 19 51 30 6'12 6'h(C) 4'12 + 2'12 232 17 61 22 6'12 6'h(C) 6'h(C) 4'12 + 3'1, 87 14 65 21 6'12 4'12+ 4'12 + 113 15 44 41 6'12 + 6'12(8) 3'12 6'12+ 9'h(C) 4'12+ Sex 7'12 8'12(8) 8'12(8) 4'12 + Male 54 4 50 46 7'12 7'12(8) 4'12 + Female 79 13 54 33 8'12 7'1,(C) 8'h(C) 4'12 + , Region: X" = 43.0, P < 0.001 8'12 4'12 + Age: X" 18.8, P 0.01 9'12 9'12(8) 9'12(8) 4'12 + = < Sex: X" = 4.5, P > 0.10 , Incisors sectioned and aged and codes assigned by Matson's Commercial Microtechnique, Milltown, Montana, 1978: A = "Codes assigned by Matson's Commercial Microtechnique; results nearly certain; 8 = some error possible; C = error likely. A = results nearly certain, 8 = some error possible, C = error likely. " Age assignments made at Illinois Department of Conservation check station, Crab Orchard National Wildlife Refuge, 1977 and 1 Matson assigned individual age classes to all deer; those 1978. 4'12 and older combined for this comparison.

TABLE 2. Independent age assignments for 30 white-tailed deer from cementum annuli counts of both incisors. * Thus, 70% of all known-age 2V:t-year-olds Incisor NO.2 Incisor were accurately identified (70 of 100); yet, the NO.1 1'12 2'12 3'12 4'12 5'12 6'12 7'12 8'12 9'12 10'1211'1212'1213'12 assigned age group was only 50% correct (70 1'12 2 of 140). Another problem involves confusion 2'12 2 3 1 of accuracy with precision. Bell (1974) stated 3'12 1 5 that the square of the coefficient of correla­ 4'12 3 1 1 5'12 2 1 tion (r") between known and estimated age 6'12 1 for a particular sample was a measure of ac­ 7'12 1 curacy. Hackett et al. (1979), perhaps 8'12 1 understandably, interpreted these reported 9'12 3 values as the percentages of accuracy 10'12 1 achieved during Bell's study. In actuality, the 11 '12 data showed considerably less agreement 12'12 than suggested by the r" values which were a 13'/, measure of precision and not accuracy. For 14'12 1 example, if all estimates were exactly 1 year , Incisors collected and labeled by the Cooperative Wildlife Research Laboratory, greater than actual age, perfect correlation Southern Illinois University-Carbondale. Material sectioned and analyzed by would result(r 1.00) yet accuracy would be Matson's Commercial Microtechnique, Milltown, Montana. = 0%.

76 Tooth Eruption and Wear vs. the laboratory as opposed to jaws in situ tionable yearlings" based on this technique under check station conditions (Roseberry contained 6% 21;)-year-olds according to Dental Cementum Analysis and Hardin 1978). In general, private analyses dental cementum analysis. The proportion of of cementum structure produced somewhat all yearlings considered "questionable" by Comparisons of independent age better agreement with eruption and wear external incisor criteria was not reported. assignments by tooth eruption and wear and than did commercial efforts; however, the Rice (1977) also recorded a 6% error for 261 cementum annuli counts have been reported only direct comparison (Gwynn 1978) showed mule and white-tailed deer from South by several authors. Lockard (1972) found commercial annuli counts superior in identi­ Dakota designated 11;) by incisor crown 74% agreement between the 2 techniques for fying "known-age" fawns and yearlings. The wear. Unlike the Kansas sample, "ques­ a large sample of O. virginianus from various poor results using annuli counts in Mississippi tionable yearlings" were not included in this locations. Similar rates of agreement were may have reflected physiological anomalies test. Karl Menzel of the Nebraska Game and recorded for this species in Indiana (Olson of the sample. Hackett et al. (1979) suspected Parks Commission (pers. comm. 1979) tested 1967) and Alabama (Boozer 1969). Low and that their animals had experienced a summer external incisor and tooth eruption and wear Cowan (1963:470), however, noted "many drought-induced stress period and, therefore, criteria on 537 mule and white-tailed deer discrepancies" in mule deer. With the excep­ a double annuli deposition. and found 86% and 79% agreement for 1 'h tion of Campbell (1967) and Thomas and and 2Yz+ categories, respectively. Accor­ Bandy (1975), results of other comparisons of External Incisor Characteristics ding to Lee Gladfelter of the Iowa Conserva­ tooth eruption and wear and cementum an­ tion Commission (pers. comm. 1979), 15% of nuli counts have been generally disappoint­ Use of external incisor characteristics as 74 whitetails designated yearlings by external ing, especially for deer 21;) + years (Table 4). criteria of age is not well documented in the incisor characteristics were called 21;) by a Closer agreement with annuli counts was literature. Working with Kansas white-tailed commercial laboratory using annuli counts, generally achieved when eruption and wear and mule deer, Brown and Peabody (1972) and "many" deer aged as 21;) + by the was evaluated from clean, excised material in reported that a group of deer aged as "ques­ former method were called 1Vl by the latter.

TABLE 4. Some comparative age assignments by tooth eruption and wear vs. cementum annuli counts for unknown age Odocoileus. Percent Agreement" Procedure>" No. Location Source Deer V, 1V, 2V, 3V, 4V, 5V, + TEW CAC British Columbia Thomas and Bandy (1975) 769 100 100 94 68 44 100 Excised Private Colorado Erickson et al. (1970) 116 95 ---47 In situ Private tlltnols'l Campbell (1967) 128 100 100 100 87 75 100 Excised Private Illinois Roseberry and Hardin (1978) 801 88 51 25 72(4V, +) In situ Commercial Illinois Roseberry and Hardin (1978) 413 92 60 32 77(4V, +) Excised Commercial Kansas Brown and Peabody (1972) 300 100 72 67 44 212 Excised Private Maine Gilbert and Stolt (1970) 682 94 66 52 32 92 In situ Private Minnesota Kuehn and Karns (1970) 854 70 39 25 87 Variable3 Private Missouri Porath and Torgerson (1979) 1125 60 25 20 89 In situ Commercial Mississippi Hackett et al. (1979) 404 0 17 27 In situ Commercial Pennsylvania Hill et al. (1975) 199 65 58 55 69 Excised Private Virginia Gwynn (1978) 312 50 42 Excised Private Virginia Gwynn (1978) 190 100 84 Excised Commercial " Age classes based on TEW except for Maine and Kansas studies; 5V, + category created for this comparison. ""TEW = tooth eruption and wear; CAC = cementum annuli counts.

1 Sample contained some known-age animals. 2 Mean percentage agreement for individual age classes past 4v'. 3 62% in situ, 38% excised. Implications of Incorrect Age presumably because of regional variation in CONCLUSIONS Assignments tooth wear. In III inols, rates of agreement be­ tween tooth eruption and wear and annuli Most state agencies expressed satisfaction It is obvious that techniques currently counts did not vary regionally (Roseberry and with their procedures for aging deer; yet, available for age determination of Hardin 1978). research has shown that the methods current­ Odocoi/eus are not infallible. Of primary Some workers have concluded that age ly used often produce conflicting results. concern is whether errors are compensatory assignment by tooth eruption and wear Lack of a common means for assigning ages and if not, are their consequences trivial or causes underestimation of the strength of the to midwestern deer is regrettable though important. Robinette et al. (1957) felt that 2'11-year cohort and overestimation of the perhaps understandable. Deer management tooth eruption and wear errors should be 3 V, -year age class (Ryel et al. 1961, Lockard programs operate at various levels of fund­ compensatory and therefore the technique 1972). In contrast, Erickson et al. (1970) and ing; and, the problems associated with collec­ satisfactory for determining population age Roseberry and Hardin (1978) felt that the tion of harvested material vary greatly from structure. Caughley (1977), however, pointed 2V,-year cohort was inflated by this aging state to state. More importantly, however, out that aging errors are not necessarily com­ method. The strength of this cohort in­ research has thus far failed to produce a pensatory even when they occur at a con­ fluences calculated survival rates for 1 '11-, definitive, unequivocal aging procedure for stant rate for all age classes. For example, 2 '11- and 3'I1-year-olds. In fact, aging Odocoileus suitable for large samples of suppose that a sample contained 80 2'h-year­ discrepancies may significantly alter age­ harvested animals. The lack of a common old deer and 60 3'I1-year-olds and the prob­ specific survival schedules (Kuehn and Karns system for aging deer hampers the interstate ability of mistaking one for the other was 1970, Roseberry and Hardin 1978, Hackett et exchange of information and the interpreta­ exactly the same (e.g., 0.30). Such aging errors al. 1979). tion of age-specific research results. Needed would result in an estimated ratio of 74:66. is a well-planned, coordinated, and definitive The rate and direction of aging errors ap­ evaluation of aging techniques using ade­ pears highly variable among age classes, quate known-age material representing all techniques, and geographic regions. The AGE DETERMINATION geographic and ecological regions of the cementum analysis technique has been PROCEDURES USED IN Midwest. Such research should focus on age suspected of overaging deer from Virginia THE MIDWEST classes 1 V, to 4V" as fawns can be satisfac­ (Gwynn 1978) and Mississippi, but underaging torily identified by a number of easily applied animals from Texas (Hackett et al. 1979). Procedures for acquiring samples and criteria and older animals are numerically Moen and Sauer (1977) bel ieved that tooth assigning ages to harvested deer from less important in exploited populations. eruption and wear underaged deer at New midwestern states are summarized in Table 5. Several aspects of the aging question need York's Seneca Army Depot. In contrast, Most agencies operate some type of check particular attention. Use of external incisor Brown and Peabody (1972) reported that Kan­ station (either mandatory or voluntary) where characteristics to separate fawns, yearlings, sas biologists using tooth eruption and wear hunters may bring their animals for registra­ and adults has not received adequate scien­ tended to overage deer through 6'11 years tion and/or data collection. Age determina­ tific testing. Another problem is occasional (based on annuli counts). Gilbert and Stolt tions are generally based on criteria of disagreement between dental cementum (1970) felt that Maine biologists using tooth molariform tooth eruption and wear. Several analysis and tooth eruption and wear regard­ eruption and wear were inclined to of the less populated states do not operate ing the separation of yearlings from older overestimate the ages of younger deer check stations, but rely on hunters to volun­ animals. This situation is especially (1'11-2'11) and underestimate those 3'11 and tarily submit incisors from harvested animals. troublesome when the techniques are used in older. Bell (1974) noted the same tendency Typically, biologists from these states at­ combination as suggested by Gilbert and among 4 Louisiana biologists. Thomas and tempt to segregate fawns, yearlings, and Stolt (1970) and Kuehn and Karns (1970). The Bandy (1975) also noted a trend toward adults by external incisor characteristics; role of commercial laboratories in providing underestimation of deer 3'11 to 5'11 but high then, incisors from adults and "questionable cementum analysis services to wildlife agen­ accuracy for younger animals. According to yearlings" are sectioned and ages estimated cies needs additional evaluation. This tech­ Kuehn and Karns (1970), use of the tooth by dental cementum analysis. Two of these nique is apparently going through a eruption and wear technique resulted in age states send material to a commercial developmental stage and further improve­ being underestimated in some regions of Min­ laboratory for cementum analysis and 2 use ment may follow the acquisition of more nesota and overestimated in others, in-state facilities. known-age material from various geographic

78 TABLE 5. Procedures for acquiring samples and assigning ages to harvested white-tailed deer in the Midwest. Approximate Sampling Intensity Method of Age Categories State (in Percent) Source of Material Aging" Defined Personnel Illinois 100 Mandatory check stations TEW v,-4V, + College Students Indiana 10 Special hunt check stations TEW '.1,-4'.1, + Professionals Iowa 25 Incisors voluntarily submitted EIC v,,1V',2v,+ Professionals CAC·· 2'.1,-5'.1, + Commercial Lab Kansas 90 Incisors voluntarily submitted nc ';',1v,,2V,+ Professionals CAC·· 2V, . Commercial Lab Michigan 10 Voluntary check stations TEW v"1V,, 2'.1,+ Nonbiologists TEW v, ... Professionals

Minnesota 10 Mandatory registration stations Size '.1,-1 V, + Nonprofessionals Mandatory registration stations TEW v,,1V,,2V,+ College Students Voluntary check stations TEW '.1,,1 '.1,,2';' + Professionals Incisors voluntarily submitted nc V,,1V,,2V,+ Professionals and collected at stations CAC·· Yo. Professionals Missouri 10 Selected mandatory check stations TEW V.. Professionals Nebraska 50 Mandatory check stations TEW '.1,-7'.1, + Professionals Incisors voluntarily submitted EIC '.1,,1 V" 2';' + Professionals Ohio 15 Mandatory check stations TEW V,-5V, + Professionals South Dakota 50 Incisors voluntarily submitted EIC '.1,,1 '.1" 2V, + Professionals CAC·· 2'.1, . Professionals

Wisconsin 10 Selected mandatory check stations TEW Yo, 1'.1" 2V" 3';', Professionals & 4V,-5v"6V,-8V, , Nonprofessionals 9v,-12V" 13';' + • TEW = tooth eruption and wear; flC = external incisor characteristics; CAC = cementum annuli counts. • ·Used to assign yearly ages to adults segregated by other methods.

regions (Gwynn 1978). One approach to the eruption and wear method, i.e, its inherent flashlights, wash bottles, and jaw spreading problem of regional variation in cementum accuracy will not improve. However, better and cheek splitting devices. structure would be for agencies to make their application of the technique can provide Cost and personnel requirements of the own age determinations from commercially more reliable results. Whenever possible, age various aging techniques must be considered prepared slides. Several experienced inter­ assignments should be made from clean, ex­ in addition to their accuracy and precision preters could be assembled to concurrently cised jaws by experienced professionals with (Erickson et al. 1970). Equipment and facility examine sections projected on a screen, with immediate access to representative known­ needs are much less for tooth eruption !nd ready access to known-age slides. age material. Attempts to age jaws in situ wear than for annuli counts, although the Whereas the preparation and interpreta­ should probably be limited to segregating training and experience necessary for proper tion of tooth sections may be a developing fawns, yearlings, and adults. Even then, interpretation. are not. Hunter-supplied in­ art, the same cannot be said for the tooth liberal use should be made of such aids as cisors are the least expensive source of

79 material; but, additional biases are intro­ LITERATURE CITED CONNOlLY, G.E., M.L. DUDZINSKI, and W.M.. duced. Check stations are labor-intensive LONGHURST regardless of whether tooth examinations or ANDERSON. AE. and D.E. MEDIN 1969. An improved age-lens weight regression incisor collections are involved; however, 1969. Antler morphometry in a Colorado mule for black-tailed deer and mule deer. J. deer population. I. Wildl. Manage. Wildl. Manage. 33(3):701-4. they accommodate public relations and col­ 33(3):520-33. lection of additional biological information. COWAN, I. McT. Cost of private annuli counts depends on ARMSTRONG, R.A. 1~36. Distribution and variation in deer (Genus availability of equipment and personnel; but, 1950. Fetal development of the northern white­ Odocoi!eus) of the Pacific Coastal tailed deer (Odocoi!eus virginianus Region of North America. Calif. Fish and commercial rates for histological sectioning borealis Miller). Am. MidI. Nat. Game 22(3):155-246. (Matson 1978) seem competitive, even when 43(3):650-66. compared to less time-consuming procedures DOWNING, R.L. and R.W. WHITTINGTON such as grinding sections (Brown and BELL, B.e. 1964. Eyelens weights - valuable deer manage­ 1974. A comparison of age determination ment tools. Proc. Conf. Southeast. Assoc. Peabody 1972). techniques for white-tailed deer Game and Fish Comm. 18:17-20. Assuming that further research with (Odocoi!eus virginianus Zimmerman) ERICKSON, J.A, A.E. ANDERSON, D.E. MEDIN, and D.e. known-age material improves or confirms the from differing Louisiana soils. La. State BOWDEN dependability of commercial work, the Univ., Baton Rouge. MS Thesis. 156 pp. 1970. Estimating ages of mule deer - an evalua­ following might be an appropriate data col­ BOOZER, R.B. tion of technique accuracy. J. Wildl. lection scheme for many midwestern states. 1969. Cementum annuli versus tooth wear Manage. 34(3):523-31. Check station operators could separate aging of the white-tailed deer of ERICKSON, I.A and W.G. SELIGER fawns, yearlings, and adults by tooth eruption Alabama. Auburn Univ., Auburn. Ala. 1969. Efficient sectioning of incisors for and wear, and then collect adult incisors for PhD Thesis. 102 pp. estimating ages of mule deer. J. Wildl. commercial processing and local age deter­ BROWN, E.R. Manage. 33(2):384-88. mination. This approach would: (1) maintain 1961. The black-tailed deer of western the benefits of check stations, (2) exploit the Washington. Wash. State Game Dep. FRIEND, M. advantages of tooth eruption and wear and BioI. Bull. No. 13. 124 pp. 1967. A review of research concerning eye-lens weight as a criterion of age in animals. bypass the disadvantages, (3) utilize relative­ BROWN, W.S. and W.e. PEABODY N.Y. Fish and Game J. 14(2):152-65. ly cost-efficient commercial preparation of 1972. Modification and field use of the incisor slides, and (4) employ local expertise for their annuli count technique for deer. Pap. 1968. The lens technique. Trans. North Am. interpretation. If necessary, data costs could presented at First Prairie Deer Wildl. and Nat. Resour. Conf. 33:279-98. be further reduced by processing only adult Workshop, Rapid City, S.D. 15 pp. GASAWAY, W.e., D.B. HARKNESS, and R.A. RAUSCH female incisors. Little information would be CAHALANE, V.H. 1978. Accuracy of moose age determinations sacrificed by simply designating males as 1931. Age classes of whitetail bucks killed in from incisor cementum layers. J. Wildi. fawns, yearlings, or adults. northern Michigan in 1929. J. Mammal. Manage. 42(3):558-63. 12(3):285-91. 1932. Age variation in the teeth and skull of GILBERT, F.F. the whitetail deer. Cranbrook Inst. 1966. Aging white-tailed deer by annuli in the Acknowledgments. Appreciation is express­ Science, Sci. Publ. No.2. 14 pp. cementum of the first incisor. J. Wildl. ed to midwestern deer biologists and managers Manage. 30(1 ):200-2. who provided information regarding their pro­ CAMPBELL, G.R. grams and, in some cases, unpublished data from 1967. The dental cementum technique for age GILBERT, F.F. and S.L. STOlT their files. All state agencies contacted responded determination of white-tailed deer positively to requests for information. (Odocoi!eus virginianus). South. III. 1970. Variability in aging Maine white-tailed Univ., Carbondale. MA Thesis. 27 pp. deer by tooth-wear characteristics. J. Wildl. Manage. 34(3):532-35. CAUGHLEY, G. 1976. Wildlife management and the dynamics GWYNN, r.v. of ungulate populations. Pages 183-246 1978. Tooth wear and replacement vs, cemen­ in T.H. Coaker, ed. Applied biology. Vol. tum analysis age assignments of yearling 1. Academic Press, New York. 358 pp. Virginia white-tailed deer. Pap. 1977. Analysis of vertebrate populations. lohn presented at 14th Northeast Deer Study Wiley & Sons, New York. 234 pp. Group meet., Moodus, Conn. 5 pp. 80 HACKETT. E.J., D.C. GUYNN. JR.. and H.A. JACOBSON LAWS. R.M. MclEAN. D.D. 1979. Age structure differences produced by 1952. A new method of age determination for 1936. The replacement of teeth in deer as a annuli aging and wear and replacement mammals. Nature 169(4310):972-73. means of age determination. Calif. Fish aging. Pap. presented at 2nd Southeast and Game 22(1):43-44. Deer Study Group meet., Mississippi LEOPOLD. AS., T. RINEY. R. MCCAIN. and L. TEVIS. JR State, Miss. 10 pp. 1951. The Jawbone deer herd. Calif. Dep. Nat. MOEN. A.N. and P. SAUER Resour. Div. Fish and Game Bull. NO.4. 1977. Population predictions and harvest HAUGEN. AO. and D.W. SPEAKE 139 pp. simulations. Pap. presented at the Joint 1958. Determining age of young fawn white­ LEWALL. E.F. and I. McT. COWAN Northeast-Southeast Deer Study Group tailed deer. J. Wildl. Manage. meet. Blackstone, Va. 36 pp. 22(3):319-21. 1963. Age determination in black-tail deer by degree of ossification of the epiphyseal MORELAND. R. HILL. J.R. plate in the long bones. Can. J. Zool. 1952. A technique for determining age in 1973. Dental annuli for aging deer at the 41(4):629-36. black-tailed deer. Proc. Conf. West. Rachelwood Wildlife Research Preserve. LOCKARD. G.R. Assoc. State Game and Fish Comm. W. Va. Univ., Morgantown. MS Thesis. 1972. Further studies of dental annuli for aging 32:214-19. 39 pp. white-tailed deer. J. Wildl. Manage. OLSON. J.c. HILL. J.R., D.E. SAMUEL. and A WOOLF 36(1 ):46-55. 1967. A study of tooth cementum layers for 1975. Dental annuli for aging deer receiving LONGHURST. W.M. age determination in Indiana white­ supplemental food. W. Va. For. Notes 1964. Evaluation of the eye lens technique for tailed deer and Isle Royal moose. Purdue 4:10-13. aging Columbian black-tailed deer. J. Univ., Lafayette, Ind. MS Thesis. 61 pp. HOFFMAN. R.A. and P.F. ROBINSON WildI. Manage. 28(4):773-84. PARK. B.C. and B.B. DAY 1966. Changes in some endocrine glands of LORD. R.D., JR. 1942. A simplified method for determining the white-tailed deer as affected by season, 1959. The lens as an indicator of age in cotton­ condition of white-tailed deer herds in sex and age. J. Mammal. 47(2):266-80. tail rabbits. J. Wildl. Manage. relation to available forage. U.S. Dep. Agric. Tech. Bull. No. 840. 60 pp. HUDSON. P. and L.G. BROWMAN 23(3):358-60. 1959. Embryonic and fetal development of the 1962. Aging deer and determination of their PORATH. W.R. and O. TORGERSON mule deer. J. Wildl. Manage. nutritional status by the lens technique. 1979. White-tailed deer population measure­ 23(3):295-304. Proc. 1st Natl. White-tailed Deer Dis. ment, harvest analysis and season HUNTER. G.N. Symp. 1:89-93. recommendations. Mo. Dep. Conserv. 1947. Physical characteristics of Colorado Low. W.A and I. McT. COWAN Perf. Rep. Pittman-Robertson Proj. mule deer in relation to their age class. 1963. Age determination of deer by annual W-13-R-32. 76 pp. Colo. Game and Fish Dep. 38 pp. structure of dental cementum. l. Wildl. RANSOM. AB. Manage. 27(3):466-71. JONES. F.L. 1966. Determining age of white-tailed deer 1953. Aging the Inyo mule deer. Proc. Conf. LUDWIG. J.R. from layers in cementum of molars. J. West. Assoc. State Game and Fish 1967. Comparison of age determination Wildl. Manage. 30(1 ):197-99. Comm. 33:209-19. techniques for the white-tailed deer of RICE. L.A southern Illinois. South. III. Univ., Car­ KELLER. C}, and L.M. LANDRY 1977. Age determination of South Dakota deer 1976. Eye lens weight as an age indicator of bondale. MA Thesis. 30 pp. populations by dental cementum layers, white-tailed deer in central . LUDWIG. J.R. and R.W. DAPSON 1973-1976. S.D. Dep. Wildl., Parks, and Trans. Ky. Acad. Sci. 37(3-4):98-102. 1977. Use of insoluble lens proteins to For. Compl. Rep. Pittman-Robertson estimate age in white-tailed deer. J. Proj. W-75-R-19. 46 pp. KUEHN. D.W. Wildl. Manage. 41(2):327-29. 1980. Influences of irregular dental cementum 1970. An evaluation of tooth wear as a layers on aging deer incisors. J. Wildl. LUETH. F.X. criterion for estimating age in white­ Manage. 44{1):266-68. tailed deer. Univ. Minn., St. Paul. MS 1963. A comparison of some aging techniques Thesis. 79 pp. for Alabama deer. Proc. Conf. Southeast. RINEY. T. Assoc. Game and Fish Comm. 17:31-37. 1951. Standard terminology for deer teeth. J. Wildi. Manage. 15(1):99-101. KUEHN. D.W. and P.D. KARNS MATSON. G. 1970. An evaluation of tooth wear as a 1978. The preparation and analysis of tooth ROBINETTE. W.L. and W. JENSEN criterion of aging white-tailed deer. Pap. sections to determine age. Matson's 1950. A simplified method for determining the presented at 32nd Midwest Wildl. Conf., Commercial Microtechnique Newsl., age of mule deer. Utah Dep. Fish and Winnipeg, Manit. 10 pp. Oct. 1978. 5 pp. Game Inf. Bull. NO.1. 4 pp. 81 ROBINETTE, W.L, D,A. JONES, G, ROGERS, and ),5. SCHEFFER, V.B. SEVERINGHAUS, C.W., H.F. MAGUIRE, R.A. GASHWILER 1950, Growth layers on the teeth of Pinnipedia COOKINGHAM, and J.E. TANCK 1957. Notes on tooth development and wear as an indication of age. Science 1950. Variations by age class in the antler for Mountain mule deer. ). Wildl. 112(2907):309-11. beam diameters of white-tailed deer Manage. 21(2):134-53. SCOTT, R.F. related to range conditions. Trans. North Am. WildL Conf. 15:551-70. ROSEBERRY, ).L. and ),W. HARDIN 1954. Population growth and game manage­ 1978. A comparison of alternate methods of ment. Trans. North Am, Wild/. Conf. TABER, R.D. age determination for white-tailed deer. 19:480-504. 1969. Criteria of sex and age. Pages 325-401 in Rep. filed Coop. Wildl. Res, Lab., South. R.H. Giles, )r., ed. Wildlife management SERGEANT, D.E, and D.H. PIMLOTT III. Univ., Carbondale, 23 pp. 1959. Age determination in moose from sec­ techniques. The Wild I. Soc" Washington, D.C. 623 pp. RYEL, L.A., L.D. FAY, and R.C. VAN ETTEN tioned incisor teeth. J. Wildl. Manage. 1961, Validity of age determination in 23(3):315-21. THOMAS, D.C. and P.). BANDY Michigan deer. Mich. Acad. Sci., Arts, 1973. Age determination of wild black-tailed SEVERINGHAUS, C.W. and Lett. 46:289-316, deer from dental annulations. ). Wildl. 1949. Tooth development and wear as criteria Manage. 37(2):232-35. SAUER, P.R. of age in white-tailed deer. ). Wildl. 1971, Tooth sectioning versus tooth wear for Manage. 13(2):195-216. 1975. Accuracy of dental-wear age estimates assigning age to white-tailed deer. Proc. of black-tailed deer. l. Wildl. Manage. SEVERINGHAUS, C.W. and E,L. CHEATUM Northeast Fish and Wildl. Conf. 28:9-20. 39(4):674-78. 1956. Life and times of the white-tailed deer. 1973, Seasonal variation in physiology of Pages 57-186 in W.P. Taylor, ed. The deer URBSTON, D.F. white-tailed deer in relation to cemen­ of North America. The Stackpole Co., 1968, A comparison of several techniques for tum annulus formation. Proc. Northeast Harrisburg, Pa. and Wildl. Manage. lnst., aging white-tailed deer. Va. Polytechnic Fish and Wildl. Conf. 30:281-98. Washington, D.C. 668 pp. Inst., Blacksburg. MS Thesis. 89 pp,

l·/.i· Z ~' ' •. ~: ''; : IJ···.fr /~ ..." J'i YT ~'iJ,~?~t· ,.,.,.".

82 DEER HERD MANAGEMENT - PUTTING IT ALL TOGETHER

William A. Creed Wisconsin Department of Natural Resources Ranger Station, Box 576 Rhinelander, WI 54501 Frank P. Haberland Wisconsin Department of Natural Resources Box 7921 Madison, WI 53707

Abstract: Wisconsin's deer management pro­ short of these ratios when herd levels are above or members appointed by the Governor for stag­ gram employs a system of % deer management below specified goals. Winter severity gered 6-year terms. An important function of units with deer density goals specified for each measurementsand predicted population responses the Board is the establishment of policy for unit. On a statewide basis, the current goals have recently played an increasing role in the guiding DNR programs. The Wisconsin deer prescribe an average overwinter herd of 575,000 quota-setting process. permitting an annual gun harvest of up to 150,000 Antlerless quota recommendations are sub­ management policy states, in part, that deer. Since 1964, annual gun season harvests have jected to annual review at public hearings and by "regulations shall be designed to meet the averaged 105,000 ranging from 71,000 to 151,000. the Conservation Congress, a statewide following objective: to maintain a deer herd Bow harvests during the same period have added sportsmen's group. Recommendations by the Con­ in balance with its range and at population 3,000 to 18,000. gress, as well as by the techni.cal staff of the levels reasonably compatible with Primary inventory methods include: (1) Department of Natural Resources, are considered agricultural and forest management objec­ registered kill trends; (2) sex-age-kill population by the Natural Resources Board in establishing the tives in each deer management unit." Deer final quotas. estimates calculated from kill, ages, and observed management units are areas of similar fawn:doe ratios; and (3) pellet counts in northern habitat bounded by major roads. Currently units from 1955 to 1978, which have recently been INTRODUCTION 2 superceded by deer trail counts. Registered kill there are 96 units averaging about 1500 km 2 trends and population projections therefrom form Details of herd management policy, survey (580 mile ) in size. Each unit is surveyed on air the backbone of the program. Supplementing and design, and harvest regulations may vary photos to determine the area of deer range. complementing the 3 basic inventories are from state to state, but the basic ingredients Deer population goals are expressed as measures of winter severity, hunting pressure, and of a management program pose many deer per square mile of deer range over­ winter losses. similarities. All states must establish objec­ winter. Goals in individual management units Wildlife managers maintain ongoing histories for tives, measure populations, and establish range from a low of 2 to a high of 12 deer per each unit, tabulating annual records of popula­ regulations. The purpose of this paper is to km 2 of range (approximately 5-30 deer per tions, harvests, and hunting pressure. These pro­ describe how one major deer state in the m ile 2) and total 575,000 deer statewide. Deer vide the background for annual harvest recommen­ population goals in the forested deer range dations. Antlerless quotas commonly range from Midwest, Wisconsin, goes about the annual 60 to 80% of the antlered buck kill when popula­ business of regulating its white-tailed deer are based on long-term average carrying tion levels are near goals, but may exceed or fall resource. capacity as determined by unit population responses to past winters of varying severity. HERD MANAGEMENT POLICY Goals in the agricultural range reflect an Creed, William A. and Frank P. Haberland. 1980. Deer herd management - putting it all together. AND GOALS estimate of hunter demand balanced with an Pages 83-88 in Ruth l. Hine and Susan Nehls, eds. assessment of human tolerance to deer White-tailed deer population management in the The Wisconsin Department of Natural numbers, particularly as it relates to crop north central states. Proc. 1979 Symp. North Cent. Resources is governed by the Natural damage and the frequency of deer-vehicle Sect. Wildi. Soc.116 pp. Resources Board which consists of 7 collisions on highways.

83 The overwinter population level of 575,000 is greater. In 1979 these amounts were in­ deer is capable of providing a fall herd in ex­ TABLE 1. Wisconsin deer gun creased to 20¢ and $20. Registration data in­ cess of 800,000 deer which permits an annual harvests, 1964-78. clude breakdowns by sex, age, license type, gun harvest up to 150,000 or more depending county, and management unit. on the effects of winter severity. No. Deer Harvested Deer registration is the cornerstone of our Management policy prescribes annual Antlered Antlerless herd management program. It has convinced hunting seasons to maintain deer population Year Bucks Deer Total most doubters about the reliability of our goal levels and allows the use of buck (3 in. + 1964 65,052 28,393 93,445 deer kill figures, and more important, antler), either-sex, or buck and antlerless per­ 1965 60,994 37,750 98,744 registered buck harvest trends have supplied mit harvest strategies. Buck-only seasons are 1966 67,362 42,700 110,062 one of our most dependable measures of 1967 71,302 57,295 128,597 occasionally used to allow herd increases deer population change. Kills are pinned 1968 62,521 57,465 119,986 where numbers are below goals because of 1969 52,655 45,353 98,008 down to much smaller land units than would the effects of severe winters or continuous 1970 50,308 22,536 72,844 be possible with all but the most elaborate either-sex seasons. Either-sex seasons regular­ 1971 48,994 21,841 70,835 questionnaire schemes. ly apply in the intensive agricultural areas of 1972 49,416 25,411 74,827 southern Wisconsin. Buck hunting combined 1973 57,364 24,741 82,105 Sex-age-kill with antlerless quotas for specific manage­ 1974 67,313 33,092 100,405 ment units is the system applied over most of 1975 73,373 44,005 117,378 Used in Wisconsin since the early 1960's the state. 1976 69,510 52,999 122,509 (Creed 1964), the sex-age-kill method for 1977 82,762 49,148 131,910 The antlerless permit in Wisconsin is called calculating deer densities has become in­ 1978 87,397 63,448 150,845 creasingly important in the statewide the "party permit" because it requires the for­ management picture. It currently serves as mation of a group of at least 4 hunters to sub­ the primary index to deer populations over mit a valid application. The permit entitles most of the state. the party to harvest 1 deer of any age or sex season, in addition to the deer the hunter is Our version of the sex-age-kill method is in addition to the 1 deer allowed on each entitled to shoot during the gun season. modified after Eberhardt (1960), the prime hunter's regular license. In practice, it is an difference being that we employ a 20% antlerless permit because only about 3% of POPULATION INVENTORIES non harvest adjustment for calculating the deer harvested under the permit are populations of adult bucks. Further, we use antlered bucks. The party permit system has Unit deer populations are monitored been in effect continuously beginning with through a variety of surveys adapted to Severinghaus and Maguire's (1955) procedure for estimating adult sex ratios. And, fawn the 1963 deer season. specific range types and harvest problems. Since that time, Wisconsin gun hunters The primary inventory methods include recruitment estimates are obtained from have enjoyed an average annual gun harvest registered kill trends, sex-age-kill population routine July-September field observations by of over 105,000 deer, ranging from a low of estimates, and pellet group surveys in all DNR field personnel. about 71,000 in 1971 to a high of almost northern units. Steps in our analyses are as follows: 151,000 in 1978 (Table 1). These totals in­ (1) Adult buck populations. Minimum clude an average of almost 33,000 party per­ Deer Registration numbers of adult bucks (1'h + years) for a mit deer annually ranging from 12,500 in previous year are determined by totalling 1971 to almost 54,000 in 1978. Wisconsin has a long history (1953 to pre­ the subsequent legal harvest of adults Wisconsin became the first state in the na­ sent) of compulsory deer registration which alive (shown by age structure) in that tion to establish an archery deer season back requires hunters to present their deer to of­ previous year. To correct for non harvest in 1934. Since that time, bowhunter numbers ficial registration stations for inspection and mortality, we assume an 80% final and harvests have steadily increased. Recent tagging. These stations, including DNR of­ recovery rate. This includes an estimated bow deer harvests have increased from about fices, gasoline stations, stores, sheriff's of­ 10% crippling loss plus a 10% adjustment 3,000 in 1964 to over 18,000 in 1978. Licensed fices and similar public places, total about for bowhunting and non hunting losses. archers increased to 158,000 in 1978. Present 450 statewide and are well dispersed across This correction has worked well for heavi­ regulations allow the bowhunter to take any the state. Cooperators have been paid 10¢ per ly hunted units, but has proved inade­ 1 deer during an approximately 85-day deer registered or $10 per season, whichever quate for the wilder, relatively inaccessi­

84 ble units where nonhunting mortality (par­ population inventory for the Northern Forest pansion of replies produces hunter density ticularly winter losses) often exceeds 20%. Region from 1955 to 1978. They also provided estimates for management units by day and (2) Adult sex ratios. These are computed by the initial background for the establishment for the season. procedures described by Severinghaus of unit population goals in 1962. Dead Deer Surveys. From 1955 to 1978, and Maguire (1955). Very briefly, the Our application of the method involved dead deer searches were conducted calculation involves dividing the propor­ sampling 11 to 12 deer management units per simultaneously with pellet surveys. But in tion of yearling bucks by the proportion of year on a 3-year rotation. Estimates of deer some very severe winters, we have also field­ yearling does, after first correcting the per square kilometer of deer range were then checked randomly selected blocks with 5- to yearling buck proportion according to the computed for individual units and also pro­ 10-person crews. Such surveys inherently are male/female ratio derived from shot jected for the entire Northern Forest Region quite imprecise because of the clumped samples of fawns. (ca. 38,500 km 2 of deer habitat). distribution of deer carcasses, but they do (3) Fawn:doe ratios (net recruitment). These The main value of the pellet group survey help to evaluate impacts on regional popula­ ratios are obtained from observed was to establish modern deer density "bench tion levels. numbers of fawns and adult does annually marks". Prior to 1955, density estimates came Analysis of Winter Severity. Measures of reported by Department field personnel primarily from deer drive censuses (Swift winter severity relate to spring deer condition during July, August, a.id September. In 1946). Reliability of pellet surveys on a unit and fawn survival (Verme 1968, 1977). In any-deer season zones, ratios of fawns:doe basis was highly variable, but when projected Wisconsin (B. E. Kohn pers. comm.) a com­ in the legal kill are substituted for ob­ to the entire Region, they correlated closely (r bination of days with 46+ cm snow on the served summer ratios. = 0.83, P < 0.01) with subsequent buck kills ground plus days below -18°C temperatures (4) Buck to total population expansion fac­ (McCaffery 19768) and even better with buck correlated very closely (r = 0.99, P < 0.01) tors. Expansion factors (E. F.) are recom­ kills of the previous fall (r = 0.92, n = 15). with the immediate percentage change in the puted annually for each of 13 zones ac­ Pellet counts lost favor in Wisconsin dur­ buck kill (B. E. Kohn 1979 pers. comm.). This cording to the following formula: ing the 1970's, primarily because of increased severity index also provides an estimate of E. F. = 1.00 + (B/D) + (B/D)F costs, but also because other methods, winter losses and subsequent fawn recruit­ especially sex-age-kill estimates and trail ment. where B = corrected yearling buck pro­ counts, produced similar information at less portion, which is determined from the pro­ cost. Hence, they were discontinued follow­ ANNUAL REVIEW OF portion of yearling bucks in adult buck kill ing the spring survey in 1978. HERD STATUS divided by male/female fawns aged; D = Trail counts (McCaffery 1976b) now proportion of yearling does; and F = substitute in part for pellet surveys. Similar Each deer season begins a reappraisal of fawns:doe ratio from summer observa­ coverage can be achieved for about 1/4 the deer population levels. Buck kills especially tions. cost of pellet survey, and precision is are examined closely because we have found Zone expansion factors are updated an­ somewhat better (ca. ± 15-20%, P = 0.05). that harvest rates are relatively consistent for nually and used to calculate deer popula­ Trail counts are not as yet considered fully a given management unit and thus buck kill tions for all units with adequate age samples. operational in our state because training of trends approximate population changes. At Annual expansion factors are used where field personnel is still underway. But, results this time we also study yearling percentages samples are adequate (200 bucks aged; 200 to date look quite promising. from age samples to determine recruitment. does aged; and 200 doe-fawn observations) Upturns or stability in yearling percentages and long-term factors are used where Supplemental Surveys generally indicate satisfactory recruitment, samples are small. These expansion factors while downturns reflect herd losses since the are then converted to buck kill to total A continued search for better, less expen­ previous deer season. population factors and applied regionally to sive, or less time-consuming methods has led Preliminary kill estimates are available obtain approximate population estimates for to development of a number of surveys soon after the deer season ends and units where aging information is not which play important roles in assessing the computer-summarized data are available by available. current status of deer populations and hunt­ mid-February. ing effort. )ust recently (Creed et al. 1979), we Pellet Surveys developed expansion factors relating long­ Hunter Pressure Poll. A sample of 10,000 term-average buck kill to total population. Deer pellet surveys (Eberhardt and Van Etten license buyers are mailed a questionnaire to These factors are quickly and easily used by 1956, Olson et al. 1955) formed the basic determine where and when they hunted. Ex­ field managers to estimate approximate 85 population levels (Fig. 1). Final estimates are computed by the normal sex-age-kill method in early March. By this time, managers are able to update deer management unit MULTIPLY BUCK KILL/mill TIMES THESE histories which chart harvest, population, and FACTORS TO GET: hunting pressure data (Fig. 2). A. MINIMUM WINTER OENSITY FOR By reviewing and updating management ZONES A-H unit histories, managers quickly determine B. MINIMUM FALL OENSITY FOR the status of current population levels ZONES J-N relative to long-range goals and also assess whether past harvests were consistent with these goals. They are now able to weigh the impacts of antlerless kill rates relative to trends in the buck kill, and adjust their harvest recommendations accordingly. HARVEST RECOMMENDATIONS Deer harvest recommendations are based primarily on 3 considerations: (1) The status of the unit deer population in relation to established population goals (at the goal, or higher or lower) as deter­ mined by inventories previously de­ scribed. (2) The past effect on unit deer herds of vary­ ing antlerless harvest levels gained from 16 years experience with the variable quota system. (3) The estimated impact of the previous winter's severity on deer survival and subsequent fawn recruitment. In the beginning years of variable quota management, it was common to harvest from 15 to 25% of the estimated fall population, with heaviest kills in central and southern units and with buck harvests usually ex­ ceeding antlerless kills. These harvest rates often failed to halt herd growth when recruit­ ment was good, but were occasionally ex­ cessive in years immediately following severe winters. With the lengthening accumulation of kill histories for individual units, we have im­ proved the setting of antlerless harvest quotas, and we have leaned heavily on buck FIGURE 1. Average buck kill: total population expan­ kill as a guide. Through experience it has sion factors by summer observation lone. These are been found that most populations will sus­ based on sex-age-kill estimates. tain an antlerless harvest of from 70 to 80%

86 regular and 2 alternate delegates from each Deer Management Unit History county are elected to staggered 3-year terms Form 1300-74 Department of Natural Resources by the citizens attending the public hearings. Wildlife managers are instructed to meet IM,nagement Unit or County with their local Conservation Congress S4A I members to explain the deer quota recom­ mendations. Gross Area Popu1at; on Goa1 Wisconsin law requires public hearings on 620 SQ. Mi. Deer/SQ. Mi. 2'£ proposed administrative rules (regulations). A Deer Range 431 Sq. Mi. Total Deer /0,77$" public hearing on fish and game regulations is held in each county on the 4th Monday in 1922 1911. 19R 19],! 19..1! 19~ 19z£' 1977 19JfL 1979 April. Citizens .attending vote on the pro­ Spring Population Estimate posals, including deer quotas, and these Fall Population Estimate \S-A-K) ..* votes and other public testimony become 2".0 2'1.8" 3'2.'a '2.9.8 27." 33.s:! 31;.3 Buck Ki11 part of the official hearing record. Wildlife /Ob3 1/'2.0 /400 1572 "01 11;5'0 18/S lOtI. '"?

SEVERINGHAUS, C.W. and H.F. MAGUIRE 1955. Use of age composition data for deter­ mining sex ratios among adult deer. N.Y. Acknowledgments. Thanks are due Keith Fish and Game 2(2):242-46. McCaffery and Bruce Kohn, biologists with the SWIFT, E, Forest Wildlife Research Group, Wisconsin DNR. 1946. A history of Wisconsin deer, Wis, Con­ for use of unpublished data and comments on the serv, Dep. Publ. No. 323, 96 pp, manuscript. We are also grateful to a large number of wildlife managers, technicians, wardens and VERME, L.j, others who over the years have contributed in so 1968. An index of winter weather severity for many ways to the management program we northern deer. j. Wildl. Manage. describe. Portions of the manuscript are derived 32(3):566-74. from studies supported by the Federal Aid in 1977, Assessment of natal mortality in Upper Wildlife Restoration Act under Pittman-Robertson Michigan deer. J. Wildl. Manage. Projects W-79-R and W-141-R. 41(4):700-8. < i/ __

88 POPULATION PHENOMENA - THEIR MANAGEMENT IMPLICATIONS

George E. Burgoyne, Jr. Michigan Department of Natural Resources Box 30026 lansing, MI 46909

Abstract: The use of quantitative techniques improving the performance of decision register is nothing more than a mathematical for summarizing deer data followed close on the makers and planners (Ackoff 1970). This does model of the transactions in a checking ac­ heels of the first concerted efforts at collecting not appear to be true in wildlife manage­ count. Actual cash flows into and out of the data about deer. More elaborate quantitative ment. However, when deer management has checking account can occur in a very dif­ techniques for analyzing such data were soon reached the scale such that several states are ferent order than in the register. In fact, developed to provide additional information not depending upon how well the register is kept, obvious from a qualitative examination of the harvesting well over one hundred thousand data. Such quantitative analysis techniques have deer per year, are serving hundreds of one may have experienced the surprise of become fairly widespread. While these techniques thousands of hunters, and are generating finding that the model is imperfect - when a are an improvement over simple qualitative millions or billions of dollars worth of recrea­ check bounces:­ perusal,they are often limited by considering only tion revenue, it is difficult to argue that this is Certainly it appears that wildlife manage­ a few factors. Mathematical models can be not big business. Why have wildlife managers ment problems are more complex and less developed to include all of the data about a avoided the operations research and well defined than traditional business prob­ population incorporating both well-known and mathematical modelling techniques that lems (Halter et al. 1972). Wildlife managers hypothesized relationships. Such models need not have proven so useful elsewhere? must consider a wide variety of factors when be computer based, but they will provide a better One might argue that wildlife managers, making decisions regarding the management framework for integrating data into deer manage­ ment decisions. More importantly, the models can like most people, are unfamiliar with using of a game population such as deer. The wide provide an explicit framework for objectively in­ models. This can only be partly true since variety of subjects covered in this volume at­ corporating the feedback, the learning experience each of us uses models extensively in running test to the complexity of the decision-making provided by new data which measures the response our everyday lives. The way one drives an process. And yet, deer managers still talk a to recent management decisions. While modelling automobile is based on a conceptual model great deal about deer numbers: numbers in various forms isbeing incorporated in moredeci­ of how the automobile responds to the born, numbers surviving, and quite impor­ sion making, much still remains purely subjective various controls such as steering, braking, tantlv, numbers to harvest. Gabrielson (1951) and poorly documented. In such a system, feed­ and throttle. It is easy to see that there are oc­ pointed out that after considering all the back is much more difficult to incorporate and other factors, the wildlife manager must first managementis much slower to improve its perfor­ casions when this model and the decisions mance. based upon it are far from perfect. Almost and foremost control harvesting by humans everyone has had the shock of applying so that a breeding stock remains. INTRODUCTION power brakes for the first time; one expects a Wildlife managers should take advantage gradual stop and is surprised to have a very of everv available tool for making the task Mathematical modelling has been widely sudden stop. Many have had a more frighten­ less complex. It appears that one very useful incorporated into the business world, greatly ing example of model failure when applying tool, population modelling, has been brakes in a manner suitable for dry pavement neglected although it has the potential to in­ Burgoyne, George E., Jr. 1960. Population only to discover that the pavement is icy. crease the manager's ability to use popula­ phenomena - their management implications. While most of our models are not explicitly tion information. Pages 69-94 in Ruth L. Hine and Susan Nehls, eds. mathematical, some of them are. In fact, It has been suggested that the newness of White-tailed deer population management in the many of us use a mathematical model almost population modelling has held up its use. north central states. Proc. 1979 Symp. North Cent. Sect. Wildt. Soc. 116 pp. every day of our adult lives. A checkbook However, the idea of using population 89 models in population studies and population management decisions (Strickland 1979). problem arises in that there is no requirement management has been with the field from the Meanwhile in the Midwest, Minnesota has that these separate considerations reach con­ very beginning. Through the discussion of the developed its own modelling system and is sistent results. The model's advantage here history of population study by Allee et al. utilizing it as part of the management process then comes from the idea that by explicitly (1949), one can trace conceptual models at (Karns and Snow 1979), and Indiana uses defining a model outside the mind, it is possi­ least as far as Plato (Cole 1957). However, modelling on military reserves (J.e. Olson ble to build in checks for consistency. truly mathematical aspects of population pers. comm.). While much less publicized After the modelling work has become suffi­ study did not appear until the 17th century than the Colorado work, Pennsylvania has ciently sophisticated, one may take advan­ (Eberhardt 1969). developed its own modelling system for deer tage of certain types of mathematical for­ Looking back through the wildlife manage­ management and has been using it since 1966 mulations that permit solving for unknown ment literature, there are many discussions of (Lang and Wood 1976). By now virtually every values very like solving equations in algebra both conceptual models and explicitly state has attended Jack Gross' modelling problems (Wagner 1969). Other, more com­ mathematical models. Trippensee (1948) workshops in Colorado. In addition, 20 states plex, models can be resolved by using trial talked about the subject in his discussion of have received their own copies of the and error to explore a variety of alternatives variations in the numbers of wild animals, as ONEPOP program (T. Pojar pers. comm.). In and then to choose the one which seems best. did Gabrielson (1951) in his discussion of spite of this, only a small number of the While just the process of developing the population controls. The very first article one states have actually incorporated population model can be quite valuable, there is much finds in the first volume of The Journal of modelling in their management decision­ more value to be gained by getting the model Wildlife Management is a mathematical making process (Pojar and Strickland 1979). into actual use in assisting with decision mak­ treatment of pheasant mortalities by Contact with the North Central states reveals ing (Halter et al. 1972). Errington and Hamerstrom (1937). Even that many states hope to incorporate modell­ earlier, in Game Management, a book ing into the management process in the next MODEL DEVELOPMENT wildlife managers might think of as "The few years. However, both Ohio and Wiscon­ Bible" of wildlife management, Leopold sin utilize their own version of the age-sex Perhaps the greatest problem in incor­ (1933) dealt very explicitly with both concep­ method analysis technique developed by porating modelling into the decision making tual and mathematical models. His breeding Eberhardt (1960) while in Michigan. process is that the manager must trust a potential charts and the accompanying model's predictions before basing a decision discussion could not provide a clearer foun­ VALUES OF MODELLING on them. But before one can trust the predic­ dation for the use of mathematical modelling tions, it is necessary to understand the basic in the field of wildlife management. Since we have already concluded that all inner workings of the model (Ackoff 1970). It And yet, over a quarter of a century later, wildlife managers are already using mental sometimes appears that model developers Ruhl (1959) found it necessary to call for the models, why should one pursue the further look upon their model as the "better mouse incorporation of operational research tech­ development of models outside the brain? trap" and expect that users will beat a path to niques based on modelling to improve big Why should one attempt to incorporate their doorway. If more thought and time had game management in the lake states. While mathematical modelling in the management been devoted to the implementation stage the Iiterature contains many population of wildlife populations? there would be fewer models buried in the models, there appears to be very little use of The first benefit to be gained is that the literature or gathering dust on someone's modelling in the actual management of deer person writing down his ideas for a model will shelf. in the United States (Caughley 1976). find that it helps him to clarify and learn Our approach in Michigan has been de­ Perhaps the most widely publicized use of about his own ideas. Secondly, defining a signed to mesh model development with modelling in big game management has been model outside the manager's head makes it needs and use so that the state does not end a result of the work by Jack Gross and his available for other workers to learn from and up developing a purely theoretical model of associates in Colorado (Gross et al. 1973). to improve upon. our deer population that would go unused Pojar (1977) pointed out several examples Often touted as a major benefit of because it was not understood. We are ex­ where the ONEPOP program actually con­ mathematical models is that the models can ploring the value of theoretical models to tributed to the management decision-making handle more factors than a human brain. In contribute to our understanding of the process in Colorado. It appears that Wyom­ fact, the brain can handle a large number of dynamics of the deer populations, and there ing is also committed to utilizing ONEPOP to factors, but it appears to lump them into have been significant findings regarding the improve the use of population data in smaller groups for separate consideration. A information in age structure data (Burgoyne, 90 in press). On the other hand, we are working This is certainly not a drastic new ap­ Hence, the biologist must balance his recom­ with our field biologists to utilize the benefits proach. It seems reasonable to conjecture mended legal harvest against his concept of that the most basic forms of modelling can that deer managers have always had to do illegal losses and wounding losses (C) to provide. While our deer management pro­ something like this, at least informally, in predict the residual population (H). Finally, gram is benefitting immediately, the stage is order to put together their recommendations additional space is provided for recording an being set for the incorporation of more ad­ for harvest levels. There are probably un­ assessment of range conditions (I) and special vanced techniques to further enhance the published compilations similar to this lying considerations (J) regarding management ability to manage deer populations. around in the files of every game manage­ recommendations. Early work in Michigan led to the conclu­ ment agency. The wildlife manager may choose not to sion that 2 beliefs were at fault to a large part Appendix 1 contains the basic outline for fill in the detail if he doesn't believe he has for preventing or delaying the incorporation the summary which Michigan has employed the information, but he must explain how he of population models into the decision­ for the past 3 years. The procedure for filling gets from one step to the next. He must ex­ making process in wildlife management. The out the form is as follows: first, the biologist plain his rationale for his figures. 2 beliefs are: first, that one must be a is required to define the specific area or The biologists' responses to the request to mathematician to understand modelling and population he is describing. * The population fill in this information often fall into 1 of 2 second, that if the modelling works it will described may be a district or part of a categories: (1) the information to complete take away the wildlife manager's role as deci­ district. Then he must estimate the initial that is not available, or (2) why bother with sion maker. population in the spring prior to fawning (A). filling out a form like that since we already Neithe! of these is true but they have been Michigan compiles population estimates do that informally? encountered by people attempting to incor­ from deer pellet group surveys for the It is true that hard, scientific estimates of porate modelling into management in other districts in our major deer-producing areas; some of the losses are not available, but the states. Colorado has been actively pursuing however, the biologists are not restricted to wildlife managers must have some general work in modelling with ONEPOP longer than using these figures. Each biologist may start ideas about the relative magnitude of these many of the rest of the states and yet it ap­ with any figure that he can justify. This could events in order to have some expectation that pears that they are only now turning the cor­ be estimated on the basis of last year's buck there will be a population available to ner on these problems (T. Pojar pers. comm.). kill or total kill, or may be a modification of harvest. The biologist can learn something It appeared that experience with models the pellet survey figure. about his own ideas by working through this would be the only path which would show: (1) Then if the biologist has the information exercise with what information he has that one did not have to be a mathematician available, he may describe the age and sex available and examining the feasible ranges to understand modell ing, and (2) that modell­ composition of the deer population at this for those factors for which he doesn't have ing would not replace biologist decision point (B). Next he must estimate the produc­ good information. makers; it would only provide them a tool. tion of deer fawns, the input of new animals A major benefit of this exercise is that it into the deer population (C). In order to forces the biologist to examine the consisten­ Deer Population Dynamics predict the population size just prior to cy of his ideas about the population being Summary harvest (f), he must estimate the losses bet­ managed. As our biologists put together their ween fawn drop in June and October 1 (D). At best information and ideas about their deer We began an exercise in modelling based this point he has estimated the population populations, they discover that certain com­ on ideas no more complex than adding and from which he will be removing animals by binations of ideas or ideas and data are in­ subtracting from a check deposit book. The his recommendations for legal harvest (F). consistent. For instance, a biologist who approach is perhaps so trivial that it needs However, the idea of wildlife management is believes that his population is reproducing at little explanation and yet it can yield such im­ to leave a sufficient residual population to a high rate but is not growing, must put portant results that it merits further discus­ continue to provide a harvestable surplus and together reasonable values for drains on the sion. yet to fit into habitat and cultural limitations. population due to natural losses, legal hunt­ The idea is to have the biologist put ing losses, and illegal hunting losses in order together the most basic form of life table for to hold the population in check. A biologist is the deer population; combining, from all in­ 'Although I recognize that a biologist may be a doing more than just making the figures formation sources, the manager's best man or woman, I have used only the pronoun "he" reasonable, he is learning about his own ideas estimates of additions to and deletions from rather than "he or she" in the interest of avoiding regarding the relative effects of different the population. cumbersome reading! drains on the population (Halter et al. 1972). 91 However, an important part of the learning Computer Modelling a clear background with which to compare comes when, in the next year, he must again the new information. go through such an exercise for the same A computer-based model can substitute This system has worked quite well in population. In the interim, he has had some for the biologist's calculator and pencil and Michigan. Over the past few years, it has new data such as legal harvest results, paper. It can store many years of data and become an accepted part of the process for population estimates, or observation results. with the biologist's ideas translated into sim­ setting up deer management regulations. The These provide the feedback to help learn ple equations, the computer can quickly process is not entirely smooth. Some whether the population figures assembled recalculate to tryout new ideas. In fact, this biologists have found that it is sometimes the year before were reasonably close. If they is exactly what models such as ONEPOP do easier to try to argue that one cannot sum were not close, it might point him toward (Halter et al. 1972, Gross et al. 1973). losses and gains from a population to get a another idea which would help him come Elaborate computer programs have been reasonable resulting population than it is to closer this year. If he was close, he may be developed to make the computer easier for admit that one's ideas about the losses and able to refine the ideas to do even better this the wildlife manager to use. But the com­ gains are inaccurate or inconsistent. year. puter model is doing nothing more com­ This exercise provides an excellent method Certainly this appears to be an extremely plicated than the biologist was doing before. for learning about a population and the ef­ simplistic exercise. Yet it provides a structure fects of management practices upon it. Since for learning more about the dynamics of a it organizes all the best information about a free-living population about which there is THE MODEL IN population into a simple form, it enhances the chance of incorporating new information some information. Without this written DECISION MAKING record only a sketchy recall of last year's as feedback to help learn from experience. It data and decisions is available to compare This exercise does not end here. Only the provides a clear, concise document which with actual results. That sketchy recall in­ busy-work of certain types of calculations can be valuable for convincing others of the hibits getting full learning experience from has been taken off the biologist. The data wisdom of controversial management deci­ successes and failures. Moreover, this will set must be compiled and processed and the sions, and it builds a framework into which the stage for further development of decisions must be made each year. The new research discoveries and more powerful rmthematlcal modelling in wildlife manage­ biologist has seen that the basic structure of mathematical modelling tools can be incor­ ment. the model is no more complicated than what porated. When the biologist finds that his last year's he can understand. Similarly, he can see that Regardless of the modelling methodology estimates were not as close to what actually the model is a tool to help him understand which an organization is developing, an happened as he would have liked, he must what is happening, not a replacement deci­ ongoing exercise such as this can help bridge generate new ideas in an effort to make the sion maker. As other modelling and field the space between the theoretical model and estimates closer. However, it is time consum­ research identifies promising relationships the management decision process. ing to go back through several years' calcula­ between environmental factors and popula­ tions to find out whether the new idea would tion phenomena, it will be much easier to ex­ have made the previous figures come out plain how such relationships can be incor­ closer to what actually appears to have hap­ porated into the model to further help the pened. When the project has reached this biologist make use of the data. Quite impor­ Acknowledgments. This study was fi­ stage, the biologist can be relieved of the tantly, the simple model provides the nanced in part by the Federal Aid in Wildlife busy-work of the calculations by substituting biologist a precise framework in which new Restoration Act under Pittman-Robertson Project a computer. data or better ideas can be incorporated and FW-3-R.

92 APPENDIX 1: Deer Population Dynamics Summary

E. Herd size October 1,19___ Specific area or population description Male Female ~ulation Fawns A. Spring population before fawns 1-year-olds Pellet survey estimate 2-year-olds Starting population estimate Older Total

F. In-season legal kill B. Herd composition June 1, 19__ Male Female ~ulation Fawns Male Female Population 1-year-olds 1-year-olds 2-year-olds 2-year-olds Older Older Total Total

G. In-season illegal kill and wounding loss

C. Fawn production Male Female ~ulation Fawns 1-year-olds Productivity 2-year-olds rate Fawns Older 1-year-olds Total 2-year-olds Older Total H. Herd size January 1,19__

Male fawns Female fawns Male Female Population Fawns 1-year-olds 2-year-olds Older Total D. losses before October 1

Male Female Population I. Assessment of range conditions relative to herd size Fawns 1-year-olds 2-year-olds J. Special considerations (including crop damage and car-deer collisions). Older Pinpoint the areas affected. Total

93 LITERATURE CITED

ACKOFF, R.L HALTER. A.N., W.M. LONGHURST, G.E. CONNOlLY, and 1970. A concept of corporate planning. Wiley­ F.M. ANDERSON Interscience, New York. 158 pp. 1972. The practical applications of simulation ALLEE, W.e., A.E. EMERSON, O. PARK, T. PARK, and K.P. modeling in wildlife management. Proc. SCHMIDT Conf. West. Assoc. Game and Fish 1949. Principles of animal ecology. W.B. Comm. 52:308-20. Saunders Co., Philadelphia. 837 pp. KARNS, P.D. and W. SNOW BURGOYNE. G.E., JR. 1979. Application of population modeling to In The sex and age structure of a heavily big game in Minnesota. Pages 15-16 in press. exploited ungulate population, in e.w. T.M. Pojar and D. Strickland. A Fowler and T.D. Smith, eds. Dynamics in workshop on the status and application large mammal populations. Wiley­ of big game population modeling. Colo. Interscience, New York. Div. Wildi. 53 pp. CAUGHLEY. G. LANG, LM. and G.W. WOOD 1976. Wildlife management and the dynamics 1976. Manipulation of the Pennsylvania deer of ungulate populations. Pages 183-246 herd. Wildi. Soc. Bull. 4(4):159-66. in T.H. Coaker, ed. Applied biology. Vol. LEOPOLD. A. 1. Academic Press, New York. 358 pp. 1933. Game management. Charles Scribner's COlE. i.c. Sons, New York. 481 pp. 1957. Sketches of general and comparative POlAR, T.M. demography. Pages 1-15 in K.B. Warren, 1977. Use of a population model in big game ed. Population studies: animal ecology management. Proc. Conf. West. Assoc. and demography. Cold Spring Harbor Fish and Wildl. Agencies. 57:82-92. Symp. in Quant. BioI. Vol. 22. New York. POlAR. T.M., and D. STRICKLAND 437 pp. 1979. A workshop on the status and applica­ EBERHARDT. LL tion of big game population modeling. 1960. Estimation of vital characteristics of Colo. Div. Wildl. 53 pp. Michigan deer herds. Mich. Dep. Con­ RUHL, H.D. servo Game Div. Rep. No. 2282. 192 pp. 1959. Big game management in the lake states. 1969. Population analysis. Pages457-95 in R.H. Trans. North Am. Wildl. Conf. 24:472-79. Giles, ed. Wildlife management tech­ niques. The Wildl. Soc., Washington, STRICKLAND. D. D.e. 623 pp. 1979. ONEPOP - implementation and use in Wyoming. Pages T.M. Pojar and ERRINGTON. P.L and F.N. HAMERSTROM, JR 7-10 in D. Strickland. A workshop on the status 1937. The evaluation of nesting losses and and application of big game population juvenile mortality of the ring-necked pheasant. J. Wildl. Manage. 1(1):3-20. modeling. Colo. Div. Wildl. 53 pp. GABRIELSON. I.N. TRIPPENSEE. R.E. 1951. Wildlife management. The MacMillan 1948. Wildlife management. McGraw-Hili 0: Z Company, New York. 274 pp. Book Co., lnc., New York. 479 pp. o c GROSS, J.E., J.E. ROELLE. and G.L WILLIAMS WAGNER. H.M. .;;; 1973. Program ONEPOP and information pro­ 1969. Principles of operations research with c 8 cessor: a systems modeling and com­ applications to managerial decisions. ~ munications project. Colo. Coop. Wildl. Prentice-Hall, Inc., Englewood Cliffs, 3: Res. Prog. Rep. 327 pp. (mimeo) N.J. 937 pp. CONVINCING THE DECISION MAKERS

Merrill l. Petoskey u.s. Department of Agriculture Science and Education Administration Washington, D.C. 20250

Abstract: This paper assumes the final deci­ pie must be firmly convinced, and supportive realistic, achievable, and of interest to a ma­ sion maker is the political appointee(s) or policy­ yourselves, of the biological decision which jority of the publics that we served. making body. Decision making begins with deter­ you are offering. During our discussion on objectives with mination of objectives based first on biological Probably the best way to elaborate on this the Natural Resources Commission, we said situation, then on the social and political implica­ subject is to briefly review the period when I that with the receipt of new funds, habitat ac­ tions. It is important that all levelsof the organiza­ tion are1flvolved in the process and that the level was Chief of the Wildlife Division in tivities properly applied to summer and of knowledge of all participants be as high and as Michigan. I took over the Division during a winter range, p'i'lmarily swamp edges and the broad aspossible. The news media should be given relatively low period in its history. The intolerant tree types (aspen and jack pine), an opportunity to participate in events leading up budget was small, about $1.7 million, most of would enable us to double the preseason to the decision as should the decision makers. which was for employees' salaries; and as the deer herd in about 10 years. Those who make the final decision should realize old axiom goes, "The lower the budget, the Our claim made headlines throughout the the consequences of a wrong decision. lower the morale." Very little was being spent state, the most prominent one being "One in "on the ground" activities for the people Million Deer by 1980". We heard from some When we offer the biological recommen­ who were paying our salaries, the sportsmen of the more vociferous sportsmen - many dations for harvest to Directors, Commis­ of Michigan. However, legislation had just who had opposed a license increase because sioners, and those people having the final been passed which raised the deer license by "there ain't no deer", However, this was our word on the regulations for the deer season, $2.50, $1.50 of which was earmarked for deer objective and all of our wildlife biologists we are recommending them to people who habitat improvement in the northern two­ believed that it was one that could be are political appointees, by one system or thirds of the state. I give full credit for the achieved! And it was. Although the original another. The political selection causes the passage of this legislation to people like target date was 1980, mild winters, and some appointee to become very cautious in affirm­ "Curley" Davenport, Dave Arnold, John habitat improvement resulted in our goal of 1 ing decisions which may bring adverse Bye/ich, and others. As an aside, the term million deer being achieved in the summer of criticism to the appointee and/or the ap­ "deer" would have been better left out of the 1975. pointor. For the purposes of this paper, the "habitat improvement" because any habitat In setting this objective, we had to con­ decision maker{s) are the Natural Resources change is good for some creatures and not so sider the resources involved: the habitat, the Commission, appointed by the Governor and good for others. deer and their reproductive potential, and the the Director, appointed by the Commission. During the first couple of months that I interested user groups. Our habitat biologists To be convincing to the decision maker{s) was Chief, we reviewed with the Commission believed that by using commercial harvest of then, you, your staff, and all your field peo- several objectives we felt were in order for the intolerant types and specific habitat the "new" Wildlife Division. (I say "new" management on an additional 30,000 or because shortly after I became Division head 40,000 acres of key habitat annually, this goal Petoskey, Merrill L. 1980. Convincing the decision we changed the name of the Division from could be achieved. We all knew the makers. Pages 95-97 in Ruth L. Hine and Susan Game to Wildlife to better reflect the reproductive potential of the animal. Deer, Nehls, eds. White-tailed deer population manage­ broadened responsibilities in modern given adequate food during the most of the ment in the north central states. Proc. 1979 Symp. resource management.) In order to be con­ year, do remarkably well in filling their range. North Cent. Sect. Wildl. Soc. 116 pp. vincing, we set goals that we felt were We told the primary interest group, the

95 sportsmen, that with 1 million deer, they After convincing our own biologists, we range goal, we were often forced to make could expect a legal harvest annually of next had to try and convince the sportsmen. recommendations prematurely, not for 200,000 animals. A kill of this magnitude Some of them, particularly the more vocal biological reasons, but for administrative raised quite a few eyebrows, but we knew ones, did not believe, and still do not believe, ones. Acquisition of paper and printing of the what the animal can do given adequate food that you can harvest more deer and still law digest too often forced a decision before and cover. maintain a stable, high population. I will not all of the information had been gathered and To be convincing to the public, we had to go into the oft-repeated, anti-doe hunters' analyzed. It's amazing how, in bureaucracy, first be convinced ourselves. Consequently, claim that for every doe you shoot you are the tail often wags the dog. I hope this prob­ in making these decisions on habitat and on losing 2 animals, etc., ad infinitum. (That's lem has now been solved in Michigan. harvest, we involved the people who were one advantage to being employed in The policy makers wanted an early deci­ closely allied to the resource itself ­ Washington. I am spared the advice that is so sion so they could give it long consideration. biologists, officers, and many other field peo­ freely offered by the proponents of bucks­ We tried to delay as long as possible to be ple - in developing the final recommenda­ only seasons.) I am a firm believer in the sure that our offerings were sound. In reality, tions or, as we are saying here, the decision. "silent majority"! If you are open with your the final decision should have been made in We faced several challenges within the facts to all groups, their support is yours. A July or August, but it was usually demanded DNR. First, many of our younger biologists resource manager must recommend the best in Mayor June. However, these are the short­ had not lived through the 40's, 50's, or early for the resource and people. The "aginers" comings of the public process. 60's when Michigan's deer herd probably are always there and always will be, but darn Needless to say, the "deer meeting", as the numbered in excess of 3 or 4 million. Without few managers are removed because they are June Commission meeting came to be known, having the first-hand experience of seeing the right. was always an interesting one. We would reproductive potential of a well-fed deer Deer are of interest to many, many people make our presentation, hopefully as objec­ herd, these younger biologists needed con­ - not all of whom are hunters. Therefore we tively as possible, and give our recommenda­ vincing that the harvest could be increased as tried to keep our information on the status of tions for harvest. The meeting was heavily at­ the range improved. Secondly, some of our the herd flowing constantly from deer season tended by politicians, sportsmen, newspaper biologists were inclined to be a bit provincial to deer season. We gave our information to people, and those representing radio and and seemed to want to manage the animals all forms of the mass media to reach as many television. as individuals in the local situation. We tried publics as possible. Weekly papers in the Members of the Michigan United Conser­ to point out the shortcomings of this kind of rural areas, the big urban dailies, and radio vation Clubs, with a membership of over management by working together and look­ and television outlets all figured in the pro­ 100,000, would always support the biological ing at different situations to encourage cess. Field and staff people were on the road, decision. Members of the Michigan Deer broader thinking. going to meetings, helping our publics better Hunter's Association and the Northern To meet these challenges and explain the understand plant succession and the animal Michigan Sportsmen's Association (both pur­ rationale for our 1980 goal, we attempted to products, deer and other wildlife. We worked porting to represent about 3,000 members) enumerate the alternatives. With the severity hard to inform for understanding. would always criticize it. As a public servant, of winter in Michigan in some years, In the decision-making process, we tried to it is somewhat amazing to me that underharvest can be serious. In addition to eliminate as much opinion, both our own and spokesmen for a few people seemingly carry losing animals, the carrying capacity of the the vocal minority's, from the decision as we the same weight in a public meeting as range is greatly reduced. The loss of habitat could. Facts, cold, hard facts, were the impor­ spokesmen for many thousands of people. Of has a serious detrimental effect on the tant thing. It was our feeling that the decision course, the few, if supported by members of reproductive potential of the herd. offered to the decision makers (political ap­ the House of Representatives Appropriations My advice was - never gamble with the pointees) should be as purely biological as Committee, seem to carry a somewhat larger range because the effects and the rebuilding possible, realizing that politicians are subject stick. process are long term. If you are not sure of to a variety of pressures and are inclined to Usually, the anti's were able to force the the harvest recommendations, it is better to compromise under such pressure. This was Commission into a compromise, to the detri­ be liberal than conservative because, with the big question: Would the politicians let us ment of the majority of people, the range, good range conditions, the animal will do it? and the deer. Such compromises are totally bounce back in a short time. The range does When it came to the year-to-year decisions irresponsible, and have resulted in the loss of not have this potential for quick response. that had to be made to implement our long- countless thousands of animals, plus unholy

96 deterioration of the habitat. It is unfortunate Therefore, following the decision, we kept deterioration of precious range. that the millions of Michigan people who are the people informed through press, radio. and In summary, field and staff together set an­ also served by the DNR were unrecognized television of the results of the harvest from nual harvest regulations pointed toward a by the decision makers. However, few politi­ the several annual deer seasons. Currently, 1980 goal. We informed our interest groups cians or others are willing to accept the and for the past several years, there has been of the objective and the reasons leading to it. blame for such compromises. It is easier to an archery season of 3 months, a rifle season We told them what would and would not blame the DNR. of 16 days, and a muzzle-loader season of 10 happen and then after the decision was made I have found this and similar situations to days. and carried out, we told them what did hap­ be encouraging over time. The usually silent A variety of surveys following the rifle pen when the seasonsand winters were over. citizen, given the facts and an opportunity to season give an immediate estimate of the A history of this kind of procedure is convinc­ participate in a decision, most often makes harvest which, r am pleased to say, was over ing, to say the least. You must believe in what the right choice. Most resource departments 171,000 animals in 1978. Not quite as much you are doing and stand up to be sure that it are quite free of partisan politics and the as we predicted in 1970, but very close to our is done. Someday, such convincing history "spoils" system. Citizens have decided this. prediction and would have been higher can strengthen the decision-making process The "bottle bill", which was recently passed without compromise. However, if this record and help others who must make the final in Michigan and is doing much towards harvest is coupled with the measured starva­ decision. reducing litter, was decided by a referendum. tion losses of over 100,000, it makes our These are good reasons to keep your publics prediction of the early 70's look pretty good. informed on tnefects and what you plan to The sad part' is the waste .of animals for do. recreation and food, and most of all,

~I;; -~ 'Ii ~~ S ---c....::.::::::::~.:, \ »->: ~ r: '\ ~~ ~ "~-''- " to

97

SELLING MANAGEMENT TO THE DEER HUNTER

John Madson 2 Mullen Lane Godfrey, IA 62035

THE CLIMATE always ill-advised. Unless, of course, such ac­ they may reflect approval, disapproval, or of­ tions have been necessitated by some natural fer suggestions. Department personnel ex­ All state wildlife agencies have deer calamity that the public can readily associate plain and justify their recommendations-and management problems that result from lack with the need for short-notice innovations. more often than not, the various factions find of acceptance of recommended practices. Ideally, long-range shifts in basic manage­ common ground for agreement. This is of 2 general types: the unwillingness or ment are prefaced with carefully planned Several years ago in Illinois we ex­ inatlrtityqf i~.andownertoprovide habitat public relations strategies-although there's a perienced a.severe public relations problem diversity, and the reluctance of.certain rather' consistent tendericv in conservation with the issuance of deer hunting permits. publics to accept management innovations. departments to never patch the roof until it There was much bitterness about the alloca­ Almost all of the latter-which is what con­ begins to rain. tion of these licenses and their alleged use as cerns us here-is likely to emanate from The anticipation of a management prob­ political favors. An Illinois Deer Task Force hunters. The primary public that must be sold lem is part empirical and part intuitive. The was formed-composed of members of the on deer management innovations is, as agency must weigh the potential problem in conservation department, hunters, and other always, the rank-and-file deer hunter. terms of experience: what responses, from citizens. It was successful in establishing a Every deer state has its distinct climate of what quarters, were evoked by analagous in­ more equitable system of permit allocation. hunter-opinion and mood. This is the result of novations in the recent past? What action In Pennsylvania a new organization, the regional and local traditions, the political was taken, with what effect? Since then, have Pennsylvania Deer Association, has been health of the conservation agency, and the there been any shifts in public attitudes that recently formed of biologists, educators, historic relationships between hunters and could influence the issue today? naturalists, hunters, administrators, and other the agency. A public relations campaign to concerned citizens. It is designed to represent promote a management innovation must be CITIZEN ADVISORY GROUPS all technical and nontechnical factions with adapted, to some degree, to these historic interest in the Pennsylvania deer herd. One way to assess the climate of public relationships and traditions. The nuts and Management proposals and recommenda­ opinion is to solicit the response of a citizen bolts of the campaign are disarmingly simple tions are made, evaluated, and modified if group formed specifically as a sounding­ in principle. What is far less simple is the exe­ necessary. Beyond the fact that certain board. For 45 years, Wisconsin has used a cution therof. laymen act as efficient sounding boards of Conservation Congress as an advisory body. Much of the success in introducing deer public attitudes, such groups are of value as April hearings are held in each Wisconsin management innovations or changes in an­ political buffers. And maybe-just maybe­ county to air proposed regulations and other nual regulations lies in anticipating any prob­ there may be thoughtful, experienced laymen actions. People attending these hearings lems well in advance of action. Overnight who know something we don't know. crash programs to persuade the public are elect delegates to attend a statewide meeting held the weekend after Memorial Day. An PROCEDURES FOR CHANGE average of about 3 persons per county attend Madson, John. 1980. Selling management to the deer hunter. Pages 99-102 in Ruth L. H ine and this annual meeting as part of the Wisconsin But with or without such citizen advisory Susan Nehls, eds. White-tailed deer population Conservation Congress. They contribute a groups, the deer manager and his or her ad­ management in the north central states. Proc. 1979 general response to the agency's proposed ministrator are certain-sooner or later-to Symp. North Cent. Sect. Wildl. Soc. 116 pp. regulations and management innovations- undertake controversial management innova­

99 tions. The most common action of that sort is Although he was at the bottom of the office DIRECT PUBLIC CONTACT likely to be a particular restriction or re­ hierarchy, he proved to be a valuable public adjustment in deer hunting regulations. influence.) Few public relations strategies are more ef­ Let's assume such a change is being con­ And again-have similar innovations been fective than direct contact with the public, templated, and that there's little doubt it will rejected by the public? Why? When? Who led This can take several forms: open public be met with significant opposition. What the opposition? What were the results? hearings, presentations to specific groups, general course of action should be taken? To From start to finish, a good I&E section can and one-to-one contact with key individuals. begin with, such an innovation must always be the deer biologist's best friend, Bring the Open public meetings will" include the be based on good data, on findings carefully communications specialists in at the very landowners, naturalists, environmental ac­ gathered and competently assessed. This is beginning and map strategy. Develop a tivists, and some hunters, while the sports­ the most critical step in developing any media contact program. Schedule depart­ men's club contact focuses on hunters. change in current management. Nothing is mental news releases and magazine articles Beyond these are the individual "engines of more important. Furthermore, the innovation in a well-developed sequence. public opinion"-the influential local citizens. must fall within the parameters of estab­ A good I&E section maintains close per­ They may be ministers, doctors, lawyers, lished departmental policy. sonal contact with the key media-and par­ businessmen, sporting goods dealers, barbers In the initial stages, care must be taken to ticularly newspaper editors and outdoor or bartenders. overlook none of the obvious fundamentals. writers. Missouri is an excellent example of Every community has persons who exert The first steps in any management innovation this. Jim Keefe and his people cultivate con­ considerable influence on resource manage­ are critical, for they are the basis of all subse­ tacts in the Missouri Daily Press Association, ment attitudes. These key local contacts are quent effort. For example, any new regula­ meetings of that association are attended important even though they may have fought tions must be drafted with great care. I was faithfully, and close liaison is maintained your programs in the past. Over the years I've with the Iowa Conservation Commission with key editors and writers. This is the front known several local "engines of public when we held Iowa's first modern deer line of public relations and every effort must opinion" who habitually opposed departmen­ season in 1953-a 5-day, any-deer season with be made to cultivate a solid working relation­ tal policy because they had been alienated a daily bag and possession limit of 1 deer. ship with the press. Finally, there is an axiom by what they construed as arrogant, However, we neglected to set a season bag that all game managers should have tattoed overbearing behavior by public servants. limit. Several hunters killed 5 deer that week, inside their eyelids: never get into a fight with What they appeared to want more than taking 1 deer each day and legally disposing anyone who buys ink by the barrel! anything else was a measure of recognition­ of it before they went out again the next Start your public relations campaign at a and were bent on raising hell until they got it. morning. From the beginning, some modest and reasonable level-especially if Some patient, personal attention may work newspapers had been sure that outrageous the management innovation isn't initially wonders with such critics. deer season would wipe out Iowa whitetails controversial. But if and when it begins to Whatever the reason for opposition to once and for all-and their fears were con­ heat up, stand ready to call in your shock management, a personal confrontation with firmed by a Conservation Commission that troops. You should have a list of key allies on a particular editor, legislator, landowner, had given tacit permission for every hunter to whom you can count all the way-zealous resort owner or sportsmen's club president kill 5 deer! We were quite awhile digging out friends who'll throw their full weight into the can often have astonishing results. Who from under that one. fight at a moment's notice. It is important should make these contacts? Early in the process, everyone in the that you keep in regular touch with such peo­ If it's a key issue being blocked by a key department must be made aware of the pro­ ple, keeping them up to date on basic critic, the conservation director should make posed change, its rationale, and its implica­ developments. They should hear from you the contact. Even if resistance to the innova­ tions, Don't keep anyone in the dark. En­ frequently, and not just when you're in a: jam. tion is widespread, it would be wise for the forcement and I&E must be in on the ground Use them sparingly and only when really director to hit the sawdust trail. Because floor, of course, but effort should be made to necessary, and don't cry wolf. However, if it's what we're talking about here is active apprise all sections and divisions of the pro­ a choice between being undergunned' or politicking-and there's no substitute for posed change and the good reasons for it. overgunned, take the latter course. Always stumping the electorate. For one thing, it's (Doug Gilbert gives an example of a janitor in use enough gun-but not too much. Again, the flattering to a club, legislator, or newspaper a U.S. Forest Service office who happened to competent I&E section will know what publisher to be called upon by the head of be commander of the local VFW post. caliber of public relations to employ. the agency. The director may make more political hay in one hour with a publisher or end, there should be certain public relations ticular, on good deer management is most ef­ editor than a biologist could accomplish with perennials-materials that can serve from fective if that public is convinced there's an all-day seminar. Not only that, it is often year to year rather than just during annual something in it for them. Your best public the director's immediate responsibility, for campaigns. Good motion pictures are the relations, obviously, is persuading the hunter the real I&E chief of any conservation depart­ best example of this. that your proposed regulations or manage­ ment is, or should be, the director. Missouri has met this continuing need with ment innovation will benefit hunting by The biologist's prime function is to provide a series of superb motion pictures covering benefiting the deer resource. And the only his or her chief with the best field data possi­ the life histories, biology, and management way to execute that persuasion is with ble. It is up to the boss to translate that infor­ of certain major game species: cottontails, honesty, simplicity, and directness. The con­ mation into management decisions that bobwhites, , doves, turkeys, and verse of that is the worse possible public rela­ make biopolitical sense. If the director makes Canada geese. This is long-range public rela­ tions: duplicity, complexity, and evasion. Un­ a crucial management decision with the use tions at its best. Yet, I've always been fortunately, those elements have prevailed in of your data, he or she is the logical person to puzzled by the lack of a Missouri film on too many resource conservation programs. explain and justify that decision to the key deer. Obviously, the decision makers have publics in the state. One way of doing so is by not felt such a film is necessary. The official THE UPSHOT heading, say, a 3-person team that includes a wisdom must be that Missouri deer problems uniformed biologist standing ready to re­ do not warrant a major film effort-and they In his book "A Voice For Wildlife", inforce the director's statements with easily should know. However, the day will surely biologist Victor Scheffer commented: "Dur­ understood material, and a uniformed I&E come when a good deer fiJm is needed in the ing its 40 years of existence, professionat person to conduct a compact, well-organized excellent Missouri film series. As a matter of management has been weakened by in­ visual program. fact, I can't recall ever seeing what I consider breeding; in this respect it resembles the pro­ Never underestimate the effectiveness of a to be a truly effective motion picture on life fessions of education and medicine. The con­ good slide presentation. Most slide shows and times of the white-tailed deer. sequences are narrow vision, resistance to aren't much good. But there should be peo­ change, emphasis on structure at the expense ple in your I&E section who can provide top­ THE WRITTEN WORD of broad helpfulness, and a dwindling sense quality slides and advise you in making of humility." All of these shortcomings are smooth, tight narrations. In the field or in an Running a close second in perennial public dangerous, but in public relations terms none auditorium, keep the basic message simple. relations value is a comprehensive paper­ is more serious than that "dwindling sense of Always remember: the capacity of the back deer publication. The most popular in humility." Nothing can be more effective in average listener to be confused approaches our Winchester-Western series of booklets on hardening public resistance to management infinity-and so does the capacity of the major American game species has been "The innovations. average biologist to be confusing. Don't wear White-Tailed Deer", which now sells at its Yet, in broad terms of public relations with out your audience; don't tell them more than production cost of $2. It has gone through 6 the rank-and-file hunter, we have certain ad­ they want to know. printings, and we are about to order a 7th. vantages today that are greater than ever I'd also like to put in a good word for the Like the Missouri wildlife films, our deer before. More and more hunters (not just the show-and-tell field trip. Like the slide presen­ booklet is a popular translation of technical leaders, but the rank-and-file as well) are tation, it can be a boring flop that does more biological and management materials, and realizing that one of the best basic supports harm than good. It can also be immensely ef­ there's a never-ending public appetite for of their act of hunting is the fact that it's con­ fective if it offers some graphic demonstra­ such things. ducted in concert with professional game tions and if it is properly planned and Like it or not, we're all in the business of management. Without such game manage­ publicized. Field trips should focus on the product promotion just as surely as any ment we would have nothing but emotion most dramatic and easily understood situa­ toothpaste company. And there's something with which to counter emotion-and that's a tions. For lay appeal, no computer printout to learn from the toothpaste commercials. contest I do not think we could win. Our can possibly compare with a high browse line Some of the most effective are ones that modern systems of hunting are wholly depen­ or some necrotic jawbones. don't tout their product for the good of the dent on competent game management and If there's one thing we can be sure of in resource, i.e. good dental health, but because its support by the hunting public. Such game management, it's the necessity for it will improve our love lives. Selling the management is the only solid ground from perpetual repetition of fundamentals. To this public in general, and the hunter in par­ which we can defend the sport of hunting­

1Jl and the lack of it is the only solid ground precede and supersede any public relations that point on, the most basic and effective from which our enemies can attack. program. There may be an exception to that: public relations result from the good biology Convincing the hunter of this is the when an intense public relations effort is re­ they helped create. To think otherwise is to strongest kind of public relations. And again­ quired to establish a solid, professional game build a house from the attic down. first, last, and always-good biology must management system in the first place. From

...;~'..'' .' '.'., .I ,/1 .' ...... >+.':

102 DEER MANAGEMENT FOR WHAT?

Edward E. langenau, Jr. Rose lake Wildlife Research Center 8562 East Stoll Road East lansing, MI 48823 deer hunter in Michigan (Langenau 1979b). INTRODUCTION TOPICS REVIEWED These recreational uses appeared to be com­ Sponsors of the symposium requested that Research on Public Groups patible in the field because of the seasonal nature of deer hunting and the renewable some information on public aspects of deer Deer Hunters. Review of literature on the nature of the resource. More (1979) has management be included. In response, a characteristics of deer hunters showed that thoroughly discussed policy issues concern­ lengthy literature review was prepared on these individuals are not much different than ing nonconsumptive wildlife recreation. .peeple and deef (laogenau 19J9.i) and was the general public (Hendee and Potter 1976). Violators. In general, violators are younger submitted as a working paper. This review is despite stereotypes that they are less and more often of local, rather than urban, available upon request to the Rose Lake educated, rural, and blue-collar workers. residence (Melynk 1978). They are also nor­ Wildlife Research Center. The review in­ Research has also suggested that the number mally the more proficient and skilled hunters cluded 57 pages of text plus an 11-page of hunters will increase in the near future but (Jackson et al. 1979). The chance of getting bibliography and could not be reproduced in that the quality of hunters will probably caught taking a deer illegally in Michigan was its entirety in these proceedings. The purpose decline (Langenau and Mellon 1980). shown to be 1 in 200 (Stoll and Hussain 1979). of the current paper is to summarize the Landowners. About 58% of the big game literature review and to highlight key results Research has suggested that a primary effect hunting in the United States during 1975 oc­ of illegal activity is the reduction of satisfac­ of research on public aspects of white-tailed curred on private land (U.S. Fish and Wildlife deer management. tion among other hunters (Purol et al. 1978). Service 1977). Much of this hunting included Anti-hunters. Research has shown that the landowner and friends or relatives of the 30-60% of the public is opposed to hunting, INTEGRATING PRINCIPLE landowner. The percentage of landowners depending on the region of the country and who hunt deer has increased during the past wording of the question (Shaw et al. 1978, Literature on people and deer is considered 20 years (Evans 1979). It was concluded that Kellert 1978). A notable finding has been that in the context of a "multiple-satisfactions" state wildlife agencies may be over­ there is not a large group of undecided in­ policy (Potter et al. 1973). This approach identifying with the hunter who does not own dividuals to be swayed in either direction recognizes that a wide variety of satisfactions land. (Applegate 1975). Characteristics of people are derived from wildlife-oriented recreation, Literature on the trespass problem was also (Applegate 1973a), levels of wildlife know­ and that these satisfactions are measurable summarized, with the striking finding that the ledge (Wywialowski 1977), and basic percep­ and can be used to establish goals and ratio of trespassers to hunters given permis­ tions that people have about wildlife (Kellert evaluate wildlife programs (Hendee 1974). sion has been as high as 1.3 to 1 (Brown and 1978) are strongly related to anti-hunting sen­ Thompson 1976). The granting of permission timent. for strangers to hunt on private land is related to crop damage from deer (Queal 1968). Management Preferences of langenau, Edward E., Jr. 1980. Deer management However, landowners have tolerated substan­ Deer Hunters for what? Pages 103-5 in Ruth l. Hine and Susan tial losses in exchange for the presence of Nehls, eds. White-tailed deer population manage­ deer (Brown et al. 1978). The limitations and uses of opinion polls ment in the north central states. Proc. 1979 Svrnp. Nonconsumptive Deer Enthusiasts. There were discussed at some length in the North Cent. Sect. Wild/. Soc. 116 pp. are 3 non consumptive deer users for every literature review. Results of research on the

103 attitudes of deer hunters were summarized Deer Hunting Satisfaction MANAGEMENT IMPLICATIONS on several issues: paying for the right to hunt, , introduction of exotic The dimensions of satisfaction in hunting A significant number of general and species, scientific wildlife management, con­ have been thoroughly studied. Commonly specific recommendations were made in the trol of hunter densities, and antlerless deer reported factors have been nature, escapism, literature review. It was concluded that hunting. ingroup companionship, shooting, skill, sociological data have an important role in vicariousness, trophy display, harvest, equip­ deer management but that these data should ment, outgroup verbal contact, and outgroup be collected and used together with Vehicle-Deer Collisions visual contact (Potter et al. 1973). There has biological data. The consequences of using Biological and economic aspects of this been some disagreement as to the impor­ and not using research findings on public problem were presented in the review. tance of kill. Some authors have found killing behavior in program planning were dis­ Vehicle-deer collisions have significant deer to be unrelated to satisfaction (Kennedy cussed. It was argued that an outline for ac­ economic impact, with cost estimates of 1974), and others report a need for some tion must be formulated to apply this type of about $10 million annually just for the State minimum probability of success (Stankey et information. Specific recommendations were of Michigan (Hansen 1978). People who have al. 1973). made to managers and researchers on how to been involved in accidents with deer do not proceed with the management of deer for public benefits. want less deer. They do think that deer are Economic Aspects of more abundant than those not involved in vehicle-deer collisions. Deer Management Acknowledgment. Supported in part by the Federal Aid in Wildlife Restoration Act under Wildlife economics is a rapidly expanding Pittman-Robertson Project W-117-R. Deer Hunting Accidents discipline, evidenced by the review of 691 references by Leitch and Scott (1977). The Deer hunting is safer than many other out­ most commonly used figure for the value of a door recreations and 3.7 times safer than driv­ day of big game hunting is $50 (Horvath LITERATURE CITED ing to and from deer hunting sites (RyeI1973). 1974), although estimates vary by method ApPLEGATE. I.E. Results of research on hunting accidents (Hoover 1976). Economists are also beginning 1973a. Some factors associated with attitudes have shown that being hit with stray rounds to evaluate deer management programs with toward deer hunting in New Jersey was not very common: muzzle to victim benefit-cost and cost-effectiveness analyses residents. Trans. North Am. Wildl. and distances were usually 10 yd or less, most (Hansen 1978). Nat. Resour. Conf. 38:267-73. hunters were shot by members of their own 1973b. Deer and the people of New Jersey. N.J. party, and many hunters shot themselves Outdoors 23(8-9):3-9. (Kerrik et al. 1978). Studies have shown no 1975. Attitudes toward deer hunting in New difference in accident rates under bucks-only Communication and Education Jersey: a second look. WildI. Soc. Bull. or either-sex regulations and there has been 3(1 ):3-f>. Research findings iridicate that deer no statistically significant relationship be­ hunters lack basic knowledge about wildlife BENNETT, Clo, JR., G,E. BURGOYNE, JR.• l.L. COOK, J.P. tween accident rate and hunter densities management (Watson et al. 1972), but that DUVENDECK, E.M. HARGER, E.E. lANGENAU, JR., R.J. (Jenkins 1960). MORAN, and i.c. VISSER the general public is even less informed 1980. Michigan deer habitat study: preliminary (Applegate 1973b). Anti-hunters have less results, Trans, North Am. Wild/. and Nat. Social Impacts of Deer Habitat knowledge about the environment than Resour. Conf. (In Press) nonhunters and hunters (Dahlgren et al. Improvement BROWN, Ll... D,J. DECKER. and c.P. DAWSON 1977). 1978. Willingnessof NewYorkfarmers to incur The advent of Environmental Impact The literature review summarized findings white-tailed deer damage. Wildl. Soc, Statements has increased the need for on media use and potential for education of Bull. 6(4):235-39. research on social impacts of management. target groups. Education of hunters in the BROWN. Lt, and D.A. THOMPSON Studies of trail development (Thomas et al. field through interpretation of population 1976. Changes in posting and landowners at­ 1976) and c1earcutting (Bennett et al. 1980) changes and with management demonstra­ titudes in New York State, 1963-1974. have shown positive hunter response. tion areas was also stressed. N.Y. Fish and Game J. 23(2):101-37. 104 DAHLGREN, R.B., A, WYWIALOWSKI, T.A. BUBOLZ. and KENNEDY, J.J. QUEAl.L.M. V.L. WRIGHT 1974. Attitudes and behavior of deer hunters in 1968. Attitudes of landowners toward deer in 1977. Influence of knowledge of wildlife a Maryland forest. J. Wildl. Manage. southern Michigan, 1960 and 1965. Mich. management principles and behavior 38(1 ):1-8. Acad. Sci., Arts, and Lett. 53:51-72. and attitudes toward resource issues. KERRIK. J.N., P.T. BROMLEY, and R.G. ODERWALD Trans. North Am. Wildl. and Nat. Resour. RYEL, L.A. 1978. Analysis of hunting accidents in Virginia. Conf. 42:146-55. 1973. Notes on hunting accident rates. Mich. Proc. Conf. Southeast. Assoc. Fish and Dep. Nat. Resour. Surv. and Stat. Servo EVANS. R.A. Wildl. Agencies 32:840-49. Rep. No. 131. 6 pp. 1979. Changes in landowner attitudes toward LANGENAU, E.E., JR deer and deer hunters in southern SHAW, W.W., E.H. CARPENTER.L.M, ARTHUR, R.L. GUM, 1979a. Human dimensions in the management Michigan, 1960 to 1978. Univ. Michigan, of white-tailed deer: a review of con­ and D.J. WITTER Ann Arbor. MS Thesis. 45 pp. 1978. The American disposition toward hunt­ cepts and literature. Mich. Dep. Nat. ing in 1976. Wildl. Soc. Bull. 6(1):33-35. HANSEN, e.S. Resour. Wildl. Div. Rep. No. 2846. 68 pp, 1978. Social costs of Michigan's deer habitat 1979b. Non-consumptive uses of the Michigan STANKEY, G.H., R.e. LUCAS. and R.R. REAM improvement program. Mich. Dep. Nat. deer herd. J. Wildl. Manage. 43(3):620-25. Resour. Wildl. Div. Rep. No. 2808.61 pp. 1973. Relationships between hunting success LANGENAU. E.E., JR and P:M. MElLON and satisfaction. Trans. North Am. WildI. HENDEE, J.e. 1980. Characteristics and behaviors of and Nat. Resour. Conf. 38:234-42. 1974. A multiple-satisfaction approach to Michigan 12- to 18-year-old hunters. J. game management. Wildl. Soc. Bull. Wildl. Manage. 44(1):69-78. STOLL, G.D. and N.G. HUSSAIN 2(3):104-13, LEITCH, J.A. and D.F. SCOTT 1979. An estimation of illegal deer losses in HENDEE. j.c. and D.R. POTTER 1977. A selected annotated bibliography of Michigan. Mich. Dep. Nat. Resour. Prog. 1976. Hunters and hunting: management im­ economic values of fish and wildlife and Rep. Pittman-Robertson Proj. W-121-R-3. plications of research. U.S. For. Servo their habitats. N.D. Agric. Exp. Stn. Misc. 27 pp. Gen. Tech. Rep. No. SE-9. 25 pp. Rep. No. 27. 132 pp. HOOVER, R.L. MElYNK, M.J. THOMAS, J.W., J.D. cu, i.c. PACK, W.M. HEALY. and 1976. Incorporating fish and wildlife values in 1978. An annotated bibliography on the social H.R. SANDERSON land use planning. Trans. North Am. and behavioral aspects of wildlife law 1976. Influence of forest land characteristics Wildl. and Nat. Resour. Conf. 41:279-88. enforcement. Alberta Fish and Wildl. on spatial distribution of hunters. J. Wildl. Manage. 40(3):500-6. HORVATH. l.C. Div. 40 pp. 1974. Southeastern executive summary. MORE. T.A. Economic survey of wildlife recreation. U.S. FISH AND WILDLIFE SERVICE 1979. The demand for nonconsumptive Ga. State Univ. Environ. Res. Group, 1977. 1975 national survey of hunting, fishing, wildlife uses: a review of the literature. Atlanta. 68 pp. and wildlife-associated recreation. U.S. For. Servo Gen. Tech. Rep. No. Washington, D.C. 91 pp. JACKSON. R., R. NORTON. and R. ANDERSON NE-52. 16 pp. 1979. Improving ethical behavior in hunters. Trans. North Am. Wildl. and Nat. Resour. POTTER. D.R., lC. HENDEE, and R.N. CLARK WATSON. M.H., G.c. JAMSEN, and L.W. MONCRIEF Conf. 44:306-18. 1973. Hunting satisfaction: game, guns, or 1972. The Michigan deer hunter. Mich. Dep. nature. Trans. North Am. Wildl. and Nat. Nat. Resour. Res. and Dev. Rep. No. 259. JENKINS. D.H. Resour. Conf. 38:220-29. 33 pp. 1960. Hunter safety. Mich. Conserv. Suppl. 14 pp. PUROl, D.A., H.R. Hill, N.G. HUSSAIN. and G.D. STOll WYWIALOWSKI, A. KEllERT, S.R. 1978. The perceptions and attitudes of 1977. The anti-hunting movement: the people 1978. Policy implications of a national study Michigan deer hunters toward an accept­ involved, their attitude development, of American attitudes and behavioral able rate of compliance. Mich. Dep. Nat. and implications for wildlife manage­ relations to animals. U.S. Fish and Wildl. Resour. Prog. Rep. Pittman-Robertson ment. Iowa State Univ., Ames. MS Servo Rep. No. 024-010-00482-7. 124 pp. Proj. W-121-R. 67 pp. Thesis. 37 pp.

105 Wisconsin DNR WHITE-TAILED DEER MANAGEMENT IN THE MIDWEST - SYMPOSIUM SUMMARY

Robert S. Cook and Rebecca Field u.s. Fish and Wildlife Service Washington, D.C. 20240

INTRODUCTION phases in technological applications. The use ment. Now we talk about deer physiological of telemetry and computers has expanded limits as affected by temperature, and The presentations at this symposium, and and extended our management capability. numbers of days and depth of snow. Karns, the discussions which each has stimulated, There has also been a gradual shift in basic referring to Minnesota, Michigan, Wisconsin, have revealed not only some high points in management concepts, from a game concept and Ontario, has shown the use of such our current knowledge of white-tailed deer emphasizing single species, to broader con­ calculations for estimating overwinter deer manasementbut atsoreflectmmetrends. siderations of wildlife species and an losses. We now address the challenge of pulling ecosystem approach to management. It is in­ The pendulum has swungtrom food supp~y together the highlights, the significant points, teresting to look at this symposium and com­ to broader issuesof population dynamics and the threads of this symposium. As a result, we pare what we have heard here in light of conditions. The trend from habitat evaluation hope this summary will serve as a stimulator these trends. to use of population dynamics is a good for thought and action in the future. trend, one leading to a more balanced, more Eleven years have elapsed since the last PARAMETERS IN refined approach to management. It shows a white-tailed deer symposium in the Midwest sophistication in the use of tools available to (U.s. Forest Service 1970). In 1968, par­ DEER MANAGEMENT the modern wildlife manager. ticipants at the 30th Midwest Fish and Those of us who are primarily big game Burgoyne has shown that mathematical Wildlife Conference held a symposium on biologists and have been in this business for a models incorporating many factors can give nonyarding white-tailed deer populations in while have watched these shifting emphases us a better framework for management deci­ the Midwest covering issues of habitat quali­ in the basis for white-tailed deer manage­ sions. Not a paper today focused on habitat ty and distribution, range appraisal, and use ment. In the 1940's and 1950's the emphasis as a single issue. Yet, the importance of of agricultural lands. Now we have focused was on food supply and range quality; habitat habitat to the total well-being of deer cannot attention on both yarding and nonyarding surveys were "in". We measured the annual be disputed. Habitat needs will be a topic at deer populations in the Midwest, and have growth of selected browse species in the fall, the 1980 Midwest meeting in Minnesota. emphasized such diverse issues as population then measured the amount remaining in the Porath has shown us that certain herd dynamics, harvest monitoring, aging tech­ spring and determined the percentage of characteristics, such as fawn mortality, are niques, and involvement and cooperation of available browse consumed. Plant species importantly interrelated with habitat quality. various publics. composition was also figured ln. Thus Harder discussed deer reproduction as in­ Karns reviewed the changes in deer biologists could determine if the herd ex­ fluenced by the females' diet. Karns and management in the last 10 years. We have ceeded its food supply. Creed and Haberland related weather, seen advances in technology and new em- At the 1968 White-Tailed Deer Symposium, habitat, and winter severity to winter deer Verme (1970) observed that in midwinter deer mortality in the northern range. Cook, Robert S. and Rebecca Field. 1980. White­ become less active and speculated that this Costs are becoming a more important part tailed deer management in the Midwest - sym­ posiumsummary. Pages 107-10 in Ruth l. Hine and phenomenon could be a physiological of management considerations. Mooty Susan Nehls, eds. White-tailed deer population downshifting as a survival mechanism. That pointed out that a broad ecological approach management in the north central states. Proc. 1979 speculation has now been substantiated, and to management is not only biologically Symp. North Cent. Sect. Wildl. Soc. 116 pp. we are using this knowledge in our manage­ sound, but economical too. Gathering as

107 much needed data as possible during our cen­ ing illegal kill, yet obtaining reliable data on good management program: (1) sound susing efforts makes good sense and reduces this segment of herd loss is still a problem. biological data and interpretation, and (2) al the cost per unit of information. Age ratio determination is basic to herd well-planned program for engendering the Time has allowed the accumulation of management. Roseberry has shown that the understanding and support of the majority of much data on deer populations and their techniques for obtaining the basic age infor­ our constituents for our management deci­ habitats. It is the refinement of these data mation, separation of fawns, yearlings, and sions. and in our methods of using them that has led adults, are quite adequate although not in­ Most of the papers in this symposium to our present high level of herd control fallible. While much time and money has which dealt with the biological data for deer through harvest. We have heard repeatedly in been spent on aging to the year class, this management also referred to the need for this symposium that statistically reliable detailed information does not seem essential public contribution, or at least understanding population estimates are essential to deer for most management needs. However, if and acceptance of the data base. The in­ herd management. Although a number of computers are used for predictive modelling, volvement of the public was acknowledged suitable methods have been discussed, no detailed age distinctions may be worthy of at this symposium in complying with harvest single method seems to serve all needs best. more effort. As pointed out by Roseberry, it is registrations, providing reproductive tracts, Gladfelter recommended data from car­ disturbing that age data are often not com­ and locating fawns. killed deer for an accurate economic source parable between states due to different langenau recognized a special public of reproductive data. Harder pointed out that techniques and interpretation. It is re­ group, private landowners. With the high natality and reproduction provide basic infor­ grettable to think that a pooling of age data percentage of hunting occurring on private mation in estimating population size and for several midwestern states could be im­ lands, landowners can have a direct role in setting harvest quotas. While most such data peded by lack of coordination. This con­ deer management and harvest. are used to characterize past populations, the straint may prevent us from characterizing Systematic evaluation of public attitudes, application of computers now offers us op­ white-tailed deer populations on a regional especially those of user groups, must be a portunities for advanced predictive basis. How to deal with and correct such in­ factor in assessing the success of any capabilities. We can begin to ask "what if" consistencies could be addressed in future resource management program. A recent questions that arise when considering a varie­ symposia such as this one. study conducted by Dr. Stephen Kellert of ty of alternative management decisions. In other respects white-tailed deer manage­ Yale University, financed by the U.S. Fish and Burgoyne reported that Minnesota, Colorado, ment is becoming much more concise in the Wildlife Service, has sampled the attitudes and Pennsylvania have incorporated Midwest. Creed and Haberland have de­ and behavior of all Americans toward mathematical models of population data into scribed a program for Wisconsin that appears wildlife and their habitats (Kellert 1979). That their deer management. Also, several other basically sound and is strengthened with study should be helpful in improving our north central states are planning to make every year of experience. The core of the pro­ understanding of the nonhunting public. modelling systems a central part of their gram is a series of calculations that have langenau found that nonconsumptive management decision-making process. shown, over time, rather consistent correla­ users of game species have generally been ig­ In discussing legal kill and the usefulness tions leading to 3 fundamental considera­ nored by managers although white-tailed of the resulting data, Ryel showed that com­ tions upon which harvest recommendations deer are high in public preference. He re­ pulsory deer registration provides more time­ are based. These considerations are status of viewed the effects on management of hunter ly and acceptable information than do hunter the deer population in relation to the goal for and nonhunter attitudes towards wildlife. report cards and mail-back questionnaires. that population, past effects on the popula­ langenau suggested that a well-planned The illegal kill, however, as discussed by tion of varying antlerless harvest, and education and communication program can Beattie, Cowles, and Giles, seems to be an estimated impacts of the previous winter. be a useful tool for managers in influencing area of deer herd statistics that has often While Wisconsin and other states are on the user understanding and preferences. been overlooked, although there are methods way to balancing herds, habitat, and harvest, Educational research in wildlife manage­ to use in estimating illegal kill. This factor in the well-calculated results and harvest ment appears to be a badly neglected area. deer harvest regulation could become in­ recommendations still must meet the test of langenau cited several studies indicating creasingly important as the increased cost of public acceptance. that both deer hunters and the general public living enhances the attractiveness of understood little about basic deer biology poaching and legal complexities add to the INVOLVING THE PUBLIC and management. Public involvement and at­ problems of the law enforcement officers. Petoskey and Madson, as well as others, titudes in management decisions make Cooperation of the public is vital in suppress­ pointed out the 2 major components of any public education and information programs 108 all the more important. A well-planned pro­ organizations as well as individual, influen­ The Fish and Wildlife Service, as well as gram for educating and informing our publics tial decision makers. Petoskey reminded us other federal agencies, has encountered must focus not only on the finished manage­ that the more people are involved in the deci­ several problems solved only through ment recommendations, but on all phases of sion making, the better the decision will be cooperation and coordination with states and deer ecology - from an understanding of deer and the better it is accepted. That is good ad­ other interested groups. Through the shared and their habitat and winter stress to such vice for the states as well as the federal agen­ efforts of many, we were able to get a grip on things as reproduction and censusing. cies. In his 1978 Environmental Message, the chlorinated hydrocarbon problems and As Madson said, we are in the business of President Carter gave emphasis to this impor­ we are working together to save our product promotion. He encouraged tant management dimension. wetlands, our migratory waterfowl, and many managers to sell the public on the need for The Fish and Wildlife Service and many of our endangered and threatened species. solid, professional wildlife management, and states are initiating a new thrust in public in­ Faced with a growing need for well-trained then on the management recommendations volvement and understanding. We are seek­ resource managers, we established that follow. Both "sales" must be based on ing to involve the public in problem discus­ cooperative agreements between the Fish dependable biology, the management part­ sion before management decisions are made and Wildlife Service, state agencies and state nership between agencies and their public, rather than telling people about the decisions universities for graduate training of fish and and the perceived welfare of the resource after they are made. For example, we need to wildlife biologists through 50 Cooperative they both seek to perpetuate. "Selling" a pro­ let the white-tailed deer lovers of North Research Units across the country. Similar gram to the public is a vital part of any America know why deer distributions are af­ cooperative agreements are being developed publ~ martagemet'tt recommendaaon. The fected by togging in the northem limits of to enhance extension education efforts. must "buy" the idea to make the project their present range and why that habitat is more than just an intellectual exercise for the changing back to the climax stage, making it THE FUTURE managers. Advertising is based on the simple less capable of supporting deer. If that kind Looking down the road, what new prob­ goal of creating a market and promoting ac­ of educational approach doesn't help lems are facing us? we meet here again in ceptance of an idea, a product, or a program. generate understanding as well as support for If Private industry does a tremendous job-how our programs, it should at least raise the in­ the Midwest in 1989 to share the latest in well are we doing? tellectual level of the arguments. white-tailed deer management, what will be As Petoskey clearly pointed out, the selling the most pressing issues? It is our obligation as public officials to job begins in our own agencies. Our efforts We predict that the demands on deer provide education and background informa­ must be effective inside our agencies before habitat will increase and become more en­ tion to the people on the management deci­ we go "outside". It is a discomforting truth twined in the interrelationships with other sions affecting their wildlife. As Russell Train when one realizes that some of our greatest management problems. There are 2 topics we has said, "It is the American people who own antagonists may exist in our own organiza­ would like to mention that have special America's wildlife with both the Federal and tion. For every outside critic or opponent, potential for significant modification of deer state governments exercising control of that there is too often a person on the inside will­ habitat and will require major cooperative ef­ wildlife in their role as trustees of the ing to supply them with a premature draft or forts to deal effectively with them. people" (Train 1978:276). other information not yet ready for the "light The first is acid rain, of more significance of day". Good cooperation must begin at We must work among ourselves as well as currently for our fisheries resources, yet one home. Good education must begin at home. with private organizations and the public if which could lead to significant vegetative Outside of our agencies, interests in white­ we are to perpetuate our resources in this age alterations in the future. In August 1979, tailed deer are diverse and include many of accelerated consumption and economic President Carter identified the acid rain prob­ groups. Many midwestern states already have motivation. It is through single species sym­ lem as a new national initiative in his second programs designed to include the public in posia such as this that experts in specific Environmental Message. Although the pro­ the decision-making process. Madson and areas can share their research and advance blem is large, current research efforts are Creed and Haberland referred to the role of the techniques, hypotheses, and ideas of fragmented. Cooperation has extended to an the Wisconsin Conservation Congress in others. The sharing of knowledge and ex­ international level. A scientific advisory directly involving the hunting public in perience among professionals is all too often group was set up in 1978 at the mutual re­ management. Other publics to include are confined to successes and not failures; it is quest of the US and Canadian governments conservation organizations, environmental only through sharing both that our science to examine data on long-range transport of groups, legislators, and local or regional really progresses. acid precipitation and other air pollutants. 109 • The pH of certain Canadian streams has forests and their wildlife. It will be on the LITERATURE CITED already been altered to the point of inter­ agenda of the Commission's 1980 meeting in fering with spawning of anadromous Mexico City. In remarking on the 1968 Sym­ KElLERT. S.R. salmon ids. Breeding populations are being posium, Verme (1970) discussed deterioration 1979. Public attitudes toward critical wildlife established in hatcheries in order to preserve of deer habitat, especially in the northern and natural habitat issues. Phase I. U.S. the affected genetic strains. If ground water, range. He said, "A major miracle is needed to Fish and Wildl. Servo 118 pp. surface run-off, and riverine systems can be reverse the problem of serious range TRAIN. R.E. so affected, what about soil micro-organisms, deterioration, but none is in sight" (Verme 1978. Who owns American wildlife? in Council soil acidity, vegetation types and, ultimately, 1970). Perhaps energy demands on wooded on Environmental Quality. Wildlife in the deer and other vertebrates? This is a red­ areas will provide the incentive we need, for America: contributions to an under­ flag issue that must be watched carefully. deer habitat improvement. standing of American wildlife and its The second topic is habitat modification And so the trend in management con­ conservation. Washington, D.C. 532 pp. resulting from energy demands on forest pro­ tinues, expanding from a single-species focus U.S. FORESTSERVICE ducts. Such fuels as gasohol and just wood to broad ecological concerns. White-tailed 1970. White-tailed deer in the Midwest. North for home and industrial heating have great deer management will continue to include Cent. For. Exp. Stn. St. Paul. 34 pp. implications for forest inhabitants, including not only habitat appraisal, harvest statistics, VERME. L.J. our white-tailed deer. Right now the and life tables, but will also have to consider 1970. Comments on yarding and nonyarding possibilities seem remote. However, with chemical balance of rain, soils, and vegeta­ deer. in U.S. Forest Service. White-tailed face-cords of firewood priced up to $80-$125, tion and include problems of economic and deer in the Midwest. North Cent. For. the economic motivation is there and farm legal issues of land use demands. No one Exp. Stn: St. Paul. 34 pp. woodlots are close to the market. Which im­ specialist will be able to tackle all aspects of pacts are beneficial or harmful will depend white-tailed deer management in the future. on the woodlot managers and their decisions. More and more, we will have to depend on The North American Forestry Commission cooperation and shared knowledge and ef­ has formed a Wildlife Study Group to address forts. This gathering has set us on the right the possible impacts of energy demands on track for the next decade.

110 APPENDIX

A Wisconsin Deer Management Chronology (1836-1980)

Walter E. Scott 1721 Hickory Drive Madison, WI 53705

[For some time Walter Scott has been -land clearing and lumbering by settlers much opposition and laws continually are assembling new materials and consolidating was progressing in central counties (1859). violated by offenders not brought to justice and updating earlier chronologies dealing -Federal censuses showed human popula­ (1879). with management efforts on Wisconsin's tion increased from 219,456 (1847) to 775,881 -Federal census showed human population white-tailed deer. His complete paper will be (1860). rose from 1,054,670 (1870) to 1,315,407 (1880). published by the Wisconsin Department of -Deer open season was reduced from 7 to 5 -Use of poison bait was permitted for killing Natural Resources. Excerpts highlighting months. wolves and wildcats from 10 January to 10 changes during 10- to 20-year periods in land February (1866); this poisorrEait law was use, deer policy and management, public at­ 1861-1880 repealed in 1867. titudes, and related historical events are -Wisconsin produces more than 1 billion presented here to create a brief perspective -State bounty was increased to $10 each board feet of lumber (1870). for understanding the high points and low and extended also to wildcats and lynx. points of the past, and for approaching the -Slashings in the pineries along the Black, Payments indicated over 1,600 were bountied opportunities and decisions of the Chippewa, Wolf, St. Croix, and Wisconsin (1867). future.-Eds.] rivers (as well as parts of Barron, Kewaunee, -Bounties were dropped (1879). and Manitowoc counties) reported to be a -last timber wolf taken in southern Wiscon­ 1836-1860 "raging sea of flame"; Peshtigo fire (1871) sin (Jefferson Co.) (1880). -Counties of the Territory [of Wisconsin] (ex­ burned over 1,280,000 acres in Brown, Door, -Deer open season generally reduced to 3 cept Brown) may pay a bounty of $3.00 for Kewaunee, Manitowoc, Oconto, and months (1880). each wolf or "wolf's whelp" killed in the Shawano counties. Over 1,000 people per­ county (1839). ished and no doubt many deer were killed. -Hunting of deer with dogs prohibited throughout state (1876). -American Fur Company purchases 10,124 -Substantial increase in acreage under deer skins q.tCreen Bay and Milwaukee posts cultivation in farms partly due to discharged -Commercial hunting increased in the cen­ (1839). soldiers' return to agricultural life (incidental­ tral and northern counties; lumber camps also hired hunters to shoot deer for their use -Wildlife begins decline under impact of ly with their greater ability as hunters). (1866). civilization (1840). -legislature set aside over 50,000 acres of -Severe winter of deep crusted snow in state lands in Iron, Oneida, and Vilas coun­ 1856-57 permitted slaughter of deer with ties for a "state park reservation" which was 1881-1900 clubs by Indians and settlers. Considerable not to be sold or cut for timber (1878). -Of the state's original 5 million acres of deer starvation also took place. -H.C. Putnam, who made a forest census of solid pine, only 2-3 million were left by 1881 the state, urged withdrawal from sale of all and this amounted to about 40 billion feet Scott, Walter E, 1980. A Wisconsin deer manage­ state-owned timber land at this time (1880). not counting areas of mixed hardwood and ment chronology (1836-1980). Pages 111-16 in Ruth -Railroads now were extending lines across pine. l. Hine and Susan Nehls, eds. White-tailed deer population management in the north central the North and advertising it as a "sportsman's -Wisconsin ranked near No.1 in lumber out­ states. Proc. 1979 Svrnp. North Cent. Sect. Wildl. paradise" (1880). put and had about 1,000 sawmills with an Soc. 116 pp. -Protection of fish and game meets with output of around 4 million board feet annual­

111 1'1. Cut was almost entirely white pine and State and a 32-page booklet was issued this -Deer open season continued to be 20 days, turned to hemlock and hardwoods several year (1887). but there were increasing restrictions in years later (1896). About 40 pulp and paper -The office of State Fish and Game Warden southern and central counties, until, 1907-08, mills were active, employing more than 5,000 was established by the legislature (1890). 36 southern counties were closed. First deer workers (1900). -Chief Warden reports that public senti­ tag was required (1903). Approximately -Forest fires raged in the Northwest (the ment is so much against his staff that "it is ex­ 100,000 hunters were licensed (1908). For the Phillips fire in Price Co. burned 100,000 acres) ceedingly difficult to secure a conviction on first time, both residents and nonresidents (1894); a million dollars damage recorded in the strongest evidence" (1894). were limited to only 1 deer (1909). 1898, especially in Polk and Barron counties, -Federal lacey Act prohibiting interstate -Wisconsin Hunters and Fishers Association and in 1900 on the Chequamegon Bay and commerce in violation of state game laws recommended the establishment of a "one­ Menomonie River area. was passed (1900). buck" law (1910). -State Park Reservation in Iron, Oneida, and -In spite of the Federal lacey Act on Vilas counties was ordered by the legislature 1901-1910 transportation of game contrary to state laws, to be sold and the average price ran about -The legislature established a state Depart­ shipments of continued to the cities $10 per acre on the first 133,876 acres sold for ment of Forestry (1903). labeled as butter, veal, mutton, etc. (1902). cuttingc(1897). -Forester E.M. Griffith was appointed -The legislature funded the University of Superintendent of Forests and also given the Wisconsin College of Agriculture for prepara­ duties of state fire warden with the result that 1911-1920 tion of a pamphlet on farming opportunities many new fire prevention laws were intro­ -Federal figures report over 2,176,000,000 in northern Wisconsin and"A Handbook for duced and 300 fire wardens hired (1904). the Homeseeker" was published. A large im­ board feet of lumber cut in Wisconsin during -The State Forest Commission (changed to a migration to northern Wisconsin resulted 1919 and 1920. state Board in 1905) arranged for reservation from liberal state and county appropriations -A forest tree nursery was started at Trout of 40,000 acres in northern Wisconsin as a for this purpose (1896). lake for a replanting program (1911), and by nucleus of the state forest system; another -Federal census showed a population of 1919, state tree nurseries at Trout lake and 20,000 was soon added (1904), and 194,000 Tomahawk lake shipped 509,000 seedlings 2,069,042 - a 22.6% rise in the last 10 years acres in 1905. (1900). for planting and by 1920, 320,000. -The woodlot tax exemption law was -The Governor went along with Forester -State bounty on wolves was reduced from passed. Since wild fires had been stopped in $5.00 to $3.00 and on lynx and wildcat from Griffith to declare all state parks as wildlife the southern agricultural area, forest cover refuges (1913). $3.00 to $1.00 (1899). on pasture land often had increased substan­ -The State Forester now reported the state's -Severe winter of 1887-88 resulted in reports tially. Farm woods were being cut up by Forest Reserve had grown to 340,000 acres. of starvation in the far north. agricultural development so there was more He also proposed a mill tax which would sup­ -Deer open season reduced from 31/2 woodland "edge" which stimulated wildlife port forest land acquisition and manage­ months statewide, with any number of deer production (1907). ment, but this was defeated by agricultural of either sex legal without a license (1881), to -The Governor appointed a Conservation interests with powerful friends in the 20 days, 1-20 November, and a bag of 2. Two Commission on basic resources, including legislature (1912). The Supreme court held counties were closed (1895). forests (1908). the state forest reserve policy unconstitu­ -Some sportsmen were proposing a closed -Between 1904 and 1910, the State Forester tional and Forester Griffith left Wisconsin deer season for at least 5 years (1891). reported 3,789 fires which burned 2,491 ,975 (1915). -Deer hunting license was required: resi­ acres. First report of the Wisconsin Commis­ -Over 'h million acres were burned on state dent, $1 and nonresidents $30, and 2 coupons sion featured Forest Griffith's recommenda­ Forest Reserve lands. Fire protection work with each license. Possession of deer in red or tions for many new fire protection methods proceeded with the erection of fire towers spotted coat was prohibited (1897). Over and laws (1909). and telephone lines and development of fire 32,000 resident licenses were reported sold -Federal census showed state population lanes. By 1916 there were 1 V. million acres (1900). was 2,333,860 (1910). under intensive protection of forest rangers -The legislature ordered publication of the -The last record of a cougar taken in the using hundreds of miles of fire lanes, roads, state's fish and game laws by the Secretary of state was reported from Douglas Co. (1908). and telephone lines.

112 -State aid to counties for highway develop­ violations caused by "crass ignorance". He -Seventy-one local conservation clubs were ment started this year and in 4 years, over also started a training school for game affiliated with the Wisconsin Game Protec­ 5,000 miles of highways had been improved wardens (1912). tive Association which was active at this to take care of the 124,000 registered vehicles -About 155,000 hunters were licensed and time. About 500 people attended the Ap­ (1912). 992 people were arrested (of which 52 were pleton organization meeting of Wisconsin -University of Wisconsin Agricultural Exten­ hunting without a license) (1914). Division, Izaak Walton League of America. A sion personnel promoted many educational state congress called by the Conservation projects, including how to control weeds, 1921-1930 Commission in Madison attracted delegates clear lands, drain fields, and raise cattle, con­ -The federal government passed laws which from 100 interested clubs, reflecting the new centrating during these early years largely in provided more state aid for forest fire groundswell of public interest in conserva­ northern Wisconsin (1915). fighting, promotion of farm forestry, and tion (1922). -In the Town of Remington (Wood Co.), enlarging of the federal forests (1924). -WCD Superintendent of Game Wallace B. drainage and settlement were in full swing -Seventeen northern counties had less than Grange recommended a continuous survey of with many new farms, roads, and schools 1'h million acres of saw timber left on their the state's wildlife, hunter report cards call­ (1912), but by 1920 the "drainage boom" had 11 million acres; only about 6% of the ing for data such as weight of deer taken, run its course and finished anything left of cutover lands were developed as farms and establishment of a system of wildlife refuges, the tamarack and spruce swamps. much was tax-delinquent. There were 55 establishment of a permanent wilderness -First large state game refuge was estab­ wood pulp mills operating in the state (1925). park where all predators would be protected, and cancellation of the bounty system in lished - 2 townships in Forest Co. (Argonne) -The Legislature passed an initial act to (1915). favor of salaried hunters to control serious establish National Forests in Wisconsin and cases of predation (1928). -tt was reported that timber wolves were the Northern Highlands State Forest was set common in northern Wisconsin (1912). up (1925). -WCD hired 6 state trappers to work with a supervisor loaned by the U.S. Biological -C.B. Cory's book on "The Mammals of Il­ -Second vote of the people for a constitu­ Survey animal control division to clean out a linois and Wisconsin" (published by the Field tional amendment to permit special taxation "severe infestation of wolves" (1930). Museum of Natural History in Chicago), of forest lands was passed so that forest shows deer as "common" in only 21 northern management on cutover lands could be pro­ -Wallace Grange investigated reports of counties and in the northern parts of Barron, moted (1927). deer starving on Chambers Island (Door Co.), and found deer food completely unavailable. Eau Claire, and Polk counties (1912). -The Legislature enacted the "Conservation Local wardens verified deer starvation on the Act" establishing the Wisconsin Conservation -The 2o-day any-deer season continued un­ island - maybe as many as 100 in the winter til 1915, when the first one-buck law came in­ Commission, passed the Forest Crop Tax Law of 1927-28. Deer were also starving on Rock and strengthened WCD powers to set up fire to effect. The season was shortened to 10 Island (Door Co.) and damage to trees and control districts and to require people to get days (1917) and fawns were protected in shrubs was severe (1928). 1918. About 5,000 bucks were taken in 1915, permits to burn in protected areas (1927). -The experience of Michigan led the way for and 53,593 hunters took about 18,000 deer in -Legislative actions gave counties authority Wisconsin by a few years with reports of deer 1918. to engage in forestry activity (Marinette dying in some yards as early as 1926 and I.H. County Forest was established), and author­ -The nonresident deer hunting license fee Bartlett's first report on "Deer Yards in the ized a 1120 mill tax on property to finance was increased to $50 and the first settler Upper Peninsula of Michigan" was published hunting licenses were issued. Paper tags (101\:) state forestry and fire protection activities. in 1928. were required for the first time (1917), and Counties also were authorized to zone lands later, metal (1920). for agriculture, forestry, and recreation -A survey of conservation wardens and -The Conservation Commission wrote to all (1929). sportsmen indicated that at least 20 counties contained no deer (1929). state legislators urging just a buck season - or -Over 5,000 forest fires burned over 1.8 maybe even a closed season - because of an million acres in state and federal protected -The one-buck deer season was open in 27 unusually large harvest in 1919 areas, in spite of more funds and powers in counties for the last 10 days in November, (18,000-25,000). fire control work, and there were extensive and in 1926 was changed to the first 10 days -The Chief Warden launched an educa­ peat fires in central Wisconsin. All conserva­ of December. Harvest in 1930 was estimated tional program in schools to reduce game tion wardens were also made fire wardens. at 23,000. Cost of deer tags increased to 50¢

113 1931·1940 food shortage, and starvation were showing million acres under intensive forest protec­ -In the first 6 years of the decade, about up" and claimed he was one of the first tion. Of this, 1,800,000 was federally owned 12,000 forest fires burned over 1.2 million wardens to report starvation in the deer yards lands while the rest were in state and private acres. during early 1930's. ownership (1949). -A special session of legislature ap­ -During the decade, artificial feeding was -Over 25 million trees were distributed for propriated '12 million dollars for forest pro­ begun and many instances of overbrowsing planting in the state. Of these, 2'h million tection facilities and a state work relief pro­ and serious winter losses of deer in yards were planted by the U.S Forest Service on na­ gram. These funds opened and improved were reported - e.g., Flag Yard (1935), Apostle tional forests and 1'14 million on the in­ over 1,000 miles of forest roads and hundreds Islands overrun, entire NE with more deer dustrial forest planting program (1950). of miles of firebreaks in addition to many than ever before, Jackson Co. with 200 deer -The deer population over much of northern other jobs. This program was continued the per section in one area (1937); the U.S. Forest Wisconsin averaged 30 deer per square mile next year with cooperation of the federal Service urged removal of 14,000 deer from and some places at this time were as high as government (1932). the Chequamegon National Forest (where 50 and 60 (1941). Aldo leopold estimated the -In the decade after 1936, forest fires were deer mortality was reported at 1300 for the deer population at 500,000 (1943). held to less than 50,000 acres a year. winter of 1934-35); severe overpopulation of -The natural predators of deer (timber deer in Wisconsin with natural winter feeding -To encourage reforestation of the cutover wolves and cougar) were practically gone grounds on the decrease (H.W. MacKenzie, areas, the legislature increased the forestry (1941). F.N. Hamerstrom, James Blake, 1937-39). mill tax to 2/10 of a mill which greatly pro­ -Out of 60 reports, on deer damage to moted public and private forestry work. At -Aldo leopold urged that the newly formed young forest tree plantations, only 10 had no this time, county forests totaled over 1 Y2 Wisconsin Conservation league attempt to damage and 12 were more than half million acres and private land placed under bring the resort owners and deer hunters destroyed (1946). together to "argue out their conflict" - at the Forest Crop law (open to public hunting) -Deer yard conditions were rated as "poor" totaled about 160,000 acres (1937). least to try. He claimed the only successful regulators for the deer herd were the deer in 1941; 1948, 78% poor; 1949, 65% poor in -The legislature requires teaching of con­ north, 52% poor in central. An estimated predators (1940). servation in public schools and authorized 15,000-20,000 deer died of starvation in north­ state fencing for areas having continuous -On 12 June the Conservation Commission ern deer yards (1949-50). Continued deteriora­ deer damage (1935). approved several Pittman-Robertson research tion in 1950. projects, one of which was on deer (1940). -The Wisconsin Conservation Congress of -An extensive outlay of 'funds and tonnage elected citizens held its first statewide -During the decade, the deer season of artificial feed occurred (e.g., $600,000 in meeting at Madison to consider wildlife generally went from 10 days (one-year-old 1943; $73,000 in 1948 for almost 2 million regulations and discuss deer problems (1935). bucks) to 7 days (forked antler bucks) - with 3 pounds of alfalfa, hay, concentrated feeds, closed seasons and one 3-day season (1931, -About 260 conservation clubs with a total and small grains). 1933, 1935, and 1937). Numbers of hunters in­ of 40,000 members were listed for the state creased from 70,000 (1932) to 105,000 (1940) -Aldo leopold spoke at the Conservation (1936). and the estimated kill ranged from 15,000 Congress meeting on "carrying capacity" of -The Civilian Conservation Corps was (1937, 3-day season) to 35,000 (1932, 10-day the land for game and later was appointed organized in Wisconsin with 14 camps of season). Chairman of a 9-person Citizen's Committee about 200 men each and 24 additional camps -Over '14 million acres of land was closed to study the deer situation (1942). on the national forests and Indian reserva­ for protection of deer (1939). -Public relations became a big part of deer tions. They greatly improved forest protec­ -A voluntary sportsmen's license was research project work, with presentations to tion, but also made possible wildlife manage­ established for a fee of $5.00 or more and Citizen's Deer Committee, WCD Commis­ ment techniques such as deer census. With some of the money was to be used for refuges sioners and personnel, radio broadcasts, and the Works Progress Administration (WPA), and public hunting grounds (1937). field inspections - taking people "into the much forest improvement work was ac­ bush and letting them use their own eyes" complished and needed installations built (1943). (1933). 1941·1950 -Aldo leopold said that "the price of pro­ -Ernest Swift reported that "as early as -The American Forest Products Industries, gress in conservation as a whole was to soft­ 1930, isolated cases of overbrowsing, winter Inc. reported that Wisconsin had almost 18 pedal the deer problem" (1946). 114 -Leopold and other deer experts recom­ -The Conservation Congress and Conserva­ also mentioned controlled burning as a possi­ mended a more liberal season on deer, reduc­ tion Commission voted for an any-deer ble tool (1952). tion in size of large refuges, and temporary season in 1948, but Governor Rennebohm, -President Ira N. Gabrielson of the Wildlife removal of bounties on wolves (1943). who was running for re-election, refused to Management Institute spoke at the Silver An­ Leopold proposed "controlled deer hunting" sign the order and sent it back with suggested niversary Forest Conference about "the rela­ as a necessary addition to the state laws changes because the legislative majority was tion between forestry and wildlife", pointing (1945); Governor Rennebohm urged the unfavorable. out that deer management and forest legislators to pass "some form of controlled management, when planned and executed hunting law which will be adequate to solve cooperatively, gave best results (1953). our deer problem" (1949); and the Wisconsin­ 1951·1960 -A summary of winter deer yard checks Upper Michigan Section, Society of -Wisconsin's population was 4,040,000 showed that "the major change was in in­ American Foresters recommended that the (1959). crease in yards in which browsing exceeded Legislature give WCD broad authority and -The state expended $93,200 on wolf and current carrying capacity". Another report power to regulate the harvest of surplus deer coyote bounties, paying for the killing of stated that of the 114 deer found in the north in managed forest areas (1949). 4,498 adults and 324 cubs. Dr. A.W. Schorger in spring, 46% showed positive evidence of -"In Wisconsin you brawl and argue called this an "unfortunate bounty" (1960). starvation. A survey of 431 does showed that endlessly over the management of white­ -A Forest Habitat Improvement Project was about 82% were carrying young, and that tailed deer. You have more deer than any set up and financed through Federal Aid with 12% of the yearling fawn does also were state in the Union. What your state needs is the objective of improving wildlife manage­ carrying young (1956). an immediate kill of 200,000 deer," - Dr. ment techniques, many of which were -After the second statewide any-deer Durward L. Allen at a conference on Wiscon­ directed toward deer (1951). season in 1951, season type varied from sin natural resources (1949). -At the International Association of Game, forked horn to spike bucks, with many -During 1946-48, bounty was paid on 6,967 Fish, and Conservation Commissioners southern and western counties continuing coyote and wolves and 1,004 wildcat and meeting in Rochester, N.Y., Ernest Swift shorter either-sex seasons. Season length lynx - with no definite figures for timber presented a paper, "Deer Herd Control ranged from 7-16 days. Hunters increased to a wolves and lynx (1948). Methods and Their Results". Much of this by high of 349,000 (1959) and harvest peaked at -The deer season was open for bucks now was a success story, with the hunters' ap­ 105,596 (1959) after the 1951 any-deer season throughout the decade, ranging from 5-9 proval, except for failure in attempts to get (279,000 hunters and 129,000 deer). days; a split season in 1943 included authority for "controlled hunting" on deer -For the third consecutive season, Wiscon­ antlerless deer, and from then on from 3-7 (1951). sin led the nation in the white-tailed deer kill agricultural counties were opened for deer of -In the March Conservation Bulletin, and deer reduction was temporarily ac­ either sex. In 1950 the first"any-deer" season Stanley DeBoer wrote a piece called "Feed complished (1951). (7 days) since 1914 was held. Numbers of 'Em-With An Axe!" He pointed out that this -In 1954 two-thirds of the bucks taken were hunters increased to 312,000 (1950) and deer program was no cure-all, but it was less costly less than 3 years old and Washington Co. was harvest ranged from 29,000 (1944) to 168,000 than trying to feed starving deer with hay and opened for the first time since 1906. It (1950). concentrated deer foods (1952). seemed evident that the deer herd would in­ -A so-called "Save Wisconsin's Deer Com­ -Artificial feeding of deer in winter by WCD crease under buck-only seasons if nonhunting mittee" was organized in the north country to was discontinued (1955). losses were reduced, too. protest the "shameful slaughter" of does and -Important developments were deer fawns in 1943 (1944). -In the north, 50% of the deer yards still registration (1953), the Unit concept of deer -The first controlled antlerless deer hunt were incapable of carrying more deer than in management (1957), and the "party permit" was held on the Necedah National Wildlife the 1951-52 winter and in a harder winter than system (1957). Refuge (Juneau Co.) with 36 deer taken per one so mild, planned timber cutting would be square mile (out of a reported population of helpful. So said John Keener, leader of the 60 per square mile). There were 2,028 permits West Central District Forest Development 1961·1970 issued in cooperation with the federal Project in a Conservation Bulletin article on -The state's forest protection covered about government as landowner (1946). "The Need for Deer Range Management". He 33 million acres of which 17,082,290 acres

IlS were considered "critical area" (1968). -The basic deer season was a 9-day spike ciples of deer management (A Profile of -Wisconsin deer herd size is reported at buck season, with either sex open for shorter Wisconsin Hunters, Tech. Bull. No. 60, 1972). about 750,000 with some in every county of periods in several agricultural counties. Party -David Jenkins of Michigan DNR noted 2 state (1970). permits were prohibited in 1961 and 1962, but in 1963 the variable quota plan was initiated. areas of notable progress in Michigan and -There was widespread starvation during Wis:::onsin: "the slow gaining of public the winter of 1964-65 in northwestern Hunting pressure reached a high of 507,000 (1969). The highest kill occurred in 1968, with understanding and support for big game Wisconsin and scattered losses elsewhere in management and the inclination of the north (1964). The winter of 1970-71 119,986 registered deer, of which 57,465 were antlerless. legislatures to grant their commissions or resulted in the most severe losses throughout game boards increasing authority" (1977). major parts of the north since careful obser­ vations began. -DNR Secretary Earl formed a Hunting -Ernest Swift (now a director of National Ethics Committee to help solve the quality Wildlife Federation) was quoted in the Vilas 1971-1980 problems related to deer hunting pressures by County News-Review under the headline -The estimated population of Wisconsin such large numbers of hunters (1977). "Swift Reveals 'Cold War' Between Wardens, now was 4,609,000. Rate of increase in the -DNR aiming at a "wintering" population of Experts" as saying that in his opinion "it is decade from ~96O to 1.970 still was rapid ­ 575,000 whitetails, which "would mean about time for the sportsmen and citizens of 11.8% (1976). 750,000 deer in the woods come hunting Wisconsin to pause and evaluate what the -Total land open for public hunting under season". Wildlife managers figured car kills results already have been" (1961). the Forest Crop Law as of 1979 was 2,275,589 could be taking 27,000 deer each year and -The WCD goal was to keep a winter deer acres. dogs possibly as many as 10,000 in a bad year herd of 430,000 with a plan of harvesting -In view of highway car kills as high as (1980). 75,000 to 100,000 annually (1962). 12,702 in 1967 and severe losses of fawns in -License fees increased with sportsmen's -The Conservation Commission adopted a the winters of 1966-67 and 1970-71 (when license now $22.50; resident big game $11.00; deer and forest management policy recog­ from 50,000 to 60,000 deer died of starvation) nonresident big game $60.50, and resident nizing that these resources now must be it is better to let the hunters do the bow hunting $9.50 (1979). managed on an integrated basis (1962). harvesting, but this is impossible because of -Development of the Unit management "campaigning against the party permit" in -Information sheets on the so-called system resulted in more deer shot and less 1969 and 1970 (Article in September-October "Hunter's Choice Permits" were being widely range damage, with the ultimate goal being a 1971 Conservation Bulletin). distributed, a system already used successful­ more stable deer harvest and population -Based on 18,200 deer killed annually by ly in Michigan. These could replace the (1966). cars from 1976 through 1978, the total loss 4-member party permits of recent years (1980). -Research information published on borne by accident victims was $22.1 million. establishing and managing forest openings -Almost half of the hunters oppose a doe -An average of 594,000 hunters registered was an important step forward in range season under any conditions, indicating a 100,000 deer, of which 36,000 were antlerless, management for deer (1969). lack of understanding of the ecological prin­ annually from 1971-77.

116 Symposia Sponsored by the North Central Section of The Wildlife Society:

1965 Management and Research/Emphasizing Management of Forests for Wood Ducks. (James B. Trefethen. ed. Pub/. 1966. Wildlife Management Institute. 709 Wire Bldg .. Washington, D.C. 20(05) 1967 Management/Current Continental Problems and Programs. ' (Ruth l. Hine and Clay Schoenfeld, eds. Pub/. 1968. Reprints available through Student Chapter. Wildlife Society. CNR. University of Wisconsin. Stevens Point. WI 54481, $3.75) 1969 Predator Ecology and Management. (Albert W. Erickson and Clen C. Sanderson, eds. Pub!. 1972. j. Wild/. Manage. 36(2):210-404) 1973 Biology and Management of Pheasant Populations in North a: America. Z (Not pub/.) o c 'in c 1977 Waterfowl and Wetlands/An Integrated Review. o (Theodore A. Bookhout, ea. Pub/. 1979, $3.(0) :;: ~ 1979 White-tailed Deer Population Management in the North Central States. (Ruth l. Hine and Susan Nehls. eds. Pub/. 1980)