INVENTORY OF LARGE MAMMALS IN THE DEUX BALE NATIONAL PARK, UPPER VOLTA Thesis for the Degree of M. S. MICHIGAN STATE UNIVERSITY JOHN P. SIHVONEN 1977 Umvcrsxty * LEMMA?“ ABSTRACT INVENTORY OF LARGE MAMMALS IN THE DEUX BALE NATIONAL PARK, UPPER VOLTA By John P. Sihvonen During 1972-1974, the densities of large mammal populations were estimated in the Deux Balé National Park. an area of 566 km2 in the Sudan savanna zone of west-central Upper Volta, Nest Africa. Indices to animal abundance used were: I) Hahn (1949) walking-strip counts. 2) animals seen per kilometer of travel. and 3) territorial mapping. Results obtained from the first and second methods were similar. but the visibility factor in the first method made it the more reliable index from year to year. The third method was applied to species with distributions not amenable to strip-count methods. INVENTORY OF LARGE MAMMALS IN THE DEUX BALE NATIONAL PARK, UPPER VOLTA By .zi.'> John P. Sihvonen A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Fisheries and Wildlife T977 ACKNOWLEDGMENTS I would like to thank the Departement des Eaux et Foréts, Ministére de l'Agriculture, Upper Volta, and the United States Peace Corps for the opportunity to work in the Deux Balé National Park. I express my appreciation to Monsieur Jean Barry, former Directeur des Eaux et Foréts, and to members of his staff. particularly Mssrs. Casi- mir Ziba. Chef d'Inspection, and Matias Siéba, Chef du Cantonnement. I owe a deep debt of gratitude to Eaux et Foréts personnel assigned to the Deux Balé. Mssrs. Yaya Boly, Koffi Ganou, Ernest Yao, Ousmane Kagoné, Ernest Kaboré, Marcel Somé, Adama Barry, and Béton Dao for their help with the faunal inventories of the region. Their enthusiasm and encouragement were abundant and welcome. I extend heartfelt thanks to Mssrs. Simon Ouattara and Ouétian Bognounou of the Section Botanique, Centre Voltaique de la Recherche Scientifique in Ouagadougou. for their immense assistance with plant identification. I express my gratitude to Dr. N. Derek Clayton of the Royal Botanic Gardens. Kew, England, for kindly identifying the Gramineae. I pay special homage to my interpreter and companion, Kokoro Matias Tindano. who with constant good humor, willingness, competency, and a precocious perception of the foibles of man, made possible so much that seemed otherwise. ii My concluding words of thanks go to Michigan State University, especially my major advisor, Dr. George Petrides. I would like to thank the members of my committee, Drs. Donald Dickmann and Leslie Gysel, for their critical review of this thesis. TABLE OF CONTENTS Page LIST OF TABLES ....................... v LIST OF FIGURES ...................... vii LIST OF APPENDICES ..................... viii INTRODUCTION ........................ 1 THE STUDY AREA ....................... 2 Climate ......................... 4 Topography and Geology ................. 5 Mammalian fauna ..................... 5 Flora .......................... 6 METHODS .......................... 8 RESULTS .......................... l7 Accuracy of Walking-Strip Counts ............ l7 Indices to Abundance: The Hahn Method Vs. Relative Abundance Counts ................... 24 Principal Species in the Park and Surrounding Areas . . . 32 Additional Species: Territorial Mapping ........ 34 Total Populations of Large Mammals and Ungulate Biomass ....................... 38 CONCLUSIONS AND RECOMMENDATIONS .............. 42 Count Methods ...................... 42 Management Implications ................. 43 SUMMARY .......................... 45 REFERENCES ......................... 47 iv Table LIST OF TABLES man- -method population estimates and densities per for five ungulate species in February- -March, 1974, Deux Bale National Park, Upper Volta ...... Herd sizes observed in five ungulate species, December, 1973 - April, 1974, Deux Balé National Park, Upper Volta. Duplicate sightings eliminated . . . Park-wide population estimates of roan antelope and western hartebeest based on herd classes and individuals in the Deux Bale National Park, Upper Volta ........................ Numbers of herds seen for five ungulate species, Deux Balé National Park and Baporo, Dibon, and Laba Forests, Upper Volta, 1974 ........... Mean visibility distances and 95 percent confidence intervals calculated from first- -sighting distances and from Hahn-method visibility measurements, January-March,1974,Deux Balé National Park, Upper Volta ........................ Relative abundance of fourteen mammals from counts by pairs of observers on foot, February- March, 1974, Deux Balé National Park, Upper Volta ...... Relative abundance of twelve mammal species from road counts by single observers on motorbicycle, Deux Balé National Park, Upper Volta ......... Hahn-method estimates compared to animals-seen indices during walkin -strip counts, February- March, 1974, Deux Bale National Park, Upper Volta Ratios of numbers seen per kilometer traveled for five ungulate species compared to ratios of Hahn- method index estimates, Deux Balé National Park, Upper Volta ..................... Page 17 19 20 23 25 26 27 28 3O Table Page 10. Changes in mean visibilities computed from estimated first-sighting distances and from Hahn- method visibility measurements, Deux Balé National Park, Upper Volta ................... 31 ll. Comparative Hahn-method densities for five ungulate species in April-June, 1974, Baporo, Dibon, and Laba Forests, Upper Volta ............... 33 12. Hahn—method po ulation-index (as numbers and den- sities per kaY for five species of ungulates on 158 km2 areas bordering Grand Balé Creek in both the Deux Bale National Park, 25-29 March 1974, and the Dibon Forest, 1-3 April 1974 ........... 35 13. Averages and 95 percent confidence intervals for distances to which ungulates of three sizes could be seen, February-June, l974, Deux Balé National Park and Baporo, Dibon, and Laba Forests, Upper Volta . . ....................... 36 14. Populations and densities of twenty-two large mammals estimated from the results of Hahn-method counts, animals seen per kilometeE indices, and territorial mapping in the 566 km Deux Balé National Park, Upper Volta .............. 37 15. Estimated biomass of large-ungulates, Deux Balé National Park, Upper Volta .............. 40 16. Biomass values for large ungulates in several West African parks and reserves .............. 41 vi Figure LIST OF FIGURES Upper Volta, West Africa, showing the locations of its national parks and major vegetation zones (Keay, 1959) ..................... System of zones and base lines used in the random selection of walking-strip count transects, Deux Bale National Park, Upper Volta ........... Transect locations for 4-9 February 1974 large- mammal count, Deux Balé National Park, Upper Volta . . Distribution of count transects, Deux Balé National Park and Baporo, Dibon, and Laba Forests, Upper Volta ........................ Page TO IT 12 LIST OF APPENDICES Appendix Page 1. Summary of weather data from the Boromo Meteoro- logical Station, Boromo, Upper Volta: (A) Average monthly rainfall (mm) from 1921-1974, (B) Monthly rainfall from 1971-1974, and (C) Average monthly relative humidity from 1961-1970 and average monthly evaporation (mm) from 1961-1974. Taken from Begué (1937), Anon. (1966, 1970), and unpub- lished records of the Boromo Meteorological Station, 1970-1974 ........................ 50 2. List of mammals found in the Deux Balé National Park, Upper Volta, based on field observations by the author and on the bat collection of J. F. Heisterberg. Common and scientific names after Dorst and Dandelot (1970) and Rosevear (1965) (Chiroptera) ...................... 53 viii INTRODUCTION The author had the privilege, under the auspices of the United States Peace Corps and Upper Volta's Departement des Eaux et Foréts, of working from December, 1971 through June, 1974 in the Deux Balé National Park. The park was an unknown entity in 1971, having never been the subject of the faunal studies logically prerequisite to its intended development for nature preservation and wildlife-oriented tourism. The author undertook the inventory of the Deux Balé's wild- life resources to provide the Departement des Eaux et Forets with data on which to base management and developmental decisions. The inventory techniques used in these studies were limited to those possible under a very meager operating budget, but with the cooperation of an excel- lent complement of Eaux et Forets rangers. Objectives of the study were: 1) to determine the composi- tion and relative abundance of the large-mammal fauna, and 2) to assess the implications of those findings for park management. THE STUDY AREA The Deux Balé National Park has an area of 566 km2 and is located between latitudes N11°30' and Nll°50' and longitudes N2°40' and W3°10' in Upper Volta, West Africa (Figure l). The park is situated in the Sudan savanna zone as delineated by Keay (1959), and within the Meridianal Subdomain of the Sudan Domain of the Soudano-Zambesian Vegetative Region in the classification scheme proposed by Troupin (1966). In 1937, the area presently occupied by the park was included in the forest reserve system in recognition of its being one of the least altered and most heavily wooded zones in the region destined to become Upper Volta (Plaisance, 1936). The Deux Balé Forest was upgraded to national park status in 1968, but necessary legislative action remains incomplete. The park, being in a relatively densely-populated part of Upper Volta, is subject to the pressures usually associated with wild- lands in proximity to human settlements: poaching, livestock grazing, and unauthorized burning. Demand for land in the region is increasing due to recent population influxes from the drought-stricken Sahel and, in many places, bush fallow agriculture is practiced to the very boundaries of the park. Actual management of the area comprises only the mainten- ance of the 102 km tourist track, an early dry-season burning 4." - ‘ SAHEL mm: “ ~~ \‘~ \ \‘ II SUDAN SAVANNA ZONE Black fletigoATeer a Ouagadougou ‘2 N \ _' Cree Boromo ~t‘ e an Deux Balé . . Bdbo: Balé National Po gat1 i oulassd°~ _‘ Creek Park ar. NORTHE' GUINEA 00 SAVANN ' ZONE N LEGEND -——. National boundary .___ Park boundary a City of town -..__. Limits of vegetative zone (after Keay, 1959) Scale 1:4,000,000 - 62% of scale Figure 1. Upper Volta, West Africa, showing the locations of its national parks and major vegetation zones (Keay, 1959). SCI) Iou IRE yea Me: In: I91 an Se I‘d DE fr to in an schedule, and infrequent patrols to control illegal activities. Tourism in the park is not being actively promoted, and fewer than twenty groups of tourists--foreign and loca1--visit the Deux Balé yearly. Climate Rainfall records over the period 1921-1974 from the Boromo Meteorological Station, located near the northern limit of the park, indicate an average annual rainfall of 967 mm (Begué, 1937; Anon., 1966, 1972; unpublished station records, 1971-1974). The Deux Balé area is considered to have a five-month rainy season, May through September. The period from November through mid-March is nearly rainless. Relative humidity over the period 1961-1970 averaged 52 percent (Anon., 1972). Seasonally, relative humidities (percentages) average in the low 80's from July through September and in the 20's from December through March. The average annual evaporation is close to 2200 mm with annual evapotranspiration on the order of 1700 mm. During the dry season, maximum daily temperatures range from 37° to 40°C with minimums from 17° to 25°C. Daily variations are less extreme during the rainy season with maximums ranging from 30° to 32°C and minimums from 20° to 22°C (Anon., 1966, 1972). A summary of weather data is given in Appendix 1. Topography and Geology The general aspect of the Deux Balé is gently-rolling terrain with occasional lateritic prominences; elevations range from 235 to 310 m. Soils are predominantly of the tropical ferruginous group, hydromorphic soils are found along the major streams (Figure l) and on their floodplains (Leprun and Moreau, 1968). The park is underlain by three rock formations of approximately equal width which are oriented south-south-west to north—north-east. From east to west they are composed of 1) green and sericitic schists from the Birrimean period, 2) metamorphosed birrimean schists, and 3) granito-gneisses, respectively (Anon., 1968). Mammalian fauna The Deux Balé has twenty-three species of mammals which are jackal-sized or larger. Among them are five frequently-encountered antelope species: roan antelope (Hippotragus eguinus Desmarest), oribi (Ourebia ourebi Zimmermann), Grimm's duiker (Sylvicapra grimmia Linn.), western hartebeest (Alcelaphus buselaphus major Pallas), and bushbuck (Tragelaphus scriptus Pallas). Less-frequently encountered are western kob (Adenota kob Erxleben), red-flanked duiker (Cephalo- phus rufilatus Gray). defassa waterbuck (Kobus defassa Ruppell), and African elephant (Loxodonta africana Blumenbach). There are three primates all of which are commonly seen: green monkey (Cercopithecus aethiops Linn.), patas monkey (Erythrocebus patas Schreber), and 3'1: anubis baboon (Papio anubis Fisher). A total of thirty-six mammal species is known to be present in the park (Appendix 2). Flora Deux Balé vegetation is largely typical Sudan savanna with a few areas dominated by Northern Guinea savanna species (Kershaw, 1968). Numerous vegetation types are found in the area. The vast majority of the park is covered by tree savanna or open savanna wood- land types; small areas of grassland are found on floodplains and on very shallow soils underlain by laterite. Woody species ubiquitous on upland areas are: Acacia dudgeoni Craib ex Holl., Butyrospermum paradoxum (G. Don) Hepper, Combretum glutinosum Perr. ex DC., Crossop- terix febrifuga (Afzel. ex G. Don) Benth., Detarium microcarpum Guill. and Perr., Lannea acida A. Rich., Piliostigma thonningii (Schum.) Milne-Redhead, and Terminalia laxiflora Engl. Locally, Anogeissus leiocarpus (DC.) Guill. and Perr., Burkea africana Hook., and Isober- linia doka Craib and Stapf may be quite common. Important grasses on upland areas are: Andropogon ascinodis C.B. C1., A, gayanus var. bisquamulatus (Hochst.) Hack., Diheteropogon hagerupii Hitchc., Hyparrhenia subplumosa Stapf, Loudetia togoensis (Pilger) C.E. Hub- bard, Monoeymbium ceresiiforme (Nees) Stapf, and Schizachyrium sanguineum (Retz.) Alston. The forest galleries along major water- courses are mainly composed of Mitragyna inermis (Nilld.) 0. Ktze., with Acacia sieberiana DC., Cola laurifolia Mast., and Pterocarpus santilinoides L'Her. ex DC. being the most frequent component species. Shall spec: fem Smaller watercourses rarely support galleries comprised of riverine species; denser-than-average strips of upland species are normally found along their banks. METHODS Hahn (1949) described a method for censusing Texas white- tailed deer (Odocoileus virginianus Miller) with walking-strip counts. Transects two or more miles in length were surveyed across habitat representative of the vegetation and land-use on the count areas. The distances to which deer were visible along each transect line were estimated by having a second observer pace at right angles to the line at lOO-yd intervals. The distance at which a white handkerchief-- simulating a deer's tai1--in the second observer's hip pocket was obscured by brush and slope from the observer on the transect line was measured. The visibility measurements for each transect were then summed and twice the mean of the measurements times the length of the corresponding transect line was taken as the acreage on which deer could be seen. This acreage divided by the number of deer seen by a single observer while following the transect gave a figure of acres per deer. Acreage per deer divided into the total area for which the transect line was representative gave an estimate of that area's deer population. Hahn's (1949) walking-strip census was modified for use in the Deux Balé National Park and in three contiguous reserves, the Baporo, Dibon, and Laba Forests. There being no map showing the vege- tation types of the region, the situation of transects in a manner representative of vegetation and land-use was not possible. Transect locations were chosen randomly in the park from a system of zones accessible along their lengths by vehicle or motorbicycle. A base line was drawn across each zone (Figure 2), the base lines were measured, and points at lOO-m intervals along each base line were given three-digit numbers. Starting points for transects were then located on the base lines by drawing sets of three digits each from a table of random numbers. A minimum distance of one kilometer was kept between transects in the same zone (Figures 3 and 4). In the three forests, transect lines were spaced equally along base lines (Figure 4). Transect routes followed were in the form of open or closed rectangles with their long axes 4-7 km in length and usually 1 km apart (Figures 3 and 4). They were followed by two-man teams using compasses and counts were started from twenty minutes before to forty minutes after sunrise. Distances were determined by pacing. Con- siderations of vehicular access, area to be counted, time, and man- power limited the number of transects selected to thirty in the park and to five, twelve, and eleven in the Baporo, Dibon, and Laba Forests, respectively. Walking-strip counts were carried out in the Deux Balé during 4-9 February and 25-29 March, 1974, with respectively 350 and 352 km being traversed. The northern two-thirds of the 245 km2 Dibon Forest was counted during 1-3 April, 1974 by walking 151 km of transects. The 81 km2 Baporo and 188 km2 Laba Forests were inventoried during 4-7 June, 1974, with respectively 74 and 150 km of transect routes being followed. IO .aupo> Lana: .xcmm Pocopumz w—mm xamo .muummzmgu ucaou apcumimcwxpm: mo :owuumpmm sauce; ms» 2' tom: wasp, mama ucu meson mo smamxw .N «camp; I I ~ 2 x \ a. L\ p. 1.." \ \ sr\ll¢y IN (NV); .Illll Inf-a \‘I‘ I c x 1 1 1 \ \ III IIHVnIAI ”an" .\.K1s.1:\\ a d .A \ u / \A‘ D \( /l\ \\ _ IV\\\/. 4 I\ s: . LiIr oh I II in; $553. r . L IA / 38 x89 . It. \ [HI (\ x _ z I sf I Il/II\(“0~ \ / s Ijs 0 e Wfl w. .9. a 0 d 0 e n e .3.m u n o. a m H is e e 9. 1| n S a O a C Z B S LEGEND —— Park boundary ll .mapo> Luna: .xgmm Pmecwumz wpmm xzmn .msmcou pussasimmcmp eump xumagamu m e com mcowunuop pummcmgh .m mcamwu ... 2-.., \ \... ../ ., . .,./ // 3...: N52. 23 —§8 NMMUL\\ \ 1% g. ~V z e e n 0] .1. 0a y-Ik 0c rtc OS a “a 0 d 0 r 08 n t 3m "t 00a am Ct 115 es Sal e k nr 1 r a“ a 3 r0 c D PTT S N E G — E _ _ L 12 Baporo . uiiifllll ' F°\ /'\\\ IIIIIHI!) D -on 0’°s LEGEND .___. Park or forest boundary -——- Transect outline --~- Base line Scale 1:300,000 - 62% of Scale Figure 4. Distribution of census transects, Deux Balé National Park and Baporo, Dibon, and Laba Forests, Upper Volta. reg? were Hen r—f' 13 Determination of visibility along transects in the Deux Balé region varied slightly from Hahn's (1949) design. Visibility profiles were not determined for every transect. Further, visibility distances were calculated for three size-classes of antelope: 1) small, 50 cm at the shoulder corresponding to Grimm's duiker and oribi, 2) medium, 80 cm at the shoulder corresponding to bushbuck, and 3) large, 145 cm at the shoulder corresponding to roan antelope and western hartebeest. At ZOO-m intervals, the second observer walked at right angles away from the transect line. Average distances from the line of travel at which cloth strips pinned to the second observer's clothing at heights of 50, 80, and 145 cm became hidden by vegetation or slope were meas- ured by pacing. In January, 1974, a total of 208 visibility distance observations for each size-class was made in the Deux Balé along eight transect courses selected by the same method as were walking-strip count transects. Average visibility distances computed in the park were also used in the vegetatively-similar Dibon Forest. Due to visibility changes between January and the dates scheduled for counts in the Baporo and Laba Forests, 180 visibility distance observations for each size-class were made in each forest in May-June, 1974. Population estimates (P) were calculated for each count area using the formula: __A_U_ P‘sz Where A = total area of the park or forest, N = total number of animals observed, 0 = average width of the visible strip, and L = total length T4 of transects walked. The mean visibility distance was doubled so as to equal the width of the strip viewed on both sides of the line of travel. The herd-forming tendencies exhibited by roan antelope and western hartebeest violated an assumption of the Hahn census model that individual animals are randomly distributed throughout the area. Robinette, Loveless, and Jones (1974) believed that some herding could be tolerated by walking-strip censuses if area coverage was high. The area visually covered in counting roan and hartebeest, how- ever, though comprising approximately 20 percent of the park area, was found to be insufficient to overcome herding effects. To diminish the effects, herd numbers were approximated by eliminating duplicate sightings from all observations of roan and hartebeest made from December, 1973 through April, 1974. Observations of each species were then divided into two classes depending on whether the number of animals in the herd was more than or less than the median number in all herds. A herd was defined as one or more animals. The average number of animals per observation for each class was defined as a herd-class average. The number (N) for one of these species was then taken as the sum of the products of the numbers of observations in the two classes made during a count times their corresponding herd-class averages. Simple averages (total animals seen divided by total observa- tions) were not used. The difficulty of distinguishing duplicate sightings of the smallest herds (one or two animals) could conceivably cause a greater distortion of such averages--with resulting effects on population estimates-~than could occur with herd-class estimators. 15 Only nineteen of fifty-nine bushbuck seen during dry-season counts in 1973 and 1974 were farther than 0.5 km from a streambed. Wilson and Child (1964), Jacobsen (1974), Waser (1974), and Green (1976) all found bushbucks to be strongly associated with riverine areas. In accord with these observed dry-season habitat preferences of bushbucks, walking-strip counts in the Deux Balé were stratified with the 261 m2 within 0.5 km of watercourses considered as high- density habitat and the remainder of the park as low-density habitat. Relative abundances of mammals in the park also were calcu- lated per kilometer traveled on a motorbicycle and per kilometer walked during counts. Observations were made from a motorbicycle moving 25 km/h along the tourist track by single observers during the period December through March over 644 km in 1971-72 and over 366 km in 1973- 74. Observations on foot were made by pairs of observers along 702 km of transects in February-March, 1974. The distributions of anubis baboon, green monkey, patas mon- key, and red-flanked duiker in the Deux Balé rendered walking-strip counts infeasible for determining their populations. Their populations were estimated using modifications of territorial mapping techniques described by Overton (1969) and elaborated by Green (1976). All observations of these species were plotted on a 1:S0,000-scale map of the park. Plotting revealed that observations of baboon, green monkey, and patas monkey were sometimes clustered. Home range boundaries were subjectively drawn around adjacent clusters and these areas, presumed to be occupied by different troops of baboon and patas monkeys, were measured. The average area of the home ranges was divided into the 16 area of the park to estimate the numbers of possible baboon and patas monkey home ranges. Clusters of green monkey observations were always near gal- lery forest along streams. Dorst and Dandelot (1970) and Heisterberg (1975) also found green monkeys to be closely associated with riverine areas. To estimate the number of green monkey home ranges, the stream frontages thought to be occupied by oft-observed troops were measured and then averaged. The length of streams in the park which were bor- dered by gallery forest was determined. This latter length divided by the average stream frontage per troop gave the possible number of green monkey home ranges. Red-flanked duikers also exhibited strong affinities for riverine areas. The average stream frontage occupied by a duiker was estimated from repeated observations along a 5-km strip of the Grand Balé Creek. The length of stream frontage considered suitable for red- flanked duikers was longer than for green monkeys as the former species was occasionally observed along streams without galleries. Elephant, African buffalo (Syncerus caffer Sparrman), and lion (Panthera 1eo Linn.) populations were estimated on the basis of observations made from December, 1973 through June, 1974, and from track counts begun in 1972. The spotted hyena (Crocuta crocuta Erxle- ben) population was estimated from track counts. The population of hippopotamus (Hippgpotamus amphibius Linn.) was estimated from conver- sation with fishermen on the Black Volta River. Populations of eight other large mammals were judged from sightings and spoor. RESULTS Accuracy of Walking-Strip Counts Only roan antelope, oribi, Grimm's duiker, western hartebeest, and bushbuck were encountered frequently enough during walking-strip counts in the Deux Balé to justify Hahn—method population estimates (Table 1). Table 1. Hahn-method population estimates and densities per km2 for five ungulate species in February-March, 1974, Deux Balé National Park, Upper Volta. Population Estimates Averages Species February March Numbers Densities Roan antelope 942 1546 1244 2.20 Oribi 470 485 478 0.84 Grimm's duiker 294 286 290 0.51 Western hartebeest 269 251 260 0.46 Bushbuck 147 130 139 0.25 Totals 2122 2698 2410 4.26 The distribution of individuals of the different species directly influenced the consistency of Hahn-method estimates. Espe- cially oribi, but also Grimm's duiker and bushbuck, were seen at or 17 18 near the same places on different dates. Dupuy (1968) and Monfort and Monfort (1974) believed oribis to be sedentary and strongly attached to their particular areas. Child (1974) noted a similar behavior by Grimm's duikers, and the same phenomenon was reported for male bush- bucks by Jacobsen (1974). Herd size in these three species was almost always less than four animals (Table 2). These factors, sedentary behavior and small herd size, produced the relatively uniform distribu- tions necessary for consistent results with the Hahn method. In contrast, roan antelope and western hartebeest presented very uneven distributions. These species formed large herds (Table 2), ranged over great distances, and roans in particular were concentrated in favorable areas. Scattered rains in the park on 13 March 1974 caused localized flushes of green grass which markedly affected roan distribution. Of 367 roans observed during the March count, 303 were seen on five transects that crossed the grass-flush areas. Only 204 roans were observed along thirty other transects during the February and March counts. Consistency in successive Hahn-method estimates of roan and hartebeest populations based on the numbers of animals seen during counts was not probable. The variability of roan and hartebeest estimates was appre- ciably diminished by employing herd-class averages to compute N (Table 3). The main assumptions in the use of averages were: 1) population structures remain constant during the period of observation, and 2) the distribution of herds in the two classes is more random than the dis- tribution of individual animals. These assumptions were in part 19 Table 2. Herd sizes observed in five ungulate species, December, 1973 - April, 1974, Deux Balé National Park, Upper Volta. Dupli- cate sightings eliminated. II Numbers of Observations by Species (Percentages of Observations) Herd Roan Grimm's Western Size antelope Oribi duiker hartebeest Bushbuck 1 46 40 44240 104(92) 15 28; 39 74) 2 13 ll 43 39 9 8) 5 10 14 26) 3 5 4 21 19) - 510) - 4 1 1 1 l - 3 6) - 5 3 3) 1(1 - 3(6) - 6 5(4) - - 3(6) - 7 5 4) - - 2 4) - 9 4 3 - - 2 4) - 10 2 2 - - 1 2) - 11 3 2 - - 3 6) - 12 3 2; - - 2(4) - 13 2 2 - - - - 14 1 1 - - 1(2) - 15 3 2) - - 3(6) - 16 1 1 - - - - 17 2 2) - - - - 18 2§2 - - - - l9 2 23 - - 1(2) - 20 1(1 - - - 21 3 2) - - - 22 1(1) - - 1(2) - 25 1 1) - - - - 27 - - - 1(2) - 28 1(1) - - - - 29 1(1) - - - - 31 1 l - - - - 37 2§2§ - - - - 38 1 1 - - - - 4O 1 1 - - - - Totals 116(100) 110(100) 113(100) 51(100) 53(100) 20 Table 3. Park-wide population estimates of roan antelope and western hartebeest based on herd classes and individuals in the Deux Balé National Park, Upper Volta. Roan antelope Number of individuals = I 2 Population estimate = (25:60km63xkéjL Average estimate Number of herds = N Small herds (i = 1.2) = N1 Herds of 3+ animals (i = 14.1) = N2 Number estimated (N1 x 1.2) + (N2 x 14.1) = Ne 555 km2 x N; (2 x 0.153 km)L Population estimate = Average estimate Western hartebeest Number of individuals = I 2 566 km x I Population estimate = (2 x 0 163 km)L Average estimate Number of herds = N Small herds (i = 1.6) = N1 Herds of 4+ animals (x = 10.4) = N2 Number estimated (N1 x 1.6) + (N2 x 10.4) = Ne 2 Population estimate = (36: 3m16§ Nfi)L Average estimate Feb Mar 1974 1974 140 357 714 1855 1285 25 40 13 20 12 20 185 305 942 1545 1244 Feb Mar 1974 1974 50 35 305 184 245 11 7 4 53 50 259 251 250 Change +160% + 64% Change - 40% 21 supported by the approximately equal numbers of observations in both classes of the roan and hartebeest populations (Table 3). Oribis, Grimm's duikers, and bushbucks, when well-hidden, often permitted an observer to approach within 50 m before taking flight. On occasion, too, they were seen to remain in position but to lower themselves to the ground to escape detection. The tracks of freshly-departed roans and hartebeests were observed on several tran- sects indicating that some animals had escaped unseen and unheard. It seemed certain, therefore, that a sizable proportion of these five species was not counted and that populations and densities calculated from walking-strip data were underestimates. The magnitudes of the errors stemming from uncounted animals were undetennined; hence, no correction of the estimates was possible. Hahn-method counts, as employed in this study thesis, were judged to be inadequate as complete censuses. The consistency of population and density estimates calcu— lated from counts in February and March, 1974 (Table 1) indicated, however, that the Hahn method using two-man teams of observers was satisfactory as an index to oribi, Grimm's duiker, hartebeest, and bushbuck abundances. It could probably serve also as an index to roan antelope numbers, barring gross changes in their distribution patterns such as that caused by rain-induced grass flushes during the March, 1974 count. The accuracy of the Hahn method as an index to population trends depends on reliable year-to-year assessments of the visibility of animals. Lamprey (1964) and Hirst (1969) found Hahn-method 22 visibilities and actual disappearance distances to be closely correlated for the species in their studies. Heisterberg's (1977) comparisons of Hahn-method visibility distances to the actual disap- pearance distances of roan antelope, oribi, Grimm's duiker, western hartebeest, and bushbuck indicated a rough correlation, the Hahn-method distances being from 3 percent to 51 percent shorter than the actual disappearance distances. Lavieren and Bosch (in press) reported simi- lar variability in their comparisons of Hahn-method profile measure- ments to actual disappearance distances. The correlations noted between Hahn-method and actual visibility distances suggest that the Hahn method measures the visibility of animals in a consistent, though not an exact, manner. Robinette, gt_al. (l974) believed that visibility measure- ments based on the distance of animals from the observer either where first or last seen were preferable to Hahn-method measurements. This may be true in some areas, but in the Deux Balé region the paucity of animal observations during counts (Table 4) yielded sample sizes which were unlikely to reflect year-to-year visibility differences accurately. Based on variances calculated from the sets of 208 Hahn- method measurements, 200 measurements would be required to calculate a mean visibility distance with a 95 percent confidence interval less than 10 percent of that mean for large antelopes; 295 measurements would be needed to meet the same criterion for the small antelopes. Such high numbers of animal observations were not available for any counting period. Further, the 95 percent confidence intervals for mean visibilities computed from the distances of animals from observers 23 Table 4. Numbers of herds seen for five ungulate species, Deux Balé National Park and Baporo, Dibon, and Laba Forests, Upper Volta, 1974. Kilo- Western Area and meters Roan Grimm's harte- Bush- Census date walked antelope Oribi duiker beest buck Deux Balé 4_9 February 350 25 33 32 11 20 Deux Balé 25-29 March 352 40 27 33 9 13 Dibon Forest ]_3 April 151 8 14 ll 6 6 Baporo Forest * 4_5 June 74 11 5 4 4 Laba Forest * 5_7 June 150 4 5 6 6 *Though known to be present, no western hartebeest were seen in the 4-7 June counts of the Baporo and Laba Forests. 24 where first seen were much wider than those for the Hahn method (Table 5). It was felt that the work involved in calculating Hahn- method visibility distances--about 50 man-hours for 200 measurements-- was necessary to provide the desired accuracy. Indices to Abundance: The Hahn Method’Vs. ReTative Abundance Counts The Hahn census method, as stated above, was thought to underestimate population numbers but to be a useful index to animal abundances. The results of relative-abundance counts were computed (Tables 6 and 7) and data for the five common antelope species were compared to Hahn-method results to determine which type was better. For the herd-forming species, the relative abundance figures calculated from walking counts yielded less consistent results (Table 8). The difference in consistency was due to the use of herd-class averages in the Hahn estimates; their employment in the relative— abundance calculations would result in similar variations. Also, stratification of the relative-abundance counts for bushbuck by the same criteria used in computing the Hahn estimates would give a vari- ation identical to that of the Hahn counts. The order of abundance of the five ungulate species deter- mined by the two methods also were in disagreement, the abundances of hartebeest and Grimm's duiker being reversed in February (Table 8). The order of abundance as determined by the Hahn method was considered to be correct because it seemed certain that a smaller proportion of the smaller, less visible species would be seen per kilometer walked. Tab1e 5. 25 Mean visibility distances and 95 percent confidence intervals calculated from first-sighting distances* and from Hahn- method visibility measurements, January-March, 1974, Deux Balé National Park, Upper Volta. Distance by size-c1ass** First-sighting First-sighting Distance by species distance Hahn method Measure- Species Sightings Mean(m) Sightings Meanon) ments Mean(m) Grimm's 63 84:13 du'ker 121 91:13 208 103:5 Oribi 58 98:22 Bushbuck 33 101:20 33 101:20 208 132:7 Western hartebeest 20 171:38 89 180:15 208 163:8 Roan antelope 69 183:16 *First-sighting distances were ocular estimates **See text 26 Table 6. Relative abundance* of fourteen mammals from counts by pairs of observers on foot, February-March, 1974, Deux Bale National Park, Upper Volta. Count (Distance Walked) Average February_(350 km), March (352 km) (702 km) Number Number Number Number Number Species seen per km seen per km per km Roan antelope 140(25)** 0.400 367(40)** 1.043 0.722 A"“°'S 54(11) 0 183 245(24) 0 595 0 440 baboon ° ' ‘ Oribi 59(33) 0.169 61(27) 0.173 0.171 Hartebeest 60(11) 0.171 36(9) 0.102 0.137 Grimm's duiker 37(32) 0.106 36(33) 0.102 0.104 Elephant 8(2) 0.023 54(6) 0.153 0.091 Patas monkey 23(3) 0.066 28(12) 0.080 0.073 Bushbuck 24(20) 0.068 18(13) 0.051 0.060 Buffalo 2(1) 0.005 30(1) 0.085 0.045 Side-striped jackal 4(2) 0.011 2(1) 0.006 0.008 Green monkey 6(3) 0.017 0 0.000 0.008 Waterbuck 1(1) 0.003 2(2) 0.006 0.004 Western kob 3(3) 0.009 0 0.000 0.004 Warthog 1(1) 0.003 0 0.000 0.001 Totals 432(148) 1.234 879(168) 2.497 *See text **(Number of observations) 27 Table 7. Relative abundance* of twelve mammal species from road counts by single observers on motorbicycle, Deux Bale National Park, Upper Volta. Road Count (Distance Traveled) Jan-Mar, 1972 (642 km) Dec, 1973-Jan,_1974 (366 km) Number Number Number Number Species seen per km seen per km Roan antelope 212(23)** 0.330 l69(l4)** 0.462 Anubis baboon 78(6) 0.121 156(11) .426 Hartebeest 50(12) 0.078 40(4) .109 Oribi 40(21) 0.062 24(14) .066 Green monkey 26(4) 0.040 20(9) .055 Patas monkey 12(1) 0.019 11(6) .030 Bushbuck 11(8) 0.017 6(5) .016 Grimm's Duiker 11(11) 0.017 5(5) .014 Waterbuck 8(2) 0.012 6(1) .016 Red-flanked duiker 2(2) 0.003 2(2) .005 Warthog 0 0.000 3(1) .008 Western kob O 0.000 2(1) .005 Totals 450(91) 0.699 444(73) .212 *See te t . ** Number of observat1ons) 28 Table 8. Hahn-method estimates compared to animals-seen indices ’ during walking-strip counts, February-March, 1974, Deux Bale National Park, Upper Volta. Hahn-Method Population Index Animals Seen Per km Species February March Change February March Change Grimm's Oribi 470 485 +3% 0.169 0.173 +2% Bushbuck 147 130 ~12% 0.068 0.051 -25% Hartebeest 269 251 -7% 0.171 0.102 -40% Roan . antelope 942 1546 +645 0.400 1.043 +160% Totals 2122 2698 +27% 0.914 1.471 +61% 29 The observer's speed (25 km/h) and the engine noise during the motorbicycle counts (Table 9), too, introduced likely biases against sightings of the smaller antelopes. In comparison to the Hahn estimates, the ratios of hartebeests to roans were higher in the motor- bike counts while those of the three smaller antelopes to roans, parti- cularly Grimm's duiker, were lower. The order of abundance computed from motorbicycle data was also different from that of the Hahn esti- mates, but was consistent between years. The proportions of the five species in the 1972 and 1974 motorbike counts were not significantly different (x2 = 2.37) either. The relative-abundance methods do not allow for visibility changes between years and such changes can apparently be considerable in the Deux Balé (Table 10). Although not accurate measurements of visibility changes, these data (Table 10) seem to indicate that changes substantial enough to negate the usefulness of indices devoid of some assessment of the area visually covered do occur. The results of the Hahn method and the two relative abundance methods do not clearly show which is the best index. The relative- abundance counts are simpler and demand appreciably less work, but would require modification to produce results as consistent as those of the Hahn method. The Hahn method probably better approximates the order of abundance and it certainly allows better evaluation of year- to-year visibility changes. Overall, the Hahn method, as used in this thesis, would seem to be the more accurate index to ungulate popula- tions in the Deux Balé. 30 Table 9. Ratios of numbers seen per kilometer traveled for five ungulate species compared to ratios of Hahn-method index estimates, Deux Balé National Park, Upper Volta. Number Seen Per km Ratios Hahn- Observer on 2 observers method motorbieycle on foot ratios Species 1972 1974 1974 1974 Roan antelope 1.00 1.00 1.00 1.00 Hartebeest 0.24 0.24 0.19 0.19 Oribi 0.19 0.14 0.24 0.38 Grimm's duiker 0.05 0.03 0.14 0.23 Bushbuck 0.05 0.03 0.08 0.11 ”15‘3"“ 542 355 702 702 traveled (km) 31 Table 10. Changes in mean visibilities computed from estimated first-sighting distancesf and from Hahn-method visibility measurements**, Deux Bale National Park, Upper Volta. Hahn-method mean visibility (m) First-sighting distance mean visibility (m) (measurements) (sightings Species 1973 1974 Change 1973 1974 Change Grimm's duiker 54 103 76 91 and Oribi (253) (208) +91% (31) (121) +20% 78 132 78 101 Bushbuck (253) (208) +69% (8) (33) +29% Roan antelope and (£155 (:33) +47% ('33) (18890) +54% Hartebeest *First-sighting distances were ocular estimates **The criteria for disappearance differed between years. In 1973, distances to the first point of disappearance were measured, any subsequent reappearances were ignored. In 1974, distances to the last point of disappearance (no further reappearance) were measured. 32 Principal Species in the Park and Surrounding Areas Sufficient numbers of roan antelope, oribi, Grimm's duiker, western hartebeest, and bushbuck were observed in the Dibon Forest to permit Hahn estimates. Though known to be present in the nearby Baporo and Laba Forests from spoor and sightings, no hartebeests were seen during the June count; population estimates were calculated only for the other four species. Densities of all five species were generally lower in the forests than in the park (Table 11). Both the vegetation and water resources in the forests being similar to those in the Deux Balé, it was likely that human pressures--livestock grazing, poaching, fishing in waterholes, human settlement--were responsible for depressed wild- life densities. Mapped wildlife locations in the forests revealed higher animal densities away from human populations there. Certainly wildlife in the forests received much less attention from the Departe- ment des Eaux et Foréts than did the park. Negligible surveillance in the forests undoubtedly encouraged illegal acts within their bounda- ries. Possibly, they served as buffer areas for the park. Ungulate densities in the park were considered to be well below carrying capa- city (Sihvonen, 1974). The Black Volta River barred most large-mammal movements between the park and the Baporo and Laba Forests although elephants are known to cross the river with impunity even at flood-stage. The intermittent Grand Balé Creek probably permits wildlife crossings between the park and the Dibon Forest over much of the year. With the 33 Table 11. Comparative Hahn-method densities for five ungulate species in April-June, 1974, Baporo, Dibon, and Laba Forests, Upper Volta. Index-Densities Per km2 Species Baporo Dibon Laba Park** Roan antelope 2.06 0.71 1.09 2.20 Oribi 0.52 0.68 0.40 0.84 Grimm's duiker 0.33 0.36 0.30 0.51 Hartebeest * 0.75 * 0.46 Bushbuck 0.22 0.18 0.17 0.23 Totals 2 3.13 2.68 1.96 4.24 (areas in km ) (81) (245) (188) (566) *Though known to be present, no hartebeests were seen during the June, 1974 counts in the Baporo and Laba Forests. **See also Table 1. 34 exception of the roan antelope, ungulate densities on the north (park) and south (Dibon Forest) banks of the Grand Bale in March-April, 1974 were similar (Table 12). The unequal densities of roans and the restricted grass-flush area in the Dibon strongly suggested an influx of roans to newly-green areas in the park. Visibility distances in the several areas (Table 13) were similar except that the visibility of the large size-class decreased from the dry- to the rainy-season. Apparently for the small and medium size-classes, the decrease in visibility due to shrubs leafing out was offset by the collapse of dry, partially-burned grasses. Additional Species: Territorial Mapping_ Mapped territories indicated that there were twenty-eight 2 in the Deux Balé. This 2, 21 kmz, and 20 km2 anubis baboon home ranges averaging 20 km figure is essentially identical to the 20 km average home ranges found by Child (1974), Green (1976), and Heister- berg (1977), respectively. The average troop size in the park was 23 baboons, far below the averages of 40 and 45 animals found by Child (1974) and Green (1976), respectively. The latter authors believed troops split into feeding groups of about twenty baboons, but such behavior was not noted in the park. As only 4 of 88 counts of baboons in the Deux Balé in 1974 exceeded 30 animals, 23 baboons per troop was considered valid and was used to estimate the baboon population (Table 14). The average home-range size of the patas monkey was deter- mined to be 33 kmz. The average troop size was twenty. Since Dorst 35 A_N.ov Amk.ov Am_.ov Ame.ov Aem.ov “mete: eeewo N_F mp, mm mo_ Am .xeee eeeem La 0 m Apm.ev Ame.ov ANN.oV Aek.ov Amm.ov x aewmm ”we” INN mo. em om, em .xmee eeeez maopwpcm ummmnmpcm; xuancmsm wnwgo wawzv mmc< zoom :Lmummz Pesto ANEx Lma mmwuwmcmo ucmv mgmnsaz xmucHicowuopsaoa .ekmp _eee< m-_ .emeeea eeewo 5:» nee .eNmF eeeez 5N-mN .xgea Fecowpez wpmm xzmo mg» :pon cw xmmco wpmm ucmgc mcwgmvgon mewgm max mmp co mmpmpzmcz mo mowumam m>wm toe A N Ex awn mmwuwmcmn use mg¢p5== mmv xmccptcopumpznoa congestcce: .N_ m_neh 36 maoquce :moa ehemp meme_ eeme_ meme, nee 655585628: 85284 meomp NHNmF meow, NHNmF xozagmzm Ezwumz ee_o_ memo_ meeo_ memo_ eexwmwem.numew PFeEm mcsaizez cha< mcaouxez 282-58: mmwumam mmepu 8384 conwo ocoamm wpmm xamo -mem 52885: e_ eeeeemeo xuw_wewmw> gee: .mppo> Luna: .mpmmgo: can; use .coawo .ocoamm ucm xgma Pocowpmz wpmm xzmo .emmp .mczaizgmagnwm .cwmm on upzou mmem macs» we mmpepsmca cows: op mmucmummu to» m—m>gmu:w mucmcwmcou ucmugma mm was mummcm>< .m— m_neh 37 Table 14. Populations and densities of twenty-two large mammals estimated from the results of Hahn-method counts, animals seen pfr kilometer indices, and territorial mapping in the 555m Deux Balé National Park, Upper Volta. Species Numbers Density/ km2 Roan antelope 1250 2.21 Anubis baboon 650 1.15 Oribi 650 1.15 Grimm's duiker 500 0.88 Western hartebeest 450 0.80 Patas monkey 350 0.62 Green monkey 225 0.40 Bushbuck 200 0.35 Red-flanked duiker 150 0.27 African elephant 120 0.21 Side-striped jackal 100 0.18 Western kob 100 0.18 Waterbuck 75 0.13 Warthog 75 0.13 Serval 50 0.09 Bohor reedbuck 50 0.09 African buffalo 40 0.07 Caracal 20 0.04 Leopard 10 0.02 Spotted hyena 10 0.02 Hippopotamus 6 0.01 African lion 2 tr 38 and Dandelot (1970) list the average home range of this species to be 30 km2 and the average troop size to be fifteen, the Deux Balé data do not seem unusual. The average frontage per troop of the best-known green monkey home ranges was 3 km with an average of seven green monkeys per troop. A density of four red-flanked duikers per 3 km of frontage was esti- mated. Approximately 95 km of frontage suitable for monkeys and 115 km for duikers were identified in the park. Total Populations of Large Mammals andrUngulatefiBTomass The author considered the Hahn-method estimates (Table l) to be low for the reasons stated. Estimates considered reasonable by the author were made subjectively for those five ungulate species and for thirteen other large-mammal species found in the Deux Balé (Table 14). In the March, 1974 walking-strip count, few roan on the transect strips were thought to be unobserved. The resulting estimate (Table 1) was considered accurate, but was believed to include about 300 roans from the Dibon Forest (Table 12). The roan population esti- mate (Table 14) reflects these beliefs. It was thought that at least two hartebeest herds numbering four or more animals escaped observation during both the February and March, 1974 counts. The Hahn-method estimates (Table l) were increased 75 percent to give the hartebeest population estimate (Table 14). It was believed that appreciable numbers of oribi, Grimm's duiker, and bushbuck were not observed during the counts. Particularly, the young of these species were rarely encountered. Further, it was 39 felt that the mean Hahn-method visibility distances probably overestimated the visibilities of Grimm's duiker and bushbuck. The Hahn-method estimates (Table 1) of the oribi, Grimm's duiker, and bush- buck populations were augmented 35 percent, 70 percent, and 45 percent, respectively (Table 14), in consideration of these beliefs. The population estimates (Table 14) of thirteen infrequently- encountered species were based on sightings, spoor, and species behav- ior. Their orders of abundance compared to those of more commonly observed species in the foot and motorbicycle counts (Tables 6 and 7) also influenced the estimates. Because of possible reader interest, the large-ungulate bio- mass of the Deux Balé was calculated (Table 15) and compared to bio- masses in other West African parks and reserves (Table 16). The Deux Balé's biomass was intermediate to those of four other areas, three of which (Comoé National Park (Geerling and Bokdam, 1973), Borgu Game Reserve (Child, 1974), and P6 National Park (Heisterberg, 1977)) were also below carrying capacity. The Arli National Park was considered at carrying capacity only for anubis baboon, buffalo, warthog (Phacochoerus aethiopicus Pallas), and western kob (Green pers. comm.). 4O Table 15. Estimated biomass of large-ungulates, Deux Balé National Park, Upper Volta. S Estimated Average . Total Biomais pecies Populat1on We1ghts*(kg) B1omass(kg) (kg/km ) Roan antelope 1250 250 312,500 552 African elephant 120 2100 252,000 445 Western hartebeest 450 150 67,500 119 African buffalo 40 500 20,000 35 Waterbuck 75 150 11,250 20 Bushbuck 200 50 10,000 18 Hippopotamus 6 1400 8,400 15 Oribi 650 12.5 8,125 14 Western kob 100 70 7,000 12 Warthog 75 70 5,250 9 Grimm's duiker 500 10 5,000 9 Bohor reedbuck 50 40 2,000 4 Red-flanked duiker 150 10 1,500 3 Total large-ungulate biomass 710,625 T255 *Estimated from Lamprey (1964) and Geerling and Bokdam (1973). 41 Table 16. Biomass values for large ungulates in several West African parks and reserves. m Locality Biomass (kg/kmz) Comoé National Park, Ivory Coast** 566 Borgu Game Reserve, Nigeria* 703 Deux Balé National Park, Upper Vo1ta++ 1255 P6 National Park, Upper Volta+ 1995 Arli National Park, Upper Volta*** 2202 *Chde (1974) **Geerling and Bokdam (1973) ***Green (1976) +Heisterberg (1977) ++This study CONCLUSIONS AND RECOMMENDATIONS Count Methods The Hahn method for censusing white-tailed deer was found to be of uncertain value for species with varied herd size and stratified distribution. Transects, too, could not be located in a manner certain to be representative of the Deux Bale region's vegetation. Although the method was not found to be useful as an absolute census, it seemed to provide consistent results as an index to the abundance of the region's rather sparse antelope populations. Calculation of mean visibility distances independent of animal observations made the method particularly useful for assessing changes in low-density popu- lations where animal observations were limited. The continued use of the Hahn method is recommended as an index to species abundance. The use of relative-abundance counts either from a motor- bicycle or on foot as indices to population changes is feasible if the counts are amended to allow for estimation of mean visibility. The averages of the observed disappearance distances of the species encountered during counts could be used to estimate annual changes in animal visibility. In contrast to relative-abundance counts on foot, those on motorbicycle sample an area biased toward riverine and floodplain vegetation types. Both count methods tended to distort relative- 42 43 abundance data because of the unequal visibilities of the species counted. Care must be exercised in interpreting such relative- abundance results. To avoid complications due to weather such as those with roans in the March, 1974 count, successive annual counts should be scheduled at phenologically-similar dates. Meteorological data Appen- dix 1) indicate that the January-February period normally has the most consistent weather. In addition to information on population numbers, trends, and distributions, other benefits accrue from walking-strip counts. Poachers may be apprehended or their blinds located and destroyed. Other illegal entrants--fishermen and herdsmen--may be caught or their activities curtailed. The location of waterholes and scenic areas may be noted and the status of vegetation ascertained. The counts serve multiple purposes, and their costs, mainly for transportation, are nearly the same as for routine patrols. Territorial mapping has been employed by several authors (Child, 1974; Green, 1976; Heisterberg, 1977) to estimate primate populations as well as populations of species demonstrating highly- localized distributions. This method does provide a logical means of estimating population size in such species. Management Implications Low wildlife densities and the level of illegal activities-- poaching, livestock grazing, fishing, fire--in the park and forests indicate the need for regular surveillance by Eaux et Forets personnel. 44 If the protection of natural area values is to be undertaken seriously, the budget allotted to the cadre of rangers must be augmented. In addition to suppression of illegal activities, efforts to educate the local citizenry on the values of parks should be made. Without public support, wild areas in the increasingly-settled Deux Balé region will face a difficult future. The low wildlife densities indicated by counts (Tables 1, 9, 13, and 14) and the low animal numbers found along the tourist track (Table 14) signal that the Deux Bale National Park is not ready for wildlife-oriented tourism on a large scale. While habituating wild- life to the passage of vehicles, the active promotion of tourism would do little to enhance the park's reputation as a wildlife-viewing area. Only a gradual increase in the encouragement of tourism is recommended. The present vehicle track cannot accommodate many tourist! parties: it is too short, misses many scenic spots, and bypasses several areas with high wildlife densities. The road network should be redirected and lengthened. Additional viewing areas are present in the Dibon Forest since the forest's wildlife densities are similar to those of the park (Table 10). Expansion of the park to include the Dibon is recommended. SUMMARY In the Deux Balé National Park, an area of 555 km2 in the Sudan savanna zone (Keay, 1959) of west-central Upper Volta, West Africa, the populations of twenty-two large mammal species were inven- toried during the period January, 1972-June, 1974. The populations of the five most common antelope species also were estimated in three forest reserves which are contiguous to the park and which cover 514 1on2. The roan antelope, oribi, Grimm's duiker, western hartebeest, and bushbuck populations were inventoried by Hahn (1949) visibility- profile procedures. Visibility distances for three size-classes of antelope were computed and were used to calculate average strip widths. Herd-forming tendencies in roan antelope and western hartebeest resulted in non-random distributions which were partially rectified by the use of two herd-size categories. Bushbuck preference of riverine habitat necessitated the areal stratification of counts for that species. Because some animals were thought to have been missed during counts, the Hahn method was believed to underestimate population numbers. The consistency of successive estimates, however, indicated that it might well serve as a good index to species abundance. The numbers of individuals of each kind seen per kilometer of travel were inferior to the Hahn method as indices to the five com- mon antelope species. Such tallies distorted the order of abundance 45 46 in these species and gave no evidence that they would compensate for year-to-year changes in animal visibility. The populations of anubis baboon, green monkey, patas monkey, and red-flanked duiker were estimated with modified territorial map- ping techniques (Overton, 1969; Green, 1976). The number of thirteen other large-mammal populations were judged from direct observations and spoor. The Deux Balé's large-ungulate biomass was intermediate to the biomasses of four other West African parks and reserves. REFERENCES REFERENCES Anon. (1966). Apercus sur 1e climat de la Haute-Volta. Service Météorologique, Ouagadougou, ASECNA, Haute-Volta. Anon. (1968). Carte de reconnaisance photogéologique. Assistance Technique, Neerlandaise, Hollande. Anon. (1972). Tableaux climatologiques (Periode 1961-1970). Service Météorologique, Ouagadougou, ASECNA, Haute-Volta. Begue, L. (1937). Contribution a l'étude de,la végétation forestiére de la Cfite d'Ivoire. Publ. Comité Etudes Hist. Scien. Afrique Occidentale Francaise, Paris. Serie B, No. 4, 127 p. Child, G. S. (1974). An ecological survey of the Borgu Game Reserve, Nigeria. Kainji Lake Research Project, Technical Report 4, FAO, Rome. 130 pp. Dorst, J., and Dandelot, P. (1972). ’Guide des grands mammiféres d'Afrique. Delachaux et Niestle, Neuchatel. 286 pp. Dupuy, A. (1968). Le Parc National du Niokolo-koba. Mém. IFAN, 3, 443-472. Geerling, C., and Bokdam, J. (1973). Fauna of the Comoé National Park, Ivory Coast. Biol. Cons. 5, 251-257. Green, A. A. (1976). Density estimate of the larger mammals of Arli National Park, Upper Volta. MS Thesis, University of Massachusetts, Amherst. 206 pp. Hahn, H. C., Jr. (1949). A method for censusing deer and its applica- tion in the Edwards Plateau of Texas. Texas Game, Fish and Oyster Comm. 24 pp. Heisterberg, J. F. (1975). Further notes on P6 National Park, Upper Volta: Ecological surveys and development prospects. Un- published report to the Departement des Eaux et Foréts, Ouagadougou, Upper Volta. 70 pp. . (1977). Flora and fauna of P6 National Park, Upper Volta, West Africa. MS Thesis, Purdue University, West Lafayette. 132 pp. 47 48 Hirst, S. M. (1969). Road strip census techniques for wild ungulates in African woodland. J. Wildl. Manage. 33, 40-48. Jacobsen, N.H.G. (1974). Distribution, home range and behaviour pat- terns of bushbuck in the Lutope and Sengwa Valleys, Rhodesia. J. South. Afr. Wildl. Manage. Ass. 4, 75-93. Keay, R.W.J. (1959). A vegetation map of Africa. UNESCO, Oxford. 24 pp. Kershaw, K. A. (1968). A survey of the vegetation in Zaria Province, N. Nigeria. Vegetatio 15, 244-268. Lamprey, H. F. (1964). Estimation of the large mammal densities, bio- mass and energy exchange in the Tarangire Game Reserve and the Masai Steppe in Tanganyika. E. Afr. Wildl. J. 2, 1-46. Lavieren, L. P. van, and Bosch, M. (In press). Large mammal densi- ties in Bouba Ndjidah National Park, Cameroon. Terre et Vie. Leprun, J. C., and Moreau, M. (1968). Etude pédalogique de la Haute- Volta région: nord-ouest. 0.R.S.T.0.M. 6500-399. 346 pp. Monfort, A., and Monfort, Nicole. (1974). Notes sur l'écologie et le comportement des oribis (Ourebia ourebi Zimmermann, 1783). Terre et Vie 28, 169-208. Overton, W. S. (1969). Estimating the numbers of animals in wildlife populations. Page 417 in R. H. Giles, Jr. ed. Wildlife Management Techniques, 3rd edition. The Wildlife Society, Washington, D.C. 623 pp. Plaisance, G. (1936). Avant projet de Forét Classée des Deux Balé. Unpublished report to the Service Forestier, Abidjan, Ivory Coast. Robinette, W. L.; Loveless, C. M.; and Jones, D. A. (1974). Field tests of strip census methods. J. Wildl. Manage. 38, 81-96. *Rosevear, R. D. (1965). The bats of West Africa. British Museum (Natural History), London. 604 pp. Sihvonen, J. P. (1974). Rapport sur un questionaire d'enquéte sur l'historique de la région des Deux Balé. Unpublished report to the Departement des Eaux et Foréts, Ouagadougou, Upper Volta. 11 pp. Troupin, G. (1966). Etude phytocénologique du Parc National de l'Akagera et du Rwanda Oriental. Inst. Nat. Recherche Scientifique, Butaré, République Rwandaise. No. 2, 293 pp. 49 Waser, P. (1975). Diurnal and nocturnal strategies of bushbuck Tragelaphus scriptus (Pallas). E. Afr. Wildl. J. 13, 49- 631 Wilson, V. J., and Child, G.F.T. (1964). Notes on bushbuck (Tragela- phus scriptus) from a tsetse fly control area in Northern Rhodesia. Puku 2, 118-128. *Quoted in Appendix 2. APPENDICES APPENDIX 1 APPENDIX 1. Summary of weather data from the Boromo Meteorological Station, Boromo, Upper Volta: (A) Average monthly rainfall (mm) from 1921-1974, (B) Monthly rainfall from 1971-1974, and (C) Average monthly relative humidity from 1961-1970 and average monthly evaporation (mm) from 1961- 1974. Taken from Begue (1937), Anon. (1966, 1970), and unpublished records of the Boromo Meteorological Station, 1970-1974. A. Average monthly_rainfall (mm) from 1921-1974. Period of record flppth_ 1921-1932 1933-1960 1961-1974 1921-1974 January tr 0.0 tr tr February tr 3.5 0.9 2.1 March 8.7 9.8 7.7 9.0 April 23.7 32.9 48.6 34.9 May 112.1 87.8 82.4 91.8 June 136.0 115.7 116.3 120.4 July 159.6 231.3 181.3 202.4 August 249.4 283.4 249.7 265.3 September 159.1 190.0 186.9 182.3 October 70.8 43.9 36.7 48.0 November 6.4 10.6 7.9 8.9 December 0.0 1.4 4.2 1.8 232:2; 925.9 mm 1010.3 mm 922.5 mm 955.9 mm 50 51 8. Monthly rainfall (mm) from 1971-1974. Year 119.8: 1.9.7.1. .1922. 19.29. 1.919.. January 0 0 0 0 February tr tr 2 0 March 58 tr 0 tr April 82 48 46 93 May 31 53 94 78 June 77 95 54 75 July 261 146 183 206 August 334 326 205 292 September 173 132 108 191 October 19 77 27 21 November 0 O 0 * December 16 0 0 * AW“ _ 1051 mm 877 mm 719 rrm 955 mm** total *Values unknown to author. **Total incomplete. 52 C. Average monthly relative humidity_from 1961-1970 and average monthly evaporation_(mm)ffrom 1961-1974. Relative humidity % Eygporation (mm) Mgpth 1961-1970 1961-1974 January 23 288 February 22 294 March 27 320 April 42 249 May 58 190 June 70 123 July 78 82 August 84 55 September 83 57 October 67 123 November 45 200 December 29 247 Am“ 52% 2.228 mm average APPENDIX 2 APPENDIX 2. List of mammals found in the Deux 8816 National Park, Upper Volta, based on field observations by the author and on the bat collection of J. F. Heisterberg. Common and scientific names after Dorst and Dande- lot (1970) and Rosevear (1965) (Chiroptera). INSECTIVORA African hedgehog CHIROPTERA Straw-colored fruit bat Gambian fruit bat Cretzschmar's free-tailed bat PRIMATES Green monkey Patas monkey Lesser galogo Anubis baboon LAGOMORPHA Crawshay's hare RODENTIA Cane rat Striped ground squirrel CARNIVORA Marsh mongoose Egyptian mongoose Slender mongoose Side-striped jackal Spotted hyena Common genet Caracal Atelerix albiventris (Wagner) Eidolon helvum (Kerr) Epomophorus gpmbianus (Ogilby) Tadarida ppmila (Cretzschmar) Cercopithecus aethiops (Linn.) Erythrocebus patas (Schreber) Galogo senegalensis (Geoffroy) Papio anubis (Fisher) Lepus crawshayi (De Winton) Thrypnomy§_swinderianus (Temminck) Xerus erythropus (Desmarest) Atilax paludinosus (Cuvier) Heppestes ichneumon (Linn.) fl, sanguineus (Ruppell) Canis adustus (Sundevall) Crocuta crocuta (Erxleben) Genetta genetta (Linn.) Felis caracal (Schreber) 53 African wildcat Serval Lion Leopard TUBULIDENTATA Aardvark PROBOSCIDEA African elephant ARTIODACTYLA Western kob Western hartebeest Red-flanked duiker Hippopotamus Roan antelope Defassa waterbuck Oribi Warthog Bohor reedbuck Grimm's duiker African buffalo Western bushbuck 54 E, libyca (Forster) E, serval (Schreber) Panthera leo (Linn.) E, pardus (Linn.) Orycteropus afer (Pallas) Loxodonta africana (Blumenbach) Adenota kob (Erxleben) Alcelaphus buselaphus major (Pallas) Cephalophus rufilatus (Gray) Hippopotamus amphibius (Linn.) Hippotragus equinus (Desmarest) Kobus defassa (Ruppell) Ourebia ourebi (Zimmermann) Phacochoerus aethiopicus (Pallas) Redunca redunca (Pallas) Sylvicapra grimmia (Linn.) Syncerus caffer (Sparrman) Tragelaphus scriptus (Pallas) 7:.'|GANSTATIE 11111 [PSIT 1 *31111913113111 1 11 987