E. Afr. WiIdf. J. 9 : 57-65 (1971)

THE COLLECTION OF FOUND IVORY IN MURCHISON FALLS NATIONAL PARK, UGANDA

A. D. GRAHAM, Wildlife Services Ltd., Nairobi; R. M. LAWS, Department of Zoology, Cambridge.

SUMMARY

The paper concerns the circumstances surrounding the collection of ivory from dead elephants, with particular reference to Murchison Falls National Park.

The characteristics of the interval between death and complete disintegration of an elephant are described. These, combined with observations of known age skeletons, comprised the criteria used in classifying skeletons found from the air into three relative age classes.

Average annual mortality is estimated for the population north of the Nile (MFPN) at 147 animals yielding 1945 kg of ivory, and for that south of the Nile (MFPS) at 474 animals yielding 7497 kg ivory. Park-found ivory records are analyzed for the 1 1 y 1959-69. The expected age distribution of deaths is compared with the observed. For MFPN a bias in favour of large (male) tusks is present, explicable by the concentration of ranger search effort in areas of known high male density. For MFPS a bias towards small tusks is thought to be caused by elephants wounded outside the park dying inside it.

The National Park recovers an average of 27.6% of its available ivory per annum, with large annual fluctuations probably correlated with the incidence of wounding outside the park. High losses to poachers are evident.

An aerial search for ivory showed a tendency for elephants to die near watercourses. A finding rate of one carcass every 4.3 km of watercourse was obtained. As only 5 % of carcasses still had tusks the aerial searching was prematurely terminated. The results indicated a finding efficiency of 26.4% of the available current year carcasses. Comparative costing suggests that ground searching would be a more efficient method of finding ivory than aerial searching. The high value of the available ivory in Murchison and other areas justifies intensive searching. The low collection rates prevailing in East Africa are largely attributable to the alisence of appropriate search efforts.

INTRODUCTION

Every year in Kenya, Tanzania and Uganda, ivory that may be worth as much as E.A. €2 million becomes available for collection by man as a result of natural mortality of elephant (Loxodonra africana. (Blumenbach)) (Watson and Parker, in prep.). Virtually all of it falls in state land, much of it in National Parks or other reserves; all of it is legally state property. A better understanding of the factors influencing the collection of this ivory is of obvious signi- ficance to wildlife management in view of its potential as revenue. This paper describes the results of an experiment conducted in

Murchison Falls National Park between 29 March and 6 April 1970 to discover the rate at which elephants dying naturally could be found from the air, and to determine whether this method could be used to aug- ment existing methods of ivory collection. Park records of found ivory are analysed, and the theoretical quantity of available ivory as deduced from age-specific mortality is compared with that actually found. Reasons for the observed discrepancies are speculated on, and possible means of maximising collection suggested. The colloquial term “found ivory” denotes the essentially fortuitously found tusks of dead elephants, specially those dying naturally. “Natural”

57

death here means that occurring other than by the hand of man, i.e. from starvation, disease, accident, or “senescence”.

METHODS

Collecting ivory from dead elephants Insofar as it relates to the finding by man

of dead elephants the interval between the death of an elephant and the complete disintegration of the carcass can be described in terms of a fairly regular sequence of events in most cases. Firstly, various carni- vores, particularly vultures (Gyps ruppelli (Brehm), Necrosytes nionachus (Temminck), Trigonoceps occipitalis (Burchell), and Pspudogyps africanzrs (Salvadori) among others), are attracted to the carcass. (Vultures are absent from some situations, e.g. the Aberdare National Park, Kenya (F. W. Woodley, pers. comm.).) Their activities combined with those of mammals such JS hyaenas (family Hyaenidae), lions (Pan- rhera leo (L.)), and leopards (Punrhern pardus (L)), are dependent on the size of the carcass, its age (since death), and the incidence of injuries to it. The thick skin of an adult elephant usually denies access by scavengers to the carcass except at limited sites such as the anus, mouth, etc. As a result the soft tissues may persist for several weeks. Small carcasses, or those with damaged skin that permit the entry of scavengers last only a few days under the combined actions of large carnivores, insect larvae, and decom- position. After removal of the soft tissues and small edible bones, the carcass, now a skeleton, exerts little attraction to scavengers, and in this state lasts for many years. Tusks, and large bones such as skulls, pelvises and limb bones last the longest, and are often the only visible relics of an old skeleton. Scavengers and elephants (the latter for unknown reasons) tend to scatter the bones and tusks which eventually disintegrate under the action of organic decomposition, weathering, termites, and fire. Ivory is also eaten by porcupines (Hystrix spp.).

In Murchison Park, where nearly all found ivory is obtained by rangers patrolling on foot, dead elephants are mostly discovered by seeing vultures descend to feed. Sometimes the existence of a carcass can be inferred from the signs of vulture-perching on nearby trees, and hyaena tracks. The long grass and

COLLECTION OF IVORY IN UGANDA

58

undulating terrain inhibit direct long range observation of a carcass.

From an aircraft observation is facilitated by certain characteristics of the dead animal and its surroundings. Trampling and decom- position tend to form a bare patch of soil round a carcass which may persist for several months after removal of the soft tissues. Paths radiating out from a carcass usually develop. The result is a conspicuous anomaly which, when the bones begin to bleach, becomes a strikingly prominent object.

For the aerial work it was essential to assign ages (since death) to the carcasses found. Absolute ageing being impossible a system of three relative age classes was adopted, based on the changing appearance of a carcass with time, as described above. Since bones and ivory can last indefinitely in ideally sheltered conditions, while those exposed to severe climatic fluctuations and fire may only last a few years, it is impossible to place finite limits on the life of carcasses. The following four examples indicate this range in carcass life.

1. Fourteen groups of skeletons were seen of elephants killed during sampling operations in Murchison Falls Park from 1965-67. These were therefore 3-4 y old at the time of observation. The bones were white and little scattered.

2. A partly shaded carcass on an island in the Nile with bones already bleached when first seen in February 1967 yielded ivory only slightly weathered in April 1970. The bones however were grey and crumbled when touched.

3. A 30 y old female elephant that died on 24.8.66 in the Aberdare National Park, Kenya, was observed on 25.7.70, 4 y after death. It had never been burnt and the bones were scattered over an area 50 m in diameter. These showed no signs of decomposition and weathering was insignificant.

4. Woodley (pers. comm.) has described the relocation of tusks in the Aberdare National Park by a hunter who cached them 26 y previously in thick forest. The ivory showed no signs of decomposition and weathering was insignificant.

The age classes adopted aimed primarily at separating current-year deaths (i.e.

A. D. GRAHAM AM) R. M. LAWS

elephants that died in the year preceding the time of observation) from older ones that represent an accumulation over an unknown number of years. These latter were further separated into two classes.

All skeletons seen were classed as “recent”, “old”, or “very old”, using the following criteria. Recent skeletons were those judged to be less than 1 y old by the extent of soft tissue decomposition, presence of the bare patch of soil, and little bone scattering. Old skeletons were those judged to be more than 1 y old by the absence of soft tissue, absence of the bare patch of soil, with bones still white, and with a majority of the large bones present visible simultaneously. Very old skeletons were those with the bones grey, and a minority of the large bones present visible simultaneously. (Single bones were not classified as skeletons.) Clearly there could be overlap in these age classes and the criteria do not always apply decisively.

With the object of establishing the rate at which carcasses could be found from the air a series of planned flights was made. Owing to the uneconomical rate of finding ivory (as distinct from carcasses) these observations unfortunately had to be pre- maturely terminated. The flights were made between 29.3.70 and 6.4.70. This time of year was chosen in order to exploit maximally the extensive burning of the preceding months which greatly facilitates visibility of objects on the ground. This factor is critical in Murchison Park where mature stands of grass can completely conceal an elephant carcass even from aerial observation. Observation was from a Piper Cub aircraft flown at 72-96 km/h, 60 m above ground level with observations made from the right- hand side through open doors. Reflected light was employed wherever possible (Graham and Bell, 1969). Three types of object were searched for: 1. Entire carcasses, 2. skeletons and 3. isolated tusks and bones. The positions of all such objects found were marked on a 1:25O,OOO map. Finds with tusks were collected by landing the aircraft within walking distance whenever possible, or with a ground vehicle.

Since carnivores and elephants sometimes scatter bones and tusks over considerable distances the decision as to whether or not the tusks of a dead elephant were

present was often a difficult one. tf not found immediately the aircraft was flown in widening circles round the carcass up to a radius of 500 m. While failure to find tusks did not necessarily mean they had already been found, it was uneconomical to search further.

RESULTS

Potential ivory available Average annual mortality rates for each

age class at 5 y intervals have been estimated for the populations north of the Nile (MFPN), and south of the Nile (MFPS) from the age structures indicated by cropped samples totalling 2000 animals (Laws and Parker, 1968). An average population of 3000 was assumed for MFPN (Watson and Parker, in prep.) and 7600 for MFPS (Laws, Parker and Johnstone, 1970). No allowance was made for irregularities caused by the seasonal or permanent distribution of deaths inside or outside the park. From the average tusk weights at pivotal age (5 y age classes) the weight of ivory made available is calculated. For MFPN the annual mortality is estimated at 147 animals yielding 1945.1 kg of ivory. For MFPS it is 474 animals yielding 7497.1 kg of ivory. The results are summarised in Tables 1 and 2.

Park-found ivory records Found ivory records were available for

the 11 y 1959-69. These give the date of weighing, the finder, the ranger patrol area the tusks came from, tusk weights, and the reward paid. (Rewards are paid as an incentive to find ivory.) As relatively few tusks are brought in at any one time, and both tusks of a pair normally weighed consecutively, it was possible except in a few cases to pair the tusks and hence estimate the number of deaths the finds represented, and the average number of tusks per dead elephant. These records are summarised in Table 3.

Rewards are paid in three scales according to the condition of the ivory. Fresh, undamaged ivory receives the highest reward of Shs. 31- per lb; slightly weathered or badly damaged ivory the intermediate reward of Shs. 21- per lb; and old (weathered or rotten) ivory the lowest of Shs. 1/- per lb.

59

COLLECTION OF IVORY IN UGANDA

TABLE 1

.4rinual estimated natural death in MFPN. The amount of ivory contributed by each age class has been calculated by multiplying the product of the numbers dying and weight of larger tusk by 1.8 (the average number of tusksper elephant). Average population estimatedat 3000. The small amount of ivory contributed by

the 0-5 y age group has been ignored. Data derivedfrom cropped samples.

Males wt. larger

tusk (kg)

total ivory (kE)

Females wt. larger

tusk (kg)

total ivory (kg)

age group (years)

no. dying

no. dying

0-5 6-10

11-15 14 11 9

10

1 .8 3.6 5.6 8.9

45.4 71.3 90.7

10 8 7

10 9

0.8 1 .6 2 .3 3.2 4.5 6 .0 7 .3 7.7

14.4 23.0 29.0 16-20

21-25 26-30 31-35 36-40

160.2 57.6 72.9 86.4 92.0 97.0 88.6 44.3 29.5

634.7

10 8 5

13.2 18.6 25.9

237.6 267.8 233.1

8 7

41-45 46-50 51-55 56-60

Total

2 1 -

34.5 4 4 . 5 -

124.2 80.1 -

7 6 8 .2

8.2 8 .2

3 2 .-

70 1310.4 77

TABLE 2

Annual estiniared natural deaths in MFPS. The amount of ivory contributed by each age class has been calculated by niultiplying the product of the number dying and weight of larger tusk by 1.8 (the average number of tusks per elephant). Average population estimated at 7600. The small amount of ivory contributed bj,

the 0-5 y nge group has been irnored. Data derivedfrom cropped sumples.

age no. wt. larger total no. wt. larger total group dying tusk ivory dying tusk ivory

0-5 6-10 26 1.8 84.2 23 0.8 33.1

11-15 24 3.6 155.5 24 1 . 6 69.1 16-30 25 5.6 252.0 26 2.3 107.6 ? 1-25 27 8.9 432.5 35 3.2 201.6 26-30 36 13.2 855.4 34 4.5 275.4 31-35 30 18.6 1004.4 24 6 .0 259.2 36-40 23 25.9 1072.3 21 7.3 275.9 41-45 11 34.5 683.7 23 7.7 318.8 46-50 5 4 4 . 5 400.5 23 8 .2 339.5 51-55 2 56.7 ?04.1 19 8 .2 280.4 56-60 _ _ - ._ 13 8 .2 191.9

Tot;il 209 5144.0 265 2352.5

Males Females

(years ) ( k d tkg) tkg) tkg)

This permits old tusks to be distinguished in the records from recent tusks. All ivory for which the highest reward was paid was taken to be fresh ivory and classified as "current vear". All other tusks were classified "old". ?Very few receive the intermediate reward.)

The expected age distribution of deaths as derived from cropped samples has been compared with the observed distribution as deduced from Park records for the 6 y 1960-65 (Tables 4 and 5). These data have

been calculated as tusk weight in weight classes with 4.5 kg (10 lb) intervals. In MFPN the observed distribution shows a bias towards heavier tusks in all weight classes from 9.0-13.5 kg upwards, the degree of bias rising slightly with increasing tusk weight. In MFPS there is a bias in favour of small tusks of less than 4.5 kg.

The average found tusk weight is 11.9 kg on the north bank and 5.5 kg on the south bank, a reflection of the high incidence of

60

TAB

LE 3

Show

s th

e nu

mbe

r of

dea

d el

epha

nts f

ound

by P

ark

Rang

ers

in th

e 11

y 19

59-6

9. to

geth

er w

ith t

he n

umbe

rs a

nd w

eigh

ts of

thei

r tu

sks.

Ran

gers

do

not

repo

rt t

he

findi

ng of t

uskl

ess

elep

hant

s, th

ose

too

youn

g to

hav

e re

cove

rabl

e ivo

ry, o

r th

ose

with

tusk

s al

read

y ta

ken.

Cur

rent

yea

r fin

ds

Wei

ght o

f ivo

ry (c

urre

nt y

ear)

N

umbe

r of

tus

ks (c

urre

nt y

ear)

A

vera

ge tu

sk w

eigh

t (ru

rren

t yew

) O

ld fi

nds

Wei

ght o

f ivo

ry (o

ld fi

nds)

N

umbe

r of

tus

ks (o

ld fi

nds)

A

vera

ge tu

sk w

eigh

t (ol

d fin

d<)

YE

AR

S

1959

19

60

1961

19

62

1963

19

64

1965

19

66

1967

19

68

1969

36

158

112

104

224

298

288

177

190

186

121

708.

5 20

19.4

19

28.7

20

10.3

29

07.6

33

86.1

26

63.5

24

46.3

27

04.8

23

53.3

16

11.2

67

30

3 21

4 20

3 43

3 58

4 56

0 32

8 35

1 34

6 22

1 10

.6

6.7

9.0

9.9

6.7

5.8

4.8

7.4

7.7

6.8

7.3

66

5 9

6

9 11

7

8 13

4

4 99

2.0

50.3

21

3.2

31:3

18

8:2

206:O

44.0

7711

13

718

83.9

41

.7

88

6 15

6

11

17

7 10

14

6

4 11

.3

8.2

14.2

5.

2 9.

8 11

.8

6.4

7.7

9.8

14.0

10

.4

P

TA

BL

E 4

Con

pare

s the

annu

al es

timat

ed w

eigh

t dis

trib

utio

n of t

usks

from

natu

rally

dyi

ng

elep

hant

s in

MF

PN

(da

ta fr

om c

ropp

ed sa

mpl

es) w

ith th

e ob

serv

ed d

istr

ibut

ion

of f

ouna

ii o

ryfo

r th

e 6

y 19

60-6

5. (T

usk

wei

ghts

bas

ed o

n la

rger

tusk

of p

air;

tr

ue d

istr

ibut

ion

is to

war

ds s

light

ly l

ower

val

ues)

.

tusk

w

eigh

t cl

ass

(kg)

0- 4

.5

4.5-

9.

0 9.

0-13

.5

. . -

- -

. -

13.5

-18.

0 18

.0-2

7.0

27.0

-36.

0 36

.0-4

5.0

45.0

+

Exp

ecte

d

num

ber o

f tu

sks

mal

e fe

mal

e

50

88

38

66

20

16

10

-

4 2

-

-

-

-

Obs

erve

d

tota

l %

N

umbe

r of

%

tu

sks

of

tusk

s fou

nd

of

per

y to

tal

1960

-65

tota

l 13

8 46

.9

143

32.9

10

4 35

.4

112

25.8

20

6

8

54

12.4

16

5.

4 40

9.

2 10

3.

4 43

9.

9 4

1.4

22

5.1

0.7

13

3.0

-

7 1.

6 2 -

P P

2

0 P 3

t-

Tot

al

140

154

294

434

(721

~)

COLLECTION OF IVORY IN UGANDA

male tusks found on the former, and large number of small tusks found on the latter.

Over the 11 y 1959-69 annual average found-tusk weights ranged from 5.6-1 1.4 kg, with no trends apparent. The overall average for current year finds in the same period was 7.5 kg, and for old finds 9.9 kg which is as expected in view of the greater durability and visibility of large tusks.

Aerial search for ivory Before starting routine flights a number

of random transects was flown, and several tracts of country subjected to a total search. All carcasses found on these preparatory flights were within 200 m of a watercourse: a phenomenon previously commented on by Parker (pers. comm.). It was therefore decided to test for the validity of the assump- tion that a majority of elephant deaths occur in or near watercourses. Accordingly a rectangular area near the western boundary of the park totalling 184.3 k m 2 was searched first with transects, and then by following watercourses. As expense precluded additional samples i t was assumed that the finding rate in 2.6 h transecting of one carcass every 33 min was significantly lower than the finding rate in 4.75 h following watercourses of one carcass every 10 min. Further, all carcasses found transecting were in watercourses. As a result all further searching was along watercourses. Only one side was searched at a time, and in all 392 km of watercourse were examined. A total of 91 carcasses was found by this method, a finding rate of one every 4.3 km of water- course (both banks). The total number found by all methods was 119. Of these, 21 (17.7 %) were classed as recent, 57 (48 %) as old, and 41 (34.4%) as very old. One or both tusks were found accompanying six finds (579, and isolated old tusks from a further four elephants were also found which could not be associated with their originating skeletons. The length of watercourse flown related to area gave an estimate of 1 kni of watercourse per km2 of park, which gives an estimate of 4109 km of watercourse for the whole park. (This is watercourse !ength as defined by the flight path of an aeroplane; actual watercourse length is of course greater.) These results suggest that the watercourse search if extended over the whole park would have revealed 954 carcasses

of which 17.7% (164) would have been classed as recent, i.e. 26.4% of the estimated 621 available.

DISCUSSION

The observed distribution in found ivory from MFPN towards values which were higher than expected may be accounted for as follows. Some bias towards larger tusks is to be expected as a result of the physical circumstances surrounding the finding of dead elephants by humans. Large carcasses, i.e. of older animals, take longer to decom- pose than small ones with correspondingly greater chances of being found. Secondly, the bias towards larger tusks is actually towards males because the smaller tusks of females contribute nothing to the biased weight classes (Tables 4 and 5). This bias towards male tusks is due to an uneven distribution of male and female elephants in MFPN and associated prejudices in ranger search behaviour. Elephant tusks appear to exert powerful attractions to humans that are only partly explicable by the traditional value of ivory. The larger the tusk the stronger the attraction, which is bolstered by the increased reward. Since rangers are unfamiliar with elephant biology they base their search efforts on purely subjective grounds. Legends tend to develop following the finding of large tusks, which in MFPN are based on fact since the Pakuba area and Tangi valley consistently support a high proportion of mature males. This means a high local death rate and high average tusk weight and as a result a dispro- portionately high search effort is applied to this part of the park. The observed distribution of deaths therefore probably reflects the population structure in the areas searched, but not the whole of the north bank.

In MFPS a different situation prevails. A bias in favour of small tusks less than 4.5 kg is apparent (recalculated on the basis of percentage composition by weight for tusks over 4.5 kg the expected and observed distributions are similar). The high incidence of small tusks could be due to several factors. As on the north bank there are legends about the best areas in which to find tusks. Most tusks found on the south bank come from the lower Weiga-Weiringo river systems which are subjected to a

62

A. D. GRAHAM AND R. M. LAWS

greater search effort than the rest of the south bank. The apparently high death rate on these river systems is a probable consequence of the known high incidence of wounding from Game Department control shooting in the nearby Budongo forest. (Laws, Parker and Johnstone, 1970). Since this shooting is indiscriminate a high propor- tion of wounded animals will be young ones.

Of an estimated 621 elephants dying annually in the park an average of 172 (27.6%) with ivory are found by park rangers. However they do not report tuskless elephants (an insignificant fraction), those too young to bear ivory, or those with tusks illegally taken (probably a large fraction). The real finding rate is therefore likely to be much higher than the apparent rate of 27.6%. An average of 13 old finds are made per year, representing 187.3 kg of ivory. Owing to the high rate of illicit collection Murchison Park is unlikely to have a large accumulation of old ivory. Although the chances of finding a carcass fall sharply once it ceases to attract vultures, it seems unlikely that this alone accounts for the failure to find approximately three-quarters of the available ivory. Discontinuities in ranger search effort (many areas receive little or no attention) probably account for a large fraction of unfound ivory. The size of this fraction must however remain conjec- tural in the absence of data on the pattern of search effort. A third factor is poaching. Annual illegal entrants to the park are probably to be numbered in the thousands

(Wheater, pers. comm.), and compared to the approximately 30 rangers available at any one time to patrol, can be expected to put in more man hours searching for ivory. That poachers take a significant amount is suggested by the fact that in Kenya National Parks, where illicit collection is negligible, at least 26% of their found ivory is old (Watson and Parker, in prep.), whereas in Murchison it constitutes only 7.4%.

Despite this, the possibility that much simply goes unnoticed and rots must not be minimised. A number of tusks will be lost as a result of deaths occurring in streams or in soft mud. Parker (pers. comm.) reports the fortuitous finding of two tusks that had sunk out of sight in mud, and for the existence of which no visible signs existed. The pattern of ranger search effort relative to the pattern of burning will also exert an influence in that burning temporarily enhances visibility.

Some peculiarities of the Park records of found ivory deserve comment. A striking feature is that the number of elephants found dead (excluding old finds) increased from 36 in 1959 to a peak of 298 in 1964. Thereafter numbers declined, and in 1969 only 121 were found. This fluctuation may represent, in part at least, variations in annual mortality, or in the distribution of deaths relative to the distribution of search effort. The collection of ivory by poachers may have increased or that by rangers decreased. The latter is improbable in view of the reward incentive and absence of any

TABLE 5

Compares the annual estimated weight distribution of tusks from naturally dying elephants in MFPS with the observed distribution of found ivory for the 6 y 1960-65. (Tusk weights are based on larger tusk ofpair; true distribution

is towards slightly lower valuer).

tusk weight class (kg) 0- 4.5

4.5- 9.0 9.0-13.5 13.5-1 8.0 18.0-27.0 27.0-36.0 36.0-45.0 45.0+ Total

Expected

Number of tusks

male female 100 284 104 246 72 - 60 - 46 - 22 - 10 - 4 -

41 8 530

total % tusks of per y total 384 40.5 350 36.9 72 7.6 60 6.3 46 4.9 22 2.3 10 1.1 4 0.4

948

63

Observed

Number of % tusks found of 1960-65 total 1079 59.3 470 25.8 132 7.3 61 3.4 50 2.4 18 1 .o 10 0.6 6 0.3

1819 (3031~

COLLECTION OF IVORY IN UGAKDA

(official) change in patrol routines. That illegal collection may have increased is highly likely. The park is surrounded by man and subjected to continuous poaching, which is likely to increase with the human population expansion. The reduction in the total population of elephants by 2000 as a result of the sampling in 1965-7 will have caused a temporary reduction in the total number of natural deaths. However the number of finds is probably influenced more by wounding outside the park than by any other factor. An average of four times as many finds are made on the south bank as on the north, yet the MFPS popula- tion is only slightly over twice the size of the MFPN population, and the age distri- bution of deaths similar. Numbers of elephant shot annually in the Budongo forest and its vicinity increased fivefold from 1961 to 1967, but thereafter decreased. The incidence of wounding however is thought to have decreased from 1965 onwards with the assumption by the Forest Department of supervision of the shooting (Johnstone, pers. comm.). These changes in the likely number of wounded animals correlate wjth the overall trend seen in the Park records. The more open country of the south bank also may contribute to the higher rate of finding there.

The apparent concentration of elephant deaths in Murchison Park along water- courses is unclear. It may be the result of a tendency for injured or diseased animals to stay near water and green vegetation and eventually die there: or it may be that progressively weakening animals take the line of least resistance over the terrain and consequently accumulate at the lowest points.

Several factors probably influenced the finding of carcasses from the air, among them observer errors, bias arising out of deaths occurring far from watercourses, and discontinuities in patterns of death.

COKCLUSIOXS

In yiew of the great potential value of found ivory, attempts to maximise its collection are of obvious importance to wildlife management. Some indication of the quantity of ilory made available annually in East Africa is given by Watson and Parker (in prep.). These authors estimate total

East African elephant numbers at I ,085,OOO. The potential ivory from natural mortality in these populations would exceed K.&2 million per annum in value. In practice it is probably much less owing to man-inflicted mortality modifying natural mortality patterns, e.g. by reducing the older age classes and therefore their potential contri- bution. In 1969 total East African ivory exports were valued at K.&459,000 which in view of the low collection rates of found ivory is obviously derived mainly from killed animals. Thus the estimation of the amount of ivory yielded by natural mortality in a given elephant population will depend on its biological status and degree of interaction with man, and will vary considerably from population to population.

Certain areas free of illicit collection also have potentially vast accumulations of ivory from past years’ mortality.

The average annual collection rate in Murchison Park of 27.6% of the available current year ivory is higher than that known for any other conservation unit. Kenya National Parks average 3.8% and the Serengeti National Park 1% (Watson and Parker, in prep.); figures for other units are not available. In general the prevailing low collection rates are attributable to the absence of appropriate search efforts.

The aerial searching in Murchison Park resulted in a finding efficiency of 26.4% of the available current year deaths. Aerial searching is a skill and therefore susceptible to a wide range in individual proficiencies; conditions in Murchison are different from those in other important elephant situations. Nevertheless it must be concluded that aerial searching shows little promise in terms of efficiency as a method of routine collection, although in areas with accumulations of ivory the short-term rewards could be large.

The cost of aerial searching in Murchison Park can be approximately assessed as follows. The average finding rate of one carcass every 10 min would require 160 h flying to find 26.4% of the elephants dying annually. The time required to fix the position of each find for subsequent ground collection is likely to equal 50% of the finding time. Two hundred and forty h at €10 per h is E2400, plus the expense of ground work. The flying would occupy at least one month and calls for specialized

64

A. D. GRAHAM AND R. M. LAWS

personnel. The value of the ivory found (assuming the elimination of poaching) at 7.5 kg per tusk, worth f 2 per kg, would be €4920.

A more promising alternative is intensified ground searching. A line of 100 men at 20 m intervals could comb the entire park in 40-50 d, and experience suggests that this method would reveal most carcasses. At Shs. lo/- per d, per man this would cost €2500 plus transport which would be comparatively little as finds would be brought in on foot to central points. The work could all be done by Park staff. The practicability of ground searching coupled with the expected high finding rate strongly suggest its application.

ACKNOWLEDGEMENTS

R. J. Wheater for hospitality, assistance and making available certain Park records. I. S. C . Parker made many useful comments on the manuscript, and the rangers of Rabongo Ranger Post put in considerable time and effort recovering ivory found during the study.

REFERENCES GRAHAM, A. and BELL, R.H.V. (1969). Factors

influencing the countability of animals. E. A.fr. agric. For. J., 24 (special issue): 38-43.

LAWS, R. M. and PARKER, I. S. C. (1968). Recent studies on elephant populations in East Africa. Sytnp. zool. SOC. Lond., 21:319-59.

LAWS, R. M., PARKER, I. S. C. and JOHNSTONE, R. C. B. (1970). Elephants and habitats in North Bunyoro, Uganda. E. Afr. Wildl. J., 8:163-180.

WATSON, R. M. and PARKER, I. S. C. (in prep.). The interaction between man and elephants. E. ~ f i . Wildl. J .

We are gratefbl to the late F. x. for permission to make this study, and to

Authors’ addresses: A. D. Graham, Wildlife Services Ltd., P. 0. Box 30678, Nairoti; R. M. Laws, British Antarctic Survey, Monks Wood Experimental Station, Abbots Ripton, Huntingdon, U.K.

(Received for Publication February, 1971)

65