Decaying Raphia farinifera Palm Trees Provide a Source of Sodium for Wild Chimpanzees in the Budongo Forest, Uganda

For some years, chimpanzees have been observed eating the pith of decaying palm trees of Raphia farinifera in the Budongo Forest, Uganda. The reasons for doing this have until now been unknown. An analysis of the pith for mineral content showed high levels of sodium to be present in the samples. By contrast, lower levels were found in bark of other tree species, and also in leaf and fruit samples eaten by chimpanzees. The differences between the Raphia samples and the non-Raphia samples were highly significant (p<0.001). It is concluded that Raphia provides a rich and possibly essential source of sodium for the Budongo chimpanzees. Comparison of a chewed sample (wadge) of Raphia pith with a sample from the tree showed a clear reduction in sodium content in the chewed sample. Black and white colobus monkeys in Budongo Forest also feed on the pith of Raphia. At present, the survival of Raphia palms in Budongo Forest is threatened by the use of this tree by local tobacco farmers.


Introduction
The chimpanzees (Pan troglodytes schweinfurthii) of the Budongo Forest Reserve, Masindi District, Western Uganda, were initially studied by one of us (VR) in 1962 [1], and have been studied continuously since 1990 [2]. The Budongo Forest habitat consists of moist semi-deciduous tropical forest, and contains a number of forest types, notably Cynometra Forest, Mixed Forest, Colonising Forest, and Swamp Forest [3]. The study community of chimpanzees is named the Sonso community after the River Sonso which runs through its range. At various places along the river, Swamp Forest prevails. Swamp Forest contains several Raphia farinifera palm trees. After a single flowering and fruiting, the trees of this species lose their foliage and die. The dead bole remains standing and is 15-30 ft high. The bole rots down until it consists of a hard outer bark with a soft, moist, fibrous woody pith. Chimpanzees make a small hole in bark at the base of the dead tree with their teeth, widen it with their fingers and later their hands. Through the hole, they extract dead pith, chew it thoroughly, swallow the juice and some particles of woody matter, and finally spit out a fibrous wadge of chewed pith. There is frequently competition for Raphia pith. Eventually trees are abandoned, often with large holes and much inner pith removed (Fig. 1). Until now, the reason for consumption of the dead Raphia pith has not been known. Table 1 shows the results of the chemical analysis expressed as mg per kg of dried material for each sample. In this table, results for elements for which no more than trace amounts were obtained have been omitted. With one exception (samples 7 and 8) samples were independent of each other, coming from different trees or at different collection dates. In the case of samples 7 and 8, sample 8 was a wadge (a spit-out chewed sample) of sample 7 and was omitted from the analysis. For purposes of analysis, the samples were placed into two groups: Raphia pith (n = 10), and all other samples (n = 16). Statistical tests were done using SPSS version 7. Data were not normally distributed (Lilliefors test p,0.000). Mean mineral contents of the two groups were therefore compared using nonparametric Mann-Whitney U-tests. Results are shown in Table 2.

Results
As can be seen from Table 2, significant differences between group means (Raphia vs. all other samples) were found for four of the minerals tested, sodium, calcium, manganese and zinc. In the case of calcium (p = 0.004), higher values occurred in the 'all other samples' group and reflect the normal high level of calcium in leaves and fruits. In the case of sodium (p,0.000), manganese (p,0.006), and zinc (p = 0.007) higher values occurred in the Raphia group. Bark, leaves and fruits from other tree species did not contain the high levels of sodium found in the Raphia samples. These results are illustrated by box-plots (Fig. 2).

Discussion
Since study of the Sonso community of chimpanzees began in 1990, individuals and small parties of the animals have been seen occasionally eating the pith of dead Raphia trees, but until now the reason for doing so has not been known. The first report showing high levels of sodium in samples of dead wood eaten by apes was that of Rothman et al for mountain gorillas (Gorilla gorilla beringei) in Bwindi Impenetrable National Park, Uganda [4]. In that study it was estimated that the sodium obtained from dead wood constituted 95% of the sodium in the gorillas' diet. Using similar methods to those employed here, a mean dry sodium weight of 810.7 mg kg 21 , with a range of 100-1920 mg kg 21 was measured.. This compares with the mean dry sodium weight for Raphia samples in the present study of 5432 mg kg 21 , range 1431-14616 mg kg 21 . Raphia palm trees in Budongo Forest therefore offer higher concentrations of sodium to the chimpanzees than dead wood does to the Bwindi gorillas.
Rothman et al found that sodium levels were significantly higher in dead wood than in other items in the diet [4], and this was also the case in the present study. As they point out, sodium is an essential item of diet for apes, lack of which has wide-reaching effects on health [5]. In the Budongo Forest too, there appears to be little sodium in the other plant parts eaten and tested here (bark, leaves and fruits), although sodium is present in small quantities. In particular, the greater part of the chimpanzees' daily diet (up to 95%) consists of leaves and fruits [6] which contain only small quantities of sodium. There is however one important dietary difference between chimpanzees and mountain gorillas. The Budongo chimpanzees eat meat sporadically, mainly in the form of colobus monkeys (Colobus guereza) which are hunted, and this provides an additional source of sodium. The interaction between Raphia eating and meat eating was not explored in this study.
As in the present study, Rothman et al found significantly lower values for calcium and phosphorus in decaying wood than in the rest of the diet [4]. They also found lower values for manganese which was not the case in the present study, in which manganese was higher in the Raphia samples (p = 0.059). The reason for this may be the high levels of manganese in groundwater along the Albertine Rift [7]. Mahaney et al found high levels of manganese in clay eaten by chimpanzees living in the Mahale mountains, Tanzania, also along the Rift Valley [8]. Significantly lower values for other minerals, magnesium, potassium, zinc, and copper found by Rothman et al were not found in the present study [4].
Black and white colobus monkeys (Colobus guereza) are also known to consume Raphia pith in Budongo Forest (pers. comm., field assistants at BCFS) and during the present study hairs of this species were twice found at Raphia feeding sites. Oates found high levels of sodium, iron, manganese and zinc in swamp plants eaten by black and white colobus monkeys living in Kibale Forest, western Uganda, also along the Albertine Rift [9]. High mineral content was also found in some clays collected from stream-banks in the forest. Concentrations for swamp plants and clays were higher than in dry-land leaves, buds and fruits constituting the major part of the monkeys' diet.
An interesting feature of the samples collected concerns samples 7 and 8. These were the only two non-independent samples collected, coming from the same tree during the same observation period. Sample 7 consisted of Raphia pith collected with a knife from inside the tree. Sample 8 consisted of a discarded wadge of Raphia pith from the same tree, collected on the ground at the feeding site. Comparison of the sodium content of these two samples shows that the uneaten Raphia pith (sample 7) contained 7194 mg kg 21 , whereas the chewed and spat-out wadge (sample 8) contained 1095 mg kg 21 , evidence that the individual which chewed this sample did indeed ingest sodium.
In recent years Raphia farinifera, a tree we now know to provide the chimpanzees of Budongo Forest with essential sodium, has become scarcer. Besides natural predators such as baboons and pigs that eat its shoots, it provides humans with two products. The living trees are occasionally felled and the trunk opened to allow air into the pithy centre, which then ferments and produces an alcoholic palm wine beverage which is bottled and sold. This, however, is not thought to be the main cause of the decline in numbers of Raphia. The main danger to this species comes from local tobacco farmers, who kill the tree during its growth period, before flowering and fruiting, in order to strip its leaves for Raphia string ('raffia'), which is then used to tie tobacco leaves during and after the the drying and curing process [10]. As a result, Raphia palms are becoming scarce in Budongo Forest. Tobacco farmers and British American Tobacco, a company which buys much of the crop, will be approached about this problem and it is hoped a solution may be found.

(a) Field collection
Between 16 Feb and 2 April 2008, samples of pith (including one wadge) from Raphia trees were collected, in each case when chimpanzees had been observed feeding on the trees just before collection. Samples weighed ,50 g. Pith samples were obtained from the inside of trees with use of a knife. The wadge was collected from where it was dropped, beside the hole in the tree. Each sample was placed into a sample tube using gloves or tweezers, tubes were closed and marked with sample no., date, tree location, species of tree, name(s) of chimpanzees feeding on the tree, nature of the sample (bark, pith or wadge), initials of collector. Collectors were in all cases trained field assistants of BCFS or in one case VR. During the same time period, samples of rotten wood, bark, leaves and fruits seen being eaten by chimpanzees were collected in the same way, in adjacent forest, to be analysed for comparison with the Raphia samples. All samples were taken within a period of 2 hours to the field base where they were removed from the bags with tweezers and dried separately in a warm dry space at room temperature. Samples were dry in 1-3 days and all dried satisfactorily. The dry samples were placed in polythene bags which were sealed, marked, and taken by air to the UK for analysis. In the UK prior to analysis they were kept in a dry room at room temperature. No samples were spoiled during the pre-analysis period.

(b) Laboratory analysis
The samples of material were dried to constant weight in an oven at 105uC. The samples were then ashed at 550uC in a pyrolysis oven. The total mass of the ashed material was determined before digesting a sample (circa 0.1 g) of each material in 3 mL of aqua regia in a water bath at 100uC for 2 hours. The digested samples were diluted to 10 mL using distilled water before serially diluting to obtain 1 in 10, 1 in 100 and 1 in 500 dilutions. The elemental content of each sample was then determined using a Perkin Elmer Optima 2100 DV Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES) using 5 and 10 mg L 21 elemental standards.

Further Information
Still images and video clips of Raphia eating are available from the first author and from C. Hobaiter, School of Psychology, St Andrews University, email: clh42@st-andrews.ac.uk