QH 1 El IX JOURNAL IF THE EAST AFRICA NATURAL HISTORY SOCIETY AND NATIONAL MUSEUM September 1985 Volume 75, No. 182 LITTORAL HERMIT CRABS (Decapoda: Anomura) FROM PORT MOMBASA P.J. REAY Department of Biological Sciences Plymouth Polytechnic, Plymouth PL4 8 A A, U.K. ABSTRACT Ten species of hermit crabs belonging to the genera Coenobita (2 species), Calcinus (2), Clibanarius (5) and Dardanus (1) were collected and identified from the supralittoral and intertidal zones of Port Mombasa during February and March 1983. A key to the identification of the species is presented. The species pair, Calcinus laevimanus and Clibanarius virescens were numerically co-dominant and occurred throughout the intertidal area. Actual densities of up to 82 crabs nr2 were recorded, but aggregations of Clibanarius virescens (eg. 74 in a 10 cm2 area) could give rise to much higher density values locally. The two dominant species were small in size (up to 14 mm total carapace length), and samples contained up to 36.2% ovigerous females. They occurred in 14 different shell types, and there was evidence of both species-and size-related shell selection. I also concluded, from the scarcity of empty shells and from the behaviour of crabs provided with empty shells, that suitable shells were a scarce resource for hermit crabs in the area. Compare to the dominant species, the other eight were either rare, or restricted to a particular habitat. For example, Coenobita species were mostly found in the supralittoral zone of sandy beaches. Both Dardanus lagopodes and Calcinus latens were restricted to the shallow sublittoral or lower shore pools, whereas all other species were usually found out of water at low tide. The results are compared with those from other areas, and some suggestions for further work are made. INTRODUCTION In spite of being conspicuous and often abundant members of the intertidal and supralittoral communities, hermit crabs do not appear to have been studied along the Kenya coast. Even in the Indian Ocean as a whole, little more than species lists occur in the literature, and studies on the semi-terrestrial Coenobita rugosus from Somalia and Aldabra (eg. Vannini 1976, Vannini and Chelazzi 1981) appear to be the only detailed work undertaken. Species lists, either of crabs or of intertidal marine fauna including hermit crabs, have been given for the Red Sea by Hughes (1977); for Aldabra by Taylor (1971a); for Diego Garcia by Taylor (1971b); for Madagascar and the Comores by Dechance ( 1964); for Tanzania by Hartnoll (1976); for Pakistan by Tirmizi and Siddiqui (1982); for South Africa by Barnard (1950), and for Somalia by Lewinsohn (1982). The latter is the most relevant paper as far as Kenya hermit crabs are concerned, and provides information on habitat, distribution and diagnostic features. Outside of the Indian Ocean area, and particularly around the coasts of North and Central America, the ecology and behaviour of hermit crabs has been extensively studied. Much of this work has concentrated on the mechanics and implications of the shell selection behaviour (eg. Fotheringham 1976, Spight 1977), agonistic behaviour (eg. Hazlett 1970, Dunham 1978) and interspecific co-existence (eg. Vance 1972a, Bertness 1981). General reviews have been provided by Reese (1969) and Scully (1983), and the broader role of gastropod shells in benthic communities is discussed by McLean (1983). Hermit crabs are of no direct commercial importance, but their abundance in some areas suggests Page 2 No. 182 that they may play an important role in littoral communities, and they are potentially in direct competition with man for empty gastropod shells. Indeed, they are referred to in this context in a poster issued by the Kenya Ministry of Tourism and Wildlife which advises that one reason for not gathering empty shells is that it deprives hermit crabs of a home. With this application in mind and, in the absence of any published data on Kenya hermit crabs, the present paper seeks to bring together observations made during February and March 1983 from the littoral zone of Port Mombasa. Figure 1 The study area on the eastern shore of Port Mombasa, showing specific area A — J referred to in text. No. 182 Page 3 STUDY AREA, MATERIALS AND METHODS The study area, with major features, is shown in Figure 1. Hermit crabs were observed and collected at low tide from exposed intertidal rock surfaces, pools, and the shallow sublittoral, and from the supralittoral zone of sandy beaches. Some samples were taken to the laboratory for examination and measurement, and some were maintained alive in captivity. All crab measurements refer to total carapace length (from rostrum tip to one of the posterior tips of the posterior carapace). Gastropod shell measurements refer to the maximum linear dimension. Further details of methods are given in the appropriate section below. RESULTS The Species Present and Their Identificaton From Barnard ( 1950) and Tirmizi and Siddiqui ( 1982), it is clear that four genera are represented in the study area. These are all from the superfamily Coenobitoidea, as the bases of the third maxillipeds are proximate (close together); in the other pagurid superfamily, the Paguroidea, they are widely separated. Within the Coenobitoidea, two families are represented: the Coenobitidae (with genus Coenobitus ) and the Diogenidae (with genera Calcinus, Clibanarius and Dardanus). A key to the identification of the ten species found in this study is given as Appendix 1 , and some of the key features are illustrated in Figure 2. A list of the species is as follows: Coenobita rugosus Coenobita cavipes Calcinus laevimanus Calcinus latens Clibanarius virescens Clibanarius merguiensis* Clibanarius longitarsus* Clibanarius eurvsternus Clibanarius laevimanus Dardanus lagopodes All identifications have been confirmed by Professor L.B. Holthuis, who also provided primary identifi- cation on species marked with an asterisk. H. Milne Edwards Stimpson (Randall) (Randall) (Krauss) de Man (de Haan) Hilgendorf Buitendijk (Forskal) Distribution and Abundance The different species were distributed as follows (referring to areas denoted in Figure 1): Coenobita rugosus and Coenobita cavipes. Restricted to the strand-line and above, on sandy beaches and the maritime (terrestrial) habitats on cliff-tops. Most specimens are obtained from area J and the beaches towards, Nyali, and Coenobita rugosus was the more common of the two species. Calcinus laevimanus. Widespread and abundant throughout the littoral zone and on most flat, rocky substrates within the study area. It was particularly characteristic of shallow pools and ridges towards the top of the shore, where algal growth was sparse (eg. areas B, C and I). This was the most widespread species encountered and, with Clibanarius virescens. was also the most abundant. Calcinus latens. Restricted to pools low on the shore and the shallow sublittoral. Nowhere common, but it was most consistently found in areas G and H. Clibanarius virescens. Widespread and abundant, probably the most abundant species overall. Present throughout the littoral zone and on most flat, rocky substrates within the study area, it was most closely associated with those sites having a turf of macro-algae, eg. area D and I. Here, unlike most of the other species it often formed dense aggregations. Clibanarius merguiensis. Present in very small numbers among Clibanarius virescens in area D. Clibanarius longitarsus. Only one specimen recorded, in area E, but probably more common on muddy substrates associated with mangroves. Clibanarius eurysternus. Widespread, but nowhere common, usually occurring singly in areas dominated by the other species, particularly in area I. Clibanarius laevimanus. Only recorded from the upper shore in areas E and F, but here it was the most common species, occasionally forming loose aggregations. Its habitat was distinctive and consisted of limited algal growth, shallow silty pools and irregular rocky projections. Dardanus lagopodes. Two specimens were found, one in the shallow sublittoral of area A and one in Page 4 No. 182 Figure 2 Schematic diagrams (based on freshly dead material) to show colour patterns on me outer surfaces of third right pereiopods of Calcinus and Clibanarius species, and on the left chela of Calcinus laevimanus. Bristles on pereiopods have been omitted, and only propodus and dactylus are shown. z z z ■x. * 1 z o Z z H c < a Z u | < I ^ , < Uc i PC >- 1 -i* cj i tJ CO X :z) z PC OS < < f— , < ■ — • z — z o U Z r* X 4-J • r—< • r-* 4-1 Z c . f-H •- 1 — 1 Cl X 1 — O ”U c c. o >> a. r3 o 4-1 *H Pj OJ Ol -C 'S. ac c c x « c » ' 1 — 1 GJ Z3 "D " O ac c X 2 01 "O op cc l— OC &3 DS < z C PC i— ai z z o I S z U as > I z No. 182 Page 5 a deep pool in area H, but Dardanus species are primarily sublittoral and are therefore not likely to be common on the shore. A limited amount of information on the relative abundance and density of Caleinus laevimanus and Clibanarius virescens was collected. This can be summarised as follows: 1 . In a 1m2 quadrat set adjacent to one of the pilings in area D, repeated counts were carried out on 8, 1 4 and 1 8 February. As the hermit crabs were removed form the area on each occasion, the latter two counts reflect invasion rates. The results were: February 8 64 Caleinus laevimanus 16 Clibanarius virescens 2 Clibanarius merguiensis February 14 24 Caleinus laevimanus 10 Clibanarius virescens February 28 35 Caleinus laevimanus 14 Clibanarius virescens The quadrat was selectively placed in an area know to be favoured by Caleinus laevimanus, and the initial count can be used as a rough approximation to the maximum density observed for this species. 2. A second lm2 quadrat was selectively placed in an area of algal turf in a typical Clibanarius habitat of area D. In this quadrat, on February 1 1 , 59 Clibanarius virescens, 1 2 Caleinus laevimanus and 1 Clibanarius merguiensis were counted. This is typical of the densities encountered in such areas, but dense aggregations of the former species can result in much higher values. For example, an aggregation of Clibanarius virescens in area D occupied 10 cm- and contained 74 individuals. Such aggregations were quite common, and occured on open surfaces as well as under small overhangs, and occasionally under stones. 3. In two transect lines set parallel to and, respectively, 2m and 5m from, and to the west of, the pilings (area C), a mean density of 1 7.3 hermit crabs m'2 was found between low water mark and the sandy beach. The distribution was very patchy with many of the 0.25m2 quadrats containing no crabs and with the maximum number per quadrat being 117. Population Structure The population structure is described only in terms of size, but the proportion of berried females was also noted. Data on population structure are restricted to the species Caleinus laevimanus and Clibanarius virescens from samples obtained in area D (Figure 3). In Caleinus laevimanus I noticed that berried females had a smaller area of white on their enlarged left chelae compared to those taken to be males, which were occasionally entirely white (see Figure 2). This was assumed to be a straightforward sexual difference, but an isolated captive individual which moulted showed a ‘white’ pattern on the moulted chela, but a ‘black’ pattern on the new one. Such apparent changes in chela pattern require further investigation. Shells Inhabited The inhabited shells were characterised by the gastropod species, by shell size and by the intactness of the shell. It is relevant to ask: (a) whether there is any selection for preferred shell types by different species or sizes of hermit crabs; (b) whether shells are a limited resource, and (c) whether shell selection and availability are relevant to the competition and coexistence of species within the hermit crab guild. The following observations have some bearing on these questions: 1. A wide range of shell species is used, as shown by the genera recorded for Caleinus laevimanus and Clibanarius virescens in two m2 quadrats (Figure 4). 2. Very few empty shells were found in the areas inhabited by hermit crabs. The few empty shells found were usually either broken, blocked or belonging to the opisthobranch genus Bulla (the latter were sometimes occupied by crabs, but were probably not ideal because of the thinness of the shell and the large aperture). 3. The were some indications of species differences in shell selection. Although both Clibanarius virescens and Caleinus laevimanus were commonly found in Carithium spp. shells, Cypraea spp. and Pyrene spp. shells were almost always occupied bv Clibanarius, and Nerita spp. by Caleinus (Figure 4). Coenobita spp. also seemed to favour Nerita spp., whereas the very flattened Clibanarius eurystemus was typically found in Conus spp. shells which have a long narrow aperture. Page 6 No. 182 4. Thefe is a general relationship between the size of the crab and the size of the shell, particularly if the data are analysed species by species (Table I). Table 1 The relationship between shell length (SL) and crab (carapace) length (CL). Clibanarius virescens in Cerithium Regression Equation Correlation coefficient spp. Clibanarius virescens in Cypraea SL= 3.21 + 2.60 CL r= 0.92 spp. Calcinus laevimanus in Cerithium SL = 6.48 + 2.06 CL r= 0.87 spp. SL = 4.45 + 2.53 CL r= 0.67 Calcinus laevimanus in Nerita spp. SL= 3.11 + 1.97 CL r = 0.80 5. Hermit crabs kept in captivity showed little tendency to change their shells when presented with empty shells. One of the few to change was a Clibanarius virescens in a damaged Cerithius sp. shell, which it exchanged for the intact Nerita sp. shell presented to it in its individual container. Another experiment demonstrated that Calcinus laevimanus is capable of interspecific shell exchange by force. This involved removing two individuals (one larger and behaviourally more dominant than the other, with a correspondingly larger shell) from their shells by gentle heating. The smaller crab was then given the larger shell and the larger crab the smaller one in separate containers. Within two minutes of being placed together again, the larger crab had removed the smaller from its shell by a series of violent jerking movements, thus effecting a shell exchange which resulted in the two crabs occupying their original homes. 6. Fourteen empty Nerita sp. and Certhium sp. shells were marked with a spot of red paint on 13th March and added to one of the small pools in area B occupied by at least 24 Calcinus laevimanus and 13 Clibanarius virescens. Immediate interest was shown by the hermit crabs in these shells and, within five minutes, all the new shells were occupied, but only by Calcinus laevimanus. These diverse and very limited observations suggest some evidence for crab-specific shell preferen- ces, for a shortage of suitable shells in the natural environment, and for the ability of crabs to occupy empty shells and to rob shells from other individuals. Diurnal Behaviour Patterns A conspicuous feature of the distribution of species other than Calcinus latens and Dardanus lagapodes was their occurence on open rock and algal turf substrates during the intertidal period. Not only were the crabs often out of rock pools, but they were commonly found lying with the aperture of the shell uppermost and facing the sun. It appeared that at least some of the rock pools provided an unfavourable environment, and that exposure to the air best enabled the animals to survive the hazards of the intertidal period. Whether the stress factor was increasing temperature or decreasing oxygen content was not determined, but a single set of observations on a small pool (c. 0.3m in diameter) containing Clibanarius virescens showed a clear movement onto the rim of the pool as the temperature increased (Table I). In addition to moving out of pools, Clibanarius virescens also formed aggreg- Table 2. Movement of Clibanarius virescens onto the rim of a small pool in area C on 1 8th February, the tide left the pool at about 10. 15h and returned at 16.00h (low water at 13. 16h; air termperature 29.5°C). air temperature 29.5%C). Time Pool Temperature (°C) Number of C virescens on 10.00 30.0 0 11.15 31.5 7 12.15 36.0 32 13.15 37.5 68 15.15 37.0 82 16.15 31.0 2 No. 182 Page 7 Figure 3 Figure 4 Length-frequency histograms for Calcinus laevimanus and Clibanarius virescens (hatched area = ovigerous females). Samples I and 2 describe the hermit crabs obtained from the two m2 quadrats in area D on 8th and 1 1th February (see text). Sample 3 describes those collected from a 10 cm2 aggregation in area D on 18th February. Hermit crab carapace length (mm) The frequency of shell genus (or family) occupancy in hermit crabs from the two m2 quadrats in area D on 8th and 1 1th February (see text). Also shown below x-axis are the numbers of living gastropods in quadrat 1 (none were present in quadrat 2). KEY: I > 5 1 SHELL GENUS (OR EAMILY) Page 8 No. 182 tions several invididuals deep, as previously described. Some low-tide aggregations were also found under stones in area A, but throughout most of the study area, large loose stones were scarce. When the intertidal area was flooded by the incoming tide, underwater observation revealed very few exposed hermit crabs in their intertidal positions, but they were clearly present among the bases of the macroalgae and in crevices. Limited observations suggest that they begin to emerge again on the late ebb tide. DISCUSSION The species of intertidal hermit crabs identified in the present study were also recorded by Lewinsohn (1982), apart from Clibanarius merguiensis. In addition, Clibanarius eurystemus and Clibanarius laevimanus appeared to be less uncommon at Mombasa than Lewinsohn ( 1982) deduced from his Somalian material. However, he recorded 26 species in total, and the list included one species of Coenobita, six of Calcinus. two of Diogenes, five of Dardanus, one of Paguristes and three of Pagurus which were not recorded in the present study . The discrepancy between this list, and that of only ten species for Port Mombasa, can at least in part be explained by the wider latitudinal range sampled, and the closer attention paid to pools and areas of coral rubble in the Somalian study. It can, however, be deduced from Lewinsohn’s ( 1 982) work that, as on the Kenya coast, Calcinus laevimanus and Clibanarius virescens formed a dominant species pair over at least part of his study area. The species of intertidal hermit crabs identified from rocky substrates at Mombasa have been recorded from other areas in and beyond the western Indian Ocean. Thus Clibanarius virescens is present from Pakistan (Tirmizi and Siddiqui 1982), to Australia (Abrams 1982) and South Africa (Barnard 1950). Calcinus laevimanus is present in both Hawaii (Reese 1969) and South Africa (Bernard 1950) as well as Tanzania (Hartnoll 1976), Aldabra (Taylor 1971a) and Diego Garcia (Taylor 1971b). Calcinus latens occurs in the Red Sea (Hughes pers. comm.), Pakistan (Tirmizi and Siddiqui 1982) and South Africa (Barnard 1950). The three species mentioned so far have also been recorded from Madagascar and the Comores by Dechance ( 1964), as have Clibanarius eutvsternus, Clibanarius laevimanus, Clibanarius longiiarsus, Clibanarius merguiensis and several other species. Coenobita species are widely recorded from the Indo-Pacific region (eg. Taylor 1971b, Dechance 1964, Page and Willason 1982), both Coenobita rugosus, Coenobita perlatus and Coenobita cavipes could all be expected to occur on the Kenya coast. So far, the precise hermit crab guild, or community of species, found at Port Mombasa has not been recorded elsewhere, but this is most likely to be a simple reflection of the lack of comparable studies. The results have already been compared with Lewinsohn (1982), but Hartnoll’s(1976) study at Dar es Salaam was more comparable because of its geographical proximity and restricted study area. However, this author probably overlooked several species in recording only Calcinus laevimanus in his list of littoral fauna. From the discussion so far, it is perhaps sufficient to conclude that: (a) the genera identified from the present study are typical of intertidal rocky shore habitats in the Indian and Pacific Oceans, and (b) the dominance of the guild by a Calcinus / Clibanarius species-pair is not unexpected (Bertness 1982). The habitat and general behavior of the species at Mombasa is similar to that described for the same or related species elsewhere but, as Bertness (1982) has clearly shown in his comparison of Calcinus/ Clibanarius pairs from the Pacific and Caribbean coasts of Panama, abiotic (in his case, tidal amplitudej and biotic (type and intensity of predation) factors can act as important modifiers to the basic pattern. In the studies ot species pairs reported by Bertness (1981, 1982) and Reese (1969). it appears that Clibanarius spp. occur high on the shore, are behaviourally subordinate, and are most active as the tide comes in. Although there is a broad overlap in distribution, Calcinus spp. occur somewhat lower on the shore, are more dominant, and, at least where tidal amplitude is low, tend to follow the water’s edge. Overall, Clibanarius spp. spend more time emersed than Calcinus spp., and this may also be correlated with the aggregating tendencies which, as Snyder-Conn (1981) has suggested, appear to have some advantage in relation to the problems of desiccation and thermal stress. To a large extent, the present study supports the above conclusions but, Calcinus laevimanus appeared to occur more towards the top of the shore than Clibanarius virescens. The habit of lying emersed well above the water line of rock pools with the shell aperture uppermost, was commonly observed, as it iwas in Hawaii by Reese (1969). Clearly, however, more precise observations on No. 182 Page 9 distribution, separating the effects of tidal height and substrate characteristics, and including the remaining species, are required before definite conclusions can be drawn. Although Reese (1969) concluded that hermit crabs had relatively few predators other than a few specialised crabs (such as Calappa and Eriphia species) and octopus, it is clear that certain fish of the order Tetraodontiformes are capable of crushing even the hardest shells inhabited by hermit crabs, and that this may influence their distribution and behaviour underwater. Thus, Levings and Garrity(1983) estimated that 20% of nerites (Nerita species) in an open microhabitat in Panamanian waters would be eaten within 24 hours by fish such as Diodon sp. (porcupine fish). Similar conclusions were reached by Bertness (1982) and, in relation to hermit crabs, he found that on the Pacific coast of Panama, where fish predation is intense, Nerita sp. shells are not preferred by either Calcinus obscurus or Clibanarius albidigitus, in spite of their wide availability. This appears to be due to the fact that Nerita sp. is the preferred prey of Diodon and that these low-spire shells also offer hermit crabs little resistance Irom thermal stress. Nerita sp. shells do, however, provide good protection against stomatopods, the main predators in the Caribbean, and on this coast Nerita shells are preferred by Calcinus tibicen, but not by Clibanarius antillensis. The latter requires a tail-spire shell in its upper shore habitat, where thermal stress is the major factor. In the light of these findings, the apparent preference for Nerita sp. shells by Calcinus laevimanus at Mombasa could be associated either with the more limited tidal range compared with the Pacific coast of Panama, with other habitat differences, with a lower fish predation rate, or with real differences between the relevant hermit crab species. It would be rewarding to investigate this further. The maximum density figures recorded for both Calcinus laevimanus and Clibanarius virescens at Mombasa were of the same order as those for intertidal hermit crabs reported in the literature (Hazlett 1970, Hughes 1977, Abrams 1982). It is generally accepted that gastropod shells are a limiting resource for many hermit crab populations. The evidence and implications are discussed, for example, by Fotheringham ( 1976) and Scully ( 1983). Much of the evidence, as in the present study, comes from the scarcity of empty shells in areas inhabited by hermit crabs, although Spight ( 1977) has drawn attention to the fact that empty shells are rapidly removed by physical processes anyway, and soon become inaccessible to most hermit crabs. Vance ( 1972a), however, succeeded in increasing the density of hermit crabs by supplying empty shells to wild populations. The implications of shells being a scarce resource are that (a) there will be a discrepancy between actual shell size and preferred shell size, and (b) there may be interspecific competition affecting the structure and dynamics of the hermit crab guild. It is well established that hermit crabs in shells above or below the optimal size are disadvantaged with respect to predation (Vance 1972b), reproduction (Fotheringham 1976) and tolerance to desicca- tion (Taylor 1981), and of course the extreme effects will be either zero mobility or zero protection. It was not possible to establish preferred shell sizes for the hermit crabs found at Mombasa, but the rapid shell exchanges recorded in the natural environment suggests that at least some of these crabs were in less than preferred shells. Several studies have been carried out on habitat and niche separation in sympatric hermit crab species. Kellogg (1977), for example, working on sublittoral species off California, concluded that co-existence was made possible by a combination of habitat differences, shell size partitioning and shell species partitioning in descending order of importance. To some extent, this would also seem to apply to the Mombasa species. Bertness (1981) found that, in the Calcinus obscurus/ Clibanarius albidigi- tus/Pagurus sp. guild of the Pacific Panama coast, active competition for a limited supply of shells directly influenced shell quality and spatial distribution. Clibanarius inhabited less than preferred shells in sympatric populations of Clibanarius and Calcinus and showed distinct avoidance responses when encountering Calcinus obscurus. It also had a superior ability to locate and therefore exploit empty shells. Shell exchanges between individuals have been extensively documented (Scully 1983). The extent to which they occur is known to depend on the species (Hazlett 1970) and shell adequacy (Abrams 1982), but most exchanges appear to be intraspecific rather than interspecific. Certainly, no interspe- cific interaction was observed in mixed groups kept in captivity in the present study, whereas intraspecific interactions in both Clibanarius virescens and Calcinus laevimanus were common. In the case of the former species, the interactions were more aggressive, and casual observation supported the conclusion of Dunham ( 1 978) that the relative amount of white on the right chela was of considerable social significance. The observation that the patterning could change abruptly over one moult appears to be a new one, but further data are needed to confirm this and assess its significance. It is clear that the two dominant species at Mombasa differ not only in their habitat and shell Page 10 No. 182 preferences, but also in their activity patterns and social behaviour. Further work is likely to reveal other differences, and provide some indication of how these two species are able to live sympatrically at similar density levels. CONCLUSIONS This study provides only a preliminary account of hermit crab biology on the Kenya coast, and it should also be noted that only a limited range of habitats was examined. Thus, collections from soft substrates and sublittoral regions would undoubtedly reveal different species. The species found on intertidal rocky substrates were, however, similar to those reported in the literature from other tropical areas, and the limited amount of data collected indicates that the biology of these Kenyan hermit crabs is also similar to their counterparts in other study areas. Much work remains to be done, and fortunately hermit crabs provide a very convenient and rewarding subject for marine ecological and behavioural research. The most productive approach would be to carefully repeat recent studies such as those of Bertness (1981, 1982). This would be a valuable contribution both to rocky shore ecology in Kenya, and to hermit crab biology worldwide, since there appears to have been no substantial work carried out in the Indian Ocean area. The crabs studied are small and the shells they inhabit are not normally those utilised by shell collectors. Nevertheless, both they and their shell resources are intertidal, and thus are particularly vulnerable to human interference; this should be recognised in the context of their likely importance in rocky shore ecological processes. This importance, including a study of feeding biology, would also be a useful topic for research. ACKNOWLEDGEMENTS The author is grateful to Plymouth Polytechnic for the Fellowship which enabled him to visit Kenya, and to Mr S O Allela, the Director of KMFRI, for his invitation to work at the Institute and for providing facilities. He also thanks Mr. R.K. Ruwa of KMFRI for his stimulation and help in identifying the gastropod shells, and Professor L B Holthuis of the Rijksmuseum van Natuurlijke Histone, Leiden, for assistance with identifying the hermit crabs. REFERENCES Abrams, P. 1982. Intraspecific shell exchange in the hermit crab Clibanarius virescens (Krauss). J. Exp. Mar. Biol. Ecol. 59:89-101. Barnard, K. H. 1950. Descriptive catalogue of South African decapod Crustacea (crabs and shrimps). Ann. S. Afr. Mus. 38:1-864. Bertness, M. D. 1981. Competitive dynamics of a tropical hermit crab assemblage. Ecology 62:751- 761. - — — — 1982. Shell utilisation, predation pressure and thermal stress in Panamanian hermit crabs: an interoceanic comparison. J. Exp. Mar. Biol. Ecol. 64:159-188. Dechance M. 1964. Sur une collection de crustaces pagurides de Madagascar et des Comores. Extrait des Cashiers ORSTOM-Oceanographie 1 1 (2):27-45. Dunham, D.W. 1978. Effect of chela white on agonistic success in a diogenid hermit crab (Calcinus laevimanus). Mar. Behav. Physiol. 5:137-144. Fotheringham, N. 1970. Hermit crab shells as a limiting resource. (Decapoda, Paguridea). Crusta- ceana 31 : 193-199. Hartnoll, R.G. 1976. The ecology of some rocky shores in tropical East Africa. Estuarine Coastal Mar. Sci. 4:1-21. Hazlett, B.A. 1970. Interspecific shell fighting in three sympatric species of hermit crabs in Hawaii. Pacific Science 24:472-482. Hughes, R.N. 1977. The biota of reef-flats and limestone cliffs near Jeddah, Saudi Arabia J. Nat. Hist 11:77-96. Kellogg, C.W. 1977. Co-existence in a hermit crab species ensemble. Biol. Bull. 153:133-144. Levings, S.C., and S.D. Garrity, 1983. Diel and tidal movement of two co-occuring neritid snails: differences in grazing patterns on a tropical rocky shore. J. Exp. Mar Biol. Ecol. 67: 261-278. Lewinsohn, C. 1982. Researches on the coast of Somalia The shore and dune system of Sar Uanle. No. 182 Page 11 33. Diogenidae, Paguridae and Coenobitidae (Crustacea, Decapoda, Paguridea). Monitore Zool. Ital. NS Suppl. 16( 2): 3 5-68. McLean, R. 1983. Gastropod shells: a dynamic resource that helps shape benthic community structure. J. Exp. Mar. Biol. Ecol. 69:151-174. Page, H.M.and H.M. Willason, 1982. Distribution patterns of terrestrial hermit crabs at Enewatok Atoll, Marshall Islands. Pacific Science. 36.107-118. Reese, E.S. 1969. Behavioural adaptations of intertidal hermit crabs. Amer. Zool. 9:343-355. Scully, E.P. 1983. The behavioural ecology of competition and resource utilisation among hermit crabs, pp. 23-56. In Rebach, S. and Dunham, D.W. (Editors) Studies in Adaptation; the Behaviour of Higher Crustacea. Wiley. Snyder-Conn, E. 1980. Tidal clustering and dispersal of the hermit crab Clibanarius digueii. Mar. Behav. Physiol. 7:135-1 54. Spight, T.M. 1977. Availability and use of shells by intertidal hermit crabs. Biol. Bull. 152:120-133. Taylor, J.D. 1971a. Intertidal zonation at Aldabra Atoll. Phil. Trans. Roy. Soc. Lond. B, 260:173- 213. — — 1971b. Crustacea: Brachyura and Anomura from Diego Garcia. Atoll. Res. Bull. 149:93-101. Taylor, P R. 1981. Hermit crab fitness: the effect of shell condition and behavioural adaptations on environmental resistance. J. Exp. Mar. Biol. Ecol. 52:205-218. Tirmizi, N.M., and S.A. Siddiqui. 1982. Marine Fauna of Pakistan: I. Hermit Crabs. Mar. Invert. Ref. Coll. Centre, University of Karachi, 103p. Saad Publications, Karachi. Vance, R.R. 1972a. Competition and mechanism of co-existence in three sympatric spp. of interti- dal hermit crabs. Ecology 53:1062-1074. - — - — - 1972b. The role of shell adequacy in behavioural interactions involving hermit crabs. Ecology 53:1075-1083. Vannini, M. 1976. Researches on the coast of Somalia. The shore and the dune of Sar Vanle. 7. Field observatons on the periodical transdunal migration of the hermit crab Coenobita rugosus Milne Edwards. Monitore Zool. Ital. NS Suppl. 7(3): 1 45- 1 85. M. Vannini and G. Chelazzi, 1981. Orientation of Coenobita rugosus (Crustacea: Anomura): a field study on Aldabra. Mar. Biol. 64:135-140. Received 7 November 1983 Editors: J.J. Hebrard, H. Beentje Page 12 No. 182 APPENDIX 1 A KEY TO THE IDENTIFICATION OF HERMIT CRABS FOUND AT MOMBASA 1 — Antennular peduncle longer than carapace ; antennular flagellum blunt at tip; terrestrial and semi-terrestrial hermit crabs (Family: Coenobitidae; Coenobita) 2 — Anlennular penduncle shorter than carapace; antennular flagellum ending in a filament; littoral and sublittoral hermit crabs (Family: Diogenidae) 3 2 — Carapace and pereiopods red or with red markings. No stridulating mechanism (row of teeth) on outer surface of manus of left cheliped Coenobita cavipes — Carapace and pereiopods pale to dark grey, without red markings, stridulating mechanism on outer surface of manus of left cheliped Coenobita rugosus Coenobita rugosus 3 — Tips of chelipeds pointed and calcareous (white); left cheliped markedly larger than right Calcinus 4 — Tips of chelipeds spoon-shaped and corneous (black); both chelipeds usually of ths same size 5 4 — Left cheliped with sharply-contrasting black and white areas; right cheliped without strongly- serrated dorsal margin Calcinus laevimanus — Left cheliped pale distally, gradually darkening towards base of the propodus: right cheliped with strongly-serrated dorsal margin Calcinus latens 5 — Fingers of chelipeds opening and closing obliquely or nearly vertically; left cheliped larger than right; very bristly Dardanus lagopodes — Fingers of chelipeds opening and closing horizontally; chelipeds equal or sub-equal Clibanarius 6 6 Pattern on walking legs includes longitudinal stripes 7 — Pattern on walking legs does not include longtiludinal stripes . 9 — Basic colour pale with brown or reddish-brown markings: dactylus of walking legs equal to, or shorter than the propodus 8 — Basic colour dark blue / green (but may be pale pink in juveniles); dactylus of walking legs longer than the propodus Clibanarius longitarsus 8 — Thorax markedly flattened; longitudinal pattern on walking legs of irregular shapes; joints on walking legs not conspicuously pale Clibanarius eur vsternus — Thorax not markedly flattened; longitudinal pattern on walking legs of regular stripes; joints on walking legs conspicuosly pale Clibanarius laevimanus 9 — Walking legs with white (or yellow) and black (or brown) bands Clibanarius virescens — Walking legs with white, orange and blue colouration (this may only be a juvenile pattern, Clibanarius merguiensis NB 1. Walking legs 1 and 2 = Pereiopods 2 and 3. 2. The Diogenid genera, Paguristes and Diogenes, may also be encountered in the area. In Paguristes the fourth pereiopods (first pair of reduced thoracic appendages) are simple, not chelate. In Diogenes, which shares pointed, calcareous-tipped chelae with Calcinus. the rostrum is replaced by a moveable scale or spine, and the antennal flagellum is setose rather than bare. 3. As a result of the work of Lewinsohn (1982) it is clear that up to six additional species of Calcinus and five of Dardanus could occur on the Kenya coast. Published by The East Africa Natural History Society, Box 44486, Nairobi, Kenya. Phototypeset- ting by Kul Graphics Ltd., P.O. Box 18095 and printed by Manhattan Displays, P.O. Box 30434, Nairobi. JOURNAL OF THE EAST AFRICA NATURAL HISTORY SOCIETY AND NATIONAL MUSEUM member 1985 Volume 75, No. 183 Cyclic sequences of vegetation in the plant communities of the Aberdares Mountains, Kenya ADQ Agnew University College of Wales, Aberystwyth. ABSTRACT Cycles of vegetation change are described for two vegetation types on the Aberdares Mountains, Kenya. The shrubby Alchemilla argyrophylla undergoes a cycle of growth and degeneration rather similar to Calluna vulgaris in Western Europe; that is the phases are not synchronous unless forced to be so by fire or disturbance. Bamboo (Arundinaria alpina) however is monocarpic, patches of stems 0.5 to 5 hectares in extent flower synchronously and thus initiate a series of vegetation changes characterised by Sambucus africana which eventually lead back to Bamboo forest. These cycles are of importance in management of montane game reserves in East Africa. The nature of cyclic successions is discussed. INTRODUCTION There is no doubt that all vegetation is unstable in the short term. Individual plants die and are replaced; herbivores remove parts of selected species and these species must maintain their numbers by different strategies compared with unpalatable ones; geomorphological and geochemical processes impose slow cycles and trends on the plant cover of the earth. Grubb (1977) draws attention to the overall importance of cyclic change-over of species composition in many plant communities as an explanation of the diversities of species populations. Some plant communities predictably fluctuate around the life-history of the dominant whose death leads to a cycle of events which finally repeats itself. It is possible to refer to these as cyclic sequences and their appreciation gives insight into the dynamics of such communities as well as help in management problems. In East Africa, plant communities are seldom dominated by one species unless disturbed by man’s activities. In the montane zone, however, dominance becomes more and more pronounced and Hedberg (1964) delimits zones of characteristic species up to the alpine zone above the forest level. In general the progression of physiographic vegetation types can be said to pass from montane mixed forest to gymnosperm forest (Juniperus procera or Podocarpus latifolius*) thence into bamboo (Arundinaria alpina). Above the tree line there is tussock grassland (of Fesluca pilgeri, Andropogon amethystinus and Carex monostachya) which alternates with what is clearly heath vegetation of various types. It is the purpose of this paper to describe cyclic changes in the bamboo, and the heath vegetation dominated by the Rosaceous Alchemilla argyrophylla. The following account is based on floristic analyses made in 1976, while investigating the boundary zones between plant communities. Two line transects of 30 contiguous 50 cm x 50 cm quadrats were taken at right angles to the boundary between easily discernible plant communities. Analysis of these data by reciprocal averaging allowed identification of the quadrats from the homogeneous ends of the transects These were then taken as representative of the communities, and mean cover could be assigned to each species, as well as to the pH of surface soil from each quadrat. Clearly this resulting figure is akin to a phytosociological releve, and since it is based on non-random sampling the citation of confidence bands would be meaningless. In the case of the bamboo analyses, 10 1 m x 1 m quadrats were roughly randomised to give a basis for the stem variance figures. Nomenclature of flowering plants follows Agnew (1975). Formerly P. milanjianus. Page 2 No. 183 THE ALCHEMILLA ARGYROPHYLLA CYCLE Description This shrubby silvery-leaved Alchemilla is one of the characteristic plants of the heath vegetation of the moorlands about 3000 m on Mount Kenya and the Aberdares (replaced by Alchemilla elgonensis on Mount Elgon). It occurs in burnt bushland edges as well as heathland but in the latter it frequently dominates. The shrub grows from 20 to 70 cm tall in a complex community in which a mosaic of mounds and hollows may be associated with small open patches of grassland and herbs, and in which mole-rats (Tachyorycles spp.J disturb large areas. Alchemilla heath can burn, but these heathlands of the central Aberdares plateau do not seem to have been burnt for some time, and it is possible to find stands of apparently differing ages. I have made the general observation that on the extensively burnt shrubland of Mt. Kenya (Sirimon track head, 3700 m) the regeneration in 1976, two years after burning, included a dense population of Alchemilla argyrophylla seedlings. There was no extensive stand of Alchemilla heath at that altitude before the fire and it is possible that fire stimulates widespread germination. Figure 1 is a diagram of the supposed course of the Alchemilla cycle and incorporates part of a measured transect through a bush, and Table 1 gives the floristic composition of sample areas of the proposed sequence. Cross-sections of representative stems from three stages in the cycle are shown in Fig. 2. The growth zones are rather tantalizing, for young stems appear to be ageable but the oldest stem seems to have no clear growth zones towards the older wood. The rainfall pattern on the Aberdares is of two principal rainy periods (as in the rest of Kenya to the East of the Rift Valley) but they are hardly separated at higher altitudes. Indeed there were heavy showers during the period of my observations in July 1976 which in the lowlands is invariably dry. The main dry period is January and February and under this regime recognisable annual growth zones could be expected, but note how the new “spring” wide-vessel growth seems to have recently started on these sections. Obviously the phenology of growth on these equatorial mountains would repay investigation. In the following account of cyclic events the recognisable growth zones are used as indices of annual growth to estimate the duration of various stages of the cycle, but future work should check this against long-term observations. Pioneer (3-4 years). The establishment phase takes place either as colonisation after disturbances by fire, mole-rats, human interference or as part of the cycle within a mosaic of phases of Alchemilla heathlands. Growth rates are unknown but measurements gave a mean of 10 cm of new shoot, leafy to ground level. The flora of this phase depends on its origin. If the result of disturbance, grasses and ruderals are abundant. Creeping plants such as Carex conferta, Oxalis corniculata and Uebelinia crassifolia are common when the full cycle of growth phases is present. Building (4-7 years). Here there are erect shoots with naked woody bases, showing 30cm of new growth of the current year, with abundant flowering. The flora consists typically of abundant bryophytes, and a very diverse assemblage of flowering plants which are frequently creeping or rhizomatous. Mature (4-8 years). The older shoots lean and finally fall, and the basal bark peels in scales, while new growth averages 25 cm with abundant flowring. Maximum floristic diversity is present but bryophytes are reduced and grasses may be abundant between the spreading bushes of Alchemilla. Degenerate (possibly lOyears). Fallen shoots surround the base of the old plant, while the 10-20cmofnew growth attest the decreased performance here. The ground may be covered with bark scales from the stem bases and the flora is greatly reduced, particularly in bryophytes and grasses. In general there seems to be no difference between the soils of the study areas exemplified in Table 1 in so far as they were investigated. Discussion of the heathland cycle Cyclic sequence in vegetation are characteristic of heaths, where a dominant woody plant has a limited life span; the best known example is that of Calluna in Britain (Gimingham, 1972), the study of which was part of research into ecological management for grouse stocking. No. 183 Page 3 The principal heath species of the Aberdares range in Kenya are Erica arborea and Philippia excelsa which form woodland, and the low shrubby Alchemilla argyrophylla. This woodland is dominated by true members of the heath family, and although its soils are sufficiently acid and peaty for it to be regarded as a heathland community its long term dynamics are still unknown. Possibly there is a cycle involved but if so it must be so interrupted by fire, and by invasion by pyrophilic plants such as Stoebe kilimandscharica that there can be little evidence remaining as to its characteristics today. Possibly intensive investigation of such sites of pure Erica arborea as remain will reveal a cycle. However, the heath shrub Alchemilla argyrophylla does give evidence of a cyclic behaviour very analagous to that described for Calluna vulgaris in Scotland. It is difficult to ascribe any evolutionary advantage to the shrub consequent on the development of the cyclic vegetation pattern described. It is possible that this is another feature of the effect of the important mole-rat populations in alpine Africa, as described by Hedberg (1975). There are numerous afro-alpine plant species capable of rapidly colonising bare ground and these may have evolved as much in response to selection by solifluction as by mole-rat disturbance. THE BAMBOO CYCLE Description Bamboo occurs from 2750-3500 m over extensive areas of all East African mountains (Hedberg 1964). Wimbush (1947) has described the bamboo cycle in general terms. It is a dominant, monocarpic plant which flowers and dies back in patches throughout the forest. Flowering tends to be almost synchronised over large areas. For instance it was easy to find flowering specimens in 1 966 over wide areas of the Aberdares bamboo forests but flowering had apparently taken place in the previous decade on Mt. Kenya and flowering specimens were not found there although flowered patches with dead bamboo were common. By 1969 it was difficult to collect herbarium specimens on the Aberdares but a flight over the Mau forest revealed extensive areas of current flowering. Observations in 1968 and 1976 allow the following account to be prepared. Floristic stages are easy to observe and are described below; they are summarised in Tables 2 and 3 and Fig. 3. Pioneer Regeneration of bamboo appears to take place from the rare revitalisation of a section of fallen culm or even part of the rhizome system. Very occasional small living shoots can be found even within the dead tangle of stems 2-4 years after flowering but take a long time to become conspicuous clumps. I have not been able to find seedlings in the field, and on the Aberdares at least, I have not found viable seeds, although Dr. P. J. Greenway (pers. comm.) has assured me that they do exist. The tussocks of bushy growth which result from this revitalisation grow slowly for 3-5 years forming thickets of many-stemmed plants 1-3 m tall amongst the Sambucus of the previous phase. pH is at its highest and litter cover of the soil is low. The bushes of bamboo enlarge until they join up when the Building Phase begins. Building I define this as the stage when the height of the growing stems exceeds the general height of the stand. In fact a rapid increase in general height is obtained and this phase appears to last but a short time, from 5-7 years, and the stems retain the clumped pattern imposed during the pioneer stage. Floristically this period is the poorest, for the light-demanding species of the pioneer stage have been eliminated and the shade-loving species of the mature phase apparently migrate and invade only slowly. 1 suspect that most of the shade-loving species that are found are survivors from the previous cycle and examination of Table 2 shows this to be true of Pieris catopiera and Cyperus dereilema, for instance. Even Sanicula and Selaginella kraussiana can survive in odd pockets in the Sambucus stage. Floristic changes are accelerated by a change in pH of the soil during this period. Since the major event is a fast increasing bamboo biomass, more and more plant nutrients become locked into it and are not available for circulation. We also have to consider the effects of a change in the litter type from predominantly broad-leaved to the strongly sclerenchymatous leaves of the bamboo. Both these events are reflected in a lowering of pH during this period. Table 2 gives an analysis of sampled stem densities during this and subsequent stages. The dense clumping of this building phase can be seen and needs no further comment. Page 4 No. 183 Mature As soon as the annual new growth of stems equals the height of the stand (no further overall height increase taking place), the stands can be said to be mature and only slow changes take place for the next 10-15 years. These changes are: a) the separation of bamboo culms on the forest floor until a more regular, less clumped, pattern is produced; b) the final change of the forest floor species to those of deep shade, and the elimination of many lianas which were established during the Sambucus and Pioneer stages; c) soil pH continues to decline but not to very low levels. There is often heavy grazing of young bamboo shoots by rhinoceros and elephants, which could be important in hastening or delaying the onset of flowering. Flowering and Sambucus stage: Flowering takes place during one season only but 2-3 years pass before the culms fall, and this is most important for the whole system. The culms fall in a haphazard way, making flowered areas extremely difficult to enter. There is a sudden increase in light-demanding broad-leaved species and climbers, dominated by Sambucus africana, so that from this to the thicket pioneer stage can be called the Sambucus stage. Herb diversity increases slowly, the initial invasion leading to rampant single-species patches of vegetation, particularly climbers, but this is soon broken up by the entry of large mammals. They break the vegetation down into small patches between which trails meander allowing entry of ruderal plants. Soil pH increases but not immediately because the bamboo culms may take a number of years to rot away. However, as the process continues, more and more of the total nutrient pool is returned to the soil and thence to a much more rapid cycle within the soft-leaved dicotyledonous plants of the Sambucus stage, so that by the end of this period the maximum pH is reached. Ultimately, just before the thicket (pioneer) bamboo stage with which this account began, there is maximum heterogeneity in the vegetation because of the network of large trails which breaks it up, and the persistence of large bamboo culms, some of which may even be upright. The diversity is further increased by uneven invasion by opportunistic ruderal species, and uneven elimination of the old forest floor flora. Tree species such as Nuxia, Podocarpus and Dombeya goetzenii have their greatest opportunity for establishment and growth during the Sambucus phase, for their saplings are suppressed under the high bamboo canopy but are able to grow more quickly after this has flowered. 1 have not been able to show that the present distribution of these trees is associated with past flowering sites but it is clearly a possibility which could repay investigation. The possibility of internally operated cycles must be borne in mind when discussing vegetation changes through any cause, and in the National Parks of Kenya today a principal preoccupation is with environmental damage by elephant. This is well documented in Tsavo (Glover 1963), but not so well in the Aberdares where there is considerable evidence of recent elephant damage, which was not there in 1 969. It is increasingly difficult to find undisturbed stands of bamboo at low or high altitudes, and some Cliffortia, or even Hagenia trees on the moorland woodlands are beginning to suffer as well. This is despite the very great area of bamboo/ Sambucus edge which now exists and which was probably not available before 1965. The Sambucus areas of the bamboo cycle hold high potential for grazing and browsing animals and are heavily used so that feeding trails of rhino and elephant form a network within them, and there is nearly always a distinct marginal trail either for passage or for feeding on the marginal bamboo fronds. The stands of bamboo which I visited in July 1976 on the West of the escarpment (Kinangop side) were under great pressure from elephant. Large culms had been pulled down and trampled and I could find very few undamaged emerging bamboo shoots. Conditions on the East (Mweiga and Nyeri) side were much better. There was less damage to mature bamboo and new shoots were plentiful. In the relatively non-seasonal climate of the tropics, biological events are based on obscure triggers. Tweedie (1965) has documented periodic flowering in upland Kenya Acanthaceae, but we are a long way from an understanding of the phenomenon. The bamboo cycle creates a very high heterogeneity in the area as a whole and is of central importance to its carrying capacity, and so reserve management would gain from knowledge of the physiology and ecology of flowering. Therefore the bamboo cycle itself is a worthwhile subject for conservation, so that if the main flowering area of the bamboo became totally disturbed by big No. 183 Page 5 game pressure or by exploitation through agriculture or forestry, a sufficient area for the full development of the natural cycle should be set aside lest an outstanding example of a natural biological process be lost. Janzen (1976) discussed periodic bamboo flowering and its evolutionary significance which he suggests could be due to a need for overcoming seed predation by spasmodic over-production. In my observation of bamboo regeneration, 1 have never seen a seedling. All regenerating plants have been traceable to fallen culms, and Wimbush (1947) agrees with this observation. This of course does not argue against Janzen’s thesis, but may simply be a feature of a species marginal to the central range (East Asia) of that genus. In any case my observations suggest that the young shoots are the most grazed part of the life cycle. Excessive predation of these could possibly eliminate the bamboo. To follow Janzen and suggest an alternative evolutionary potential for the cycle of bamboo regeneration is not difficult. For surely the diversification of the habitat into mosaics, where the Sambucus patches hold prime browse material, would reduce pressure on those very vulnerable 60 cm spikes of high nutrient content on the bamboo woodland floor. The bamboo is part of a more diverse woody vegetation at the lower part of its range. Groves of Podocarpus latifolius, Dombeya goetzenii and Nuxia congesta grow amongst the bamboos in favoured sites. Based on my general observation I tend to regard those favoured sites as also the sites of most frequent flowering of bamboo, although I have no quantitative evidence for this. I have also observed seedlings of these tree species most frequently at the edges of bamboo flowered areas, and it it not unlikely that tree species distribution is indeed associated with the flowering pattern of the bamboo. GENERAL DISCUSSION Cyclic successions have been recognised for some time. Watt (1947) drew attention to their presence in certain British vegetation, while Kershaw (1973) dealt with them at length and extended the ideas to many situations where a plant species shows a cycle of performance at any one spot without affecting the vegetation as a whole. More recently Grubb (1977) has suggested that this process is responsible for a great deal of the diversity in plant communities in that the “regeneration niche”, which is made available on the decline of an individual after its cycle of growth and maturity, is the major potential site for invasion of a closed community. It is clear that communities which are characterised by cyclical successions are simply those where dominance is such that the major plant species eventually replaces itself but during this process the entry of many different species constitutes an alternating community structure. In the examples presented in this paper the bamboo cycle is more impressive in this respect than the Alchemilla because there is a far greater disparity between environments in that cycle than in the Alchemilla. But it may be that there is a general principle involved here because long lived dominants must show a certain synchrony in their phases of growth for the cycle to be recognized. If the bamboo culms on the Aberdares flowered and died one or two at a time, replacement by Sambucus for a few years would be unremarkable and the whole community would present the aspect of mixed Arundinaria alpina/ Sambucus africana with abundant lianas. It is the synchrony which gives the major effect and this is lacking in the Alchemilla, where the recognition of the various phases in the stands reported on here is probably due to differing ages since their last firing or burning; old stands show a mixture of recognisable phases. This is very similar to the situation in Calluna vulgaris in Scotland (Gimingham 1972) where burning is a recognized management tool and even-aged stands are therefore produced which can be said to be predominantly pioneer, building or mature. It is clear therefore that longevity of the dominant and synchrony of life cycle behaviour are both responsible for elevating the diffuse “regeneration niche” into a definite cycle of community events. The ultimate may be the type of cycle imposed on soil nutrient conditions within the ecosystem as suggested by Florence (1965) for Douglas Fir and oak woodland with a time scale of c. 2000 years. Obviously such cycles must be increasingly hard to recognise as more natural vegetation becomes disturbed by man’s influence. Page 6 No. 183 REFERENCES AGNEW, A.D.Q. 1974. Upland Kenya Wild Flowers: a flora of the ferns and herbaceous flowering plants of Upland Kenya. Oxford. FLORENCE, R.G. 1965. Decline of old growth Redwood forests in relation to some soil microbiological processes. Ecology 46: 52-64. GIMINGHAM, C.H. 1972. Ecology of heathlands. London. GLOVER, J. 1963. The elephant problem at Tsavo. East African Wildlife Journal I: 30-39. GRUBB, P.J. 1977. The maintenance of species richness in plant communities: the importance of the regeneration niche. Biological Reviews 52: 107-145. HEDBERG, O. 1964. Features of afroalpine plant ecology. Acta Phytogeographica Suecica 49: 1-144. — — — 1975. Studies of adaptation and speciation in the afro-alpine flora of Ethiopia. Boissiera 24: 71-74. JANZEN, D.H. 1976. Why bamboos wait so long to flower. Annual Review of Ecology and Systematics 7: 347-392. KERSHAW, K.A. 1973. Quantitative and dynamic plant ecology. London. TWEED1E, M E. 1965. Periodic flowering of some Acanthaceae on Mt. Elgon. Journal of the East African Natural History Society and National Museum 25: 92-94. WATT, A.S. 1947. Pattern and process in the plant community. Journal of Ecology 35: 1-22. WIMBUSH, S.H. 1945. The African alpine bamboo. Empire Forestry Journal 24: 33-39. — — — 1947. The African alpine bamboo. East African Agriculture and Forestry Journal 13: 56-60. Received: January 1984 Editors: HJ. Beentje & J. Hebrard No. 183 Page 7 TABLE I Mean cover percentage of species in three stands of dominant Alchemilla argyrophylla in the Aber dares Moorland. Only species with 4 or more records are cited. Others (with stand number in parentheses) are: Anagallis serpens (2), Ardisiandra wettsteinii (I), Carduus chamaecephalus (1), Carex greenwayii (3), Haplosciadium abyssinicum (2), Helichrysum ellipticifolium ( 1 ), Helictotrichon milanjianum (3), Heracleum abyssinicum (2), Hypericum lanceolatum (3), Lobelia duriprati (I). Lysimachia ruhmeriana (3), Mariscus kerstenii (2), Ranunculus oreophytus (2), Sambucus africana (3). Senecio schweinfurthii (2). Stages of Cycle Present in Stand 1 2 3 Pioneer Building Mature and and and Building Mature Degenerate No. of Quadrats 23 20 16 Forbs: Alchemilla argyrophylla 28.73 35.0 56.24 A. johnstonii 10.22 12.75 A. rothii 4.13 Ajuga remota 0.25 0.94 Carduus keniensis 2.83 0.75 Cerastium afromontanum 1.50 Conyza subscaposa 0.87 Dichondra repens 3.70 3.25 Euphorbia schimperiana 5.22 3.75 Galium kenvanum 0.43 0.25 Geranium arabicum 0.50 1.87 Helichrysum cymosum 0.65 2.00 Hypericum peptidifolium 2.83 1. 00 0.62 Luiula johnstonii 4.78 1.00 0.94 Oxalis corniculala 4.78 4.25 7.50 Pimpinella keniensis 5.00 3.75 Polygonum afromontanum 0.25 0.93 Satureia kilimandschari 0.86 1.00 Siblhorpia europaea 0.60 0.25 4.07 Sonchus afromomanus 0.22 0.25 Slachys alpigena 5.25 3.75 Swertia crassiuscula 0.80 1.00 Tolpis capensis 1.09 0.75 Trifolium burchellianum 3.69 4.00 T. cryptopodium 0.43 Uebelinia crassifolia 1.08 0.75 Veronica glandulosa 1.74 1.75 Viola eminii 3.13 V. nannae 5.00 2.00 Grasses and Sedges: Agroslis kilimandscharica 5.87 1.75 1.24 Anthoxanthum nivale 3.12 Carex conferia 5.00 3.00 4.69 Festuca abyssinica 8.25 F. pilgeri 4.00 Koeleria capensis 13.91 1.24 Mariscus sp. 2.75 Pentaschislis borussica 5.21 15.0 0.31 Poa leptoclada 3.05 3.0 2.18 Other species (Number) 2.39(4) 3.00(7) 2.87(5) Bryophytes: Acrocarpous mosses 8.70 2.0 Pleurocarpous mosses 2.17 1.75 1.25 Leafy liverworts 3.04 1.75 2.19 Mean pH (aq.) 4.62 4.67 4.65 Diversity (H) 3.039 2.732 1.746 Cover Total 138% 137% 99% Page 8 No. 183 TABLE 2 Features of successional stages in the Aberdares Bamboo (Arundinaria alpina) Pioneer bamboo 1 Arundinaria alpina (culms, m ) 3.7 Arundinaria culm variance 4.14 pH mean 5.66 Acalypha cf. volkensii S 0.20 Acritochaete volkensii w 2.20 Australina acuminata w Cardamine africana w Cynogtossum lancifotium s 0.40 Cyperus dereilema w 1.55 Cyphostemma kilimandscharicum c 0.40 Dichrocephala inlegrifolia s Droguelia iners w 0.20 Galium chloroionanthum c Geranium arabicum s Girardinia bullosa w 0.40 Hydrocolyle mannii w Mikaniopsis clematoides c 0.60 Oxalis corniculata s Pilea spp. w 0.60 Poa schimperiana s Polystichum fuscopaleaceum 1 0.60 Pseudocarum eminii c 0.60 Pteris catoptera I 0.20 Sambucus africana s 1. 00 Sanicula elata w 0.20 Selaginella kraussiana w Senecio moorei s Senecio syringifolius c 1.80 Solanum terminate 1 Veronica abyssinica s Viola abyssinica I 4.0 No. of quadrats 20 Additional species (number) 3 Additional species (Cover) 1.0 Total Ground Flora Cover 11.95 Assemblage diversity (H) 2.581 Building Mature Flowered Mature Old bamboo bamboo bamboo- Sambucus Sambucus young Sambucus 3.4 1.1 trace 1.76 1.20 5.24 5.13 5.07 5.06 6.38 0.73 0.36 8.71 11.80 14.73 9.00 1.71 0.67 1.14 2.80 0.36 0.67 0.40 2.45 4.50 5.00 10.40 4.73 1.45 3.83 3.45 0.67 0.40 1.14 5.64 2.0 0.67 0.36 1.14 3.27 0.67 2.91 4.67 2.29 0.36 0.72 0.67 0.67 0.57 8.10 9.36 3.27 14.50 5.57 1.50 0.73 3.18 2.67 1.71 5.27 20.82 42.50 2.86 13.00 0.67 6.71 0.67 0.36 0.36 2.50 1.14 2.40 12.55 4.18 2.00 1.71 0.80 2.18 2.18 0.73 3.83 0.80 1.1 3.83 7 10 II 1 1 6 2 2 1 7 9 4.57 1.60 0.36 8.89 26.00 49.97 53.6 63.46 62.58 118.86 2.581 1.923 2.201 2.385 2.465 Mean cover is given for species occurring in more than one stand; data are given for six characteristic stands. The major habit of each species is given as: C, climber; 1, indeterminate; S, ruderals of disturbed ground, W, woodland floor herbs. These categories have been decided from general observation and herbarium specimens. No. 183 Page 9 TABLE 3. Cover contributions (as percent of total herbaceous cover) oj 4 habitat classes from Table 2. Successional stage Pioneer bamboo Woodland Flora plants (W) 43.1 Plants of disturbed ground (a) including Sambucus 13.4 (b) excluding Sambucus 11.7 Climbing plants (C) 28.4 Azonal species (1) 6.7 Building Mature Flowered Mature Old bamboo bamboo Young Sambucus bamboo Sambucus Sambucus 59.1 71.6 31.2 15.5 2.8 3.4 0 30.8 53.7 67.2 0 0 22.5 20.6 31.5 5.7 19.6 34.5 19.5 17.7 22.68 5.8 8.6 2.8 6.0 Page 10 No. 183 Oxalis corniculata Sibthorpia europaea Two diagrammatic representations of transects through changing phases of the Alchemilla argyrophyUa cycle. In A, a mature bush is suppressing Agrostis and Oxalis in favour of the shade tolerant Sibthorpia. In B, a degenerating bush is allowing entry of Oxalis and Agrostis both in the centre, where perennial herbs (Viola, Stachys) are establishing, and at the margin where Sibthorpia is being replaced. big. i No. 183 Page 1 1 Sections through stems of Alchemilla argyrophylla taken at ground level from (A) pioneer and building, (B) mature and (C) degenerate phases of the successional cycle. Fig. 2 Page 12 No. 183 Sambucus africana % dlfc Cyperus dereilama 0.05) (Table 1). No. 184 Page 3 Table 1. Seasonality of ecotzoochory by hares in Queen Elizabeth National Park, Uganda, October 1981 — September 1982. Each season had three sampling periods. Total Number of Hares Examined Number of Hares Carrying Disseminules Total Number of Disseminules Mean Disseminule Per Hare 'S.E. Total Plant Spiecies Represented Seasons 1. Short 28 10 70 2.6±0.0 15 Wet 2. Long 21 19 181 9.913.2 12 dry 3. Long 22 9 39 2.010.7 7 wet 4 Short 25 24 144 5.710.8 12 dry TOTALS 96 62 434 4.511.5 23 The incidence of disseminules on the hares was 93% in the dry seasons and 38% in the wet seasons (Table 1). There was a nonsignificant relationship between the number of hares examined and those that carried disseminules per month (r = 0.33, P>0.05). This meant that the dispersal of disseminules by hares was density independent. There was also nonsignificant relationship between mean number of disseminules per hare per month and number of species represented per month (r = 0.55, P> 0.05). This showed that some plant species are better adapted for hare dispersal or that some species produced more disseminules than others (Table 3). Table 2. Numbers and species of plant disseminules by seasons, on female and male hares in Queen Elizabeth National Park, Uganda, October 1981-September 1982. MALES FEMALES Number of Hares Carrying Disseminules Total Number of Disseminules Total Disseminule Species Number of Hares Carrying Disseminules Total Number of Disseminules Total Disseminule Species 1. Short 5 * 31 6 5 39 9 2. wet Long 5 17 2 14 163 10 3. dry Long 6 29 3 3 10 4 wet 111 4. Short 10 33 5 14 / dry TOTALS 26 110 8 36 323 14 Table 3. Plant species, number and type of disseminules involved in ectozoochory with their frequency of occurrence based on 96 hares shot in Queen Elizabeth National Park, Uganda, October 1981 - September 1982. The nomenclature of shrubs and herbs follows Lind and Tallantire (1975) and for grasses, Clayton (1974). Plant Species Frequency of Number Type % of Total Occurrence of of Disseminules Number of Species Disseminules (Months) A. GRASSES 1. Hyparrhenia filipendula (Horst) Stapf 5 65 seeds 15.0 Page 4 No. 184 2. Tragus berteronianus Schult. 6 54 seeds 12.4 3. Cenchrus ciliaris L. 7 44 seeds 10.1 4. Themeda triandra Forssk. 5 29 seeds 6.7 5. Heteropogon comorius (L.) Roem. & Schult. 5 28 seeds 6.5 6. Eragroslis cilianensis Lutati 3 26 seeds 6.0 7. Chloris gayana Kunth 4 19 seeds 4.4 8. Brachiaria platynota (K. Schum.) Robysn 4 17 seeds 3.9 9. Aristida adoensis A. Rich. 1 1 1 fruits 2.5 10. Eragroslis tenuifolia (A. Rich) Steud. 3 8 inflorescence 1.8 11. Harpachne schimperi Hochst. ex. A. Rich. 1 7 spikelet + seeds 1.6 12. Eragroslis exasperaia L. 2 4 seeds 0.9 13. Panaicum brevifolium L. 1 3 inflorescence 0.7 14. Microchloa runthaii Dev. 4 4 inflorescence 0.2 B. HERBS/SHRUBS 15. Tribulus terrestris L. 3 43 fruits 9.9 16. Oxygonum sinuatum (Meisn.) Dammer 4 39 fruits 9.0 17 . Alysicarpus rugosus (Willd.) DC. 2 14 seeds 3.2 18. Achyranthes aspera L. 2 7 seeds 1.6 19. Clematis hirsuta Guill. & Perr. 3 7 seeds 1.6 20. Urena lobata L. 1 4 fruits 0.9 21. Triumfeiia macrophylla K. Schum. 1 2 fruits 0.5 22. Sanicula elaia D. Don 1 1 seeds 0.2 Forty-five female hares made up 46.9% of the total collected. These carried 324 disseminules accounting for 74.7% of the total (Table 2). The highest number of disseminule species on an individual hare was 10, 43.5% of the total number of species, on a female hare in the long dry season. Twenty percent of the females carried no disseminules as compared to 49% of the males. Females show nonsignificant (x2 =3.7, P>0.5) seasonal differences in number of disseminules as well as non-significant differences (x2=3.3, P>0.05) in the species of plant disseminules they carried (Table 2). Fifty-one (53. 1%) of the hares collected were males, and these carried 1 10 disseminules, accounting for 25.3% of the overall total (Table 2). The highest number of disseminules recorded on a male hare was 17 at the end of the short dry season and the highest number of disseminule species was 6 accounting for 26. 1% of the total annual disseminule species (n = 23) at the beginning of the long wet season of 1982. Twenty-two plant species were identified on the hares, with a total of 436 disseminules. Fourteen of the species were grasses which made up 64% of the species involved with a total of 3 19 disseminules or 73% of the disseminules recorded. Herbs and shrubs made up 36% of the species involved with 8 species represented. The herbs and shrubs totalled 1 17 or 27% of the disseminules (Table 3). Among the grass species, Hyparrhenia filipendula was the most common, with 65 seeds (15% of the total disseminules) found in 5 months. Cenchrus ciliaris was found in 7 months. Tragus berteronianus in six and Panicum brevifolium, Aristida adoensis and Harpachne schimperi each occurred in only one month (Table 3). Tribulus terrestris, a grassland herb, was the most common of the herbaceous plant disseminules encountered, with 43 fruits representing 10.0% of total disseminules and occurred in 3 months of the study. The commonest herb, Oxygonum sinuatum occurred in 4 months and contributed 39 fruits, 9% of the total disseminules (Table 3). DISCUSSION Ectozoochory by hares as demonstrated in this study could have tremendous impacts on local distribution of some plant species in the Park. Sixty-four percent of the hares examined carried disseminules on their fur regardless of the season. These disseminules are dropped off daily as the hare grooms its fur. Some areas of the Park at certain seasons of the year carry from 5-12 hares/ha, and thus their role in the dynamics of the vegetation could be significant. The hares in this study area carried a higher number of species than those studied by Agnew and Flux (1970) in Kenya savannahs, although the Kenyan animals carried a higher total number of disseminules No. 184 Page 5 than those from this Park. The higher number of species reported here could be because of differences in plant communities (and possibly soils and climate) in the two areas, and more important still are the behavioural and niche differences between Lepus capensis and Lepus crawshayi (Agnew and Flux 1970). Lepus crawshayi prefers thicker bush as well as grassland (Eltringham and Flux 1971), therefore it is capable of picking up and transporting disseminules within the ranges of the two or more habitats. The presence on L. crawshayi of disseminules of herbs, e.g., Triumfetta macrophylla, common only along swamp edges, and Sanicula elaia, a forest herb (Lind and Tallantire 1975), suggests that this species may have a large home range or wide habitat preferences. The high frequency of Hyparrhenia filipendula corresponds to its wide distribution in the Park and its adaptive features for zoochory. The less commonly observed species, e.g., Harpachne schimperi (Table 3) were not rare in the Park, but rather lacked the modifications for ectozoochory such as barbs or hooks. This finding agrees with that of Friedman and Stein (1978) who reported the seed dispersal mechanisms of Anastatica hierochuntinca to have profound influence on its ecological dispersion through zoochory. Herbs like Tribulus .terrestris and Oxygonum sinuatum are very highly modified for zoochorus life and are seasonally abundant. The highest incidences of the disseminules on hares occurred in the dry seasons and the lowest registered in the wet seasons (Table 1). The most logical explanation for these seasonality effects is that most annual and perennial plants in the region that have their growth cycles synchronized with the wet and dry seasons, thus have more mature and abundant disseminules in the dry season (Lind and Tallantire 1975). Possibly the disseminules are removed only “accidentally” as the hare scratches off itching ectoparasites from its body. Studies have reported that burning (which in this Park occurs extensively in the dry seasons) in dry seasons reduces the intensity and extent of infestations of ectoparasites (Bendell 1955, Flux 1972, Clifford et al. 1976). Fewer parasites in the dry seasons as a possible result of bush fires, could mean that hares groom less and hence accumulate plant disseminules on the body. Flux (1967) reported that moulting in Lepus europaeus Pallas in New Zealand occurs in seasons similar to the rainy seasons in the tropics, and later Agnew and Flux (1970) reported that Lepus capensis and Lepus crawshayi in the tropics moult in the wet seasons. If this is the timing for moulting of savannah hare, this could further explain the seasonality of the number of disseminules observed. The species of the disseminules were found to vary little with seasons. This could have been because of the widely scattered collection locations in the Park or due to the local variation in microclimate, vegetation and soils (Langdale-Brown et al. 1964). The higher incidence of disseminules on female hares of Lepus capensis reported by Agnew and Flux (1970) was also the case in Lepus crawshayi in this study. The explanation put forward by Agnew and Flux (1970), i.e., that the female hares have less time to groom or tend to live in thicker cover as they tend the young where they encounter more burrs, is plausible but it is open to further investigation. A further explanation could be that females tend to range farther afield, especially when breeding and lactating, to acquire enough food. ACKNOWLEDGEMENTS We would like to thank the following for any assistance or guidance they may have provided: Dr. E.L. Edroma, Uganda Institute of Ecology; Dr. Ross Wein and Mohammed El-Bayoumi, Fire Science Centre, U.N.B.; D.L. Mugabe, E.N. Sabiiti and Peter Thomas, Biology Department, U.N.B.; and the Trustees of the Uganda National Parks. The senior author wishes to recognize the financial support provided by the Canadian International Development Agency. We sincerely thank Lisa Haindl for typing the manuscript. The field work was supported by a research grant from the Uganda Ministry of Regional Cooperation through the Uganda Virus Research Institute, Entebbe. LITERATURE CITED AGNEW, A.D.Q. & J.E.C. FLUX. 1970. Plant dispersal by hares (Lepus capensis L.) in Kenya. Ecology 51 (4): 735-737. BENDELL, J.R. 1955. Disease as a control of a population of blue grouse, Dendragapus obscurus fuliginosus (Ridgway). Can J. Zool. 33:195-223. CLAYTON, W.D. 1974. Flora of tropical east Africa: Gramineae (Part 1). London, Crown Agents. 176 p. CLIFFORD, C.M., J.E.C. FLUX & H. HOOGSTRALL. 1976. Seasonal and regional abundance of ticks Page 6 .No. 184 (Ixodidae) on hares (Leporidae) in Kenya. J. Med. Entom. 13(l):40-47. EDROMA, E.L. 1975. The influence of grazing and burning on the productivity and dynamics of grassland types in Ruwenzori National Park, Uganda. D.Sc. Thesis, University of Giessen. ELTRINGHAM, S.K. & J.E.C. FLUX. 1971. Night counts of hares and other animals in East Africa. E. Afr. Wildl. J. 9:67-72. FIELD, C.R. 1968. The food habits of some wild ungulates in Uganda. Ph.D. Thesis, University of Cambridge. FLUX, J.E.C. 1967. Reproduction and body weights of the hare (Lepus europaeus Pallas, in New Zealand. N.Z.J. Sci. 10(2):357-401. 1972. Seasonal and regional abundance of fleas on hares in Kenya. J.E. Afr. Nat. Hist. Soc. Nat. Mus. 136:1-8. FRIEDMAN, J. & Z. STEIN. 1978. The influence of seed-dispersal mechanisms on the dispersal of Anastatica hierochuntica (Cruciferae) in the Negev Desert. Israel J. Ecol. 68:43-50. GRIME, J.P. 1979. Plant Strategies and Vegetation Process. John Wiley and Sons, New York. LANGDALE-BROWN, I., M.A. OSMASTON & J.G. WILSON. 1964. The Vegetation of Uganda and its Bearing on Land Use. Government Printer, Entebbe, Uganda. LIND, E.M. & A.C. TALLANTIRE. 1975. Some Common Flowering Plants of Uganda. Oxford Univ. Press. 2nd Ed. 257 p. LOCK, J.M. 1967. Vegetation in relation to grazing and soils in Queen Elizabeth National Park, Uganda. Ph.D. Thesis, University of Cambridge. MCCLINTOCK, D. 1965. The transport of plant propagules by mammals. Bull. Mammal. Soc. British Isles 24:12-13. OSMASTON, H.A. 1971. The vegetation of Murchison Falls and Queen Elizabeth National Parks. In: Uganda National Park Hand Book, 5th edt. pp. 95-112. Trustees of the Uganda National Parks, Kampala. RIDLEY, H.N. 1930. The Dispersal of Plants Throughout the World. Reeve and Co. Ltd., Ashford, Kent. 744 p. SPINAGE, C.A. 1968. The autecology of the Uganda waterbuck (Kobus defassa ugandae Neumann) in Queen Elizabeth National Park, Uganda. Ph.D. Thesis, University of London. STEEL, R.G.D. & J.H. TORRIE. 1960. Principles and Procedures of Statistics with Special Reference to the Biological Sciences. MCgraw-Hill Book Co., Inc. New York. 481 p. TOMICH, P.Q.N. & C.H. LAMOUREUX. 1968. Ecological factors on Manama Island, Hawaii. Pacific Sci. 22:352-68. VAN DER PIJL, L. 1972. Principles of Dispersal in Higher Plants. Springer- Verlag, Berlin. Received l October 1985 Editors: J.J. Hebrard, H. Beentje No. 184 Page 7 NOTICE TO CONTRIBUTORS Papers on natural history subjects will be considered on the understanding that these are being offered exclusively to the Society and have not been submitted, at the same time, to any other journal. They should be addressed to the Editors, Journal of the East Africa Natural History Society and National Museum, P.O. Box 44486, Nairobi, and be accompanied by a covering letter from the author responsible for correspondence and decisions regarding the typescript. Two copies of the contribution should be forwarded to the Society; a further copy should be kept by the author for reference. The typescript should be on A4 size paper. Use one side only, double-spaced with margins of at least 5 cm at the top and left-hand side of the pages to allow for editorial instruction to printers. The pages must be numbered and all measurements are to be in metric units. The title page should contain the title of the article followed by the name of the author(s) and institution(s) where the research was carried out. A present address, if applicable, should be included as a footnote. Abstract to be on the second page and contain not more than 150 words. State the purpose of the study, procedures, main findings and conclusions. This should be intelligible in itself without reference to the paper. Nomenclature In cases where there are well known and accepted English names, e.g. for birds and larger mammals, these should be used throughout the text, accompanied by the scientific name on first mention Authorities should be given beside the generic and specific names, when quoted for the first time, forall groups, with the exception of ornithological papers not dealing with taxonomic discussion. Text This is usually, but not necessarily, divided into parts with the headings Introduction, Methods, Results and Discussion, these being broken up with sub-headings where necessary. Main heads should be centred on the page and typed in capital letters. Subheads should be ranged to the left margin and typed in capitals. Any further headings should start with a capital letter followed by small letters on the left margin. No headings should be underlined. Underlining is an instruction to the printer to use italic type and should be reserved for all generic and species names, foreign words not adopted into English and for descriptions of bird or animal vocalisation. Acknowledgements As acknowledgements may imply endorsement of the data and conclusions contained in the authors work, it is advisable to obtain permission before quoting persons who have made contributions to the study. References Care should be taken to see that all references quoted in the text are given in full, in this section. To avoid confusion, references to books and periodicals should not be abbreviated and neither should they be underlined. For example: (in text) Cave & Macdonald (1955) or (Cave & Macdonald 1955) (in reference list) CAVE, F.O. & MACDONALD, J.D. 1955. Birds of the Sudan, Oliver & Boyd, London. Tables Each table must be typed, double-spaced on a separate sheet and numbered. Leave good spaces between columns and omit internal vertical and horizontal rules. Cite tables in the text in consecutive order. Illustrations Submit two sets of figures which should be well, drawn on bristol board or heavy tracing paper with a photographed or xeroxed copy. Lettering must be of first quality (e.g. lettraset) and consideration must be given to size when reduced. Typewritten or freehand lettering will not be accepted unless put onto an overlay which can be followed by the printers for typesetting. Each figure should have a label pasted on its back giving the number of the figure, caption details and accreditation Cite each figure in the text in consecutive order. A separate caption sheet should be added to the typescript. Photographs are acceptable if they have good definition and contrast and should be produced on glossy paper. Proofs The author will be sent one set of galley proofs for correction of typographical errors. Authors may be charged for any changes from the original typescript made by him at this stage, and which result in additional printing expense. Plates or page proofs will not be sent Published material The author is entitled to 25 copies free of charge. Further copies can be supplied at cost price, provided the order is placed when the typescript is submitted if there is more than one author, 30 copies will be sent to the corresponding author for distribution The society is supported entirely by membership subscriptions and may not, for financial reasons, be able to accept all submissions of value. Wherever possible an author should seek a grant, from his institution or other source, to help with costs of publication. Page 8 No. 184 Published by The East Africa Natural History Society, Box 44486, Nairobi, Kenya. Phototypeset- ting by Kul Graphics Ltd., P.O. Box 18095 and printed by Manhattan Displays, P.O. Box 30434. Nairobi. Ellx m — JOURNAL f OF THE EAST AFRICA NATURAL HISTORY eiix SOCIETY AND NATIONAL MUSEUM NH ™.mber 1985 Volume 75, No. 185 VEGETATIVE KEY TO THE ALPINE VASCULAR PLANTS OF MOUNT KENYA Truman P. Young Department of Biology University of Miami Coral Cables, Florida, USA and Mary M. Peacock* Department of Botany University of California Davis, CA, USA INTRODUCTION In recent years there has been an increasing interest in several aspects of plant biology in the alpine zone of Mount Kenya. To my knowledge, at least a dozen research projects were carried out between 1977 and 1984. Fortunately, the flora of the region has been the subject of a fine monograph by Hedberg (1957), and the currently available volumes of the Flora of Tropical East Africa (FTEA hereafter) contain a majority of the alpine species. Although reproductive individuals are generally more prevalent on the mountain throughout the year when compared to the drier lowlands, many species are only rarely found in the reproductive state (personal observations). As part of a comprehensive study of the vegetation of the upper Teleki Valley on Mount Kenya, we produced this key based on vegetative characters. It has since proven useful (in manuscript form) in studies by various other researchers. It is hoped that its publication will facilitate and encourage future biological research on Mount Kenya. A lower elevational limit of 3500 meters was chosen to eliminate a number of forest species that occur sporadically above the timberline. Several species not listed by Hedberg are included. These represent either new records for Mount Kenya, such as Helictotrichon umbrosum and Cystopteris diaphana, or new altitudinal ranges discovered in our studies, such as Kniphofia thomsonii and Asplenium £(Agnew 1974). A separate paper will document the distribution, frequency, and ecology of the approximately 70 species found in the upper Teleki Valley. As an additional aid to identification, three short reproductive keys for difficult groups are appended to the main vegetative key. The first of these covers three species of Helichrysum, the second covers herbs with opposite entire leaves, and the third covers the grasses. For three genera (Pua, Colpudium, and Cerastium) of two species each, no reliable vegetative distinguishing traits could be found. These genera are included in the reproductive keys. One other genus deserves special mention. Hedberg (1957) and Clayton (1970) distinguished Pentaschistis minor and P. borussica by panicle shape. The former reportedly has a linear panicle, and the latter an open panicle. In addition, their altitudinal distributions on Mount Kenya were thought to be disjunct (Hedberg 1957). We have found that not only can Pentaschistis spp. be found at intermediate elevations, but that panicle shape in P. minor varies with plant age and air temperature (T.P.) Young, personal observations). In addition, Clayton (1970) reports the existence of intermediates between P. minor and P. borussica. We have found no consistent vegetative differences, and both key out here as P. minor. present address: Department of Biology, University of Nevada, Reno, NV, USA page 2 No. 185 Hints to Using the Key It is particularly helpful to use this key in conjunction with Hedberg (1957), Agnew (1974), and relevant volumes of the Flora of Tropical East Africa (cf. Clayton 1970), sections of whose keys we have used here. After reaching one or more possible species identities in the key, compare your specimen to the descriptions in these texts, paying attention also to habitat and elevation. U nfortunately, the first two of these references are difficult to obtain at this time in East Africa. One hopes that this will not always be the case. This key is designed for use with living vegetative material. Dry material may differ, especially in color. Reproductive material is often very different from vegetative material; for example, several small rosette species produce long leafy stems at reproduction, and leaves on reproductive stems may differ from those on vegetative plants. As in any vegetative key, there is likely to be some confusion concerning seedlings and young plants. For example, very young shaded Lobelia telekii individuals look similar to mature Limosella africana plants. In this case the latex of the former is indicative, but in others it must be left to the reader’s own development of a ‘feel’ for the species. In particular, young plants of woody or shrubby species often appear herbaceous. We have tried in this key to reduce such ambiguities, relying on invariable characters as much as possible. In order to make this key accessible to as broad a readership as possible, we have tried to minimize the use of botanical jargon. Nonetheless, some specialized terminology is unavoidable, particularly for the grasses. 1) 2) 3) 4) 5) 6) 7) 8) 9) 10) ID Plant a fern; leaves (fronds) thin, compound, glabrous (but may have scales); ultimate segments dentate; leaves arising from a rhizome; plant not a rosette 2 Plant not a fern; if leaves compound and glabrous, then ultimate segments entire or the vegetative plant a rosette 3 Leaves multipinnate; final frond segments (pinnae) fan-shaped; rachis with coppery scales Asplenium sp. £. (Agnew) Leaves pinnate; pinnae lanceolate, rachis without scales Cyslopleris diaphana (syn. C. Jragilis) Plant glabrous; erect stems densely covered by numerous single-veined leaves less than 2 cm long (a nonseed plant that often has sporangia in its leaf axils) Lycopodium saururus Plant not as above if leaves small, glabrous and entire, then not densely covering erect stems (Angiosperms) 4 Leaves with parallel venation, entire, simple, often grasslike, not succulent, >2 cm long; if <2 cm long, then with distinct ligules ( Monocotyledons) 5 Leaf veins net-like; leaves entire to deeply lobed, simple or compound, if somewhat grasslike, then succulent or woody or with liguless leaves<2 cm long (Dicotyledons) 32 An aquatic plant with long internodes and leaves >1 cm broad Potamogeton schweinfurthii Plant terrestrial, or with narrow leaves and short internodes 6 Leaves >1.5 cm wide 7 Leaves <1.0 cm wide 9 Leaves often >3 cm wide, always narrowing near the base. Disa stairsii Leaves 1.5 to 2.5 cm wide, linear 8 Leaves with a raised midvein on both upper and lower surfaces; leaf blades fiat Gladiolus watsonoides Leaves without a raised midvein, or only on the lower surface; Leaf blades often V-shaped .... Kniplwfia thomsonii Leaves triangular to rectangular in cross section, with notched corners (in cross section) . . . Romulea keniensis Leaves flat, folded, rolled, round, or V-shaped, not notched 10 Leaves>4 mm wide, tinged red and densely long hairy Luzula abvssinica Leaves < 4 mm wide, if greater, then not tinged red and densely long hairy II Leaves with a distinct ligule, either membranous ora fringe of hairs (G rami nae).... 14 (see also 1 34) Ligule indistinct or absent (Carex) 12 No. 185 page 3 12) Leaves (culms) round Carex runssoroensis Leaves flat to V-shaped 13 13) Leaves <4 mm wide, strongly V-shaped Carex monostachya Leaves >4 mm wide Carex sp. (prob. bagaertii) 14) Ligule a fringe of hairs 15 Ligule membraneous 16 15) Backs of leaves with single raised mid veins Pcniaschistis minor Backs of leaves with a number of equal ribs Andropogon ameihystinus 16) Leaf forming two right angles between the sheath and the blade Koeleria capensis Leaf forming a single acute angle at the ligule 17 17) Leaves flat or folded, not readily rolled between the fingers 18 Leaves (tightly) rolled or subulate, readily rolled between the fingers 28 18) Leaves >20 cm long, >5 mm wide Andropogon ameihystinus (inch A. longipes) Leaves <20 cm long, if longer then <5 mm wide 19 19) Leaves flat 20 Leaves flolded; if open, then with a distinct crease 23 20) At least some leaves >4 mm wide 21 All leaves <3.5 mm wide 22 21) Leaves glabrous or (sparsely hairy) when crushed not having a distinct aromatic smell or taste Colpodium spp. (see 139) Leaves usually long hairy; when crushed smelling and tasting of cumarin Anthoxanthum nivale 22) Leaves sparsely pubescent; some hairs >2 mm long Helictoirichon umbrosum Leaves glabrous or with a few hairs <1.5 long 4 grosiis quinqueseta 23) Folded leaf >1.5 mm wide 24 Folded leaf < 1.5 mm wide 26 24) Stem with distinct internodes >2 cm long Calamagrostis hedbergii Plant tufted; internodes <1 cm long 25 25) Upper leaf sheath of two distinct parts -a membraneous extension of the ligule inside, and a green leafy lip outside Colpodium spp. (see 1 39) Upper leaf sheath not of two distinct parts Poa spp. (see 146) 26) Ligule < 1.5 mm long with dark glands at its base, especially in older leaves Festuca abyssinica Ligule > 1.5 mm long, without dark glands at its base 27 27) Leaf bases, sheathes, or blades tinged red; blades flexuous Deschampsia flexuosa Plant not tinged red; leaf blades straight Agrostis sclerophylla 28) Leaves smooth or only slightly rough to the touch 29 Leaves scabrous, distinctly rough to the touch 30 29) Culm bases white, not grey or brown or reddish Agrostis gracifolia Culm bases grey or brown or reddish 30 30) Ligules<1.5 mm long; culm bases often reddish Festuca pilgeri Ligules > 1 .5 mm long; culm bases not reddish 31 3 1 ) Leaves striate Agrostis volkensii Leaves estriate Agrostis traehyphylla 32) Leaves producing a milky latex 33 Leaves not producing a milky latex 35 33) Latex white; leaves never >4 cm long; leaves entire, often emarginate . . Dianlhoseris schimperi Latex cream colored; leaves usually >4 cm long; leaves shallowly crenate, not emarginate .... (Lobelia) 34 34) Midvein glabrous in smaller plants; in larger plants, rosette retaining a reservoir of water Lobelia deckenii ssp keniensis Lower midvein pubescent on the underside, rosette not retaininga reservoir of water(note: hybrids between these two species occur rarely) Lobelia telekii 35) Leaves or stem armed with stout spines, not merely barbed 36 Plant not armed with spines, although some leaves may have weak barbs 38 page 4 No. 185 36) Only stems armed; plant a woody shrub Helichrysum citrispinum Leaves armed, plant a rosette (Carduus) , . 37 37) Leaves compound; undersides white with pubescence Carduus keniensis Leaves dentate, undersides green Carduus chamaecephalus (syn. C. platyphyllus) (note: hybrids occur rarely) 38) Leaves distinctly compound and plant herbaceous or woody only at the base 39 Leaves simple, entire to deeply lobed; or if leaves compound, then plant distinctly shrubby (see 106) 49 39) Leaves with three leaflets; leaflets entire or minutely toothed Trifolium multinerve Leaves with more than three leaflets, or if three then distinctly dentate 40 40) Leaflets ovate, with acuminate teeth Cardamine obliqua Leaflets dentate, lobed, entire, sometimes filiform; not ovate 41 41) Leaves <10 cm long and leaflets >5 mm wide (Ranunculus) 42 Leaves >10 cm long; or if less, then leaflets <5 mm wide 44 42) Leaves multipinnate Ranunculus oreophylus Leaves trifoliate 43 43) Leaflets deeply lobed Ranunculus keniensis Leaflets merely dentate Ranunculus aberdaricus 44) Leaflets <2 mm wide, filiform (Peucedanum) 45 Leaflets >3 mm wide, dentate 46 45) Leaf rachis glabrous Peucedanum friesiorum Leaf rachis sparsely pubescent Peucedanum kerstenii 46) Leaves densely pubescent, white to silvery in appearance 47 Leaves sparsely pubescent, greenish 48 47) Leaflets pinnately lobed or leaves bipinnate Anthemis tigrensis Leaflets entire to 1-2 lobed Cotula abyssinica 48) Leaflets <5 mm wide Haplosciadium abyssinicum Leaflets >7 mm wide Heracleum inexpectatum (syn. Heracleum elgonense) 49) Plant a rosette. internodes<5 mm long (alt hough leafy stolons or reproductive shoots may be present) 50 Vegetative plant with distinct internodes 73 50) Leaves entire 51 Leaves dentate to deeply lobed 58 5 1 ) Leaves spatulate, >1.5 cm long 52 Leaves not distinctly spatulate; if slightly so then <1.5 cm long 54 52) Leaves <5 mm wide, not purple tinged Limosella aquatica (syn. Limosella africana) Leaves >5 mm wide, or purple tinged (Swertia) 53 53) Plants producing stolons Swertia crassiuscula Plants not producing stolons Swertia volkensis 54) Leaves succulent Subularia monticola Leaves not succulent 55 55) Leaves glabrous 56 Leaves pubescent 57 56) Leaves >3 mm wide Dianthoseris schimperi Leaves <3 mm wide Sagina afroalpina 57) Underside of leaf apex with a distinct gland; leaf hairs not glandular Myosotis keniensis Underside of leaf apex without a white gland; leaf hairs glandular Cerastium spp. (see 1 33) 58) Leaves dentate to lobed less than halfway to the midvein 59 Leaves lobed more than halfway to the midvein 70 59) Leaves robust, thick ( I mm), with stout midveins and incurved margins, dentate 60 Leaves thin, with thin margins, dentate or not * 66 No. 185 page 5 60) 61) 62) 63) 64) 65) 66) 67) 68) 69) 70) 71) 72) 73) 74) 75) 76) 77) 78) 79) 80) 81) 82) 83) Leaves white woolly underneath 61 Leaves green underneath, sometimes light green due to a thin layer of hairs 62 Upper leaf surfaces relatively smooth, plant becoming megaphytic Senecio brassiea Upper leaf surfaces rugulose; plant a small flat rosette Haplocarpha rueppellii Leaves < I cm wide; plant not becoming megaphytic (see 121) 63 Leaves >2 cm wide; plant becoming megaphytic 64 Plant glandular sticky Senecio schweinfurthii Plant not glandular sticky Senecio keniophytum Leaves green beneath Senecio keniodendron Leaves greenish-white beneath, due to a thin layer of hairs 65 Megaphytic rosette plant growing to several meters; absent from the Teleki Valley, occurs along rocky courses elsewhere Senecio battiscombei Megaphytic rosette plant never reaching much taller than 1 m; only found along the ecotone between adjacent Senecio brassiea and Senecio keniodendron populations, not uncommon in these situations Senecio keniodendron x 5. brassiea hybrid Leaves lobed or crenate, >2 cm long 67 Leaves toothed, <2 cm long 68 Leaves >5 cm long, often deeply lobed Scabiosa columbaria Leaves <5 cm long, crenate Conyia subscaposa Hairs simple or leaves glabrous . Wahlenbergia pusilla Hairs forked or stellate 69 Leaves densely covered by stellate hairs usually <.5mm long; (silique >.7 mm broad) Arabis alpina Leaves sparsely to moderately covered by stellate and simple hairs, some hairs at the bases of leaves up to 7 mm long; (silique <.7mm broad) Arabidopsis lhaliana Basal leaves usually thrice ternately lobed Anenome thomsonii Leaves pinnately, bipinnately, or palmately lobed 71 Leaves palmately lobed; not longer than wide; sometimes reddish (Geranium) 112 Leaves pinnately to bipinnately lobed; longer than wide, not reddish 72 Leaves >5 cm long Scabiosa columbaria Leaves <5 cm long Oreophyton falcalum Leaves succulent 74 Leaves not succulent 77 Plant woody at base Sedum ruwenzoriense Plant herbaceous 75 Leaves alternate Sedum crassularia Leaves opposite (Crassula) 76 Plant restricted to shallow soil on dry ledges; leaves distinctly succulent, nearly spherical Crassula alba Plant of seasonal boggy flats; leaves weakly succulent Crassula granvikii L.eaves entire or with barbs or small acuminate teeth 78 Leaves distinctly dentate to lobed to compound 104 Plant woody, at least at the base 79 Plant herbaceous throughout 93 Leaves broader than 7 mm, never sticky 80 Leaves narrower than 5 mm, or glandular sticky 81 Leaves opposite Hypericum keniense Leaves alternate Protea kilimandscharica Leaves clasping the stem (Helichrysum) 82 Leaves petiolate, not clasping the stem 85 Leaves >8 mm wide, glandular sticky Helichrysum formosissimum Leaves <5 mm wide, not glandular sticky 83 (see also 125) Stems usually >6 mm in diameter; upper and lower leaf surfaces distinctly different in color; plant a shrub to 2m Helichrysum chionoides Stems <5 mm in diameter; upper and lower leaf surfaces similar; plant >.5m high 84 page 6 No. 185 84) Leaves on young vegetative stems appressed, spreading when older; dark apical glands inconspicuous Helichrysum cymosum Leaves on vegetative stems spreading; dark apical glands conspicuous .... Helichrysum brownei 85) Leaves >1 cm long 86 Leaves < 1 cm long 87 86) Leaf bracts on uppermost leafless parts of stems large, dense, covering most of the stem Euryops brownei Leaf bracts small, sparse; the stem clearly visible Hehenslrelia angolensis (previously H.dentaia) 87) Leaves in clusters, silvery grey, linear Stoebe kilimanJscharica Leaves not in clusters, if linear, then deep green (Ericaceae) 88 88) Young leaves densely pubescent, <3 times long as broad (Blaeria) 89 Young leaves glabrous or nearly so, >3 times long as broad 90 89) Leaf hairs >.5mm; plant sparsely branched, the side branches much weaker than the main stem, leaves usually >3 mm long Blaeria jilago Leaf hairs <5mm long; plant richly branched; leaves usually <2.5 mm long Blaeria johnstonii 90) Young stems densely pubescent 91 Young stems glabrous or nearly so 92 9 1 ) Leaf blade <5 times the length of the petiole, petiole > 1 mm long Phillipia trimera Leaf blade > 5 times the length of the petiole, petiole usually < 1 mm long Erica arborea 92) Leaves <5mm long; plant a shrub to several meters Philippia excelsa Leaves >5mm long; plant a small woody herb to .5m Erica whyteana Note: The genera Erica and Philippia can be distinguished most reliably in flower by the relative size of the dry stigma: Dry stigma >3 times the width of the style Philippia Dry stigma <3 times the width of the style Erica 93) Leaves in whorls of four or more (Galium) 94 Leaves opposite or alternate 96 94) Leaves 4-6 in a whorl Galium glaciale Leaves (6-)8-10 in a whorl 95 95) Leaves barbed Galium ruwenzoriense Leaves not barbed Galium ossirwoense 96) Leaves alternate 97 Leaves opposite 98 (see also 127) 97) Leaves >4 times long as broad Senecio jacksonii Leaves <3 times long as broad Anagallis serpens 98) Leaves distinctly spathulate and some >2 cm long 99 Leaves not distinctly spathulate, or less than 2 cm long 100 99) Leaves > 1.5cm wide Swertia kilimanJscharica Leaves < 1.5cm wide Swertia subnivalis 100) Leaves with thickened margins ( Satureja ) Satureja bijiora (inc. 5. punctata) Leaves without thickened margins 101 101) Stem with glandular hairs (Cerastium) 133 Stem without glandular hairs 102 102) Opposite leaf bases united Crassula granvikii Opposite leaf bases not united 102b 102b) Stems succulent, reddish; leaves rarely >3 times long as broad Anagallis serpens Stems non-succulent, not reddish if alive; leaves often >3 times as long as broad 103 103) Leaves narrowing at the base Callitriche stagnalis Leaves broad at the base Montia font ana 104) Leaves with distinct stipules 105 Leaves exstipulate 113 105) Plant woody, at least at the base 106 Plant herbaceous 109 No. 185 page 7 106) 107) 108) 109) 110) HD 1 12) 113) 1 14) 1 15) 116) 117) 1 18) 1 19) 120) 121) 122) 123) 124) Leaves bipinnatc Artemisia afro Leaves trifoliate or simple 107 Leaves bipinnate Adenocarpus mannii Leaves or leaflets dentate or serrate 108 Leaves densely pubescent, usually trifoliate; stipule membraneous, with undivided apex Alchemilla argyrophylla Leaves sparsely to moderately pubescent, simple; stipule foliaceous, with a dentate apex Alchemilla johnstonii Leaves with stinging hairs Urtica massaica Leaves without stinging hairs 110 Stipules dentate; plant erect Cineraria grandijlora Stipules entire; plant spreading Ill Leaves sharply dentate Alchemilla cyclophyla Leaves without acute teeth (Geranium) 112 Leaf blades reniform (kidney shaped) Geranium kilimandscharica Leaf blades pentagonal Geranium arabicum Leaves opposite, at least near the base 114 Leaves alternate 120 Leaves more than twice as long as broad 115 Leaves less than twice as long as broad 117 Leaves deeply lobed toward base, entire at apex Valeriana kilimandscharica Leaves crenate-dentate throughout (Bartsia) 116 Leaves usually <3 times long as broad, rarely rolled (flowers purple) Bartsia abvssinica Leaves usually >3 times long as broad, often rolled (flowers yellow) Bartsia decurva (syn. Bartsia kilimandscharica) Leaves minty, stems hairy (Satureja) Leaves not minty, stems glabrous (Veronica) Leaf bases cordate Leaf bases cuneate to truncate All stems prostrate Some stems ascending to erect Leaves glabrous Leaves pubescent (Senecio, see Hedberg 1957, page 225) Plant glandular sticky Plant not glandular sticky Plant woody at the base Plant herbaceous Plant woody, at least at the base Plant herbaceous Leaves petiolate, dentate Leaves apetiolate, mostly entire 118 119 Satureja kilimandscharica Satureja simensis Veronica gunae Veronica glandulosa Hebenstretia angolensis (see at 86) 121 122 123 Senecio rosei/lorus Senecio purtschdleri Senecio schweinfurthii 124 Senecio keniophytum Senecio jacksonii SPECIAL REPRODUCTIVE KEYS TO DIFFICULT GROUPS Unarmed Helichrysum spp. 125 Involucre bracts appressed, inconspicuous; capitula diameter <4mm . . . Helichrysum cymosum Involucre bracts open, showy; capitula diameter > 15mm 126 126) Heads 1-5 in each corymb, 2. 5-3. 0cm wide, white with faint reddish tinge in bud Helichrysum brownei Heads usually 5-10 or more in each corymb, 1 .5-2. 5cm wide, pure white or with a brownish tinge Helichrysum chionoides Herbs with opposite, entire, glabrous leaves 127) Ovary of four separate carpels 128 Carpels united 129 128) Flowers unisexual Callitriche stagnalis Flowers bisexual Crassula granvikii page 8 No. 185 129) Flowers irregular Satureja biflora Flowers regular 130 130) One style and stigma 131 3-5 styles or stigmas 133 131) Petals small, <4mm long Montia fontana Petals longer than 6mm 132 132) Petals with indistinct nectaries without ciliation . Swertia suhnivalis Petals with distinct ciliate nectaries Swertia kilimandscharica 133) Petals often inconspicuous, with a narrow slit at apex: capsule teeth erect with reflcxcd margins octaiulrum Petals emarginate at apex, capsule teeth backwards or spirally jCerasliam a/romontanum Cramineae 134 Inflorescence [two to] several digitately arranged spikes Andropogon amethystinus Inflorescence an open or contracted panicle 135 135 Ligule a fringe of hairs Pentaschistis minor Ligule membraneous 136 136 Spikelets with one floret 137 Spikelets with more than one floret (sometimes only one fertile, but then with more than one awn per spikelet) 144 137 Floret with long hairs longer than the spikelet (sometimes deciduous upon drying) Calamagrostis hedbergii Floret without hairs or with only short hairs 138 138 Leaves >3mm wide; florets awnless (Colpodium) 139 Leaves <3mm wide, if greater then florets awned (Agrostis) 140 139 Spikelets 4-6. 5mm long; leaves to 12cm long Colpodium chionogeiton Spikelets 2. 5-3. 5mm long; leaves to 6.5cm long Colpodium hedbergii 140 Florets awnless Agrostis sclerophylla Florets awned 141 141 Leaves flat, >2mm wide Agrostis quinqueseta Leaves rolled, <2mm wide 142 142 Leaves smooth Agrostis gracifolia Leaves rough 143 143 Leaves striate Agrostis volkensii Leaves estriate Agrostis traehyphylla 144 Florets awnless 145 Florets short to long awned 147 145 Glumes enclosing the florets Koeleria capensis At least the upper florets exserted (Poa) 146 146 Panicle contracted Poa leptoclada Panicle open Poa schimperiana 147 Upper florets distinctly exserted; lemma straight-awned from the tip (Festuca) 148 Upper florets not distinctly exserted; awns dorsal or bent 149 148 Leaves rough Festuca pilgeri Leaves smooth Festuca abysinica 149 Spikelets enclosed by the glumes, 2-8mm long Deschampsia Jlexuosa Spikelets exserted from the glumes, 8-16mm long 150 150 Leaves usually >4mm wide, smelling of cumarin Anthoxanthum nivale Leaves <4mm wide, without a distinctive smell Helictotrichon umbrosum No. 185 page 9 BIBLIOGRAPHY Agnew, A.D.Q. 1974. Upland Kenya Wildflowers. Oxford University Press, London. 827 pp. Clayton, W.D. 1970. Graminae (Part I). Flora of Tropical East Africa. Milne-Redhead, E. andR.M. Polhill, eds. Crown Agents for Overseas Governments and Administrations, London. Hedberg, O. 1957. Afroalpine vascular plants. Symbolae Botanicae Uppsalensis 15. ACKNOWLEDGEMENTS We thank the Government of Kenya, the National Herbarium, A.D.Q. Agnew, M. Otieno, and the Arthur K. Gilkey Fund of the American Alpine Club. Received June 1984 Editors: H.J. Beentje, J.J. Hebrard Page 10 Page 11 Page 12 No. 185 NOTICE TO CONTRIBUTORS Papers on natural history subjects will be considered on the understanding that these are being offered exclusively to the Society and have not been submitted, at the same time, to any other journal. They should be addressed to the Editors, Journal of the East Africa Natural History Society and National Museum, P.O. Box 44486, Nairobi, and be accompanied by a covering letter from the author responsible for correspondence and decisions regarding the typescript. Two copies of the contribution should be forwarded to the Society; a further copy should be kept by the author for reference. The typescript should be on A4 size paper. Use one side only, double-spaced with margins of at least 5 cm at the top and left-hand side of the pages to allow for editorial instruction to printers. The pages must be numbered and all measurements are to be in metric units. The title page should contain the title of the article followed by the name of the author(s) and institution(s) where the research was carried out. A present address, if applicable, should be included as a footnote. Abstract to be on the second page and contain not more than 150 words. State the purpose of the study, procedures, main findings and conclusions. This should be intelligible in itself without reference to the paper. Nomenclature In cases where there are well known and accepted English names, e g. for birds and larger mammals, these should be used throughout the text, accompanied by the scientific name on first mention. Authorities should be given beside the generic and specific names, when quoted for the first time, forall groups, with the exception of ornithological papers not dealing with taxonomic discussion. Text This is usually, but not necessarily, divided into parts with the headings Introduction, Methods, Results and Discussion, these being broken up with sub-headings where necessary. Main heads should be centred on the page and typed in capital letters. Subheads should be ranged to the left margin and typed in capitals. Any further headings should start with a capital letter followed by small letters on the left margin. No headings should be underlined. Underlining is an instruction to the printer to use italic type and should be reserved for all generic and species names, foreign words not adopted into English and for descriptions of bird or animal vocalisation. Acknowledgements As acknowledgements may imply endorsement of the data and conclusions contained in the authors work, it is advisable to obtain permission before quoting persons who have made contributions to the study. References Care should be taken to see that all references quoted in the text are given in full, in this section. To avoid confusion, references to books and periodicals should not be abbreviated and neither should they be underlined. For example: (in text) Cave & Macdonald (1955) or (Cave & Macdonald 1955) (in reference list) CAVE, F.O. & MACDONALD, J.D. 1955. Birds of the Sudan, Oliver & Boyd, London. Tables Each table must be typed, double-spaced on a separate sheet and numbered. Leave good spaces between columns and omit internal vertical and horizontal rules. Cite tables in the text in consecutive order. Illustrations Submit two sets of figures which should be well, drawn on bristol board or heavy tracing paper with a photographed or xeroxed copy. Lettering must be of first quality (e.g. lettraset) and consideration must be given to size when reduced. Typewritten or freehand lettering will not be accepted unless put onto an overlay which can be followed by the printers for typesetting. Each figure should have a label pasted on its back giving the number of the figure, caption details and accreditation. Cite each figure in the text in consecutive order. A separate caption sheet should be added to the typescript. Photographs are acceptable if they have good definition and contrast and should be produced on glossy paper. Proofs The author will be sent one set of galley proofs for correction of typographical errors. Authors may be charged for any changes from the original typescript made by him at this stage, and which result in additional printing expense. Plates or page proofs will not be sent. Published material The author is entitled to 25 copies free of charge. Further copies can be supplied at cost price, provided the order is placed when the typescript is submitted. If there is more than one author, 30 copies will be sent to the corresponding author for distribution. The society is supported entirely by membership subscriptions and may not, for financial reasons, be able to accept all submissions of value. Wherever possible an author should seek a grant, from his institution or other source, to help with costs of publication. Published by The East Africa Natural History Society, Box 44486, Nairobi, Kenya. Phototypeset- ting by Kul Graphics Ltd., P.O. Box 18095 and printed by Manhattan Displays, P.O. Box 30434, Nairobi. JOURNAL OF THE EAST AFRICA NATURAL HISTORY SOCIETY AND NATIONAL MUSEUM May 1989 Volume 7^, Number 186 REPORT ON ACTIVITY IN THE NORTHERN CRATER OF OL DOINYO LENGAI, 24TH JUNE TO 1ST JULY 1988 Sr Cb.fi Celia Nyamweru, Department of Geography, Kenyatta University, P.0 Box 43844, Nairobi, Kenya. INTRODUCTION 01 Doinyo Lengai, the only active volcano in the Gregory Rift Valley of East Africa, is also the only active carbonatite volcano in the world. Several times this century it has erupted lava and ash composed largely of sodium carbonate minerals. Early accounts of its eruptive activity were given by Hobley (1918) and by Reck and Schulze (1921); later accounts were provided by Richard (1942), Guest, (1956), Dawson (1962) and Dawson, Bowden and Clark (1968). The most recent widely known eruption was the explosive eruption that occurred between August and October 1966. Few people, whether in East Africa or elsewhere, are aware that the volcano erupted in early 1 983, and that small scale activity in the northern crater has continued since then. A report on the activity from 1 983 to 1987 has been published by the present author (Nyamweru 1988), and several short accounts have appeared in the Scientific Event Alert Network Bulletin between 1983 and 1988. The account that follows is the record of activity between 24th June and 1st July 1988, during which period the author was a member of a party that camped in the inactive (southern) comer of the north crater (area S in figures 1,2 and 3). DIARY OF ERUPTIVE ACTIVITY 24-06-88, 1630 hours: First view into the crater when party members arrived on the crater rim. No lava flowing out on crater floor. Liquid lava present in two (connected) small lava lakes in the area T4T7 (see figures 1 , 2 and 3). The larger lake lies to the east, and is about 8 to 10 metres in diameter, with overhanging sides 2 to 3 metres high. The liquid lava is black, with very low viscosity and bubbling actively, breaking in bubbles over 50cm in diameter which throw ‘spray ’ over 4 metres high. Small drops of lava fall outside the vent and form lapilli on the surrounding slopes. Plate I is of this lava lake and shows two bubbles of lava breaking on the lake surface, below the overhanging north wall of the vent. The second lava lake lies about 20 metres to the west, and is also active; a vertical fan-shaped spray of lava is rising from a narrow pinnacle in the centre of the lake, building it up. At Page 2 No. 186 1630 hours there is no other activity visible in the crater, although steam and sulphurous fumes are, emitted from a number of vents on the crater floor, its walls and its rim (see figure 1). 24-06-88, circa 1 800 hours: lava begins to overflow from the southern side of the eastern lava lake of T47T, and flows rapidly southwards; Flow 1 in figure 4. 24- 06-88, between 1900 and 2000 hours: The flow (mainly of a lava) reaches the southern wall of the crater and is stationary but still hot at 2000 hours. During this period lava probably also flows eastwards, where it reaches the base of the crater wall. After dark there is an intermittent orange- yellow flare of burning gas from the top of the western ventof T4T7, and a dull red glow of lava from the eastern lava lake and its outflow. 25- 06-88, 0700 hours: no new flows during the night: whitening of rock by chemical change beginning to occur around the margins of Flow 1 . Lava is bubbling gently in the eastern lake of T4T7 but not splashing high or flowing out. The western lake is quiet, with little or no bubbling sound. 25-06-88, circa 1200 hours: lava begins to spill over the lip of the eastern lake, rapid flow begins towards the south, very liquid black lava flowing in a stream about 1 to 1.5 metres wide, spreading out and slowing down as it flows (Flow 2 in figure 4). 25-06-88, 1345 hours: liquid lava overflowing to the east in a very narrow layer, probably less than 5 cm thick but several metres wide; this ‘sheetflow’ lasts only for a few minutes (Flow 2 in figure 4). 25-06-88, 1400 hours: active outflow has largely ceased. 25- 06-88, 1630 hours: lava moving in the lava lakes on T4T7 but not flowing out or bubbling over the edges of the lakes. 26- 06-88, 0730 hours: small outflow of lava from the southeast side of T4T7, escaping from a little crack below the rim of the lava lake. This has ceased before 1000 hours, when the lava is solid but still very hot (Row 3 in figure 4). 26-06-88, 1000 hours: lava bubbling actively in both lava lakes, but not spattering out or overflowing. 26-06-88, 1030 hours: small tricklings of fluid lava emerging from within the edge of Row 2, near the northeast crater wall. Thickness of edge of flow is 50 to 60 cm; surface is solid but still warm. 26- 06-88, 1600 to 1730 hours: no flowing lava on crater floor but lava bubbling at a relatively low level in T4T7 east. During this period, the noise of moving lava is heard from the open holes on the upper slopes of T5, from which shimmering heat and steam are rising; looking deep into the holes, it is possible to see liquid lava moving around below. At 1630 hours a small patch of pahoehoe lava is still warm on the north slope of T5 and probably formed within the last one or two days, though it was not observed flowing; this little flow appears to have originated from a crack low down on the slope of T5. 27- 06-88, 0800 hours: lava is bubbling at a very high level in the lava lake T4T7 east. 27-06-88, circa 0900 hours: a crack opens on the lower southwest side of the T4T7 ridge and very fluid lava flows out in a narrow stream; Row 4 in figure 4. 27-06-88, 0940 hours: lava has spread along a bearing of about 210 degrees from its source, towards the southwest comer of the crater. Black, smoothly flowing lava is in a stream about 50 cm wide, spreading out into tongues with a small-scale a surface; it cools to a dark chocolate brown colour. 27-06-88, 1 000 hours; flow is about 112 metres long and reaches to within 27 metres of the southern crater wall at the western end of the saddle. Its thickness is about 20 to 30 cm on the lobes at its lower end. During this period the lava remains bubbling actively at a high level in T4T7 east. 27-06-88, 1115 hours: flow is continuing, spreading out between T1 and T2 but not extending further south. No. 186 Page 3 27-06-88, 1 145 to 1200 hours: flow continuing, lava lake in T4T7 east bubbling actively. 27-06-88, 1300 hours: lava bubbling actively in T4T7 east. At the head of the new flow (F4) a homito (HI) about 50 to 70 cm high has formed within about the last hour, by bubbling of highly gaseous lava from a narrow slit near its top. The homito (see Plate' II) is being built up by wrinkled cords and tiny lobes of smooth lava, while below it lava is still flowing out rapidly in streams 30 to 60 cm wide. In places the liquid lava disappears into a tunnel and then emerges further downstream; at this stage steady movement over quite a wide front is continuing towards the lower end of the flow. Slabs of pahoehoe form on top of the flowing lava, then are carried on by the moving rock, broken and tilted; lines of tilted slabs mark the flow lines (see Plate III). 27-06-88, 1340 to 1350 hours: homito still bubbling; volume of lava escaping from beneath it is gradually decreasing, as the flow is contained within the limits of the earlier flows. 27-06-88, 1430 hours: homito no longer bubbling but flow continues. 27-06-88, 1515 hours: flow continuing in a narrow (20 to 30 cm wide) channel within the earlier channels. Flow is on surface near source, then goes into a tunnel. 27-06-88, 1615 hours: source of flow has moved a few metres downstream, bubbling out under a narrow corded skin of lava that is few hours older. There is a gentle spattering in the west side of T4T7, and more active bubbling in T4T7 east. 27-06-88, 1700 hours: flow is continuing; there is active bubbling in T4T7 east and a gentle noise of moving lava deep below T5. 27- 06-88, later that night flow ceases along F4; bubbling continues in T4T7 east. The central pinnacle in T4T7 west, which was enduing a spray of lava on the evening of 24-06-88, collapses. 28- 06-88, 0800 hours: no new flows on crater floor, no outflow of lava, but lava bubbling actively in T4T7 east. 28- 06-88, 1630 hours: active continuous bubbling from T4T7 east, little orno sound from any other vent. 29- 06-88, 0750 hours: a new large flow has formed in the night, overflowing from a groove cut in the lip around the lava lake, T4T7 east; this is Flow 5 which has gone mostly towards the south and southwest (Figure 5). 29-06-88, 0810 hours: flow very active from the vent, with overflow from two places on the rim of the lava lake; lava has also flowed round to the east and northeast and has reached the north wall of the crater below the rim cone Cl . On the northwest side of T4T7 there is spattering of very gas- rich lava from the top of one of the little cones that border the vent; new trails of dark lava are form ing down the pale grey slopes of the cone. 29-06-88, 1010 hours: lava bubbling in both the eastern and western lakes of T4T7; overflow still continuing from the south side of T4T7 east. 29-06-88, 1 100 hours: lava bubbling in both lakes; occasional spattering from the top of the western cone. 29-06-88, 1230 hours: active bubbling and overflowing from T4T7 east; line of small homitos building up along the line of flow where liquid lava is breaking out of tunnels below the older surface. 29-06-88, 1400 hours: lava bubbling and overflowing; southwestern comer of flow has reached within 16 metres of the southeast comer of T2. 29-06-88, 1430 to 1530 hours: spray of highly gaseous lava from the top of a homito, the highest of a line of about 7 homitos that lie along the axis of Row 5; the spray from the homitos has stopped by about 1 530 hours but extension of the lava flow further downstream continues until later in the day. 29-06-88, 2000 hours: surface outflow of lava has ceased before this, but the lava lake on both sides of T4T7 is bubbling actively, with adull orange glow and an occasional pale orange flare of gas. When the molten lava spatters on the rim of the vent, it glows dull red for some seconds before it cools. Page 4 No. 186 30-06-88: active bubbling and splashing of lava from both the eastern and western sides of T4T7 continues throughout the day, but there is no surface overflow during daylight hours. 30-06-88, approx. 1949 hours: a new vent opens on the east side of T4T7 and lava begins to flow rapidly to the east and southeast, quickly reaching the east wall of the crater (figure 5). 30-06-88, 2030 hours: the lava has spread to the east of T5, covering the earlier (Flow 1) lava in this area. A lava cone is soon built up at the source of the lava (T8) which at this time appears to be 2 or 3 metres high and about 10 metres in diameter. A spray of dull red lava is thrown a further 3 metres above the top of the cone. At this time the level of lava in T4T7 has fallen by as much as 10 metres and it is possible to see a large hollow below the arch that joins the two lava lakes (east and west sides of T4T7). 30-06-88, 2200 hours: lava from T8 fountaining from the cone and flowing actively. 01-17-88, 0230 hours: occasional spattering from the top of T8. 01-07-88, 0630 hours: occasional spattering from the top of T8. The cone is now about 10 to 12 metres high, with a smooth slope to the southeast, a break in slope on its northwest side. The lava below T4T7 can be heard moving around, though at some depth The last members of the party left the crater rim at about 0700 hours. LIST OF LAVA FLOWS (see figures 4 and 5) Flow 0: This was not observed actually flowing but at noon on 25-06-88 was still warm and had undergone very little whitening; it had originated from the northwest slope of T4T7 and had flowed to within a metre of the north crater wall; see figure 4. This flow probably formed on 24-06-88; according to M. Krafft it was already in existence when he visited that part of the crater late that day. The flow was about 35 metres long, width at downstream end (across 3 lobes) circa 30 metres, width near base of cone circa 14 metres. Thickness between 30 to 50 cm. It was largely composed of a lava, but it was overlain with little lobes of pahoehoe that seem to have resulted from the closing stage of the eruption. Flow 1: This flow came from the eastern lava lake of T4T7 and formed between about 1800 to 1900 hours on 24-06-88. Its dimensions were as follows: from the vent to the east-northeast crater wall, about 50 metres long; from the vent to the southern crater wall, about 180 metres long; maximum thickness 50 to 70 cm (see figure 4; this flow is also visible in the centre of Plate IV). Flow 2: This flow formed between about 1200 and 1400 hours on 25-06-88, by overflow of very liquid lava from the rim of the eastern lava lake. Its dimensions were not measured but it was a much smaller flow than Flow 1 and only covered parts of it, within a few tens of metres of the vent (see figure 4). Flow 3: This formed between about 0730 and 1000 hours on 26-06-88, it was another very small flow which escaped from a crack on the southeast side of T4T7, below the rim of the lava lake (dimensions not measured, but see figure 4). Flow 4: This flow started at about 0900 hours on 27-06-88 from a crack that opened low on the southwest slope of T4T7. It continued to flow in gradually narrower channels until after 1700 hours on the same date; between about 1200 and 1430 hours a small homito formed at its upper limit, by effusion of highly gaseous lava. The final dimensions of the flow were: length about 120 metres; width ranging from 1 metre close to its source to about 40 metres at its downstream end; thickness maximum 50 cm. At its final extent, the toe of the flow reached to within 12 metres of the south wall of the crater (see figure 4; this flow is also visible in Plate IV, where it is the narrow, dark flow on the left side of the crater floor). No. 186 Page 5 Flow 5: This flow probably started during the early hours of 29-06-88; by our first visit to the crater at 0750 hours that day, the flow had extended far to the south of its source; it was overflowing actively from a notch cut in the southern rim of the eastern lava lake of T4T7. The flow also extended to the east and northeast of its source, covering a distance of about 80 metres from the vent to reach the crater wall (see figure 5). Flow continued until about 1600 hours and at its final extent the flow was over 1 50 metres long and had reached to within 20 metres of the south crater wall at the saddle, and to within 6 metres of the north side of Tl. Flow 6: This flow began to form at about 1940 hours on 30-06-88, from a new vent low on the east side of T4T7. Movement was very rapid in the early stages and by 2033 hours the new flow had extended round the slopes of T5 and reached the eastern crater wall (see figure 5). The thickness of the flow ranged from 20-30 cm to a maximum of about 50 cm. At the source of the flow, lava fountaining built up a new cone (T8; see figure 5) which by 0630 hours on 01-07-88 was about 10 to 12 metres high.The flow of lava probably ceased soon after 2200 hours, but spattering from the top of the cone continued throughout the night and was still taking place at a reduced level the following morning. DIMENSIONS OF CRATER Pacing from west to east on a diameter just to the north of T2, between T4T7 and T5: 244 metres. Pacing from north to south on a diameter to the west of T4T7, past the eastern comer of T 1 : 222 metres (average of 2 pacings). Pacing on a diameter at bearing 1 36 degrees, from the base of D, passing just to the south of T5: 253 metres. Height of northeast wall (estimated) 40 to 50 metres. Northwest wall could be slightly higher (but not more than 60 metres). East wall is slightly lower (30 to 40 metres). The level of the crater floor around T5 and T4T7 is markedl y higher than the level to the south , at the base of the saddle, and may even be higher than the lowest point of the saddle. DESCRIPTION OF THE CRATER; THE ERUPTIVE CENTRES (see figures 1 to 3). Southern crater: this feature, to the south of the central summit, was a shallow, gently sloping depression floored by rather coarse grey ash, with patches of green vegetation on its floor and sides. Tl: (the ‘igloo’) inactive and largely unchanged since April 1988 and even December 1987; rounded pale grey slope (length about 45 metres around base of slope) facing southwards towards the saddle. Overhang facing northwards into the crater; distance across base of overhang was about 10 metres. Occasional slight emission of steam from base of its inner slopes. T2: this collapsed centre also largely unchanged since April 1988; in plan it was approximately circular, about 33 metres in diameter, with inward facing walls (best preserved on its north and south sides) and a central cone with an open hole on its south side. This hole was more than 6 metres deep and inside it was a jumble of blocks of rock; some warm fumes were rising from it. Considerable sulphur staining was visible on its inner slopes and fumes were emitted from the central holes from the base of the outer walls. There were ‘stalactites’ along the base of the inner walls. T5: little change in this group of pinnacles since April 1988. When first seen at 1630 hours on 24- 06-88 there was no steam visible, no sound of molten rock at depth, but later during the week there Page 6 No. 186 was emission of steam, and shimmering heat rising from the holes that were still open towards the top of the pinnacles. At 1630 hours on 26-06-88 a member of the party heard and saw liquid lava moving around deep below T5, and at mid-day on 30-06- 88 it was possible to hear moving lava below this feature. A small flow of pahoehoe lava was still hot on the north slope of T5 atcirca 1600 hours on 26-06-88, and may have formed within the last 24 to 48 hours, though it was not actually observed flowing. T5 is visible on the right of the crater floor in Plate IV. T4T7: this was the centre of activity throughout our stay. It changed greatly after December 1987 (when there were two quite small separate vents, T4 and T7) and according to D. Peterson also changed noticeably between his visit on 27-05-88 and the end of June 1988. By then it had developed into a continuous ridge, broadly aligned east- west, with an uneven Crestline rising to a series of grey lava pinnacles; its length (parallel to its long axis) was about 80 metres. The two lava lakes (eastern and western) were connected under an arch of older lava, which supported a lava pinnacle over 3 metres high. When the level of lava was low (as it was late on 30-06-88) it was possible to look through and under the arch from one lake to the other. The lava pinnacle in the centre of the western lava lake that was emitting a spray of lava on 24-06-88 collapsed on 27-06-88. T4T7 can be seen on the far (north) side of the crater floor in Plate IV. A3: these inactive vents, small cones and lava flows were still clearly visible on the crater wall, though rather paler in colour than they had been in late 1987; largely grey or white, with a crumbled, rather powdery surface; erosion by running water had begun to cut little gullies down parts of this slope. There were at least 5 small vents on the northern slope, as well as others to the west and (possibly) the east The cones were of basically the same form and dimensions as comparable features on the crater floor; the three larger ones stood between 3 to 6 metres above the crater wall and there were two smaller ones. The largest cone lay to the west of the cluster, and had a steep top with an open vent, though the bottom of the hole was not visible. A4: on the northern wall of the lowest part of the saddle, just behind T1 , there was an area of soft, powdery pale grey/brown material, with two or three small, partly in filled pits which appear to be heavily weathered small vents comparable to those on the north wall of the crater, opposite them. The top of the saddle for several metres was made of soft powdery pale grey/brown material rather than of the mid-grey, harder and coarser material (which includes pieces of rock up to 10 cm long) that made up the rest of the saddle and much of the upper rim and summit of the volcano. D: this inactive vent, with a small lava flow below it, was still clearly visible on the crater wall, though pale brown and heavily weathered. There was possibly an equally old, smaller vent to the south of D, close to the top of the western crater wall. Cl: when first seen at 1630 hours on 24-06-88, was not emitting steam. The cone rose over 4 metres above the crater rim and had steep slopes of soft, crumbly grey material. An estimate of its dimensions was 1 0 metres long and 8 metres high. Later during the week occasional emission of steam from this cone was observed. Cl lies at about 1 o’clock on the crater rim in Plate IV. Crater floor: the areas not covered by very recent flows were various shades of pale grey, pale brown and white; different generations of flows could be distinguished by overlapping lobes and tongues of weathered lava. Despite the intensive weathering of the lava surfaces, the pahoehoe and aa structures, as well as some lines of small homitos, were clearly distinguishable. For example the tilted slabs of pahoehoe and the small homito observed forming on Row 4 on 27-06-88 could be recognized in a weathered state on nearby older flows. Figures 1 and 2 show FF, two medium grey No. 186 page 7 flows which had longitudinal blackened fissures following their long axes. The blackening marked a line of fumaroles and was due to chemical activity of the fumes. Viewed from the summit on 28- 06-88, the oldest parts of the crater floor appeared to be on its southeast side (beyond T5; here there was a small alluvial fan of material brought down from the slopes above, overlying the weathered lava) and on the southwest side beyond Tl. M: the saddle was very low on its north side; its height there was about 4 metres. To the south, towards the inactive southern section of the crater, there was a drop of about 12 metres; thus the floor of the southern section was at a a lower elevation than that of the active northern part of the crater. DESCRIPTION OF THE CRATER; THE FUMAROLES (marked xxx on Figure 1 ; not marked on the other figures) East rim: Strong sulphurous smell, fumes and yellow staining along cracks on rim. West and northwest rim: deeply cracked in many places, both parallel to the rim and across it. Some cracks emit mainly steam and support local patches of vegetation, including mosses and lichens; others emit sulphurous fumes and the surrounding areas are largely bare of vegetation. Tl: occasional gentle emission of steam and slight smell of sulphur. T2: sulphurous smell, yellow staining, fumes from inner slopes on both sides. T5: occasional emission of steam (see notes above). Northeast wall: patches of bright yellow sulphur, local gende emission of steam. M: the saddle; the east side of the saddle, where it begins to rise steeply, is heavily altered and there is emission of sulphurous fumes. On the west side of the saddle, at a rather lower elevation, there is an open crack (less tnan 5cm wide) that crosses the saddle,with active emission of fumes and some sulphur staining (see note b. below). Southern section of the crater: the crack referred to above crosses the saddle and extends into the southern ‘inactive’ section of the crater. Occasional emission of steam was also observed from a spot near the centre of the southern section of the crater. GENERAL OBSERVATIONS AND CONCLUSIONS a. Over the period that we observed the crater, activity was concentrated in the area T4T7, where there was a lake of liquid lava over 40 metres long (including the eastern and western lakes and the connection between them). However liquid lava was also seen and heard moving around deep below T5 and there was evidence of recent small flows from the northern slope of T5, so this centre is not entirely extinct. Broadly, however, it appears that most activity is concentrated towards the northeastern and eastern sides of the crater. b. A line of weakness seems to cross the whole crater, on a bearing about 42 degrees to 222 degrees. On the northeast wall it is marked by a small grey vent near the top of the wall, with grey material below it On the crater floor it is marked by fumaroles that have formed long lines of black staining that run to the west of Tl. On the saddle it is marked by the big crack which emits sulphurous fumes. The new cone (T8) that formed the night of 30-06-88 lay very slightly to the northwest of this line. This line of weakness could well form the site of future eruptive vents. Page 8 No. 186 c. Eruptions have taken place at several locations on the crater walls, in particular on the north wall (A3 and Cl), but also apparently along the saddle (A4). Considering the fumaroles that extend across the saddle, and the high level of the floor of the northern section, it seems quite possible that lava could in the future reach the surface in the ‘inactive’ southern section of the north crater. Equally, it is quite possible that f uture eruptions could occur on the crater walls or even on its rim. d. The newly erupted lava weathers extremely rapidly, changing colour almost completely; the surface goes from dark grey or dark chocolate brown on newly cooled lava to pale grey within less than 48 hours. This makes it virtually impossible to recognize a ‘new’ lava flow unless one observes it within one or two days of its formation and thus creates serious limitations to the detailed picture of changes in the crater that may be drawn. The rate of change of the lava may depend to some extent on the weather conditions; when the atmosphere is very moist (and windy) one might expect the change to be more rapid than when it is still, hot and dry. However even the flows that erupted on the sunny days that we were in the crater (e.g. Flow 4) changed colour very rapidly. e. During the seven days during which observations were made, liquid lava was always present close below the surface; at least 4 quite substantial flows were produced and a cone over 1 0 metres high was built up. It is not possible to tell whether this represents the general level of activity over the last five years or not, since no other continuous observations over seven days have made. However the intermittent observations made over the years since 1983 indicate that liquid lava was frequently present, bubbling close below the surface if not actually flowing out. The topography of the crater floor and walls has changed strikingly over this period, evidence of the emission of quite a large volume of lava. f. In particular the northern ‘pit-crater’ that existed in the late 1960’s and during the 1970’s has become much shallower; shallower even than it was before the explosive eruption of 1966. One may ask how much shallower it will get before another major explosive eruption occurs and recreates a deep, vertical walled pit crater at the summit of Ol Doinyo Lengai? ACKNOWLEDGEMENTS I am extremely grateful to Maurice and Katia Krafft, of the Centre de Volcanologie Vulcain, B.P. 5 - 68700 Cemay, France, for making it possible for me to visit 01 Doinyo Lengai for a whole week, and also to David, Thad and Michael Peterson of P.O. Box 2534, Arusha, Tanzania, who handled all the arrangements for our visit. My thanks are also due to the Vice-Chancellor, Kenyatta University, for permission to be away from the university during our expedition. GLOSSARY AA a lava flow with a rough, jagged, clinkery surface. (A term of Hawaiian origin). See the lava in the foreground of Plate III. HORNITO a small mound built up on top of a lava flow by clots of very fluid rock escaping from openings in the roof of an underlying lava tube. Two homitos are shown to the right of the standing figure in Plate II. PAHOEHOE a lava flow with a smooth, billowy or ropy surface. (A term of Hawaiian origin). See the lava in the foreground of Plate II. BIBLIOGRAPHY No. 186 Page 9 Dawson, J.B. 1962. The geology of 01 Doinyo Lengai. Bull. Volcanol. 24: 349-387. Dawson, J.B., Bowden, P. and Clark, G.C. 1968. Activity of the carbonatite volcano Ol Doinyo Lengai. 1966. Geol. Rundschau 57: 865-879. Guest, NJ. 1956. The volcanic activity of Ol Doinyo L’Engai, 1954. Records. Geological Survey of Tanganyika 4: 56-59. Hobley, C.W. 1918. A volcanic eruption in East Africa. Jour. E. Africa and Uganda Nat. Hist. Soc. 13: 339-343. Nyamweru, C.K. 1988. Activity of 01 Doinyo Lengai volcano, Tanzania, 1983-1987. Jour, of African Earth Sciences 7: 603-610. Reck, H. and Schulze, G. 1921 . Ein Beitrag zurkenntnisdes Baues und der jungsten Veranderungen des L’Engai- Vulkans in nordlichcn Deutsch-Ostafrika. ZciLschrift fur Vulkanologie 6: 47-71. Richard, J.J. 1942.01 Doinyo Lengai - The 1940-41 eruption. Volcanological observations in East Africa. Jour. E. Africa and Uganda Nat. Hist. Soc. 16: 89-109. Scientific Event Alert Network Bulletin, 1983 (Vol. 8), 1984 (Vol. 9), 1987 (Vol. 12) and 1988 (Vol. 13). Published by the National Museum of Natural History, Washington, D.C. Editor: H.J. Beentje Figure 1 : The crater of 01 Doinyo Lengai sketched from a point on its southwest rim at 1 520 hours on 25th June 1988. xxx: fumaroles. sss:Sulphur. For other letters and numbers, see text. Page 10 No. 186 Figure 2: The crater sketched from the same point as figure 1 at 1600 hours on 28th June 1988. No. 186 Page 11 Figure 3: The crater sketched from the same point as figures 1 and 2 at 1545 hours on 29th June 1988. Page 12 No. 186 No. 186 Pa8e 13 Figure 4: Plan of the crater of 01 DoinyoLcngai showing the approximate positions of flows FO, FI, F2/F3 and F4. Approximate diameter of crater is 240 metres. >z Figure 5: Plan of the crater of 01 Doinyo Lengai showing the approximate position of flows F5 and F6. Page 14 No. 186 Plate I: Lava bubbles breaking in the eastern lava lake of T4T7. In the background is the crater wall with the cones labelled A3 in figures 1 to 3. Plate II: The source of Flow 4 at 1615 hours on 27th June 1988; two homitos and several small channels of pahoehoe lava. Behind the standing figure arc the pinnacles forming T4T7 an in the background stands the crater rim with the rim-cone Cl. No. 186 p^e 15 Plate III. Surface of Flow 4 on 30th June 1988; tilted slabs of lava already turning white. The broken wall of the old vent T2 is in the background. Plate IV: View of crater from the southeast on 28th June 1 988, showing the saddle to the left, the rim- cone slightly east of north and the eruptive vents T1 , T2, T4T7 and T5 on the crater floor. ( all photographs taken by the author) Typesetting by The Print Shop, P O Box 24576, Nairobi. Printed by AMREF, Wilson Airport Nairobi. JOURNAL OF THE EAST AFRICA NATURAL HISTORY SOCIETY AND NAJKHIto>MUSEUMS September 1987 Volume 75, No. 187 THE DIET OF THE BARN OWL, TYTO ALBA (SCOPOlT), IN NAIROBI, KENYA. Cecilia M. Gichuki, Department of Ornithology, National Museums of Kenya, P.O. Box 40658, Nairobi. ABSTRACT The food of the Barn Owl, Tyto alba (Scopoli), was studied by the analysis of undigested remains of prey found in regurgitated pellets. About 1200 pellets were obtained from a nest site at Karen i(36° 12'E; 1°2T), in Nairobi, during the period January 1977- July 1979. Analysis of the skeletal remains found in the pellets revealed that rodents made up 63. 8% of the diet, with the multimammate rat Mastomys natalensis being the principal single species. Anurans made up 18.8%, crociduras 12.7%and birds 4.6%of the tatal prey items. Bats, lizards and invertebrates formed a minor proportion. Thus, diverse and varying quanties of non-rodent prey were taken by Barn Owls at Karen. Consequently, it appears that the owl would readily switch from the preferred rodent prey to other items during difficult periods. The owl seemed to be an opportunistic feeder with a relatively small hunting range and a preference for hunting in open habitats. INTRODUCTION The diet and ecology of the Barn Owl, Tyto alba (Scopoli), has been extensively studied in southern Africa (Kolbe 1946, Davies 1959, Hanney 1962, Winterbottom 1966, Vernon 1972). In East Africa, Barn Owls are reident and widely distributed in urban and peri-urban areas (Britton 1980), but little detailed research has been carried out on their food. Laurie (1971) investigated their diet in Serengeti National Park (Tanzania) and found that rodents were the principal prey. According to Norris (1972), pellets collected in Nairobi National Park (Kenya) were found to contain a wide range of prey items, with rodents predominating. Apart from vertebrate prey, pellets in Nairobi and Serengeti were found to contain diverse but minor invertebrate prey items. Apart from the work of Norris (1972), there seems to have been no serious study on the food of Barn Owls in Kenya. The purpose of the present study was to determine the prey taken by a pair of Barn Owls at Karen (36° 12'E, 1° 21'S), 13 kilometres west of Nairobi. Since Karen is near Nairobi National Park, this work supplements the previous study reported by Norris in 1972. STUDY AREA Nairobi lies on the northern edge of the Athi-Kapiti plains and at an average elevation of 1770m above sea level. The distribution pattern of rainfall is bimodal, with the long and short rains occurring in the periods March-May and November-December, respectively. The mean annual rainfall is 1048mm, with a maximum of 1077mm and a minimum of I0I8mm. Karen is a residential area with medium human population density. Prior to the establishment of human settlement, the area was covered by a dry type of tropical semi-evergreen forest, with tall trees such as Croton megalocarpus and Shrebera alata dominating. Greenway (1943) reported that understorey shrubs and lianes are abundant in this type of forest, but the actual number of tree species is quite limited. Except for scattered patches, most of the original forest at Karen has been cleared and replaced with residential plots ranging in sizes from 5 to 7 hectares. A wide range of exotic species of trees, shrubs and herbs have been planted in the plots amongst the remaining native vegetation. Hedges of Kei Apple (Aberia caffra ), Cupressus spp. and other ornamental shrubs enclose the plots. Page 2 No. 187 In Fig. 1 , the principal habitats in which Barn Owls are likely to hunt for prey are shown relative to the position of the owls’ nest site at Karen. There is a golf course and other areas of open land where various grasses grow, mainly grazing and farmland species. There are also semi-permanent swamps and man- built ponds which serve as breeding sites for amphibians. Typhaceae (bulrushes) and Cyperacae (reeds) overgrowing areas with stagnant water were inhabited by rodents and served as suitable roost sites for many bird species, particularly ploceids. Some plots at Karen had livestock, open paddocks and stables which were attractive to rodents and grain-eating birds. Figure 1. Karen area showing environs of owl’s nest site. MATERIALS AND METHODS Barn Owl pellets were collected from an attic of an occupied building at Karen during the period January 1977-July 1979. A pair of owls had been intermittently nesting in the attic since 1976. The pellets were not obtained regularly, but were collected during the general cleaning of the house by the owner. Measurements of complete and compact pellets were taken using sliding calipers, after which the pellets were carefully broken up using a scalpel. The skeletal remains obtained from the pellets were placed in water to which some detergent had been added. After a day, the skeletal material was rinsed in clean water and later bleached in hydrogen peroxide for a period of 24-48 hours. Bleaching of the skeletal material facilitated their identification by comparison with standard museum specimens. A total of 1 200 Barn Owl pellets were obtained from the nest site at Karen. They ranged in size from 49 x 25 mm to 26 x 16 mm, and except for a few loose ones, most of them were coated with dried saliva and were therefore compact. There was a tendency for pellets to contain complete skulls of the prey, however some contained no skulls at all. Some of the pellets analysed contained a single prey item but the bulk of them contained skeletal remains of different prey species. Out of the 1200 pellets, 4470 prey fragments were recovered, however only 2262 prey items were identified. The balance, consisting of 231 cranial fragments, 1042 right and 935 left mandibles of vertebrate prey, were not identified due to excessive fragmentation and loss of specific diagnostic characters. No. 187 Page 3 Table 1 shows the numerical percentages of the identified prey items and fragments excluding the unidentified material. Rodents were the principal prey, with frogs and shrews making secondary constituents in the diet. The main prey groups will be dealt with in more detail in the following sections. Prey Group Number of Items % total items Rodentia 1499 66.3 Insectivora 269 1 1.9 Chiroptera 1 .04 Aves 97 4.3 Reptilia 1 .04 Anura 395 17.5 Arthropoda + + Total 2262 100.1 Note: + = Present but not countea Table I . The composition of the diet of the Barn Owl at Karen, based on identified items and fragments of the major prey groups. Arthropoda Some arthropod prey remains were found but were not counted. Insect remains were found in the form of limbs, elytra, mandibles and fragments of the cuticle. There was evidence of termites (Order Isoptera), beetles (Order Coleoptera) and crickets (Order Orthoptera). Some of the insects may have been taken by prey species such as amphibians, shrews and birds that were subsequently killed by the owl. Anura Frogs were detected in the pellets by their characteristic astragalus, and by the pelvic and pectoral girdles, but most anurans were identified from skulls. The identity of Xenopus spp. was confirmed from the examination of the pelvic girdles. Anurans formed 18.8% of the total prey and, as shown in Table 2, Hemisus guineensis was the most common. This species was also common at Karen. It burrows in the ground feeding on the surface at night and appearing in large numbers at the onset of the rains. Bufo gutturalis is also widespread at Karen but no specimen was recovered from the pellets. The species is known to be very toxic to mammals (Duff-Mackay pers. common.), and this may also be the case for birds. Because 40.5% of the anuran prey were unidentified (Table 2), the number of species eaten by the owl may have been more than four. Prey species Number of items % total items Hemisus guineensis 221 55.9 Pvxicephalus delalandii 5 1.3 Ptvchadena spp. 1 0.3 Xenopus borealis 8 2.0 Unidentified fragments 160 40.5 395 100 Table 2. The frequencies of anuran (frog) prey, based on the identified skills, pelvic girdles and the unidentified fragments. Aves Bird remains found in the pellets mainly consisted of skulls, skull fragments and feathers. As an overall element of the diet, birds represented 4.3% of the total prey. As shown in Table 3, the bird prey was diverse, with 1 2 genera identified. Diversity may have been higher if all the bird prey fragments were Page 4 No. 187 identified to species level. Most of the bird genera belonged to the family Ploceidae, which may have been preyed upon whilst in their communal roosts. The skeletal remains of other bird families were occasionally found. Prey species Number of items % total items Colius spp. 1 1 10.1 Pycnonotus spp. 2 1.8 Mira fra spp. 1 0.09 Lonchura spp. 6 5.5 Malaenornis spp. I 0.9 Lanius spp. 1 0.9 Zosterops spp. 1 0.0 Passer spp. 1 1 10.1 Ploceus spp. 14 12.8 Quelea spp. 13 1 1.9 Euplectes spp. 5 4.6 Vidua spp. 2 1.8 Unidentified spp. 41 37.6 Total 109 99.8 Table 3. The frequencies of avian prey, based on identified skulls and the unidentified skull fragments. Chiroptera and Reptilia Bat and lizard remains were infrequent in the pellets, with only one fragment each. Because of the loss of specific diagnostic characters, I could not identify the fragments further. fnsectivora Shrews (family Soricidae) were third in abundance among the animals taken by Barn Owls at Karen. They were an easy prey for the owl since they are slow-moving and nocturnal. The bulk of skulls and mandibles of shrews recovered from the owl’s pell-ets were apparently from young animals and were thus difficult to dientify to species level with certainty because according to Dr. Weib Spitzenberger (pers. comm.) the crania! characteristics do not provide a definite diagnosis and external features and measurements are necessary. Roden t ia Rodents were the commonest prey. Two families, Muridae and Rhizomydae, both with a total of 14 different species, were identified, but there may have been a few rodents species that could not be positively identified. The Muridae formed the greater proportion of the diet of the Barn Owl. A total of 442 skulls (including their mandibles) were recovered from the pellets. The skulls of individual rodent species were of varying sizes, indicating that small rodents of varying ages and sizes were preyed upon by the owl. Table 4 shows the percentage composition of rodent prey taken. Of the murid rodents that were positively identified from skull and dental characters, Otomys angoniensis made up 36.2% of the total number of skulls, Mastomvs nata/ensis made up 29.2% and Dendromus sp. and Mas sp. combined made up 24.7%. Other rodent species made up only 9.9% of the skulls. However, skulls alone would not show a complete picture of the total rodent prey taken. No. 187 Page 5 Prey species Number of skulls Total number of items Skulls all items % grand total Rattus rattus 1 1 0.2 0.5 Rhabdomvs / Lemniscomys spp. 6 40 1.4 2.7 Mastomys natalensis 129 855 29.2 57.0 Dendromus 1 Mus spp. 109 109 24.7 7.3 Otomys angoniensis 160 325 36.2 21.7 Oenomys hypoxanthus 1 7 0.2 0.5 Dasymys incomtus 1 7 0.2 0.5 Arvicanthis niloticus 12 79 2.7 5.3 Grammomys dolichurus 6 40 1.4 2.7 Pelomys fallax 1 7 0.2 0.5 Tatera spp. 0 7 0 0.5 Tachyoryctes splendens 16 16 3.6 1.1 Total 442 1499 100 100.3 Table 4. The composition of the rodent prey of the barn owl based on (a) skulls alone and (b) all the identified items. Apart from the skulls, there were 1057 loose mandibles, some of which were similar to those of the identified skulls. Also, there were teeth and other cranial fragments, some of which matched the skulls already identified. As shown in Table 4, combination of whole skulls and loose skulls fragments resulted in an increased number of prey items and gives a better picture of the rodent prey than skulls alone. When ail the identified skeletal fragments are considered M. natalensis made up 57.0% of the total rodent prey. This species was widespread in the study area, especially in open grassland and cultivated areas. Otomys angoniensis was the next most important rodent species, constituting 21.7% of the total rodent prey. This rodent is crepuscular and inhabits swampy areas, and its habit of feeding in open sites might have made it vulnerable to predation despite of its low density in the study area. Arvicanthis niloticus was a common species in grassland and cultivated areas at Karen, but only a few remains of the spies were recovered from the pellets, probably due to its largely diurnal nature. Dendromus spp. and Mus. spp. are small rodents and as many as four skulls were recovered from a single pellet. Dendromus spp. were identified from their characteristic grooved incisors, while those of Mus spp. are ungrooved and have a notched surface. Because their skulls were gragile, these were excessively fragmented and the incisors were in most cases lacking. Loose teeth might also have been demaged or lost during the cleaning of the material. Without the incisors, only 4 skulls of Dendromus spp. and 3 of Mus spp. were positively identified. The balance of 102 skulls could have belonged to either genus. The Root-rat Tachyoryctes splendens was common at Karen but it formed a very low percentage of the Barn Owls’ diet. Only 16 skulls were recovered from the pellets, and all of them belonged to young animals. An adult Root-rat may weigh up to 250 g (Kingdon 1974), and may therefore be too large for the owl to kill or carry. Creek Rats (Pelomys fallax) are both diurnal and nocturnal feeders, and though suitable habitats for them were available at Karen, only one specimen was recovered from the pellets. Similarly, only one specimen of the Rusty-nosed Rat Oenomys hypoxanthus was obtained from the Barn Owl pellets. The Swamp Rat (Dasymys incomtus ) occupied tall grasses growing in and adjacent to semi-permanent swamps at Karen. However, only one specimen of the species was found in the pellets. Other rodent species of minor importance in th diet included the Black Rat, Rattus rattus, probably associated with stables and farm buildings, which was uncommon in the pellets. Tatera spp. were uncommon at Karen, but their numbers probably increase in dry periods of the year. A few skeletal parts belonging to this genus were present. Both grass mice Lemniscomys spp. are diurnal feeders and were abundant in the Karen area. They were present in the pellets, but could not be identified to species level because the specimens were apparently old and their dental cusps heavily worn out. According to Kingdom (pers. comm.), such old specimens with worn out dental cusps are difficult to differentiate with certainty. However, it is likely that most of the skulls were those of the more common species Lemniscomys spp. D1CUSSION AND CONCLUSIONS The size of the Barn Owl pellets collected at Karen were comparable to those of pellets collected in different localities in southern Africa by Skead ( 1963) and Vernon (1972). The pellets collected at Karen Page 6 No. 187 tended to contain complete skulls of one or more species, but this appeared to depend on the size of the prey animals. However, as Brooke (1967) reported, not yet pellet contained whole skulls. The skeletal remains of some prey animals, especially Praomys spp., Dendromus spp. and Mm spp. were often so fragmented that only the molar teeth could be detected in the pellets. Similarly, the molar teeth and incisors of young animals were more frequently recovered than whole skulls. Analysis of the pellets indicated that a wide range of small vertebrates was preyed upon by the owl, which appears to be well adapted to feeding on rodents and shrews that are nocturnal or crepuscular. Other prey groups, notably arthropods, bats, frogs and lizards seem to be supplementary components in its diet. The relative abundance of bird and frog skeletal remains in the pellets implies that Barn Owls readily switch to them if shrews and rodents are not easily available. The multimammate Rat ( Mastomys natalensis) is abundant at Karen (Gichuki 1978), and was the dominant single prey item in the diet. The species may have been preferred and easily available in the owl’s hunting range. Further, this prevalence may be ascribed to the semi-commensal nature of both this rat and the owl with man. In South Africa (Vernon 1972) and Nigeria (Demeter 1978), the Barn Owl has been reported feeding almost exclusively on the multimammate rat in areas where the latter was abundant and easily available. The Swamp Rat Otomys angoniensis was second in abundance to the Multimammate Rat as Barn Owl prey. Both O. angoniensis and Dasvmys incomtus occupy similar habitats. The two rodent species were present in semi-permanent swamp vegetation and in reeds bordering man-made ponds and dams at Karen. However, only one specimen of D. incomtus was recovered from the pellets, compared with 160 specimens of O. ongoniensis. Misonne (1963) compared trapping records of the two rodent species with their skeletal remains recovered from Barn Owl’s pellets obtained at Kigezi (Uganda) and concluded that O. angoniensis displaces D. incomtus from its optimum habitat. My study showed a similar situation at Karen where 0.5% of D. incomtus and 21.7% of O. angoniensis were recorded from Barn Owl’s pellets. A wide range of other rodent species were taken, but in small numbers. Creek Rats (Pellomys fallax) and the Rusty-nosed Rat ( Oenomys hypoxanthus ) occupy swampland and forest edges (Delany 1975), but were only minor components in the diet of the owl at Karen. In West Africa, Heim de Balsac and Lomotte (1958) reported that Barn Owls seemed to select against Creeks Rats and Rusty-nosed Rats, though the two species were widespread within its hunting range. The two rat species and the Root Rat (Tachyorvctes splendens) may have been selected against by the pair of Barn Owls at Karen. Furthermore, no forest rodents found at Karen, such as Graphiurus spp. and Lophuromys spp., were recovered from the owl’s pellets. It would thus appear that the owl avoided hunting for prey in well wooded areas, preferring less enclosed habitats. Hunting by the Barn Owl is a function of its food requirements and the difficulties involved in procuring the food. Thus the owl compensates for poor conditions by hunting for a longer period and in a larger range (Smith and Marti 1976). In an Israeli desert, Bodenheimer (1949) estimated Barn Owl hunting range to be 2-25km2. At Karen, the habitat conditions favour establishment of a rich small mammal fauna, so that there was probably an abundant food supply in the area. Therefore, the hunting range of the owls was likely to be small. From the distribution of the identified prey species, both recovered from the pellets and trapped in the Karen area, it would appear that the hunting range of the pair of Barn Owls was probably not greater than 5km2. The hunting range of the studied pair of Barn owls may have overlapped with the range of at least two other pairs known to have roosted and nested in the Karen area (Eric Risley, pers. comm.). Competition from the other two pairs and the changes in the abundance and availability of suitable prey may have caused the pair studied to alter their hunting range from time to time. ACKNOWLEDGEMENTS I wish to acknowledge with thanks the assistance accorded to me by the following officers of the National Museums of Kenya. Dr. G. R. Cunningham-van Someren allowed me to use his Barn Owl pellet material and carefully read through the manuscript, and Mr. A.D. Mackay identified the anurans, reptiles and rodents. Mr Issa Aggundey and Mrs. Nina Mudida provided a reference collection of bone material that was used in identifying my specimens. I am also grateful to Prof. D.E. Pomeroy and Dr. G.H. Martin for reading and criticising the manuscript. Mr. Nathan Gichuki made his rodent data from Ngong and Karen available to me. Drs C.J. Vernon, and P. Ballman, and Professors J. M. Winterbottom No. 187 Page 7 and Lester Short found time during the Fifth Pan-African Ornithological Congress in Malawi to discuss the manuscript with me. Finally, 1 greatly appreciate the assistance and co-operation given to me by Mr. Eric Risley, the Lord family at Karen, Mrs. Louise Fordyce and Sebastian Ballard, during the writing up of the paper. REFERENCES BRITTON, P.L. (editor) 1980. Birds of East Africa. Nairobi. East Africa Natural Flistory Society. Bodenheimer. F.S. 1949. Problems of vole populations in the Middle East. Reports on the population dynamics of the Levant Vole Microtus guenteri. Research Council of Israel, Jerusalem. Brooke, R.K. 1967. On the food of the wood owl. Ostrich 38( l):55. Davies. D.H.S. 1959. The Barn Owl’s contribution to ecology and palaeoecology. Ostrich Suppl. 3:144-153. Delany. M.J. 1975. The Rodents of Uganda. London: British Museum (Natural History). Demeter. A. 1978. Food of a Barn Owl Tvto alba in Nigeria. Nigerian Ornith. Soc. Bull. 14:45. Foster. B. and A. Duff-Mackay. 1966. Keys to the genera of Insectivora, Chiroptera and Rodentia of East Africa. J.E.Afr. Nat. Hist. Soc. Nat. Mus. 25:3(112): 189-209. Gichuki. N.N. 1978. A study of small mammal populations in Ngong Forest. Unpublished under- graduate final year Report, University of Nairobi. Greenway, P.J. 1943. Second draft Report of Vegetation Committee. Pasture Research Conference. (Typescript 52). Nairobi East Africa Herbarium. Printed in K.R. CIA, Vol. 9 part 1, 1973. Hanney. P. 1962. Observations on the food of the Barn Owl Tvto alba in Southern Nyasaland. Ann. Mag. Nat. Hist. Ser. 13(6): 305-313. Heim de Balsac, h. and m. lamotte 1958. La reserve naturelle integrate du Mont Nimba, 15 Mammiferes. Mem. Inst. Fr. Afr. Noire 53. Kingdon. J. 1974. East African Mammals. An Atlas of Evolution in Africa. Vol. MB. London, Academic Press. Kolbe. F.F. 1946. The case for the Barn Owl. E. Afr. Wildl. J. 1:69-73. Laurie. W.A. 1971. The food of the Barn Owl in the Serengeti National Park, Tanzania. J.E. Afr. Nat. Hist. Soc. Nat. Mus. 28(125): 1-4. Ml SONNE. X. 1963. Les rongeurs du Ruwenzori et des regions voisines. Explorat. Parc. Natn. Alberta. 14. Inst. Parcs. Natn. Congo Beige, Brussels. Norris. C.E. 1972. Barn Owl pellets. EANHS Bull. 1972:129. Skead. C.J. 1963. Contents of pellets of the Barn Owl Tvto alba at a rural roost. Ostrich 34(3): 171-172. Smith. D.G. and C.D. Marti. 1976. Distributional status and ecology of Barn Owls in Utah. Raptor Res. 10:33-44. Vernon. C.J. 1972. An analysis of owl pellets collected in Southern Africa. Ostrich 43:109-123. Winterbottom, J.M. 1966. Notes on the food of Tvto alba in the Etosha pan. Ostrich 37(2) : 1 39. Received: July 1983 Editors: J.J. Hebrard, H.J. Beentje Page 8 No. 187 NOTICE TO CONTRIBUTORS Papers on natural history subjects will be considered on the understanding that these are being offered exclusively to the Society and have not been submitted, at the same time, to any other journal. They should be addressed to the Editors, Journal of the East Africa Natural History Society and National Museum, P.O. Box 44486, Nairobi, and be accompanied by a covering letter from the author responsible for correspondence and decisions regarding the typescript. Two copies of the contribution should be forwarded to the Society; a further copy should be kept by the author for reference. The typescript should be on A4 size paper. Use one side only, double-spaced with margins of at least 5 cm at the top and left-hand side of the pages to allow for editorial instruction to printers. The pages must be numbered and all measurements are to be in metric units. The title page should contain the title of the article followed by the name of the author(s) and institution(s) where the research was carried out. A present address, if applicable, should be included as a footnote. Abstract to be on the second page and contain not more than 150 words. State the purpose of the study, procedures, main findings and conclusions. This should be intelligible in itself without reference to the paper. Nomenclature In cases where there are well known and accepted English names, e.g. for birds and larger mammals, these should be used throughout the text, accompanied by the scientific name on first mention. Authorities should be given beside the generic and specific names, when quoted for the first time, for all groups, with the exception of ornithological papers not dealing with taxonomic discussion. Text This is usually, but not necessarily, divided into parts with the headings Introduction, Methods, Results and Discussion, these being broken up with sub-headings where necessary. Main heads should be centred on the page and typed in capital letters. Subheads should be ranged to the left margin and typed in capitals. Any further headings should start with a capital letter followed by small letters on the left margin. No headings should be underlined. Underlining is an instruction to the printer to use italic type and should be reserved for all generic and species names, foreign words not adopted into English and for descriptions of bird or animal vocalisation. Acknowledgements As acknowledgements may imply endorsement of the data and conclusions contained in the authors work, it is advisable to obtain permission before quoting persons who have made contributions to the study. References Care should be taken to see that all references quoted in the text are given in full, in this section. To avoid confusion, references to books and periodicals should not be abbreviated and neither should they be underlined. For example: (in text) Cave & Macdonald (1955) or (Cave & Macdonald 1955) (in reference list) CAVE, F.O. & MACDONALD, J.D. 1955. Birds of the Sudan, Oliver & Boyd, London. Tables Each table must be typed, double-spaced on a separate sheet and numbered. Leave good spaces between columns and omit internal vertical and horizontal rules. Cite tables in the text in consecutive order. Illustrations Submit two sets of figures which should be well, drawn on bristol board or heavy tracing paper with a photographed or xeroxed copy. Lettering must be of first quality (e.g. lettraset) and consideration must be given to size when reduced. Typewritten or freehand lettering will not be accepted unless put onto an overlay which can be followed by the printers for typesetting. Each figure should have a label pasted on its back giving the number of the figure, caption details and accreditation. Cite each figure in the text in consecutive order. A separate caption sheet should be added to the typescript. Photographs are acceptable if they have good definition and contrast and should be produced on glossy paper. Proofs The author will be sent one set of galley proofs for correction of typographical errors. Authors may be charged for any changes from the original typescript made by him at thisstage, and which result in additional printing expense. Plates or page proofs will not be sent. Published material The author is entitled to 25 copies free of charge. Further copies can be supplied at cost price, provided the order is placed when the typescript is submitted. If there is more than one author, 30 copies will be sent to the corresponding author for distribution. The society is supported entirely by membership subscriptions and may not, for financial reasons, be able to accept all submissions of value. Wherever possible an author should seek a grant, from his institution or other source, to help with costs of publication. Published by The East Africa Natural History Society, Box 44486, Nairobi, Kenya. Phototypeset- ting by Kul Graphics Ltd., P.O. Box 18095 and printed by Manhattan Displays, P.O. Box 30434, Nairobi. JOURNAL THE DISTRIBUTION AND ECONOMIC IMPORTANCE OF THE MANGROVE FORESTS OF KENYA J.O. Kokwaro Department of Botany, University of Nairobi ABSTRACT The mangroves form a group of higher plants which form a unique ecosystem, in that they grow in that part of land which is neither in demand for human settlement nor for agricultural use. They are also unique in their adaptation to both soil and water conditions. They are useful as a source of timber, for building poles, fuel, dyes, tannins, and are also known to provide both shelter and food for part of the marine fauna. Their value to the country, therefore, calls for proper utilization and conservation of all the available mangrove forests along the coast. The demand for forest products, including those from the mangroves in Kenya, is greater than the available resources from the forests, and unless proper and prompt planning for their protection is implemented our mangroves will soon be among the endangered ecosystems in the country. IN TRODUCTION The Kenyan coast runs from the Somalian border at 1°40' S southwest to 4°40' S at the border with Tanzania (Map 1). It lies in that hot tropical region where the weather is primarily controlled by the great monsoon air currents of the Indian Ocean. It is the southeast monsoon which brings the long rains from April to June, when most of the annual precipitation is received. The short rains begin around October or November, and both long and short rains occur mainly in the morning or mid-day hours. The mangroves form the type of vegetation collectively referred to as mangrove swamps, mangrove forests or mangrove thickets. They are a common feature of tropical shores and are usually formed around the mouths of rivers and creeks where there is a gradual slope of the land towards the sea as well as a large tidal range resulting in a broad inter-tidal zone, consisting of a mixture of sand and silt. These communities are generally confined to tidal estuaries and lagoons, as they are protected from the force of the open sea in these localities. The supporting soil is primarily heavy mud which is mostly saline, though frequently influenced by freshwater streams and rivers. Though the mangroves can withstand seawater with high salinity, their communities are usually most prolific in areas not far from the mouths of coastal streams. ADAPTATION TO A UNIQUE ENVIRONMEN f Lagoons behind tropical shores as well as creeks influenced by freshwater streams contain brackish water, and their shores are mostly muddy. Such shores may support a growth of mangroves, which are adapted to this habitat — an environment which is unique due to the following factors: (1) Fluctuating salt content. The mangrove swamp is essentially tidal, receiving water of low salinity from the river and water of high salinity from the sea at different times each day. The plants and animals in the mangrove system will thus have to be adapted to withstand such changes in salinity. (2) Aeration. The soil in the mangrove swamp is saturated by water and hence almost completely lacks the oxygen required by the plants for root respiration. (3) Mobility of the soil. The soil level is unstable, as the streams bring down alluvial soil which is deposited, only to be washed away again by sea currents. This makes it difficult for seedlings to establish themselves. Page 2 No. 188 To be able to withstand the fluctuations in salt content most plants of the mangrove swamps are halophytes, i.e. plants with high osmotic pressure in their cell-solutions. In order to obtain oxygen for respiration some of the plants have pneumatophores (breathing roots or aerial roots); and by means of stilt roots they are able to withstand the mobility of the soil. MAP I Distribution of Mangrove forests along the Kenya Coast line DISTRIBUTION I he mangrove swamps along the Kenyan coast cover approximately 52.980 hectares (Table 1 ). The largest stands occur in the Lamu area including the islands of Manda and Patta, and also along the Vanga-Funzi system near the Kenya-Tanzania border. The former area receives its low-salinity water from the Doduri and Tana rivers, while the latter receives this from the Ramisi, Mwenaand Umba rivers. No. 188 Page 3 Other areas along the coast with less extensive mangrove stands are Mtwapa, Kilifi and Mida creeks to the north of Mombasa; the Mombasa-Port Reitz area; Gazi to the south of Mombasa; and the Ngomeni-Fundi Isa area to the north of Malindi (Map 1). The border between the mangroves and the non-halophytic vegetation is found to be well-defined everywhere along the coast, except where fresh- water from the rivers comes into the ocean. Where natural vegetation is disturbed, an impenetrable, evergreen, usually thorny bush dominated by Baobab trees (Adansonia digitata)'\s found. However, since most land adjacent to the mangroves is cultivated, plantations, especially of coconut, are prevalent. TABLE 1: Distribution of the Mangroves (from Doute et al. 1981) LOCALITY DISTRICT AREA IN HECTARES Kiunga Lamu 3,025 Lamu Lamu 30,475 Kipini (Witu) Tana River 1,595 Mto Tana (Witu) Tana River 250 Mto Kilifi (Formosa Bay) Kilifi (1,515), Tana River (820) 2,335 Mto Fundisa (Formosa Bay) Kilifi 330 Ngomeni Kilifi 1,815 Mida Creek (Malindi) Kilifi 1,600 Takaungu (Malindi) Kilifi 30 Kilifi Creek Kilifi 360 Mtwapa Creek Kilifi (410), Mombasa (115) 525 Tudor Creek Mombasa 1,465 Port Reitz Mombasa (380), Kwale (1195) 1,575 Maftaha Bay Kwale 615 Ras Mwachema Kwale 5 Funzi Bay Kwale 2,715 Vanga Kwale 4,265 Distribution by districts: Lamu District Tana River District Kilifi District Mombasa District Kwale District 33,500 2,665 6,060 1,960 8,795 Total 52,980 ECONOMIC IMPORTANCE Small trees and shrubs are the most important plants of the mangrove swamps. There are five important genera of widely distributed woody plants in the mangrove vegetation of the Kenyan coast, each genus containing one species. Bruguiera gymnorrhiza, Ceriops tagal and Rhizophora mucronata belong to the family Rhizophoraceae, Sonneratia alba to the Sonneratiaceae and A vicennia marina to the Verbenaceae. They are all viviparous except Sonneratia, and often have stilt roots and pneumatophores (breathing roots). Avicennia and Sonneratia are the first colonizers of the swamps. Once established, mud can accumulate among their breathing roots, producing conditions favourable for Ceriops and Rhizophora. Rhizophora is the commonest and most important constituent of the mangrove swamps. It usually occupies the most favourable sites between Sonneratia and Avicennia on the creek edge, and Ceriops on the landward side. Bruguiera is normally found scattered in Rhizophora stands. For a long time, the Coastal Kenyans have exploited the rich natural products of the mangroves to supplement their marginal producing agricultural land. They use the mangrove plants in many ways, and these are discussed below and listed in Table 2. Page 4 No. 188 Poles The most important product of the mangroves is poles for export and for local house-building. Annually, half a million poles were exported from Kenya during the 1930s. About 300.000 headloads of withies were obtained from the mangrove forest annually during the same period (Rawlins 1957). The majority of poles and withies are from Rhizophora mucronata. Vegetable Tannins These are generally considered as minor forest products in Kenya. However, during the mid-1950s, the mangroves were yielding tan bark exported at the rate of 3,500 tons per annum (Rawlins 1957). In many ways the mangrove bark is a unique tanning material, the supply of which is virtually inexhaustible. There is no need for planting or weeding, as the mangrove trees propagate themselves and no other trees are able to establish themselves in this special environment. The common tannin-yielding genera of the mangroves are A vicennia, Bruguiera, Ceriops, Heritiera, Rhizophora , Sonneratia and Xylocarpus. Of all these, Rhizophora mucronata is the easiest to strip and prepare for both local use and for export, and the tannin content of its bark (12-50 %) is among the highest. There are several reasons why tannin from mangroves has not come to the forefront in Kenya as a tanning material. Firstly, the mangrove forests of Kenya are not very extensive, compared to those of Tanzania. The second reason is the difficulty in collecting the bark from the swamps. Finally, the differences in tannin content between the various genera precludes indiscriminate felling of trees if a product of consistent quality is to be obtained. A possible additional disadvantage is that of the unwillingness of leather-buyers to use dark-coloured sole leather — and mangrove tannin is dark red. Research aimed at removing or bleaching the coloured components of the mangrove tannins will definitely increase their use and consequentially the commercial value of Kenyan mangroves. Fuel Coastal Kenyans have for a long time used different mangrove species as a source of fuel. We find that the kind of raw material used in traditional fuel depends more on accessibility than on the quality of the plants used. Those who live close to the mangrove forests therefore have the tendency to use the wood, frequently as firewood and occasionally for charcoal production. Charcoal produced along the coast is generally exported to the Middle East and was an extremely lucrative trade until the late 1970s when the Kenya Government had to intervene to prevent the complete destruction of forests, including the mangroves. It was estimated that the charcoal export from Kenya to Kuwait alone was at a rate of 35.000 tons a year (East African Standard 08.03. 1971). The bulk of charcoal for export is still produced from the coastal forests. Apart from Bruguiera gymnorrhiza and Rhizophora mucronata whose poles (boriti) and bark (for tanning) are of high value commercially, the rest of the mangrove species are utilized in one way or another as a source of fuel. Both firewood and charcoal are obtained from A vicennia, Ceriops, Heritiera, Lumnitzera, Sonneratia and Xylocarpus species, and most of these yield high quality fuel since they have hard and compact wood. For the coastal people, charcoal is a major source of income whenever they can produce and export it, as the Middle Eastern demand for Kenyan charcoal is ever present; charcoal, even when imported, is cheaper than oil as a source of energy, and certain industrial work is better done by using charcoal as fuel than by electric or oil energy. Fortunately, the rate of regeneration of the mangroves is high when they have been harvested, since most of the species produce fruits and seeds which easily establish themselves. In Malaya, mangrove seedlings are collected as they drop, and planted in rows after the trees have been harvested and the swamps cleared. Proper planting has the advantages of ensuring that the seedlings are not washed away by the sea currents; of making harvesting easier, since the trees grow in lines; and of making increases in production possible as required. It is clear that the demand for charcoal will continue to rise in Kenya. It is therefore the responsibility of the Forestry Department to encourage plantation of mangroves for the production of tannin, building poles, charcoal and firewood for both local consumption and export. It should be noted that most mangroves do not coppice when felled; this in itself will create some employment in fields like tanning, charcoal production and timber, primarily for export. The production of charcoal can be carried out by using modern and more economical methods such as the CUSAB charcoal kiln. Summary of Economic Use of Mangroves No. 188 Page 5 < z U 5 u uj ^ 2 UJ UJ p u. ■JS | O 3 ^ O o ■e % S2 511 (D S3 5 £ -g -a o o ii <-> ■a 2 O o u O ? 2 c o c y — r S ^ >>p .2 c o 3 — , C/3 U i_ UJ > ^ > — O yi P < 3 5 .■£ 3 jSs o O C _e e 3 -a ^ ^ u, 3 3 "O z Jr < a VO c - '3 3 C >> 3 Q H 8? £8 .5 • a 3 e >v S3 Q H ^.S (N C tj- a Tj* a v 3 Q H c c 3 H c c 3 H 0? UJ aa C/3 - 1> S ’S 3 °- 00 a> C C ,, o •- J2 •r o. o -S OQ S =3 a, cd e >,JH .•£ o 13 ^ §• g> 5 I O 3 2- _o c/3 UJ < Z P < u o ■o c 3' iH '3 u2 2 S S S 3 Cd 03 *0 c 03 ^ -c ^ c« C o 3 3 -^0 3 ‘35 P U/ 2 2 2 cd cd cd cd -o -a c c cd cd 2 cd £ >, C cd cd ’EL cd s;t 3 o 2 -* J3 o 33 -x 2 2 cd 00 c o cd S o <0 k r~ ?o VOLUME 76, No. 190 ALPINE PLANT COMMUNITIES OF MT. ELGON. - AN ALTITUTDINAL TRANSECT ALONG THE KOITOBOSS ROUTE E. Beck1, H. Rehder2, E.D. Schulze3 and J.O. Kokwaro4 ABSTRACT The afroalpine vegetation of Mt. Elgon was studied along an altitudinal transect ranging from the montane forest up to the summit of koitoboss (13,880 ft.). Since, due to recurrent fires, the boundary between the lower Ericaceous and the alpine belt is an interlocking one, both vegetation girdles were encompassed in the studies. The Ericaceous belt can be divided into a woodland and bush storey. The former, with respect to the species encountered represents a transition zone from the montane forest to the microphyllous bush vegetation. The latter was designated as Ericaceous bush and exhibits various indications of former fires, thus rather reflecting different stages of regeneration than a climax vegetation. Within the alpine belt tussock grassland and two types of Carex bog form a lower storey. The vegetation of the upper zone is characterized by a field layer of predominantly Alchemilla elgonensis and thus could be address as Alchemilla scrub. However, according to the accompanying species several types of this crub were distinguished: a Helichrysum scrub which was found on shallow soil covering volcanic rock slabs; an Euryops bush, bordering the so-called Dendrosenecio woodland vegetation. The latter was divided into a Dendrosenecio e/goueusw-Community on moist soil and a Dendrosenecio barbatipes-C ommunity covering the well-drained steeper slopes of the caldera rim. INTRODUCTION Mt. Elgon ( 14, 1 78 ft.) is one of the seven East African high mountains exhibiting alpine vegetation. As described in detail by Hedberg ( 195 1) the vegetation characteristic of the alpine belt is usually found above the so-called Ericaceous belt, i.e. above 1 1,000 to 1 1,500 ft. However, probably due to recurrent burning, a broad altitudinal zone exists at Mt. Elgon, in which alpine grassland or scrub vegetation has penetrated into the Ericaceous belt thus narrowing that girdle considerably and even bordering the montane forest belt in a small area south of the caldera (Hamilton, 1981). In the present communication this zone of mutual interference was included in the alpine belt. With respect to the upper border of this belt, Mt. Elgon is not high enough to exhibit a nival zone; therefore the summits, too, are covered by alpine vegetation. Since this paper describes the plant communities along the Koitoboss route, i.e. along the access from the south- east, it should be mentioned that the East of Mt. Elgon gains considerably less precipitation than the South 1 Lehrstuhl fur Pflanzenphysiologie, Universitat Bayreuth D-8580 Bayreuth, Postfach 101251 2 Institute fur Botanik und Mikrobiologie, TU Munchen, Arcisstra/3e 21, D-800 Munchen 3 Lehrstuhl fur Pflanzenoekologie, Universitat Bayreuth, Postfach 101251 4 Botany Dept. University of Nairobi, P.O. Box 30197 Nairobi, Kenya. Page 2 No. 190 and especially West (Hamilton 1981). Thus the southeastern slopes may be intermediate with respect to humidity and rainfall. Although no exact data were available, there is no doubt that the maximum of rainfall occurs in the montane forest belt (Hamilton 1981), and that the alpine region, although more often covered by clouds than the montane girdle, gains considerably less precipitation. The soils along the Koitoboss route are mostly of the alpine Brown Soil or Brown Soil-Gley type which have been described for Mt. Kenya (Beck et al. 1 981), but mires and boggy areas outside and inside the caldera were also recorded. The soil of these swamps consists of a material classified as “radicellous peat”, i.e. an organic soil which had developed predominantly from roots and rhizomes of higher plants as well as from mosses, but without any contribution of Sphagnum species. Hedberg (1964) also stated that he could not find one Sphagnum plant at Mt. Elgon. In particular close to the caldera rim, and also at the summit of Koitoboss, large slabs of volcanic rock crop up which are covered by an only very shallow humus layer. TOPOGRAPHY Easy access to the alpine region of Mt. Elgon is provided by a driveway leading from the Kitale Gate of Mt. Elgon National Park to an altitude of about 10,800 ft. It ends about 260 ft. above the sharply notched Kimothon valley. From that point a faint foot tract which enters the caldera rim between Koitoboss and Lower Elgon was used to establish the altitudinal transect. From the sadle the tract turns eastward to climb Koitoboss summit via its W face. Some bogs of the caldera, in particular the so called Koitoboss swamp were included in the transect. For mapping the route and localizing the plots the Sheet NA-36-12 (Kapenguria), Series Y 503, printed by Survey of Kenya was used and a sketch of the relevant area was drawn magnifying the pertinent area (Fig. 1). No. 190 Page 3 CLASSIFICATION OF THE VEGETATION Conspicuous, repeatedly occuring combinations of plant species were listed and species quantification was performed according to the method of Braun-Blanquet (Mueller-Dombois and Ellenberg, 1974). Unknown plant species were collected and identified using the “Flora of Tropical East Africa” (Hubbard and Milne-Redhead), “Upland Kenya Wild Flowers” (Agnew 1974), “Kenyan Trees and Shrubs” (Dale and Greenway 1961) and “Afroalpine Vascular Plants” (Hedberg 1957). To the Dendrosenecios the nomenclature proposed by Nordenstam (1978) was applied. Identification of difficult speceies was confirmed by comparison with material of the East African National Herbarium at Nairobi. Unfortunately, the mosses have not yet been identified. For abstraction of vegetation units, the proposals of Hedberg (1964) were used as guiding principle. THE VEGETATION FORMATIONS Ericaceous belt Having passed several types of dense montane forest along the driveway to the Koitoboss area, an open type of Ericaceous woodland is entered at an altitude of a bout 10,200 ft. This woodland is characterized by tree-like representatives of the Ericaceae as are the broad-leaf species Agauria salicifolia and the microphyllous species Philippia keniensis ( and excelsa) and Erica arborea. It still comprises montane forest species, too. Typical respresentatives of the latter are the trees of Olea kilimandscharica, Juniperus procera, Rapanea rhododendroides, Hagenia abyssinica and Hypericum keniense as well as some herbs, e.g. Viola eminii. This lower storey of the Ericaceous belt thus may be regarded as natural transition from the forest to the Ericaceous bush, the latter representing the upper storey of this belt. Two analyses of such a woodland vegetation are provided by Table 1 (plots 1 and 2). Clearing on rocky outcrops exhibit an open vegetation type which, in analogy to A7otzli’s classification on Mt. Kilimanjaro (1958) might be described as Exotheca abyssinica - Agrostis spp. - Community (plot Nr. 3). Apart from the replacement of Agrostis volkensii at Kilimanjaro by A. gracilifolia and A. kilimandscharica at Mt. Elgon, this community differs from that described by Klotzli in that mosses are dominating on the rocky soil instead of grasses. In contrast, e.g. to the western part of Mt. Kilimanjaro (Beck et al. 1983) the Ericaceous bush on Mt. Elgon is an open vegetation, appropriately characterized by Cotton (1932) as “subalpine grasslands with scattered Ericaceous bushes and suffruticose plants of temperate genera”. There is no distinct border between the lower storey of the Ericaceous belt and the upper one. In the latter the trees are more and more replaced by bushes of microphyllous species, such as Philippia keniensis, Anthospermum usambarense and Stoebe kilimandscharica. In the visited area between Kimothon and Kassowei (Fig. 1) the Ericaceous bush more or less everywhere exhibits signs of burning, such as charred Philippia stems, from the basis of which new shoots had developed. Another direct indication of former fire is the absence of the cylinders of dry lea- ves around the stems of Dendrosenecio elgonensis. Thus the plant communities found in these areas (plots 28 to 30) on the one hand by the absence of the tree-like differential species listed in Table 1 can already be distinguished from those of the Ericaceous woodland, but on the other hand rather appear to reflect regeneration stages of the bush than the climax plant community. According to Hedberg ( 1 964) luxuriant growth of Carduus keniensis, as it is indicated by plots 29 and 30 can be taken as to show recolonization of burnt-over areas. According to our own observations on Mt. Kilimanjaro (Beck et al. 1983) the same is true for the predominance of Artemisia afra (plots 28 to 30). A Ipine belt The vegetation formations and plant communities typical of the alpine belt are shown in Table 2. a) Tussock grassland In the Koitoboss area the upper boundary of the Ericaceous belt varies between 10,800 and 1 1 ,800 ft (cf. Hamilton 1981). Usually it extends to the higher altitudes merely on steep slopes and rocky crests while on the valley floors it is replaced by a tussock vegetation. According to Hedberg (1951) tussock grassland is the most important vegetation type of the alpine belt of Mt. Elgon, covering more than half of the area. This author further emphasized ( 1 964) that fires are at least partly responsible for the large extension of the tussocks inside and outside the crater and that it is in particular the frequency of burning that favours the spread of grassland at the cost of the scrub vegetation. This is undoubtedly true for grassland covering well-drained soils, especially in the lower zone of the alpine belt. In plots 24 and,25 such plant communities Page 4 No. 190 have been analyzed. The herb layer is dominated by the tall tussocks of Festuca pilgeri, the gaps' in between being occupied by other grass species as Pentaschistis borussica, Agrostis kilimandscharica and Koeleria capensis. The inconspicuous occurence of Everlastings in such tussock grasslands may indicate the above mentioned fire-triggered competition of grassland with scrub communities. A dense moss layer was found covering patches of bare soil in between the grass tussocks. b) Carex bogs Another type of tussock vegetation is represented by the so-called Carex bogs (Hedberg 1964) which cover water-soaked peaty soils on flat depressions without any or with reduced drainage. Naturally those bogs occur more frequently at lower elevations than in the upper alpine regions. At Mt. Elgon two types of Carex bogs may be differentiated by the presence (plots 23 and 26) or absence (plots 1 3 and 1 4) of tree- like groundsels. Carex bogs (type 2 in Table 2) characterized by the latter (Fig. 2) by phenotype very much resemble those described from the southeastern slopes of Mt. Kilimanjaro (cf. Fig. 86 in Hedberg, 1964) which have been described by Klotzli (1958) as Senecio cottonii-Carex monstachya Community. At Mt. Elgon the dominant sedge is Carex runssoroensis, and Dendrosenecio elgonensis was found instead of D. cottonii (plot 22 and Fig. 2). In the Ericaceous bush storey Carex ninagongensis and two other Cyperaceae, presumably Scirpus species (which could not be identified due to the lack of reproductive organs) appear to replace Carex runssoroensis (plot 26) at otherwise comparable bogs. The other type of Carex bog (type 1 in T able 2) which on the transect was found exclusively inside the caldera (plots 1 3 and 1 4) at altitudes around 13,000 ft. appears to be characteristic of more or less permanent flooded areas. It is dominated by the tussocks of Carex runssoroensis (Fig. 3) and like the first bog type contains considerable numbers of the attractive giant rosette species Lobelia elgonensis. Another similarity results from the occurrence of Alchemilla johnstonii in between the hummocks. However, the occurrence of hydrophilic herbs as Ranunculus volkensii, Ericaulon volkensii, Crassula granvikii or Swertia crassiuscula and the above mentioned absence of giant groundsels render easy differentiation of the two types of Carex bogs which physiognomically may be addressed as mires (plots 23 and 26) and swamps (plots 13 and 14), respectively. It should be noted that the absence of Alchemilla elgonensis from the typical Carex bogs and from the tussock grassland clearly distinguishes these vegetation formations from all other plant communities recorded in the alpine zone of Mt. Elgon (Table 2). With a few cuts the same holds true for the absence of Helichrysum ambliphyllum, Lobelia telekii, Valeriana kilimandscharica and Festuca abyssinica. Thus, next to the tussock grassland, a scrub vegetation in which in particular Alchemilla elgonensis is of great importance characterizes the alpine belt of Mt. Elgon. c) Alchemilla scrub A pure Alchemilla scrub, established by Alchemilla elgonesis, A. johnstonii and Valeriana kilimandscharica, as has been described by Hedberg (1964) from an area near Maji ya Moto, could not be found along the Koitoboss transect. This difference may be rather due to the sizes of plots employed by Hedberg (1964), 1 x lm)and in the present work ( 10 x 10m), than to clearcut changes of the vegetation. Inspite of a more or less closed and monotonous layer of Alchemilla scrubs various plant communities belonging to the vegetation formation Alchemilla* scrub could be differentiated by the presence of other conspicuous species. Where Everlastings other than H. amblyphvllum were as frequent as the Alchemilla species (plots 15 and 20) the vegetation was designated as Helichrysum scrub which predominantly covers rocky ground. It should be noted that the Helichrysum scrub described in this work differs from that mentioned by Hedberg (1964) in that the ubiquitous H. amblyphyllum is of minor importance and by the absence of Alchemilla johnstonii, Dendrosenecio elgonensis, Peucedanum kerstenii and Carduus keniensis. However, even the pure Helichrysum scrub described here for Mt. Elgon is almost identical (allowing for vicarious species such as Alchemilla argyrophylla and elgonensis) with a mixed vegetation of Helichrysum scrub, Alchemilla scrub and Ericaceous bush (cf. Table 2 in Beck et al. 1983) on Mt. Kilimanjaro. The Helichrysum scrub shown by Hedberg for Mt. Elgon (cf. Fig. 89 in Hedberg 1964) will be described in the present work as Dendrosenecio elgonensis-Com mu n i t y . Another plant community abstracted from the Alchemilla scrub was designated as Euryops bush. It is characterized by the occurence of the conspicuous bushes of Euryops elgonensis together with both Alchemilla species, as well as by the absence of a moss layer plots 17, 18, 21) and was usually observed bordering the Dendrosenecio c/gonemi.y-Community versus mires or tussock grassland. The Euryops bush of Mt. Elgon is completely different from the Festuca abyssinica-Euryops dacrvdiodes Community (Beck et al. 1983) reported for Mt. Kilimanjaro, as grasses are of minor importance or even lacking. The occurrence of Peucedanum kerstenii and Carduus keniensis in the Euryops bush as well as the localization of this community close to the grassland suggest that its origin may be linked to occasional fires. No. 190 Page 5 d) Dendrosenecio woodlands The most conspicuous vegetation formations of the alpine belt of Mt. Elgon are those characterized by the trees of giant groundsels. According to Hedberg (1964) they were combined by the term Dendrosenecio woodlands. LikeatMts. Kenya and Kilimanjaro (Rehder et al. 1981, Beck et al. 1983), Dendrosenecios occur at Mt. Elgon in afield layer either of Alchemilla species or of tussocks. Since in the latter case the scattered occurence of giant groundsels contradicts the idea of a woodland, such a plant community was described as Carex bog (type 2) which is shown in plots 23 and 26. Thus the Dendrosenecio woodlands of Mt. Elgon, described in this work, consistently exhibit an understorey of Alchemilla scrub, and thus closely resemble the Alchemilla- Dendrosenecio johnstonii-Community described for the Shira plateau of Mt. Kilimanjaro (Beck et al. 1983). Two types of Dendrosenecio woodlands can easily be differentiated on Mt. Elgon. The lower girdle of that vegetation formation is characterized by the less branched trees of Dendrosenecio elgonensis and by the considerable contribution of Alchemilla johnstonii to the field layer. The fact that the latter species prefers at least occasionally wet soil, together with the occurrence of Ranunculus oreophvtus, Lobelia lindblomii and Swertia crassiuscula indicates that a moist substrate is a prerequisite for the Dendrosenecio elgonensis-Community (plots 16 and 22). This idea is corroborated by the capability of D. elgonensis to advance into boggy areas. Completely different from that community is the second type of Dendrosenecio Woodlands which was designated as Dendrosenecio barbatipes-Community (plots 6 - 10) and was recorded on well drained slopes. It is characteristic of the upper alpine zone of Mt. Elgon where the rather dense stands of widely branched Dendrosenecio barbatipes trees provide an aspect which is unique for the steep slopes of the caldera (Fig. 4). Since Lobelia telekii, although not specified as a differential species, was consistently found to accompany Dendrosenecio barbatipes (Fig. 5). A striking physiognomic similarity of the Dendrosenecio barbatipes-Community with the Dendrosenecio woodlands of Mt. Kenya (Relider et al. 1981) was established. The latter is characterized by D. keniodendron, L. telekii and Alchemilla argyrophylla. Vegetation patches composed of species of both types of Dendrosenecio woodlands of Mt. Elgon have also been recorded, predominantly and expectedly at intermediate levels (plot 1 1) but, interestingly, also on the large summit plateau of Koitoboss (plot 5). The latter, due to poor drainage, is inhabitated by hydrophilic species like Ranunculus oreophytus and Lobelia lindblomii. CONCLUDING REMARK The vegetation units described in this work may be typical of the area of the National Park, which apart from fire, appears as widely uninfluenced by man. In contrast to these, heavily grazed areas with different grass communities must be expected to occur on the drier hills inside the crater (Hedberg 1964). Even the Dendrosenecio woodlands appear to be of less importance outside the National Park as it is documented by a photo (Hamilton 1982) showing only scattered groundsel trees in the region between lake Kimilili and Sudek (14,140 ft.). Therefore the altitudinal transect reported herein by no means should be generalized for the whole alpine belt of Mt. Elgon. ACKNOWLEDGEMENTS The work was supported by the Deutsche Forschungsgemeinschaft. The authors wish to thank the Assistant Director, Mr. Pertet, Ministry of Tourism and Wildlife, Nairobi for his efficient promotion of the project. Thanks are also due to the Warden of Mt. Elgon National Park for providing guidance to the area. The skillful assistance with the fieldwork by Dr. R. Scheibe, Dr. M. Senser and Dr. P. Ziegler is gratefully acknowledged. The authors, in addition, are very grateful to C.H.S. Kabuye, East African National Herbarium, Nairobi for help with the identification of plant species. REFERENCES Agnew, A.D.Q. 1974. Upland Kenya Wild Flowers, Oxford Univ. Press, London. Beck, E., Rehder, H., Pongratz, P., Scheibe, R. Senser, M. 1981. Ecological analysis of the boundary between the Afroalpine vegetation types ‘Dendrosenecio Woodlands’ and ‘Senecio Brassica-Lobelia Keniensis Community’ on Mt. Kenya. J. East Afri. Nat. Hist. Soc. and Natl. Mus. 172 : 1 - I I. Beck, E., Scheibe, R., Senser, M. 1983. The vegetation of the Shira Plateau and the western slopes of Kibo (Mt. Kilimanjaro, Tanzania). Phytocoenologia 1 1 : I - 30. Page 6 No. 190 Cotton, A.D. 1932. The arborescent Senecios of Mt. Elgon. Kew Bull. 10 : 465 - 475 Dale, I.R., Greenway, P.J. 1961. Kenya Trees and Shrubs. Buchanan’s Kenya Est. Ltd., Nairobi Hamilton, A.C. 1982. Environmental History of East Africa. A Study of the Quaternary. Academic Press, London, New York. Hamilton, A.C. Perrott, R.A. 1981. A Study of altitudinal zonation in the montance forest belt of Mt. Elgon, Kenya Uganda. Vegetatio 45 : 107 - 125 Hedberg, O. 1951. Vegetation Belts of the East African Mountains. Svensk Botanisk Tidskrift 45 : 140 - 202 Hedberg, O. 1957. Afroalpine vascular plants. Symb. Bot. Upsal. 15:1-411 Hedberg, O. 1964. Features of Afroalpine plant ecology. Acta Phytogeographica Suecica49 : 1 - 144 Hubbard, C.E., Milne-Redhead, E. (eds.) 1952 - 1983. Flora of Tropical East Africa, London KlOtzli, F. 1958. Zur Pflanzensoziologie des Siidhanges der alpinen Stufe des Kilimandscharo. Ber. Geobot. Forschungsinst, Riibel 1957 : 33 - 59 Knapp, R. 1971. Einfiihrung in die Pflanzensoziologie. Ulmer Verlag, Stuttgart Mueller-Dombois, D., Ellenberg, H. 1974. Aims and Methods of Vegetation Ecology. Wiley, New York, London, Sydney, Toronto. Nordenstam, B. 1978. Taxonomic studies in the tribe Senecioneae (Compositae). Opera Botanica 44 : 1-84 Rehder, H., Beck, E., Kokwaro, J.O., Scheibe, R. 1981. Vegetation analysis of the Upper Teleki Valley (Mt. Kenya) and and adjacent areas. J. East Afr. Nat. Hist. Soc. and Natl. Mus. No. 171:1-8 Received January 1984 Editors: H.J. Beentje, J.J. Hebrard No. 190 Page 7 Fig. 1 Map of the Kenyan sector of Mt. Elgon, redrawn and magnified from the Sheet Kapenguria, Series Y 503, Sheet NA-36-12, Edition 2 - 5K, published by Survey of Kenya (Nr. 2400/8/73). The altitudinal transect was made along a faint footpath, the approximate alinement of which is given by the dotted line. The figures represent the numbers of plots of vegetation analysis which are described in detail in Tables 1 and 2. Page 8 No. 190 Fig. 2 Carex runssorensis bog (Carex bog type 2, cf. Table 2, plot 23) with Dendrosenecio johnstonii ssp. elgonensis, Lobelia elgonensis and scattered specimens of Helichrysum amblyphyllum. The slopes in the background are covered by the Dendrosenecio elgonensis-community. (Foto: E. Beck, March 15th, 1983). Fig. 3 Carex runssorensis bog (Carex bog type 1, cf. Table 2, plot 13) in the caldera of Mt. Elgon (Koitoboss swamp). Patches of Alchemilla johnstonii can be detected between the Carex tussocks. The slopes in the background are covered by the Dendrosenecio barbatipes- Community (Foto: E. Beck, March 15th, 1983). No. 190 Page 9 Fig. 5 Close up the Dendrosenecio barbatipes- Community at the western slopes of Koitoboss summit (cf. plot 8 in Table 2). D. barbatipes, Lobelia telekii and Alchemilla elgonesis and Helichrysum amblyphyllum can be detected on the photo as important representatives of this community (Foto: E. Beck, March 15th, 1983). Fig. 4 Dendrosenencio woodlands: Dendrosenecio barbatipes- Community on the southeastern slopes of Lower Elgon at an elevation of 1 3,000 ft. The field layer consists predominantly of Alchemilla elgonensis (Foto: E. Beck, March 15th, 1983). Page 10 No. 190 Table 1 Species lists of 7 sample plots of the Ericaceous belt of Mt. Elgon along the Koitoboss Track. Cover abundance numbers after KNAPP (1971): + = few individuals with small covers; 1 = numerous or scattered individuals cover up to 5% of the sample area: 2 = cover of 5 - 25%; 3 = cover of 25 - 50%; 4 = cover of 50 - 75%; 5 = cover of 75-100% of the soil surface; T = trees (more than 3 m high). Transitorial features in regeneration communities are screened. Plot 1 End of the Koitoboss driveway at an altitude of 10,750 ft.; 20 X 20 m; tree layer with 60% cover; 14th March 1983; Plot 2 50 m W of Plot 1 at 10,810 ft. altitude; 20 X 20 m; tree layer with 30% cover; 14th March 1983; Plot 3 Rocky outcrop near the End of the Koitoboss driveway at an elevation of 10,750 ft.; 20 X 20 m; volcanic rock covered by mosses; 14th March 1983; Plot 27 Regeneration stage of Ericaceous bush 50 m N of the end of the driveway at an altitude of 1 0,800 ft. 10 X 10 m; 16th March 1983; Plot 28 Footpath to Koitoboss at an altitude of 1 1 ,870 ft.; 50 m S of the valley floor; 10 X 10 m; 15th March 1983; Plot 29 Footpath to Koitoboss, upper part of the slope; 1 1,670 ft.; 10 X 10 m; 15th March 1983; Plot 30 Footpath to Koitoboss, close to Plot 29; 1 1,670 ft.; 10 X 10 m; 15th March 1983; Table 2 Species list of 23 sample plots of the alpine zone of Mt. Elgon representing an altitudinal transect from the caldera up to the summit of Kotoboss and down to the end of the Koitoboss driveway. Cover abundance numbers as in Table 1 ; n(n) indicates two layers, usually a tree and a shrub or herb layer. Typical communities are screened; Hatching in the heading indicates mixed vegetation. All analyses were made on March 15th, 1983. Plot 12 W-slope of Koitoboss, 50 ft. above the floor of the crater 13,040 ft., 10 X 10 m; Plot 14 Bog in the caldera below Koitoboss summit; 12,970 ft., 10 X 10 m; Plot 13 Swamp at the E-rim of the caldera below Koitoboss summit; 13,000 ft.; 10 X 10 m; Plot 19 Spring area close to Koitoboss track, 12,970 ft., 10 X 10 m; Plot 9 Boulder stream on the W-slope of Koitoboss summit at 13,380 ft.; 15 X 15 m; Plot 10 Boulder stream on a flattening of the W-slope of Koitoboss summit; 13,340 ft.; 10 X 10 m; Plot 7 NW slope of Koitoboss summit on a plant-covered scree stream; 13,750 ft.; 10 X 10 m; Plot 6 As in plot 7 on a shallow ledge; 13,680 ft.; 3 X 12 m; Plot 8 W-face of Koitoboss summit with 20% bare rock; 13,800 ft.; 5 X 5 m; Plot 1 1 W-slope of Koitoboss; 1 10 ft. above caldera; 13,100 ft.; 10 X 10 m; Plot 5 Plateau representing the summit of Koitoboss, 13,880 ft.; 15% bare rock; 10 X 10 m; Plot 16 Flattening on a rock table E of the saddle W of Koitoboss; 13,130 ft.; 6 X 12 m; Plot 22 Along Koitoboss footpath at 12,830 ft.; 10 X 10 m; Plot 23 As in plot 22, 12,740 ft.; border of a mire; 10 X 10 m; Plot 26 Mire at an altitude of 12,060 ft. in the grassland area; 10 X 10 m; Plot 4 100 m N of the footpath (middle of the slope) at 12,230 ft. altitude; 10 X 10 m; Plot 17 S-slope of a wide gully on the footpath at 13,070 ft. altitude; 10 X 10 m; Plot 18 SE-slope of the same gully as in plot 17; 10 X 10 m; Plot 21 50 m SW of Koitoboss footpath at 12,870 ft.; 10 X 10 m; Plot 15 Rock outcrop at the saddle W of Koitoboss summit; 13,160 ft.; pure rock 10%; 15 X 15 m; Plot 20 Rock outcrop 15 m N of Koitoboss footpath at 12,930 ft.; bare rock 10%; 8 X 8 m; Plot 25 Gentle slope on shallow soil 20 m W of Koitoboss footpath at 12,600 ft.; some charred trunks indicate former burning; 10 X 10 m; Plot 24 100 m W of plot 25 at 12,670 ft.; no direct evidence of former fire; 10 X 10 m; TABLE 1 Ericaceous Woodland Ericaceous Bush Transitio Zone Forest- Heath Exo- theca- Aqrost Ccmnui Reqeneration Types LS i) Number of plot 1 2 3 1 27 28 29 30 Altitude x 1,000 (ft.) 10.7 10. 10.8 10. 11.8 11.6 11.6 Exposition NE NE NE NE NE N N Slope (°) 10 5 10 20 5 5 5 Plant cover (%) 90 95 80 90 85 97 95 Number of species (Mosses = 1 ) 34 33 26 17 22 25 20 Philippia keniensis S. Moore 3T 1TH 1 2T+2 2 3 4 Anthospermum usambarensis K.Schum. 1 1 1 2 + 2 Helichrysum odorat issimum (L.)Less. 1 2 1 1 2 i Artemisia afra Willd. 1 2 2 2 2 2 Leucas spec. A Aqnew 1 1 1 1 Stoebe ki limandschar ica O.Hoffm. 1 T+2 1T4 3 1 1 + Hebenstretia cf. anqolensis Rolfe 1 + 1 1 Satureja biflora (D. Don . ) Benth . + 1 4 + 4 Andropoqon lima (Hack.)Stapf 1 4 + + 2 1 2 Kniphofia snowdenii C.H.Wriqht 4 1 + + 4 Olea kilimandscharica Knobl. + + T Erlanqea fusca S. Moore 1 + Hypericum kiboense Oliv. 4 1 + Luzula johnstonii Buchen. 1 1 Caucalis incoqnita Norman 1 4 Swertia kilimandscharica Enql. 1 4 Juniperus procera Hochst. ex Endl. IT Raoanea cf rhododendroides (Gilq)Mez 1T + 1 Agauria salicifolia (Cairn. ex Lam.)Hook.f 3T Nidorella arborea R.E. Fries 1 Senecio rhammatophy 1 lus Mattf. 4 Haqenia abyssinica ( Bruce ) J . F . Gmel . +T Helichrvsum formosissimum (Sch.Bip. ) A. R. 1 Erica arborea L 1 2T Viola eminii (Enql .) R. E. Fries + Cerastium afromontanum T . C . E . Fr ies&Wei 4- Hypericum keniense Schweinf. 1 Protea kilimandscharica Enql. 4 + 3 Bartsia kilimandscharica Enql. 1 + 1 Helichrysum nandense S . Moore 1 + + Lithospermum afromontanum We in. 4 4 + Pilosellcides hirsuta (Forsk . ) C . Jef f r .' 4 + 1 Exotheca abyssinica ( A . Rich . ) Anderss . 4 Agrostis gracilifolia C.E.Hubb. 4 1 Aqrostis kilimandscharica Mez 4 Dierama pendulum (L.f.)Bak. Dendrosenecio johnstonii ssp. elqonensis (T . c . E . Fries ) B . Nord . 1 2 Helichrysum qlobosum Sch . Bip. 4 1 1 Anthoxanthum nivale K.Schum. 4 4 Hypericum spec. + 2 2 ! cf Scirpus spec.B 2 1 Anaqallis serpens D.C. 2 2 2 ; Conyza subscaposa O.Hoffm. 2 + 1 Carduus keniensis R.E. Fries 4 1 2 Senecio snowdenii Hutch. 4 4 Swertia crassiuscula Gilq + + Veronica qlandulosa Benth. 1 1 cf Dichrocephala alpina R.E. Fries + 4- Cineraria qrandiflora Vatke + Galium ruwenzor iense (Cor t . ) Chiov . + + + + 4 Galium glaciale K. Krause + Sonchus bipontini Aschers. + 4 Geranium ocellatum Cambess. + + 4- 4 Geranium elamellatum Kokwaro + Satureja cf ki limandschari (Guerke)Hed + Satureja uhliqii Guerke 4 + Haplocarpha rueppellii (Sch . Bio . ) Beauv 4- cf Wahlenberqia arabidifolia (Enql.)Br + Mariscus spec. 1 Aqrostis spec. 2 Conyza welwitschii (S . Moore) Wi Id 4- Deschampsia caespitosa (L. ) PB. 1 Alchemilla cryptantha A. Rich. 4- Alchemilla elqonensis Mildbr. 2 Alchemilla johnstonii Oliv. + 2 2 2 Festuca abyssinica A. Rich. 1 + 2 1 4- Euryops elqonensis Mattf. + Helichrysum forskahlii (Gmel .) Hill . &Burt . 4 4 Lobelia elqonensis R.E.& T. Fries jr. 1 Cirsium buchwaldii 0. Hoffm. + Geranium kilimandscharicum Enql. + Sedum ruwenzor iense Bak . 1 Carex ninaqongensis (Kuk.)Rebyns + 4- Crassula pentandra (Edgeworth ) Schonl . + Mosses 3 2 1 2 Page 12 No. 190 TABLE 2 \ carex Dendrosenecio Woodlands Carex Euryops-Bush Helichrysim Ills sock bog (1) D. barbatipes D. elgonensis boa (2) Scrub Grassland ccmnunity ccrminity Type Number of plot 1 14 13 19 9 10 7 6 8 11 5 16 22 23 26 4 17 18 21 15 20 25 24 Altitude x lOOO (ft. ) 1 3.0 12.9 13.0 12 .9 13.4 13.5 13.7 13.7 13.8 13.1 13.9 13.1 12.8 12.7 12.0 11.9 13.1 13.1 12.8 13.1 12.9 12.6 12,6 Exposition - SW W W W W W - - SE E N SSW S SE E - NE E SSE Slope (°) 0 0 0 5 5 20 30 50 10 0 0 5 5 7 10 -5 <5 5 0 <5 <5 5 Plant cover (%) 9 98 100 95 80 60 95 95 80 90 85 90 95 95 95 90 90 90 95 90 90 95 90 Number of species (Mosses = 1) 9 7 5 16 12 11 10 14 9 17 14 11 9 18 27 12 14 6 23 20 15 12 Alchemilla elgonensis Mildbr. 2 2 1 5 3 2 2 3 4 5 2 2 3 3 3 2 2 Lobelia telekii Schweinf. 2 3 1 4 4 1 1 4 4 4 4 Helichrysum amblyphyllum Mattf. 1 4 4 1 4 2 2 4 2 2 1 4 1 Festuca abyssinica A. Rich. 4 4 1 4 4 4 1 4 1 1 Valeriana kilimandscharica Engl. 4 4 1 1 1 1 1 4 1 4 Dendrosenecio johnstonii ssp. barbatipes 1( +) 3(3) 5(2) 2 2 1 2(1) 2 1(1) 4 (+) (Hedb.) B. Nord. Bartsia kilimandscharica Engl. 4 4 1 1 4 1 Swertia subnivalis T. Fries jr. Senecio sotikensls S. Moore 4 1 1 2 4 4 1 Sedum ruwenzor iense Bak. 4 1 4 Poa leptoclada Hochst. 4 4 4 Xiphopteris f labellif ormis (Polret) ♦ 4 4 Schelpe Heracleum elgonense (Wolff) Bullock 1 4 Swertia uniflora Mildbr. 4 4 4 Ranunculus oreophytus Del. 1 4 Lobelia lindblomii Mildbr. 4 4 + + Dendrosenecio johnstonii ssp. 1 1(1) 1 1 2 3 + elgonensis (T . C . E . Fries ) B.Nord. Lobelia elgonensis R.E.&T. Fries jr. ♦ 2 2 Carex runssoroensis K. Schum. Carex ninagongensis (Kuk. ) Rebyns 3 5 4 5 1 1 Montia fontana L. 4 Ranunculus volkensii Engl. Eriocaulon volkensii Engl. 1 3 Lycopodium saururus Lam . Crassula granvikii Mildbr. 4 1 4 Sibthorpia europaea L. 4 Alchemilla johnstonii Oliv. 2 2 2 Euryops elgonensis Mattf. 3 2 3 4 Peucedanum kerstenii Engl. * + Carduus keniensis R.E. Fries Pentaschistis minor (Ballard & C.E. Hubbard) 4 4 1 Ballard & C.E. Hubbard Cerastium octandrum (Hochst. ex) A. Rich. 4 4 HoUrhrvanm cltrispinum Del. Crassula spec. B. Agnew 1 Geranium kilimandschar icum Engl. Senecio snowdenii Hutch- Helichrysum newii Oliv. !■ Hiern. 4 4 + 4 4 1 4 4 Helichrysum forskahlii (J.G.Gmelin) Hilliard & Burtt. Festuca pilgeri St. Y. 4 2 3 4 5 Pentaschistis borussica (K. Schum. )Pilg. 2 1 Koeleria capensis (Steud.) Nees 4 4 Cirsium buchwaldii 0. Hoffm. 1 2 Conyza subscaposa 0. Hoffm. 4 1 2 Mosses 3 3 2 5 5 3 2 4 2 Blaeria johnstonii Engl. Galium ruwenzoriense (Cort . ) Chiov . Galium glaciale K. Krause Agrostis gracilifolia C.E. Hubb. 4 1 4 Agrostis kilimandscharica Mez Luzula abyssinica Pari. Swertia crassiuscula Gilg. + 2 Sagir.a abyssinica A. Rich. Trifollum cryptopodium A. Rich. Veronica glandulosa Benth. 4 1 Haplocarpha rueppellii (Sch. Bip.) Beauv. Haplosciadium abyssinlcum Hochst. Satureja kilimandschari (Guerke) 1 Hedb. Helichrysum spec, (seedlings) Cineraria grandiflora Vatke. 4 Anthospermum usambarensis K. Schum. Piloselloides hirsuta (Forsk.) C. Jeffrey Philippia keniensis S. Moore 2 Artemisia afra Willd. Helichrysum odorat issimum (L.) Less. 1 Hebenstretia cf . angolensis Rolfe Andropogon lima (Hark .) Stapf . Scirpus spec. A Scirpus spec. B 1 2 Veronica gunae Engl. Anagallis serpens D. C. 1 Anthoxanthum nivale K. Schum. Cerastium afromontanum T. C.E. Fries 4 t Weimark Festuca mekiste W.D. Clayton Hypericum kiboense Oliv. 4 JOURNAL OF THE EAST AFRICA NATURAL HISTORY SOCIETY AND NATIONAL MUSEUM December 1988 A taxonomic revision of the grasshopper genus SPATHOSTERNUM (QRTHOPTERA, ACRIDIDAE) VOLUME 78, No. 191 J.P. GRUNSHAW, Overseas Development Natural Resources Institute, London ABSTRACT The highly wing-polymorphic, Afro-Oriental grasshopper genus Spathosternum Krauss 1977, Hemiacridinae is revised. Keys are provided to the males of the species. Seven species, prasiniferum, abbreviation, curtum, malagassum, nigrotaeniatum, brevipenne and pygmaeum are now recognised. Three species brevicorne, beninense, medium and one subspecies prasiniferum sinense are newly synonymised. INTRODUCTION This paper is the second of two revisions of hemiacridine grasshopper genera due to be included in a future handbook and field key to the grasshoppers of Eastern Africa. Reference to current assessments of the systematic problems associated with the sub-family Elemiacridinae Dirsh (1961, 1965 and 1975) were given in the first paper Grunshaw ( 1 986), and are not discussed here. The recently published “Locust and Grasshopper Agricultural Manual” (Centre for Overseas Pest Research, 1982), provides a comprehensively annotated reference guide to much of what has been previously written concerning the biology, distribution and economic importance of the genus. Reference will be made throughout the text to this manual where applicable. The genus Spathosternum was erected by Krauss ( 1 877) to accommodate Tristria nigro-taeniata Stal (1876) as its type species, being separated principally from “Oxyae and Tristriae” on the shape of the prosternal process. In 1910 Kirby considered that the genus belonged to a group of the Cyrtacant hacridinae and it had by then grown to include the species S. nigrotaeniatum, Stal ( 1 Sib), pygmaeum Karsch ( 1 893), venulosum Stal (1878) and caliginosum Walker (1871). Four years later Kirby (1914) reassigned it to the sub-family Catantopinae without giving any reason for doing so. Since then the genus has oscillated between the Cyrtacanthacridinae and Catantopinae. Tinkham (1940), for example, allocated it to the group Spathosterni of the Cyrtacanthacridinae. It was returned to the Catantopinae by Johnston (1956) as part of the group “Leptacres”, which also contained most of the present Hemiacridinae and Tropidopolinae. In the same year, Dirsh (1956), assigned the genus to the newly erected sub-family Hemiacridinae, on the basis of its divided endophallus and its sound-producing mechanism. In his study, Australian Acrididae, Rehn (1957) subsequently gave the genus tribal status by erecting the tribe Spathosternini, which also included the East Australian genus Laxabilla. More recently Dirsh (1975) erected the Spathosterninae as a sub-family of the family Hemiacrididae, but this decision has found little support amongst contemporary workers in the Acridoidea. The genus is of biogeographical interest since its representatives are Afro-Oriental in distribution. There are four species in Africa, one of which extends into Saudi Arabia. One species' is restricted to Madagascar, another to India, and a last widespread and variable species extends from Pakistan through India, Sri Lanka and Burma to South East China. Most species favour marshy habitats in which grasses predominate, typically cohabiting with Oxya spp. Abbreviations used in figures a, length of a midline drawn between the anterior and posterior edges of the lophal interspace; ad, apical diverticulum; ap, apical penis valves; b, width between the outer extremities of the lophus; cv, cingular valves; lpc, lateral process of cingulum; pd, preapical diverticulum. CO Page 2 No. 191 Abbreviations for depositories BMNH, British Museum (Natural History), London; MNHN, Museum National d’Histoire Naturelle, Paris; MNHU, Museum fur Naturkunde der Humboldt-Universitat, Berlin; NR, Naturhistoriska Riksmuseet, Stockholm; ODNR1, Overseas Development Natural Resources Institute, London. PRESENTATION The preparation and treatment of the male genitalia in this study are the same as those described by Dirsh (1956). Female spermathecae were examined by removal of posterior abdominal segments and maceration in 5 [ potassium hydroxide solution. The nomenclature of the female spermatheca used in this study is that of Dirsh ( 1957). Separation of taxa is based mainly on male genital morphology. A key is given for the males of all species and subspecies in the genus, based mainly on external characteristics. Females can only be identified reliably by association with sympatric males. Descriptions are given in the form of differential diagnoses. Main diagnostic characters together with the male genitalia, are illustrated. Three views of the aedeagus and two of the epiphallus are given. The former was drawn from dorsal, lateral and posterior aspects, each drawn at right angles to the plane of the aedeagus. The epiphallus was drawn from a dorsal aspect with the plane of the lophi at right angles to the line of sight of the observer and from a posterior view, giving maximum depth to the epiphallic bridge. In the majority of cases the oval sclerite has been omitted. Where types have been examined this is indicated. Measurements (in mm), together with ranges and means are given for the material examined. Total length is measured from the tip of the vertex to the apex of the folded posterior femur. Key to species of Spathosternum (males) 1 Fastigium of vertex with median dorsal carinulae present (Figs. 20-23) 2 — Fastigium of vertex with median dorsal carinulae absent (Figs. 17-19) 5 2(1) Tegmina fully-winged or brachypterous With (35-50) densely packed stridulatory veinlets (Figs. 42-43). Supra-anal plate comparatively large (Fig. 11), with median transverse groove. (Indo- Oriental) S. prasiniferum — Tegmina brachypterous or micropterous (Figs. 4-6). Stridulatory file present or absent, if present then vestigial and situated near apex of tegmen, usually with less than 30 stridulatory veinlets. Supra-anal plate comparatively small (Figs. 13- 1 5), with or without median transverse groove. . . .3 3(2) Stridulatory veinlets present. Apical penis valves broad with divergent apices (Fig. 53). (India) S. abbreviatum Stridulatory veinlets absent. Apical penis valves slender with divergent apices (Fig. 60), or broad with sub-parallel apices (Fig. 112). . . ." 4 4 Small micropterous species (Fig. 6). Supra-anal plate without median transverse groove (Fig. 14). Tegmina touching dorsally. Apical penis valves very slender with divergent tips (Fig. 60). (Angola) S. curtum. Small micropterous species (Fig. 4). Supra-anal plate with weakly impressed median transverse groove (Fig. 15). Tegmina not touching dorsally. Tips of apical penis valves not divergent (Fig. 1 12). (Madagascar) S. malagassum 5(1) Supra-anal plate trilobate with preapical teeth or projections (Fig. 9). (Africa, S. Arabia) S. nigrotaeniatum Supra-anal plate not trilobate, preapical teeth or projections absent (Figs. 10-12 6 6(5) Supra-anal plate triangular with median transverse groove (Fig. 10). Fully winged or brachypterous species (Fig. 3). ( W. Africa) S. brevipenne Supra-anal plate parabolic or triangular without median transverse groove (Figs. 12,16). Fully- winged or with slightly shortened wings (Figs. 66, 83). (Africa south of 1 5°N latitude) S. pygmaeum Spathosternum Krauss, 1877 Spathosternum Krauss, 1877:45. Type species: Tristria nigro-taeniata Stal, 1876:233. > No. 191 Page 3 The genus Spathosternum forms a natural grouping of closely related taxa assigned to the subfamily Hemiacridinae. Within the confines of this somewhat heterogeneous assemblage of genera, Spathosternum can be readily differentiated by a combination of characters. These are the conical, opisthognathous shape of the head, the two distinctive sinuate facial carina, the shape of the prosternal process, the conspicuous longitudinal, lateral stripe markings and by the internal male genitalia. Brachypterous and micropterous species, notably 5. curtum, S. malagassum and S. abbreviatum, can be separated from Paraspathosternum, which might be considered to be their nearest relative, by the shape of the fastigium verticis (Fig. 1), posterior margin of the pronotum and by the male genitalia (Figs. 51-52, 58-59). The female differs from the male in being larger and more robust. Note that species in the genus display a widespread incidence of intra- and inter-population wing polymorphism, perhaps implying that many are in a state of active evolution towards flightlessness. This characteristic explains why earlier authors found it difficult to make precise species diagnoses based upon purely external morphology. The genus Chrotogonus presented similar problems to Kevan (1957, 1959). Spathosternum nigrotaeniatum (Stal, 1876) (Figs. 2, 9, 17, 31-32, 24-26, 27-30, 94, 64) Tristria nigro-taeniata Stal, 1876:233. Holotype <5, S.W. Africa, (NR) [examined]. Spathosternum nigrotaeniatum (Stal) Krauss 1877:45 Diagnosis. Male. Small, integument very finely rugose, almost smooth. Head conical; fastigium of vertex (Fig. 1 7) parabolic, surface concave. Antennae filiform, shorter than head and pronotum together. Frons in profile strongly oblique; frontal ridge (Fig. 31) sulcate, marginal carina well developed, thickened, slightly narrowing below apex of vertex. Face with two distinct sinuate carinae. Eyes large, oval elongate. Dorsumof pronotum flattened; median dorsal pronotal carina well developed, sometimes intersected by second and third sulci, always intersected by principal sulcus; lateral pronotal carina (Fig. 2) becoming sinuate between second and principal sulci, slightly divergent posteriorly. Metazona shorter than prozona; posterior margin of pronotal disc obtuse-angulate. Lateral pronotal lobes more densely pitted in metazona than prozona, with two large smooth areas in upper part of prozona. Prosternal process (Fig. 32). shallowly bilobate, flattened, inclined backwards. Mesosternal interspace strongly constricted, with mesosternal lobes almost confluent, widening posteriorly. Tegmen fully developed, always surpassing apex of folded hind knees. Radian area of tegmen with series of regular parallel veinlets. Posterior femur of normal shape; external apical spine of hind tibia present. Supra-anal plate (Fig. 4) trilobate with pre-apical projections or teeth. Last abdominal tergite with pair of small, widely spaced projections. Subgenital plate obtusely conical. Cercus, unmodified simple, acute straight. Endophallus divided, with slightly elongate aedeagus. Apical penis valves protruding subapically beneath encircling cingular valves. Apices of apical penis valves narrow, divergent. Cingular valves viewed anterodorsally (Figs. 24-26) appearing as a thin membraneous, sub-circular structure, forming sub-acute, inwardly curling, slightly thickened apices, flanked by lateral, inflated processes of cingular origin. Epiphallus (Figs. 27-30) bridge-shaped with median area of adjoining lateral plates less sclerotised, almost giving a divided-bridge appearance; oval sclerites present. Ventral edge of bridge forming a small, median, upcurved projection anteriorly. Ancorae small, down-curving with acute apices. Sides of epiphallus incurved. Lophi blunt, triangular, flat, slightly upcurved. General colouration variable, often expressing green-brown polymorphism. Face light brown, stramineous or green. Dorsal surface of head, pronotum and tegmen same colour as face, or sometimes with darker medial area. A narrow, dark longitudinal stripe extends from below eye to coxae of middle legs, sometimes incomplete; above this a paler stripe extends to abdomen, marginated by a broad dark brown stripe, which continues from behind compound eye, across upper part of lateral pronotal lobes before merging with tegmen. Posterior femur light brown or dull yellow. Knee lunules dark brown. Posterior tibia light brown or grey yellow. Female. Anterior margin of vertex rounded triangular. Supra-anal plate elongate, triangulate, with raised median ridge; shallowly sulcate along its length and bisected by a median transverse sulcus; pre-apical teeth absent. Spermatheca (Fig. 94) with apical diverticulum becoming bulbous distally, apex bent and obtusely rounded; preapical diverticulum in form of a slender tube expanded proximally. General colouration green or brown with similar longitudinal lateral markings as male. Posterior femur uniformly green or brown, often showing light green on upper face. Page 4 No. 191 Figs 1-23. Paraspathosternum and Spathosternum species. __ 1 , Paraspathosternum pedestris, dorsal aspect of male head and pronotum. 2-23, Spathosternum spp. 2-8, 2, nigrotaeniatum; 3, brevipenne; 4, malagassum: 5, abbreviatum ; 6, curium: 7 , prasiniferum; 8, pygmaeum. Scale line represents 1mm. 9-16, Spathosternum spp., male supra-anal plates; 9, nigrotaeniatum: 10. brevipenne; 1 1, prasiniferum: 12, pygmaeum; 3, abbreviatum; 14, curtum; 15, malagassum; 1 6, pygmaeum (Dikwa, N. Nigeria). Scale line represents 1mm. and also applies to Figs. 17-23. 17-23. Spathosternum spp., dorsal aspect of male fastigium verticis; 1 7, nigrotaeniatum; 18, brevipenne; 1 9, pygmaeum; 20, abbreviatum; 2 1 , prasiniferum; 22, curtum, 23, malagassum. No. 191 Page 5 TABLE I. Spathosternum nigrotaeniatum: measurements. Males Females n. mean range n mean range Interocular distance (29) 0.43 (0.21-0.64) (24) 0.77 (0.60-0.95) Head width (29) 3.06 (2.60-3.35) (24) 3.71 (3.21-4.58) Pronotal width (29) 2.88 (2.46-3.63) (24) 3.64 (3.08-4.27) Pronotal length (29) 3.55 (2.97-4.23) (24) 4.30 (3.69-4.84) Posterior femur length (29) 9.54 (8.40-1 1.46) (23) 1 1,93 (9.73-13.93) Posterior femur depth (29) 2.26 (1.91-2.54) (23) 2.60 (1.85-3,02) Antennal length (6) 7.91 (6.65-8.7) (6) 7.13 (5.52-8.3) Tegminal length (23) 16.14 (12.83-20.33) (24) 17.55 (12.34-21.77) No. of stridulatory veinlets (23) 17.08 ( 1 5-20) (13) 17.92 (16-21) Stridulatory file length (23) 3.94 (3.18-5.27) (13) 4.67 (4.01-6.28) Total length (29) 19.18 (16.33-22.18) (18) 23.00 (19.19-26.54) Discussion. Despite the highly variable nature of this species particularly in relation to tegminal length and epithallic morphology (see figs. 27-30 for variation in epiphallus shape), it is nevertheless easily differentiated from all other members of the genus by the unique possession of preapical teeth or projections, which arise from the supra-anal plate. Populations from West Africa, particularly the Sahelian region, are generally larger than their East African equivalents, which approach pygmaeum in size. The available data on the life history, distribution, ecology and economic importance of this species have been reviewed elsewhere (Centre for Overseas Pest Research, 1982). Type-material examined. Holotype <5, S.W. AFRICA: Damaraland (NR) Additional material examined. SAUDI ARABIA: 1(5, Asir, Wadi Jowra, 22-25. xii. 1947 (Popov) (BMNH). YEMEN ARAB REP.: 1$, Suaid, nr. Beit el Fagih, 23-24. vi. 195 1 (Tillin) (BMNH). NIGER: 6 5$, Niamey, (Popov) (ODNRI). MALI: 3<5, 1? Diaferabe, i. 1957 fDflvcvj(BMNH), 3<5, 2 9 , Tilembeya, 10-12. xii. 1956 (Popov) (ODNRI). GHANA: ltf, 1 ? , Elmina, 20.iii. 1969 (Richards) (TDRI); ltf, E. of Shai Hills, 19.xi. 1961 (7agoJ(BMNH); 1(5, Accra Plain, 23 km. W. of Accra, 18.x. 1959 (Jago) (BM NH). NIGERIA: 1$, Gadau, vi. 1923 (Luxton & Lewis) (BMNH); 1 ? , 50 km. S. of Bama, 14.xi. 1970 (Popov) (ODNRI); 2(5, 2?, Zaria, Samaru, IARfarm, 1 l-14.xi. 1970 (Jago & Hollis) ( ODNRI); 1<5, 1 $ , Zaria, 4-14. ix. 1970 (Popov) (ODNRI). ZAMBIA: 1(5, 1$, Mweru-wa-Ntipa, Kangiri Plain, 23.x. 1957 (Fitzgerald) (BMNH). REP. S. AFRICA: 1(5, Cape Prov., Queenstown, 16. l-10.ii. 1923 (Turner) (BMNH). Spathosternum abrreviatum Uvarov, 1929 (Figs. 5, 13, 20, 53-55, 56-57, 94, 64) Spathosternum abbreviatum Uvarov, 1929:556. Holotype $, INDIA, (BMNH) [examined]. Spathosternum medium Uvarov, 1929:558. Holotype <$, S. INDIA, Masinigudi, 900-1000m., dry bush, 29. i. 1927. Syn. nov. Diagnosis. Male. Antennae (Fig. 5) much shorter than combined length of head and pronotum together, thick, distally flattened and widened. Fastigium of vertex (Fig. 20) with concave surface, divided into two by median carinulae, similar to both malagassum and curium. Pronotum relatively short, lateral pronotal carina almost straight; posterior margin of disc rounded, angular. Tegmen shortened, reaching to middle of posterior femur; stridulatory veinlets vestigial, few in number, located toward apex of tegmen. Posterior femur short, stocky. Supra-anal plate (Fig. 13) with raised, median ridge, variably sulcate along its length; posterior margin forming small elongate apex, cuticular sculpture similar to that in curtum. Genitalia (Figs. 53-55) with apices of apical penis valves divergent, like those of nigrotaeniatum and curtum, but proportionally longer and broader. Inflated lateral processes of cingular valves showing distinctive spicule-like sclerotization. Epiphallus (Figs. 56-57) similar to that of curtum, but differing in having diminutive, knob-like lophi. General colouration brown, with pale buff longitudinal lateral stripe; pattern typical for genus. Female. Antennae somewhat shorter proportionally than in male. Supra-anal plate elongate, triangular. Page 6 No. 191 Figs. 24-41, Spathosternum spp. 24-32, S’. nigrotaeniatum. 24, male aedaegus, view from above; 25, same, posterior aspect; 26 same, lateral aspect; 27, male epiphallus, dorsal aspect 28, same, posterior aspect; 29, same, showing variation in epiphallus shape, dorsal aspect; 30, same, posterior aspect; 31, frontal aspect of male head; 32, male prosternal tubercle. 33-41, S', brevipenne. 33, male aedaegus, view from above; 34, same, posterior aspect; 35, same, lateral aspect; 36, male epiphallus, shortened-wing form, posterior aspect; 37, same, dorsal aspect; 38, same, long-wing form, posterior aspect; 39, same, dorsal aspect; 40, same, brachypterous form, posterior aspect; 41, same, dorsal aspect. Scale line under Fig. 31 represents 1mm. Scale line under Fig. 37 represents 1mm and applies to all figs except fig. 31. No. 191 Page 7 Spermatheca (Fig. 97) similar to that of pygmaeum, but without accessory vesicles of preapical diverticulum; apex of apical diverticulum more bulbous, otherwise like males. TABLE 2. Spathosternum abbreviation: measurements. Males Females Interocular distance n (2) 0.40,0.51 n (2) 0.85,0.84 Head width (2) 2.13,2.27 (2) 2.75,2.83 Pronotal width (2) 2.09,2.32 (2) 3.03,3.20 Pronotal length (2) 2.26,2.38 (2) 2.93,3.09 Posterior femur length (1) 6.44 (2) 7.13,8.61 Posterior femur depth (1) 1.78 (2) 2.42,2.54 Antennal length (1) 2.58 (1) 2.44 Tegmina! length (1) 4.10 (2) 5.01,5.44 No. of stridulatory veinlets (1) 21 (1) 17 Stridulatory file length (1) 1.38 (1) 1.40 Total length (1) 12.27 (2) 15.05,16.: Discussion. The affinity between African and Indian representatives is most clearly illustrated by this species, which is externally very similar to curtutn and malagassum in appearance. Internally the aedeagus shares some characters with that of both nigrotaeniatum and curium, such as the divergent apical penis valves (compare Figs. 25, 53 and 60). In Uvarov ( 1929) a description of a new species medium was also given but unfortunately no mention of the intended depository for the type of this species was cited. As attempts to locate this type for further study have failed, this species has subsequently not been included in the key to species, although the existence of that description is acknowledged here. Curiously Uvarov (1953) made no mention of this species in his list of the species of Spathosternum. From Uvarov’s description it can be gathered that medium appeared to be nearer to prasiniferum than to abbreviatum. The wings were shorter than prasiniferum but longer than abbreviatum; the hind femora were short and thick like abbreviatum and not thin like prasiniferum. It is deduced from the above that medium is probably a wing-polymorph of abbreviatum, and it is here synonymised. Type material examined. Holotype INDIA: Nilgiris, Snowdon Peak (2426m), 6. 1 X. 1917 (Rao) (BMNH), Paratypes, 1(5, I ? , same data as holotype; 1 ? , Nilgiris, Snowdon Peak (2500m), 2.X. 1921 (Nathan) (BMNH). Spathosternum brevipenne Chopard, 1958 (Figs. 3, 10, 18, 33-35, 36-41, 95, 64) Spathosternum brevipenne Chopard, 1958: 131, Fig. 3. Holotype Q, GUINEA, (MNHN) [examined], Spathosternum beninense Popov, 1980:45. Holotype <5, BENIN (BMNH) [examined]. Diagnosis. Male. Antennae, about as long as head and pronotum together. Fastigium verticis (Fig. 18) parabolic, shallowly concave with moderately raised margins. Occiput, relatively convex. Prosternal process very broad, flat, weakly spathulate. Tegmen, shortened with wing tips just surpassing third abnominal tergite (Fig. 3), or fully developed with wing tips just reaching knee of folded hind femur. Last abnominal tergite with diminutive furculi. Supra-anal plate (Fig. 10). broadly triangular in outline with a median transverse sulcus. Genitalia (Figs. 33-35) with apices of penis valves expanded and divergent, showing spicule-like sclerotization, reminiscent to that of abbreviatum. Epiphallus (Figs. 36-41) similar in shape to prasiniferum but with ancorae more widely spaced and area of lophal interspace curvilinear. General colouration brown, stramineous, with longitudinal lateral stripes typical of genus. Posterior femur uniformally brown yellow, with outer median area sometimes being more darkly expressed. Upper lobe of hind knee black; lower lobe without black, but sometimes tinged with red. Female. Antennae, shorter than combined lengths of head and pronotum. Supra-anal plate more Page 8 No. 191 elongate than male; projections of last abdominal tergite absent. Spermatheca (Fig. 95) with apical diverticulum tapering proximally and distally; preapical diverticulum expanded proximally, apex slightly inflated, otherwise like male. TABLE 3. Spathosiernum : brevipenne measurments Males Females n mean range n mean range Interocular distance (8) 0.41 (0.37-0.45) (3) 0.94 (0.81-1.05) Head width (8) 2.92 (2.80-3.05) (3) 3.81 (3.65-3.89) Pronotal length (8) 3.12 (2.85-3.24) (3) 4.10 (3.95-4.23) Posterior femur length (8) 8.92 (8.02-9.42) (3) 12.12 (1 1.53-12.37) Posterior femur depth (7) 1.99 (1.76-2.09) (3) 2.57 (2.38-2.39) Antennal length (4) 6.71 (5.37-7.7) (1) 6.2 Tegminal length (5) 10.84 (8.92-13.47) (3) 7.77 (5.59-1 1.63) No. of stridulatory veinlets (5) 21 (17-25) (2) 18.5 (18-19) Stridulatory file length (5) 3.24 (2.62-3.86) (2) 3 ( 2.74-3.24) Total length (7) 17.23 (15.53-18.18) (3) 22.65 (22.02-23.03) Discussion. The species brevipenne, originally described from brachypterous specimens, was found to be represented by a complex of wing polymorphic populations. A single long-winged male specimen found within a collection of material, lent by M. Donskoff (MNHN), had been previously identified by Dirsh (on external morphology) as S. pygmaeum. Subsequent examination of the genitalia has shown this specimen to be a long-winged variant of brevipenne (compare epiphalli Figs. 38-41). It is also of interest to note that this specimen was collected from Nimba, Guinea (Mt. Nimba being the type locality for brevipenne). The extreme brachypterous form may be associated with adaptation to montane conditions. In addition, the recently described species beninense (Popov, 1980) was found also to be conspecific with brevipenne, this decision being based upon the following criteria; similarity in respective male genital morphology, similarity in shape and cuticular structure of the malesupra-anal plates, both being broadly triangular, obtusely pointed with a transverse median sulcus. Other characters common to both include, the possession of two small projections on the last abdominal tergite and a similar red tinge of the posterior knee lunules. A further long-winged variant was collected by Popov and Jago from the Cameroons. The distribution of brevipenne, in the light of the above, is now known to be wider than was First thought. The brachypterous form lacks stridulatory venation. It is not known whether the two forms have similar song or whether song is lacking in this morph. Nor is it known whether the faculty for song recognition has also been lost in brachypterous morphs. It is conceivable that the emission of song from a long-winged individual (still retaining stridulatory venation) could still initiate courtship with brachypterous females. Type-material examined. Spathosiernum brevipenne Chopard, holotype $, GUINEA: Mt. Nimba, Grand Nimba, 1 0.xi. 195 1 (Lamotte & Roy) (MNHN). Spathosiernum beninense Popov, holotype BENIN: Parakou, xii. 1977-i. 1978 (Popov) (BMNH). Paratypes. BENIN: 3 4ml / L 130 SOUTH EQUATORIAL SUBSURFACE WATER SPEED 0-30cm/s TEMPERATURE 9-20°C POTENTIAL DENSITY 07= 27-22- 26-9C SALINITY 3510-36 10%. OXYGEN 1 50-3 OOml/L RED SEA WATER S 700 ANTARCTIC INTERMEDIATE WATER SPEED 0 -1 5 c m/s TEMPERATURE 9°C POTENTIAL DENSITY 07=275 0-2801 SALINITY 34 95%o OXYGEN > 1 50 ml/L i Fig 3 Currents and water mass characteristics oft the Kenyan coast ► shows direction of flow. N= North. S= South. No. 192 Page 39 Figure 4. Fig 4, Morphometric measurments taken on skate and ray specimens . 1 Total length TL 2. Disc length DL 3 Disc width DW Page 40 No. 192 Figure 5. KEY q) Rojo miraletus, b) Raja miraletus egg case c) Torpedo mormorata. d) Torpedo fuscomaculata. e) Heteronarce garmam f) Gy mnura natalensis g) Taemura 1 ym n a h) Dasyalis thetidis i) Dasyatis uarnak j) Dasyatis sephen k) Myliobotis aquilo. 1). Pteromylaeus bovinus m) Aetobatis nannari . n) Manta biostris No. 192 Page 41 Figure 6 FI9 6 length and weight histogram* of the skates and Rays studied along the Kenyan Coast KEY Tl = total length. Ol = disc length. 0W= disc width. Wts weight, (a; Rojo miraletus.(b) Myliobatis oquilo.(c) Toeniuro lymnq. (d) Dasyatis thetidis.(e) 0 uornak and(f ) D sephen Page 42 No. 192 Figure 7 Fig. 7. Mean monthly catch rates of rays caught and landed along the Kenyan c oast ( vert ical bars show the standard error)- Table 1. Students t-test for the differences in mean length and weight No. 192 Page 43 Table 1 Table 2 . length/weight relationship tor skates and rays. Page 44 No. 192 No. 192 Page 45 Table 3 Table 3. Data on trawl operations along the Kenyan coast by ‘R V Ujuzi' showing number of hauls, their duration, depth of trawl and monthly variation in catches of'Rays" Month Number of Hauls Duration ( min ) Depth range Total catch kg 1980 January 4 295 40 - 68 350 February 1 0 460 35 - 65 607 March 1 1 795 40 - 328 828 April 8 870 55 - 284 486 May 2 120 28 0 - 4 00 300 June — — — — July 2 1 20 235 - 345 80 August 15 1100 5 - 270 2188 September 2 150 1 4 - 60 48 October 1 5 11 38 14 - 70 1171 November 10 697 40 - 280 838 December 1 9 985 15-270 1406 1981 January 8 475 25 - 2 80 61 6 February 8 540 1 3 - 280 427 March 2 180 35 - 60 48 April 4 27 0 30 - 280 275 May 1 30 7-15 1 0 June 3 240 9-46 175 Table 4. Table 4. Trawling areas, mean density and biomass for ‘Rays’ along the Kenya coast. (Data from RV Ujuzi FAQ KEN/ 74/023 Project Reports 1979-81.) Area Trawling Areas Number of Hauls Mean density Biomass estimate n mi kg/nm ( tonnes ) 1. North Kenya Bank IN K. B) (01°39’s to 0 26 48 72 6 43 1770 128 5 Depth contour < 10 to<3000m 12 5 S 23 7 34.706 4608 2 Between N KB to Malindi Banks I 02°48’S ) to 1 03® 2 6 5,40° 4 6 E ] 1259 23 7 34.706 4608 3 Malindi Banks to Shimom Area ( 03°4 3s 4 0° 1 1 ’E ) to(04° 47s 39*48E) 1154 30 1138 1313 Page 46 No. 192 A NOTE ON THE FISHES OF LAKE JIPE AND LAKE CHALE ON THE KENYA-TANZANIA BORDER Stephen Dadzie1 2 3, Rene D. Haller1 & Ethelwynn Trewavas 5 Lake Jipe is a shallow basin at about 3° 40' S 37° 40' E, east of the North Pare Mountains in Tanzania. Chala is a smaller lake lying in a rocky crater about 19 km north of Lake Jipe. In 1951, when it was visited by Dr. Lowe- McConnell (then Miss R.H. Lowe) Lake Jipe was about 12 miles (19 km) long and 1 .5 miles (2 km) wide and only a few feet deep. Its northern end is a swamp into which flows a stream, known in Kenya as the Lumi, from Mount Kilimanjaro. From its northwestern end issues, at least in wetter periods, the River Ruvu4 headwater of the long river formerly known by that name, but now called the Pangani, the name of the town at its mouth on the Indian Ocean. Since the formation of the barrage lake Nyumba ya Mungu (NYM) in the upper Pangani the Ruvu flows from Lake Jipe into the swampy north-east comer of this lake. Lake Jipe’ s swampy edges are surrounded by semi-aquatic grasses and reeds, and patches of water-weeds ( Najas and Potamogeton ) spread their leaves and flowers on its surface. Numerous water birds prey on the lake’s fauna and these and hippopotamus fertilize the water. The sketch by Sir Harry Johnston, reproduced here from his book of 1886, gives an idea of the lake’s appearance as it remains today, (Fig. 1). At Taveta, NE of Lake Jipe, the well-known pioneer Colonel E.S. Grogan had settled after the Second World War and established a farm including fish-ponds stocked from Lake J ipe. Lake Jipe and Chala and the Taveta ponds were visited in 1951 by Miss Lowe, who was then a member of the East African Fisheries Research Organisation at Jinja, Uganda. She collected fishes from the ponds and both lakes. The two tilapiine species of Lake Jipe proved to have been undescribed. She then travelled on to Korogwe5 on the Lower Pangani, where Major R.E. Gould, then fish culturist to the Tanganyikan (Tanzania) Government, had established fish-ponds in which he reared tilapiine species preparatory to using them for stocking dams and ponds. Among his species were the two endemic tilapias from lake Jipe and two from the River Pangani and these were the subject of Lowe’s paper of 1955, where they were named Tilapia jipe and T.girigan (Lake Jipe) and T. pangani and T.mossambica korogwe (River Pangani). The last named was later given full specific status by Trewavas (1983) and all four were included in the genus Oreochromis. The generic name Oreochromis had been proposed for O.hunteri by Gunther in 1889, a name meaning ‘mountain Chromis’ in reference to its habitat in Lake Chala on the slopes of the Kilimanjaro volcanic mass. (Chromis is a name that had been used for several cichlids, but really belongs to a marine genus.) 1 Dept, of Zoology, University of Nairobi, P.O. Box 30197 Nairobi 2 Baobab Farm, P.O. Box 90202, Mombasa 3c/o British Museum (Natural History), London SW7 5BD, England 4 The Ruvu of Bailey et al., 1987, not of Bailey, 1969. The latter is also known as the Kingani and enters the Indian Ocean at Bagamoyo. 5 Not the Korogwe at the entry of the Ruvu to NYM (Bailey et al., 1978) No. 192 Page 47 The situation in 1951 and until recently was therefore that Lake Chala contained one endemic tilapiine species, O.hunteri, and Lake Jipe two, O.jipe and O.g.girigan. Lowe (1955) recognised that fry of O.girigan were difficult to distinguish from O.pangani and Trewavas (1983) considered these two forms to be only subspecifically distinct. One of us (R. Haller), has, since May 1976, been taking Oreochromis from Lake Jipe for culture at Baobab Farm, near Mombasa. In January 1983 a third tilapiine species was found in the Lake and a brood of young about 2.5cm long was taken to Boabob Farm and reared to maturity which they reached at an age of 9 months and a total length of 16- 1 8 cm. A sample of these, five of each sex, was sent to the British Museum (Natural History) [BMNH] and identified as Oreochromis esculentus (Graham). Another of us (S. Dadzie) visited the lakes in March- April, 1985, and found in Lake Jipe, as well as the two endemic Oreochromis and O.esculentus several specimens of Tilapia rendalli Boulenger. From Lake Chala he collected T. rendalli and a large specimen of O.p.pangani , as well as the endemic O.hunteri. We have no doubt that'O.esculentus, Tilapia rendalli and O.p.pangani were introduced to these lakes from the Tanzania side, either deliberately or in the case of Lake Jipe by migration up the Ruvu River. The first two are not native to the area and no tilapine except O.hunteri was previously known from Lake Chala although this species had been collected there and specimens sent to the BMNH on five occasions, registered in 1889, 1902, 1946, 1952 and 1980. Non-tilapiine fishes also present in Lake Jipe in 1951 (Lowe, 1951) were Astatotilapia bloyeti (Sauvage), Clarias mossambicus Peters, Barbus paludinosus Peters and Petersius tangensis Lonnberg. Astatotilapia bloyeti was found also in Lake Chala by W.P. Scott in 1977 and again by Dadzie in 1985. Specimens of the introduced species and the non-tilapiines were sent to the BMNH and identified by E.T. The Introduced Species Gould (1951) published a list of the species that he had imported for culture to Korogwe. Among them wer tO. esculentus andO.variabilis, species endemic in the Lake Victoria basin, and Tilapia rendalli6 and O.macrochir from ponds in Katanga (now Shaba). O.esculentus is a mouth-brooder, now common in several dams and lakes in Tanzania, including Nyumba ya Mungu (see Bailey et al., 1987 and Trewavas, 1983), which probably acquired it from stocked ponds and backwaters in the flooded area. For food it has a strong preference for phytoplankton in open waters. This was shown by observation in its original habitat, the Lake Victoria basin (for references see Trewavas, 1983), in experimental ponds Payne, 1971) and in Nyumba ya Mungu (Bailey et al. 1978). Its weak jaws and minute teeth are unsuited either for ‘Referred to as T. melanopleura , a name no longer in use. Page 48 No. 192 scraping algal growths from stems and leaves or for biting and chewing higher plants. In NYM, although it succeeded in establishing itself as a proportion of the tilapiine fauna, its growth in length was inferior to that of the species of local origin ( O.p.pangani and O.jipe ) and also to that in its original habitat. In NYM very few specimens exceeded 22 cm in total length and 350 g in weight, whereas males of O.jipe and O.p.pangani grew to between 32 and 38 cm with a maximum weight of 1800 g (Bailey et al., 1978, figs 3 & 4). A male 0. esculentus collected from NYM in 1980 by A.I. Payne was nearly ripe at TL 17 cm; this is about the same as the length of mature fishes reared at Baobab Farm for 9 months ( 1 6- 1 8 cm) and that recorded by Lowe ( 1 955) in ponds at Korogwe (16-19 cm at an age of under 7 months). In Lake Victoria the minimum length al first maturity was recorded as 19 or 20-21 cm at different localities, but most do not breed until they are 22 cm or more in total length (Lowe-McConncll; 1956, Garrod, 1959). They go on to reach TL of 30 cm or more. Tilapia rendalli, a substrate-brooder and guarder of the young, is one of the largest of its genus. It is a voracious feeder on aquatic and semi-aquatic higher plants. Ruwet (1962 and 1963) gives a vivid descriptions of its feeding in the barrage Lake Mwadingusha in the course of River Lufira, Shaba. The lake was shallow, mostly less than 4 metres maximum depth, and in dry seasons or when the rains were deficient the draw-down at the electric power station was so great that the lake shrank to a fraction of its maximum area and semi -aquatic vegetation was able to intrude over vast stretches. When rain again flooded the lake T.rendalli immediately attacked flooded grasses. Ruet describes the event as follows (translation from the account of 1963): “I have often halted in a canoe in the middle of the immense meadows dominated by Oryza, the wild rice. I still remember the continuous cracking sound produced by hordes of (T.rendalli) tearing, cutting and browsing on the flooded stems, leaves and rhizomes of Oryza. From mid-March, under the joint activity of thousands of these fishes, the edge of the meadows from Kinshasha to Shinangwa, that is, over a distance of about 15 Km, retreated by about 8-10 metres”. Dr G.B. Bemascek, in a letter dated 24. VI. 1980 reported that the introduction in 1951 of T.rendallia to a densely vegetated lagoon near the mouth of River Lupululu in southern Tanzania had resulted in its clearance to form a lake of open water. Such are the two species now introduced into Lake Jipe. The Endemic Tilapiines What is known of the ecology of the original inhabitants is due to Lowe (1955) and Bailey et al. (1978). They are mouth-brooders, with a characteristic form of nest (mating territory). Stomachs of O.p.girigan from Lake Jipe contained fragments of water-lily plants and epiphytic algae from Najas stems and leaves. The coarse teeth of this species are well suited to such a diet. Lowe’s few specimens of O.jipe from Lake Jipe did not provide information about its natural food. In the Korogwe was restricted to ponds both this and O.p.girigan fed on detritus, but O.jipe, finer particles a difference corresponding to the difference between them in the pharyngeal dentition. In NYM both O.jipe and the indigenous O.p.pangani fed on periphyton. No. 192 Page 49 The Future of Lake Jipe One of us (S. Dadzic) has embarked on a study of the reproductive regimes of the two endemic species of Lake Jipe and their ecological relationships. The problem that now faces him is more complex, involving the impact of the introduced species on the endemics. If O.esculentus is shown to feed on phytoplankton in L.Jipe, this will remove it from competition with either O.jipe or O.p.girigan. In Lake Victoria, before the decline of tilapia fisheries, O.esculentus was favoured as food over the sympatric species O.variabilis , both for the quality of its flesh and for its ability to grow to a good marketable size, qualities that were no doubt the result of its abundant and rich planktonic diet. There is evidence however that by 1980 in Nyumba ya Mungu. O.esculentus was breeding at a small size and failing to achieve it full growth potential. In 1974 O.esculentus comprised 24% of all the tilapiines in the commercial catch (Bailey et al., 1978:122). By 1980 this species constituted the major part of the catch, even in the shallow northern part of the lake; but those measured did not exceed 17.5 cm in total length and many of them were sexually mature (A. I. Payne, personal comm and in press). Dwarfing of O.esculentus in Lake Jipe would endanger the endemic species, because nets of mesh small enough to catch adults of O.esculentus would probably catch O.jipe and O.p.girigan before they had reached sexual maturity. O.esculentus may also compete with the native species for breeding sites. These are at present unknown and difficult to find in the opaque waters of Lake Jipe. Their presence in a locality may be inferred by the concentration of males in breeding condition. Nursery sites may similarly be located by the predominance of brooding females. Tilapia rendalli and O.p.girigan may occupy the same, rather specialized, trophic niche. A relevant question is, do they sub-divide this niche? For instance, does one species favour floating plants while the other uses grasses in the swamps? Their enemies in Lake Jipe are the birds and the catfish Clarias gariepinus with which all species must take their chance together. Finally, the two introduced species in Lake Jipe may prove to be successful competitors with the local species, T. rendalli by competition for food and O.esculentus by its strategy of progenesis and dwarfing, and possibly both by competition for breeding sites. There is danger that these may oust the native tilapias, providing instead an abundant but small O.esculentus and a Tilapia rendalli the quality of which cannot be assumed to exceed or even equal that of O.p.girigan. The Fishery Lake Jipe supports a considerable fishery. Ninety-six fishermen are registered and 56 of them fish regularly. Dugout canoes are the only crafts used and gillnets of stretched mesh 2" to3.5" are set, made of nylon thread of 2,3 or 4 ply. The usual catch consists of tilapiines and some Clarias. The same species arc caught by hook and line, using hooks of sizes 14, 16 and 18. Traditional traps (migono) are set at the shallow edges of the lake and catch all species, including a haplochromine. Page 50 No. 192 Sardine traps, also set at the shallow edges, catch exclusively the small Petersius (“sardines”) that form schools in the lake (Information from the Kenya Fisheries Department). At the smaller but deeper Lake Chala there are 7 registered fishermen of whom 2 operate regularly. Fishing in both lakes is for marketing as well as for subsistence. The fields arc therefore of legitimate concern, and a factual knowledge about these resources is relevant to the economy of the local people as well as to science. Pond culture The experience at Taveta of the Late Colonel Grogan and now at Baobab Farm demonstrates the suitability of the endemic species of Lake Jipe for pond culture. Tilapia rendalli has been used in ponds, lagoons and dams both for clearing excess vegetation and as a food fish. Both this and Oreochromis p.girigan feed naturally on macrophytes. As Payne (1971) showed with Tilapia zillii at Malya (75 miles south of Lake Victoria) plant eaters can efficiently use waste from crops. The problem with Oreochromis esculentus in ponds is that it fails to grow as well as in L. Victoria. Payne (1971) found that the growth of O. esculentus virtually ceased until he fertilized the pond with ammonium sulphate, to promote a bloom of phytoplankton. This resulted in the resumption of growth. The experiments covered only seven months, however, and the maximum total length of O. esculentus at the end of this period was 20 cm, weight 140 g (mean 17 cm and 75 g); and some reproduction was taking place. In L. Victoria Lowe-McConnell (1956) estimated a growth in length to 14-17 cm for O. esculentus in the first year and Garrod’s (1959) estimates were in agreement. In ponds and probably in Lake J. i, growth is brought to a halt by the onset of sexual maturity, which in Lake Victoria is delayed until the age of 2 or 3 years and a length usually of 22 cm or more. Thus the challenge for pond culturists of this species is the familiar one with tilapias - how to delay sexual maturity. This may be done either by design of ponds to discourage territory-establishement by males (Balarin & Flaller, 1983) or by hormonal treatment to produce 100% males by sex-reversal of the females. Both methods require specialist skills and capital investment. For a summary see Mires, (1983:600-610). Acknowledgements We are grateful to Prof. M. Hyder for encouraging S. Dadzie to pursue this project and to the Director and Trustees of the British Museum (Natural History) for permitting E. Trewavas to use the facilities of the Museum since her retirement in 1961. Baobab Farm is an enterprise of the Bamburi Cement Company. On S. Dadzie’s preliminary visit to lakes Jipe and Chala invaluable assistance was given by the Fisheries Department of Kenya, which provided transport from Nairobi to the research site and also the use of fishing nets, outboard engine, fuel and manpower in the field. Mr. David Burnley, a doctorate student from Duke University, Durham, U.S.A., who was carrying out research on the palaeocology of the two Lakes at the time, assisted with logistics and the data collection. The other expenses of the visit and the cost of some apparatus for further laboratory research were funded privately. No. 192 Page 51 References BAILEY, R.G. 1969. The non-cichlid fishes of the eastward-flowing rivers of Tanzania, East Africa. Revue Zool. Bot. afr. 80:170-199 BAILEY, R.G., S. CHURCHFIELD, T.PETR & R. PIMM, 1978. The ecology of the fishes in Nyumba ya Mungu reservoir, Tanzania. Biol. J. Linn. Soc. Lond. 10:107-137 BALARIN, J.D. & R.D. HALLER, 1983. Commercial tank culture of tilapia. International Symposium on Tilapia in Aquaculture. Nazareth, Israel, May 1983,473-483. (See also references therein). GARROD, D.J., 1959. The growth of Tilapia esculcnta Graham in Lake Victoria. llydrobiologia 12:268-298. [GOULD, R.E.] 1951. Progress of Fish farming in the Territory. Pamphlets Dep’t Agric. Tanganyika. No.2, 9 pp. JOHNSTON, H.H., The Kilima-Njaro Expedition. 1886. LOWE, R.H., 1955. New species of Tilapia (Pisces, Chichlidae) from Lake Jipe and the Pangani River in East Africa. Bull Brit. Mus.nat. Hist. Zool. 2:349-368 pis 13-17. LOWE-McCONNELL, R.H., 1956. Observations on the biology of Tilapia (Pisces, Chichlidae) in Lake Victoria, East Africa. E.Afr.Fish. Research Organisation. Supply. Publ. No. 1:72 pp. MIRES, D. 1983. A technical evaluation of Tilapia cultures, pp.600-610 in Fishelson, L. & Z. Yaron (eds), International Symposium on Tilapia in Aquaculture. Nazareth, Israel 8-13 May 1983. Tel Aviv University. PAYNE, A. I. 1971. An experiment in the culture of Tilapia esculenta Graham and Tilapia zillii (Gervais) (Cichlidae) in fish ponds. J. Fish Biol. 3: 325-340. PAYNE, A.I., R.I. COLLINS ON & S. LORD (in press). Composition, trophic structure and reproductive activity of the fish fauna of the Rufigi, an eastward flowing river of Africa. RUWET, R. - CL, 1963. Tilapia melanopleura (Dum.), poissons cichlides et la lutte contre la vegetation semi-aquatique au lac barrage de la Lufira (Haut-Katanga). Bull. Soc. R.sci.R.sci. Liege 32:5 16-528. TREWAVAS, E., 1983. Tilapiine Fishes of the genera Sarotherodon, Oreochromis and Danakilia. 583 pp. London, Br.Mus.nat.Hist. Published by The East Africa Natural History Society, Box 44486, Nairobi, Kenya. Typesetting by The Print Shop, Box 24576, Nairobi and Printed by AMREF Printing Department. Page 52 No. 192 Fig. 1. Lake Jipe. From a sketch by Sir Harry Johnston in his book The Kilima-Njaro Expedition (1886). JOURNAL^ OF THE EAST AFRICA NATURAL HISTORY SOCIETY AND NATIONAL MUSEUM December 1988 Volume 76, Number 193 FIG TREES (Ficus, Moraceae) OF KENYA H. J. Beentje East African Herbarium National Museums of Kenya ABSTRACT An account is given of the wild Ficus spp. of Kenya, with a key, descriptions, distribution maps, and line drawings. Preliminary paragraphs deal with the natural history of the genus, particularly with the pollination by Fig wasps (Agaonidae, Hymenoptera), the biotic community associated with Fig trees and the cultivated species found in Kenya. Introduction The Fig trees of Kenya are a relatively poorly known group, as there are no easy keys to identify the many species, and a full treatment of their proper scientific names and synonymy still has to appear (Berg, in press). The account in Dale & Greenway (1961) was for many years the only one treating the Kenyan species, and the keys were unsatisfactory. The present article in a precursor to the new “Kenya Trees, Shrubs and Lianas” which is in preparation at the East African Herbarium. This article is based on the study of the collections of the East African Herbarium, and on some fieldwork by the author. The information on the associated biotic communities was provided by Mr. G.R. Cunningham-van Someren. Natural History of the Fig tree Many species of Ficus start life as an epiphyte on other trees. Birds and mammals eat ripe figs and excrete the seeds, often in the crooks of branches and trunks of other trees; some of the seeds will germinate in such places, and if there is some moss or plant debris in such a place the young Ficus will root, and start its life far from the ground. It sends down roots along the trunk of the ‘host’ tree, and when these reach the ground (this process may take several years) they take root there, and the root system begins to thicken. Slowly the root system will envelop the trunk of the ‘host’ tree, the branches of the epiphyte spread through the canopy of the ‘host’ tree; finally the ‘ host’ tree dies, and the Ficus stands on its own. In most of the literature the dying of the ‘host’ tree is ascribed to strangulation by the roots of the Ficus. I think that in many cases the Fig tree will be the victor in the competition for water, food and light, due to its enormous root system in the ground and its much-branched and dense crown. However, Professor Comer (pers. comm.) empha- sizes that in the Eastern Tropics many species of Ficus jl<2 strangle their ‘host’ as their enveloping root- system expands. Not all Ficus species start life as epiphytes, and many of the so-called epiphytic species may also start life as terrestrials. It is unknown whether this is true foi all Ficus spp., or if there are really obligatory epiphytes. Page 54 No 193 Figure 1. Most, if not all, species of Ficus possess an extensive root system. This allows several species, which seem to be obligatory terrestrials, to grow in localities which are too dry for most other plants: on large rocks and on lava flows (e.g. Ficus cordata, F. glumosa, F. ingens, F. populifolia, F. wakefieldii). The root system penetrates the smallest of cracks where water might accumulate. There is also a possibility that Fig trees have a capability of picking up moisture from dew, mist, and moisture-saturated air at night. They probably have a large suction-force, enabling the tree to draw moisture from its enormous root system , as well as from the aerial root system which is often present This large suction force might be the reason why no parasites such as Loranthaceae seem to grow on Fig trees. Pollination The pollination system of Ficus is a beautiful example of a highly evolved symbiosis. The fig, or syconium, regarded by most people as a fruit, is a hollow inflorescence with the flowers, and later the fruits, on the inside. The only opening is at the top, and is called the ostiole; this ostiole is partly blocked by overlapping ostiolar bracts (Figure. 2). Most species of Ficus in Kenya are monoecious, i.e. male and female flowers occur together within a single fig; the following description applies to these species. For the dioecious species, with male and female flowers in separate figs, see page 56. In most monoecious figs, the male flowers are situated near the ostiole, and are few in number. The female flowers are of two kinds: roughly 50% of them have short styles, and the other 50% have long styles; ovules of these two types differ considerably (see Verkerke, 1987). The female flowers are many, up to 2800 in F. sur (Verkerke 1988), and the short-styled ones are intermixed with the long-styled ones (Fig. 2). In F. ottoniifolia Verkerke (1986) found that there is a whole range of style-lengths, and he considers style firmness rather than length the limiting factor for wasp oviposition. The pollination system consists of several phases: PHASE 1. The female flowers become receptive; the male flowers are still undeveloped and enclosed (protogyny). The ostiolar bracts open slightly. PHASE 2. Female fig wasps are attracted by chemical odours (Barker 1985) to their ‘own’ species of Fig Page 55 No 193 Figure 2. a. longitudinal section of Fig (schematic) b. long-styled female flower c. short-styled female flower tree, i.e. that species from whose Figs they emerged. They enter the Fig through the ostiole, a long and difFicult labour, even though their bodies are adapted to it. In the process, they lose their wings and the larger part of their antennae; some female wasps even get stuck and die (Galil 1977). In some species the number of female wasps to enter may be controlled by the rapid closing of the ostiolar bracts (Galil et al. 1973a). PHASE 3. The female Fig wasps try to deposit their eggs (oviposit) through the styles of female flowers. In the long- styled flowers the style is longer than the ovipositor of the wasp, and so the attempt fails. In the short-styled flowers it succeeds, as the ovipositor is long enough to deposit an egg between the inner tegument and the nucellus of the Fig flower ovary. As the wasps are very fertile, half of all female flowers (the short-styled half) may receive wasp eggs, while the other (long-styled) half is pollinated when the wasps try to oviposit through the long style. So half the flowers may produce seeds, and half the flowers may produce wasps. PHASE 4. End of the receptive (female) phase. The fig cavity is sealed by the closing of the ostiolar bracts; the female wasps die inside the fig. PHASE 5. The interfloral period, seems to be constant for each species of Ficus (Ramirez 1974). In the sealed Fig the C02 level rises, which inhibits the fig from ripening (Galil et al. 1973b). The wasp larvae develop; the ovaries in which the eggs are present develop into galls, grow in size, and produce nutritive tissue for the larvae. PHASE 6. After a period of between 20 and 100 day the male fig wasps emerge from their galls; they are wingless and look quite different from the female wasps. They seek out the galls with the female wasps, copulate with the females while these are still in the galls, then move to the apical side of the Fig and bore and bite their way out of the fig, making one or more holes near the ostiole. The male wasps die soon afterwards. PHASE 7. Once the Fig is holed, the C02 level drops, which gives an impulse to the female wasps to emerge from their galls. They move towards the hole made by the male wasps, and on their way encounter the male flowers, which are by now fully ripe. They take pollen from the anthers and deposit it in special pollen pockets or groups of hairs on their thorax. These pollen pockets are closed with a movable lid (Galil & Eisikowitch 1974) so the pollen is not lost when the females crawl through the hole made by the male wasps. The female wasps then fly away to search for a Ficus with figs in phase 1 or 2, to deposit their eggs. Page 56 No 193 PHASE 8. As the C02 level now has dropped, the fig ripens, becoming soft and juicy; mammals and birds eat the fruit and later excrete the seeds, thereby dispersing them. In dioecious figs (In Kenya Ficus asperifolia, F. capreifolia and F. exasperata ) there are two types of figs. “Female” figs contain short-styled as well as long-styled female flowers; “male” figs contain short- styled female flowers as well as male (staminate) flowers. Female wasps emerging from their figs will enter either "female" figs, where they wil deposit their eggs in the short-styled flowers or they will enter “male” flowers, where they will do the same. Only young female wasps emerging from “male” figs will carry pollen to the fig they enter, and only if these pollen-carrying wasps enter “female” figs pollination can be achieved. It will be seen that the system with monoecious figs offers four times as many chances of pollination as the system with dioecious figs! To insure cross-pollination, the pollen-carrying female wasp emerging from a fig must be able to find a fig in phase 1 or 2, and so Fig trees must have staggered flowering seasons. In Nairobi, on 19 February 1986, six trees of Ficus thonningii were observed; three of these did not carry figs, one carried young figs, one carried figs from which female wasps were just emerging, and one carried over-mature figs. These trees were within an area of one square kilometre. Newton & Lomo (1979) observed neatly staggered flowering periods of Ficus lutea in Ghana; individual trees showed one or three flowering periods per year. In most individual Fig trees, all figs seem to be at roughly the same phase at the same time; however, I have observed trees of Ficus natalensis and F. sycomorous with only a few (not overmature) figs. The relationship between Fig tree and fig wasp is highly specific; no hybridization between figs is known. Wiebes (1979) thinks that there are 900 species of Ficus, «ach with its own species of fig wasp. Associated biotic communities Mammals eating ripe figs include several species of fruitbat (e.g. Epomops, Micropterus, Eidolon, Rousettus ), monkeys (e.g. Vervet), Baboon, Tree Hyrax, several species of squirrel. Potto, Bushbaby, Nandi Cat, and possibly genet, civet and mongoose. Fallen fruits are eaten by Bushbuck, duiker, Suni, Bush Pig, porcupine, and small rodents. Birds eating ripe figs include hombills (Silvery-cheeked, Black and White Casqued, Trumpeter, Crowned, and smaller species), most species of Turaco, pigeons (especially G^ -\, also Speckled), parrots, lovebirds ** ..u^is (many species), mousebirds, orioles, starlings (particularly the Violet-backed, which moves around following the fruiting of Ficus thonningii and F, natalensis). Yellow- vented Bulbul, several species of greenbul and thrush. Fallen fruits are eaten by francolins and several species of ground dove (Emerald-spotted, Tambourine). Woodpeckers and barbets excavate nest holes in the soft wood of Fig trees, and these holes are also used by wood hoopoes; many parasitic honeyguides lay their eggs among eggs of these species. As regards insects, several wood-boring beetles, particularly the long-homed Cerambicidae attack the wood of Fig trees. Caterpillars of moths ( Arctiidae , Eupterotidae.Lymantridae ) feed on the foliage, and are eaten by orioles, cuckoo and cuckoo- shrikes. Caterpillars of the Fig Blues, the butterflies Myrina silenus and M. dermaptera (Lycenidae) feed on the leaves. Foliage of Fig trees is often infested by scale insects (Coccidae) which are eaten by the smaller honeyguides and by many species of small warblers and sunbirds. Lichens which grow on the bark of Ficus thonningii and F. natalensis harbour great numbers of insects, caterpillars and especially spiders, which provide food for tits and the Brown-capped Weaver. Fallen and fermenting figs may attract numbers of butterflies {Char axes spp., Euphedera spp., and Melinis spp.) as well as many species of flies. Fig trees often provide support for many kinds of epiphytic plants, such as ferns and orchids. Page 57 No 193 Mr.Cunningham-van Someren observed a Fig tree in Kakamega with no less than 15 species of orchids. As mentioned on page 54, parasites do not seem to grow on Fig trees. Uses and cultural significance Under the descriptions of the species, the specific local uses are given, as recorded on herbarium labels. Some general uses, valid for several or many species, follow. Because of their soft wood, Fig tree logs are easily hollowed out and can then be used as bee hives, to be hung in trees. The Fig tree is often used as a shade tree, and for this reason is often left standing when land is cleared for agriculture. In many cultures, within as well as outside Kenya, Fig trees are considered as being special or even sacred. The Sycomore ( Ficus sycomorus) is mentioned several times in the Bible (e.g. 1 Ki. x: 27). This species was sacred to several ancient Egyptian gods, especially to Hathor, the goddess of love, and figs of this species have been found in a number of tombs dating back to the first Dynasty (Galil, 1967). Ficus religiosa is held sacred by Hindus and Buddhists: The Buddha received his enlightenment under this tree. In Kenya the Mugumo (Ficus thonningii and F. natalensis ) is venerated by the Boran, Maasai, Kiembu, Kikuyu, Kimeru and Kitaita. Meetings of the elders are often held under this tree, and cutting or damaging of such trees is strongly discouraged. The mukuyu (F. sur, F. sycomorus ) is venerated by the Kimeru and Kikuyu as the protector of springs, and several legends are associated with it (Salvadori, pers. comm.). Wood of Ficus sur is used for ritual fires during youngsters’ circumcision by the Kimeru; Maasai use the latex of this species to protect their cattle from epidemics (Salvadori, pers. comm.) Cultivated species In Kenya the following species are cultivated: F. benjamina L., a tree from India and Malaya. F. carica L., the edible Fig, originally from the Mediterranean area but now cultivated all over the world. Cultivation in Kenya has not been very successful, but in very hot areas (e.g. Garissa) this species might be a good producer of marketable figs. F. deltoidea J ack, here a shrub, but normally an epiphyte. Only recorded from City Park, Nairobi; originally from Malaya and Indonesia. F. elastica Roxb., a tree from South and Southeast Asia, formally widely cultivated for its rubber. F. macrophylla Desf., a tree from Australia recorded from the Nairobi Arboretum. F. pumila L., a climber from China and Japan (syn. F. repens Rottl.) Leaves on sterile branches are quite different from those on flowering and fruiting branches. F. religiosa L., the Bo-tree or Peepal. A large tree from India. F. vogeliana (Miq.) Miq., a tree from West Africa recorded from the Nairobi Arboretum. Page 58 No 193 These cultivated species may be distinguished as follows: Climber with dimorphic leaves Trees or shrubs F. pumila. 2 Leaves deeply lobed Leaves entire F. carica 3 Leaves rounded at apex, small; midrib forked Leaves acute or acuminate; midrib straight F. deltoidea 4 Leaf apex acute or subacute Leaf apex acuminate or caudate F. vogeliana 5 Leaf apex long-caudate; base rounded or subcordate Leaf apex acuminate F. religiosa. 6 Petiole ca. 1 cm long; leaves less than 9 cm long Petiole more than 2 cm long; leaves more than 10 cm long F. benjamina 7 Leaves white or brown beneath Leaves greenish beneath F. macrophylla F. elastica Cultivation of fig trees Most Ficus species will grow from large cuttings planted as the beginning of the rains. Indigenous spe- cies such as F. thonningii, F. natalensis and F. sycomorus are widely planted. One should never plant Fig trees close to houses, as their root systems will crack walls and raise flagstones. Key t q th1 i £i 3 Q "t Hi o § 3 NJ o a. N> < 3 o’ 9 p P § 3 O tr c P o J. E.A.N.H.S. & Nat. Mus. 79 (194) December 1989 Figure 8: view from the west rim of the crater looking eastwards across the saddle; traced from a slide taken by Alan Fowler on 26 July 1989. The width of overflow (Ml to M2) has not increased significantly and the boulder B1 is still visible. Youngest flow would seem to be F13. Dark line at x may mark persistent fumaroles. Diameter of lava patch to south (right) of saddle M1M2 is ca. 120 m. Celia Nyamweru • Ol Doinyo Lengai July 1988 - Aug 1989 13 J. E. A. N. H. S. & Nat. Mus. 79 (194) December 1989 Celia Nyamweru • Ol Doinyo Lengai July 1988 - Aug 1989 14 J E. A. N. H. S. & Nat. Mus. 79 (194) December 1989 Figure 10: aerial view from the northeast, taken by Lester Eshelman on 23 August 1989. The features visible in Fig. 9 can all be seen with little or no change. The darkest feature is the pinnacle of T13; this may be partly due to shadow as well as to fresh spatter of lava. Diameter of crater floor is ca. 230 m. Summit rises ca. 96 m above level of crater floor. Celia Nyamweru • Ol Doinyo Lengai July 1988 - Aug 1989 15 J. E. A. N. H. S. & Nat. Mus. 79 (194) December 1989 Celia Nyamweru • Ol Doinyo Lengai July 1988 - Aug 1989 i - - . ' N. H. S. & Nat. Mus “70 rWAmlv»r 1 QRQ l ■ ■ - JOURNAL OF THE EAST AFRICA NATURAL HISTORY SOCIETY AND NATIONAL MUSEUM June 1990 Volume 80, No. 195 AN ECOLOGICAL CHECKLIST OF THE PLANTS OF KIBOKO NATIONAL RANGE RESEARCH STATION, KENYA NDEGWA WA NDIANG’UI* Kiboko National Range Research Station * presently Research Fellow, East African Herbarium National Museums of Kenya P O Box 45 166 NAIROBI - Kenya ABSTRACT The flora of the Kiboko National Range Research Station was systematically sampled and collections made in each major ecological site. Belt transects were laid through the sites to serve as reference points for the collection of species during all seasons of the year. Representative specimens were collected and identified. The relative abundance, habit and longevity of each species was recorded in the field and this has been compiled to form a checklist. The most widespread families on the research station in order of prominence were Gramineae (Eragroslis, Digitaria, Chloris, Brachiaria, Setaria), Leguminosae (Acacia), Compositae, Burseraceae (Commiphora), Combretaceae (Combretum), Capparaceae, Euphorbiaceae, and Acanthaceae. INTRODUCTION Knowledge of the flora is fundamental to understanding the ecology of any region. In East Africa, several isolated studies on the plants have been conducted with various objectives. Dale & Green- way (1961), presented a pertinent account of the woody plant species in Kenya, while Lind & Morrison (1974) showed the extent to which the physiognomy of bushland is dependent upon the species comprising the woody vegetation component and the presence or absence of perennial grasses. More recently Coe & Collins (1986) gave an ecological inventory of the Kora National Reserve containing an annotated checklist of plants of the area, while Blundell (1982, 1987) dwelt on selected wild flowers of Kenya and East Africa. Since about 80% of Kenya is rangeland, there is a need to have a well documented account of the rangeland plants therein, as they are important resources for development. Range management, for instance depends primarily on a knowledge of the plants, their ecophysiology, forage value, poisonous properties, distribution and associated habitat. A floristic study of the potentially productive and unproductive range areas of the country with the objective of documenting the occurrence of forage species as well as poisonous species is necessary. Such an undertaking would benefit from an ecological checklist of the species in ecoclimatic zones four and five, as delineated by Pratt, et al. (1966). It was the purpose of this study to contribute to this data base for the Kiboko Range Research Station whose research objective in range management is to increase and sustain livestock production. The results will benefit both basic and applied research in range management in these zones. Ndegwa wa Ndiang'ui • Plants of Kiboko 2 DESCRIPTION OF THE STUDY AREA The study was conducted near Kiboko on the Kenya Government National Range Research Station which is located approximately 160 km from Nairobi, on the Nairobi/Mombasa road (Fig. 1). The station land area is approximately 30,000 h (74,000 acres). The station altitude is approximately 1000 m (3,280 ft) above sea level, and lies between latitude 2° 10' S and 2* 25' S, and longitudes 37° 40’ E and 37' 55’ E. Climate The climate of Kiboko falls under the influence of the Intertropical Convergence Zone (Whyte, 1968), characterised by a bimodal distribution of wet and dry seasons. The months of January and February characterise the short dry season, followed by a long rainy season from March to May. June to October is usually the harsh, long dry season which is followed by a rather short wet season from November to December. It is important to mention the rather off-season rains that tend to fall J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndcgwa wa Ndiang'ui • Plants of Kiboko 3 a few weeks before the expected rainy periods. These trigger early flowering and leaf formation in some woody species, as well as seed germination for the early emerging ephemerals (C. G. Gakahu, personal communication). Rainfall data for the study area over the past 43 years conform to the long-term annual average of around 600 mm (Michieka & von der Poun, 1977). Most of these data accrue from the well- established Makindu Meteorological Station. The mean maximum annual temperature for the Makindu Meteorological Station is 28.6°C and the mean minimum temperature is 16.5°C. The long-term annual average temperature is around 23°C. The long-term annual average evaporation is 2000 mm, the mean dew point is around 15.7°C and mean relative humidity is 62.5% (Michieka and von der Poun, 1977). Geological summary A large portion of this area is classified as the Basement System gneisses, and approximately one- third is covered by young lava flows of recent deposition (Michieka and von der Poun, 1977). The latter are still in the process of degradation and the soil tends to be shallow and undeveloped, rather rocky and very permeable. Michieka & von der Poun (1977), defined three major regions which can be identified in the study area with regard to the general geomorphology: (i) the area underlain by Basement System rocks of the semi-calcareous group, (ii) the area underlain by Basement System rocks of the middle, semi-pelitic group, (iii) the area consisting of recent lava flows, associated volcanic vents, cones and recent limestones. Soils of the study area The soils of the Kiboko Research Station may be broadly categorized on the basis of the gcomor-phological and geological history of the area, where influence of the landscape and geol- ogy are clearly indicated as follows (Michieka & von der Poun, 1977): ( 1 ) Soils of hills and footslopes These are on the volcanic cones of Mwailu, Duani, Dojini and Wikiamba among others (Fig. 2). These soils are predominantly shallow, well drained, black to very dark greyish brown, friable and include soils developed over pyroclastic materials. These soils may also be moderately to strongly calcareous. They are generally referred to as regosols. (2) Soils of lava flows These include both lithosols and rock-outcrops. They tend to be well drained, very shallow, black to very dark greyish brown, stony to very rocky. Some areas also have silty clay derived from “pumice” deposits, with extremes of rocky, irregular mcsorclief of many sink holes, piles of olivine basalt rocks and collapsed tunnels. (3) Soils of floodplains and bottomlands These soils generally developed on olivine basalt and include the chromic vertisols that are imperfectly drained, very deep, dark grey, firm, cracking, and moderately calcareous silty clay to clay. These form the floodplain along the Kiboko river basin, which is the northern border of the station, in Association 15 (Fig.2). Those soils in this category that have developed on recent alluvial deposits include calcic chernozem, which tend to be moderately well drained, deep, dark brown to greyish brown and strongly calcareous clay. These are found in the area bordering the private land along the Makindu river course, which is in Association 17 (Fig. 2). The imperfectly drained calcic cambisol tend to be very deep, dark grey, firm and extremely calcareous clay with pockets of soft lime. These are the soils near Makindu township (Fig. 2). J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Figure: 2 Plant associations of Kiboko National Range Research Station 1-4 Vegetation of Volcanic hills, foot slopes and lava flows 5-14 V egetation of basement system plains 15-17 Vegetation of floodplains and bottomlands Ndegwa wa Ndiang’ui • Plants of Kiboko 4 J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang'ui • Plants of Kiboko 5 The vertisols in this category are imperfectly drained, deep dark grey to black, firm cracking strongly and calcareous. These tend to occur in various pockets in the southern ranch, and also have various narrow arms stretching over this part of the study area down to Mbuinzau (Fig. 2), supporting a characteristic vegetation of Acacia drepanolobium Sjostedt and Pennisetum mezianum Leeke also Echinochloa haploclada (Stapf) Stapf. Other types of soils in this category include those vertic and calcaric fluvisol types that tend to be imperfecdy drained, very deep, strongly to moderately calcareous, firm sandy clay to silty clay sometimes cracking. These occur along the course of the Kiboko river. (4) Soils of basement system plains These soils cover the largest portion of the Research station. They form the bulk of the grazing blocks on both the northern and southern ranches. Within this rather broad classification occur those soils developed on relatively undifferentiated basement system rocks; predominantly banded gneisses. The ferrosol type, well drained, deep, with colours ranging from dark reddish brown to dark brown and yellowish red, friable to firm clay and sandy clay, come under this category. Around Shoats* boma 1 stretching down to the south east near Shoats boma 3, occurs a luvisol with characteristic red to dark reddish brown firm sandy clay to clay with topsoil of loamy sand. Bordering this soil type is a stretch of nitosol characterized by well-drained deep dark red friable sandy clay to clay which stretches to the Mombasa road, and along the bottomland soils of the Makindu river course. The soils are said to have developed on basement system rocks rich in fer- romagnesian minerals (Michieka & von der Poun, 1977). The Five-Comers area (Fig. 2) has a type of luvisol which also covers the flat expansive area to the south of Mwailu hill and patches are found to the west and southwest of Duani volcanic cone. These tend to be well-drained, deep reddish brown to dark red. They support predominantly Acacia tortilis (Forssk.) Hayne and Digitaria macroblephara (Hack.) Stapf, the dominant grass of the understory. (5) Soils under swamps These constitute a very small portion of the station, they support predominantly Pennisetum massaicum stapf as well as other grass species around the edges of the water. Plant associations The vegetation of the study area was divided into three broad categories by Michieka & von der Poun (1977), encompassing 13 plant associations as delineated by Langat, et al. (1975). These had to be further sub-divided into the now recognized 17 plant associations (Fig. 2), due to evidence of further vegetation changes, mostly arising from overgrazing and accidental fires. These factors have tended to change the rate of succession by affecting species composition, as compared to conditions in 1973-1974 when the last vegetation mapping of the study area was completed by Langat, et al. (1975). The categories have been assigned according to physiographic features and each plant associa- tion has been designated with a number as follows: A. THE VEGETATION OF VOLCANIC HILLS, FOOTSLOPES AND LAVA FLOWS The hills (volcanic cinder cones) include Mwailu, Duani, Dojini, Wikiamba (Fig. 2). Association No. 1: Chrysopogon plumosus -Sehima nervosum association on the hills and lower slopes. The association is relatively open grassland not greatly invaded by shrubs. Association No. 2: Sehima nervosum-Digitaria macroblephara - Grewia bicolor association * sheep & goats I. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang’ui • Plants of Kiboko 6 on lava flows. This type of lava flow shows some weathering of the rocks, which tend to be small and scattered, enabling some soil formation to take place. Association No. 3: Acacia brevispica - Commiphora baluensis association is on shallow, highly permeable volcanic depositions. This type of vegetation occurs on very rocky rough terrain with large volcanic boulders. The trees and shrubs tend to be well-rooted between the rocks. Grass cover, if any, is very low, usually less than 10%. Association No. 4: Bushland thicket, which is relatively undifferentiated into specific associa- tions. This tends to be a rather complex bushland which is composed of various shrub species that are well-rooted in the rock crevices. Dominants are Acacia brevispica, Harms various species of Maerua, and Asparagus falcatus L. The terrain has little or no grass cover due to a lack of soil that would support the grass root system. B. THE VEGETATION OF BASEMENT SYSTEM PLAINS This type of vegetation covers broader categories of associations and is the most Widespread and diverse on the Kiboko Station. It covers over half of the grazing blocks in the northern ranch and also over half of the total area of southern ranch. The dominant associations are as follows: Association No. 5a: Digitaria macroblephara - Acacia tortilis - Dousperma kilimandscharicum - Commiphora africana - Commiphora mildbreadii . This type of association is predominant around cattle bom a 4 on the southern ranch as well as the region between the Makindu river and the wildlife unit. This tends to be densely bushed grassland where burning would reduce the shrub species and upgrade the grazing value of these blocks. Association No. 5b: Digitaria macroblephara - Acacia Senegal - Grewia bicolor - Acacia tortilis - Acacia mellifera - Commiphora africana. In this type of vegetation, D. macroblephara and Bothriochloa insculpta (A.Rich.) A. Camus are the dominant grasses, whereas the overstory species include C. africana (A.Rich.) Engl., A. Senegal (L) and A. tortilis. The remaining are understory species. This association forms the bulk of grazing blocks both in the northern and southern ranches. There has been a change in dominance in the northern ranch close to Shoats boma 1 where D. macroblephara has been replaced by Eragrostis caespitosa Chiov. on the more closely grazed areas extending eastward toward the Makindu river. Association No. 6: Sehima nervosum - Commiphora mildbreadii - Acacia tortilis. This association intergrades with No. 5a, but S. nervosum (Rottler) Stapf has taken over the understory dominance leaving Digitaria milanjiana (Rendle) Stapf in small patches. The higher canopy also includes some A. Senegal. This association is evident in the grazing block between cattle boma 4 and boma 5 in the southern ranch. In the latter, close grazing tends to cause Eragostis caespitosa to replace D. milanjiana and Themeda triandra Forssk. which were dominants on the vegetation map of 1975 Langat, et al (1975. Association No. 7: Chloris roxburghiana - Acacia Senegal - Commiphora mildbraedii. In this association C. roxburghiana SchulL dominates the understory whereas, the overstory is shared by the woody species. This association occurs to the east of the road from boma 7, on the northern ranch, extending towards and near the Mombasa road; and on both sides of the road to boma 4 from the Mombasa road on the southern ranch, where it tends to be more wooded with taller C. mildbraedii Engl. Association No. 8: Cymbopogon pospischilii - Digitaria macroblephara- Commiphora mildbraedii. This forms sparsely bushed grassland over part of the northern ranch near the location of cattle boma 3. This is the only area where C. pospischilii (K. Schum.) C.E. Hubbard occurs in substantial amounts. The association extends towards the lava flow in its western portion where Sehima nervosum association forms a very abrupt ecotone. Association No. 9: This is a rather isolated island of Cynodon dactylon - Eragrostis superba - Adansonia digitata - Acacia tortilis. This vegetation island is about 60 h, and surrounded by lava thicket on all sides, and has a deep red clay to sandy clay soil. The herbaceous species include various legumes, notably Indigofera malindiensis Gillett and Crotalaria polysperma Kotschy. J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang'ui • Plants of Kiboko 7 Association No. 10: Bushland dominated by Cenchrus ciliaris L. in the undcrslory layer. Acacia mellifera (Vahl) Benth. at the shrub layer and A. tortilis in the overstory. It is located in the northern ranch, near Kiboko. A large portion of this area has been disturbed by clearing for a new nursery site, and Cynodon dactylon has tended to take over. The undisturbed areas still retain Cenchrus ciliaris in the understory. Association No. 11: An association of Aristida kenyensis - Cynodon dactylon - Commiphora africana - Combretum apiculatum forms part of the bushland next to Mbuinzau Market. C. dactylon (L) Pers. occurs in patches and overgrazing has tended to favour A. kenyensis , an annual grass that sheds seeds quickly after maturity leaving the area bare for the dry season. Bushland in these areas is relatively thick and in some cases impenetrable. Grass cover has greatly diminished here. Association No. 12: On the northern ranch an almost pure grassland association of Sporobolus fimbriatus Nees. - Cenchrus ciliaris surrounds the seasonal salt lake which forms next to the Mombasa road (between Kiboko and Makindu). Acacia tortilis occurs in patches close to the edges of the pond, where grass cover is very high. Pennisetum massaicum Stapf is also a prominent species. Cynodon dactylon tends to occur as one moves further away from the pond, where as Chloris roxburghiana forms part of the ecotone with other associations. Association No. 13a: A larger portion of the southern ranch is covered with Themeda triandra, and Chloris roxburghiana, Combretum apiculatum Sond. T. triandra and C. roxburghiana domi- nate the understory and C. apiculatum dominates the overstory. Illegal grazing on the area close to the Mombasa road has been severe and the species composition has changed to become predomi- nantly annuals, and A. mellifera with other shrubs of low grazing value. A. kenyensis tends to be the main grass cover, but only during the wet season. Association No. 13b: Aristida kenyensis is the understory dominant with a complex array of shrubs. Commiphora mildbraedii is present in this association due to uncontrolled grazing. T. triandra and C. roxburghiana have been grazed out and replaced by A. kenyensis and various other annual weed species, most of which sprout at the onset of the rains and persist only during the wet periods of the year. This association tends to be confined closer to the area near Mbuinzani. Association No. 14: A complex bushland dominated by Chloris roxburghiana and Digitaria macroblephara in the understory and Acacia tortilis and Grewia similis , in the shrub layer. Commiphora africana and A. Senegal dominate the overstory. This association covers a larger portion of the southern ranch and is relatively inaccessible. It has some of the longest transects laid through it and part of the bushland thicket is relatively undifferentiated. C. VEGETATION OF FLOODPLAINS and BOTTOMLANDS Association No. 15: This is classified as bottomland silty clay site which occurs along the course of the Kiboko river near Kiboko. Pennisetum mezianum dominates the vegetation with occasional annuals especially during the rainy season. Association No. 16: This vegetation type appears in isolated pockets in both the northern and southern ranches where black-cotton soil supports predominantly Echinochloa haploclada. Waterlogging is common during the rainy season due to poor drainage of this soil. Acacia drepanolobium is frequent especially where the soil is shallow. Association No. 17: This association occurs only along the Makindu river where Acacia xanthophloea Benth., and Phoenix reclinala Jacq. are the dominants. Cyperus alternifolius L. is common on the rather swampy habitat in this association. J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndcgwa wa Ndiang'ui • Plants of Kiboko METHODS Plant associations in the field were sampled systematically, such that choice of the location and number of transects depended upon the amount of variation evident on a given site. Each plant association sampled had at least two transects across it, whereas the maximum number of transects through a single plant association with great diversity was eight. The length of each transect depended on the relative distance from the centre of each association to the edges. The shortest transect of all was about 80 m long whereas some transects were up to 3.5 km in length. A total of 1 10 transects were laid in the 17 plant associations, throughout the study area. About 50 transects were spread over the southern ranch while the rest were in the northern ranch. The Five-Comers area (Fig. 2), where grazing management studies are being conducted, had various transects that were employed but not considered as part of the total number of transects. Compass bearings were used to mark the orientation of each transect. White painted wooden stakes (made of cedar to prevent destruction by termites) were placed along the length of each transect marking its alignment. In addition to the stakes, red ribbons (flags) were tied on trees that occurred along the course of each transect for easy orientation of the direction of the transect A wooden stake was placed at the center point of each plant association. Plants occurring along each transect were collected starting from the centre point to the end of the transect. A note was made regarding the longevity (annual or perennial), growth habit (herb, shrub, tree, epiphyte, climber) and relative abundance (rare, common, abundant) of each species encountered. A species was regarded as rare if it occurred very infrequendy along the transect, for instance if when walk- ing along the transect a species was seen only once or twice, with little evidence of its prominence within the association. A common species was one that was encountered frequently along all the transects within the association. The category of abundant was accorded those species that were very prominent within the association. Frequently, those species regarded as abundant had contrib- uted to the delineation of the vegetation into the different associations. The sites were revisited at two weekly intervals during the growing season to ensure the collection of any newly emergent species. Voucher specimens were deposited in the S.M. Tracy Herbarium (TAES), Missouri Botanical Garden (MO), Nairobi University Herbarium (NAI) and the East African Herbarium (EA). An ecological checklist was compiled in the following format. Families are listed in alphabeti- cal order. Within each family the genera are arranged in alphabetical order as are species within each genus. RESULTS The vegetation that develops on the Basement System plains generally is characterized by presence of species of Acacia, Commiphora, Combretum and grasses. The grasses may include species of Eragrostis, Digitaria, Chloris, Themeda and Panicum. These genera may succumb to disturbance and Arislida may become prominent with a host of other annuals as a result. Pratt & Gywnne (1977) point out that Combretum is a dominant tree genus of ecoclimatic zones III and IV. Its prominence over a large portion of the southern ranch of the station in asso- ciation with Panicum maximum Vacq. indicates a region of slightly higher in rainfall than the rest of the station. This part of the station has been regarded as falling in ecoclimatic zone IV. The location of the research station offers unique combinations of vegetation types. The lava flows, covering about 30% of the station land area, have pockets of soil that support such species as Hyparrhenia hirta (L.) Stapf, Pennisetum setaceum (Forssk.) Chiov. and P. stramineum Peter, which arc expected in higher rainfall areas in ecoclimatic zone III. Commelina africana L. var. boehmiana (K. Schum.) Brcnan, which was found in association with Cenchrus ciliaris. Acacia mellifera and A. tortilis (Association 10), is a new record from this part of Kenya as indi- cated by records at the East African Herbarium, Nairobi. J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang'ui • Plants of Kiboko 9 A summary of the number of species within each family on the station (Table 1) showed Gramineae with the highest number (90 species) followed by Leguminosae (47 species) Euphorbiaceae and Compositae with 20 species each, and the other families represented by fewer species. 27 families were represented by one genus and one species only on the station. The shrub component of the vegetation represented 18.7% of the total number of species (Table 2) whereas the non-grass herbs constituted 35.4%. The number of tree species was 12.2%, w hich was half that of the grasses. Only one epiphytic species was recorded on the research station Plicosepalus sagitlifolius (Sprague) Danser. These results indicate a high species diversity reflected on a small area, as shown also by Ndiang’ui (1984), which is characteristic of the vegeta- tion in the tropics under this type of climate. Table 1 . Summary of the number of genera and species within each family found on the Kiboko National Range Research Station. Family No. of genera No. of species %of Total species Acanthaceae 11 14 3.7 Aizoaceae 2 2 0.5 Amaranth ace ae 6 8 2.1 Anacardiaceae 3 5 1.3 Araceae 1 1 0.3 Asclepiadaceae 4 4 1.1 Balanitaceae 1 1 0.3 Bignoniaceae 1 1 0.3 Boraginaceae 3 5 1.3 Burseraceae 1 4 1.0 Campanulaceae 2 2 0.5 Capp araceae 5 9 2.4 Caricaceae 1 1 0.3 Celastraceae 1 2 0.5 Chenopodi Iceae 1 1 0.3 Combretaceae 2 5 1.3 Commelinaceae 2 4 1.0 Compositae 15 20 5.3 Convolvulaceae 2 6 1.6 Crassulaceae 1 1 0.3 Cruciferae 1 1 0.3 Cucurbitaceae 4 6 1.6 Cyperaceae 3 11 2.9 Ebenaceae 1 1 0.3 Euphorbiaceae 12 20 5.3 Geraniaceae 1 1 0.3 Gramineae 49 90 24.0 Icacinaceae 1 1 0.3 Labiatae 5 9 2.4 Leguminosae 22 47 12.5 Liliaceae 6 9 2.4 Lobeliaceae 1 1 0.3 Loranthaceae 1 1 0.3 Malpighiaceae 1 1 0.3 Malvaceae 4 12 3.2 Meliaceae 2 3 0.8 J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang'ui • Plants of Kiboko 10 Family No. of genera No. of species %Of Total species Menispermaceae 1 1 0.3 Moraceae 1 1 0.3 Nyctaginaceae 1 2 0.5 Ochnaceae 1 1 0.3 Olacaceae 1 1 0.3 Palmae 1 1 0.3 Passifloraceae 1 1 0.3 Polygalaceae 1 1 0.3 Polygonaceae 1 1 0.3 Portulacaceae 1 1 0.3 Rcsedaceae 1 1 0.3 Rhamnaceae 1 2 0.5 Rubiaceae 4 4 1.0 Rutaccae 1 1 0.3 Salvadoraceae 2 2 0.5 Sapindaceae 2 3 0.8 Scrophulariaceae 1 1 0.3 Solanaceae 3 6 1.6 Stcrculiaceae 4 6 1.6 Tiliaceac 3 8 2.1 Tumcraccae 1 1 0.3 Umbelliferae 1 1 0.3 Verbenaceae 7 11 2.9 Vitaceae 2 6 1.6 Zygophyllaceae 1 1 0.3 TOTAL 223 375 100.0 Table 2. Summary of the growth habits of species found on the Kiboko National Range Research Station. Growth habit No. of species % of Total species Grasses 90 24.1 Trees 46 12.2 Shrubs 70 18.7 Non-grass herbs 133 35.4 Climbers 35 9.3 Epiphytes 1 0.3 J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang'ui • Plants of Kiboko 11 THE CHECKLIST ACANTHACEAE Anisotes ukambensis Lindau S RARE Asy stasia charmian S. Moore AH COMMON Barleria eranthemoides C.B.C1. AH RARE B. grandicalyx Lindau AH RARE B. micrantha C.B.C1. AH COMMON Blepharis linariifolia Pers. AH RARE B. maderaspatensis (L.) Roth AH RARE Crossandra mucronala Lindau PH RARE Dousperma kilimandscharicum S ABUNDANT (Lindau) Dayton RARE Isoglossa laxa Oliv. PH RARE Justicia flava V ahl AH RARE J. matammensis Oliv. AH RARE Monechma debile (Forssk.) Nees AH RARE Rutty a fruticosa Lindau S RARE Thunbergia erecta (Benth.) Hook. A/PH COMMON RARE AIZOACEAE Corbichonia decumbens (Forssk.) Exell A/PH RARE Mollugo nudicaulis Lam. AH RARE AMARANTHACEAE Aerva lanata (L.) Schultes PH COMMON Achyranthes aspera L. A/PH ABUNDANT RARE Alternanthera p ungens Kunth AH ABUNDANT Amaranthus caudal us L. AH RARE A. spinosus L. AH RARE Digera muricata (L.) Mart. AH RARE Pupalia grandiflora Peter AH RARE P. lappacea (L.) Juss. AH COMMON RARE ANACARDIACEAE Lannea rivae (Chiov.) Sacl. T COMMON RARE L. schweinfurthii (Engl.) Engl. T RARE Sclerocarya birrea (A.Rich.) Hochst. S. birrea (A.Rich.) Hochst. ssp. caffra T RARE (Sond.) Kokwaro Ozoroa insignis Del. ssp. reticulata T RARE (Bak. f.) Gillen T RARE * Numbers 3,4,1, etc. denote associations in which the plants were found (See Fig. AH = Annual herb S = Shrub AC = Annual climber T = Tree PH = Perennial herb 3* 4 1 10 11 13 16 17 1 2 5b 3 13 14 2 5a 9 5a 9 13b 5a 11 13 6 7 17 15 5a 14b 15 16 2 8 9 4 13 2 2 2 1 2 9 12 14 15 8 15 17 3 12 2 2 2 15 2 15 3 14 5 8 14 3 9 13b 15 5 11 14 2 3 6 7 8 13b 1 2 3 6 8 9 10 8 9 2 2 5 14 PC = Perennial climber A/P - Annual or Perennial J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang’ui • Plants of Kiboko 12 ARACEAE Stylochilon puberulus N.E. Br. AH COMMON 14 RARE 5 13 ASCLEPIADACEAE Calotropis procera (Ait.) Ait.f. S RARE 5b 9 12 Diplosligma canescens K.Schum. AH RARE 2 Gomphocarpus kaessneri N. E. Br. AH RARE 2 17 Pentarrhinum insipidum E. Mey PC RARE 5b 13b BALANITACEAE Balanites aegyptiaca (L.) Del. T COMMON 5b 8 RARE 12 13b 15 16 17 BIGNONIACEAE Spalhodea campanulata P. Beauv. T RARE 2 ( ornamental) BORAGESACEAE Cordia sinensis Lam. T RARE 4 8 15 C. ovalis R. Br. S RARE 5b 7 8 10 12 16 Heliotropium sleudneri Vatke AH RARE 5b 10 15 //. subulatum (DC.) Martelli AH COMMON 5b 8 14 RARE 5 7 11 16 17 Trichodesma zeylanicum (L.) R. Br. AH RARE 9 BURSERACEAE Commiphora africana (A. Rich.) Engl. T ABUNDANT 5 6 7 8 11 14 COMMON 13 C. baluensis Engl. T COMMON 4 14 RARE 7 10 11 C. boiviniana Engl. T COMMON 4 5 14 C. mildbraedii Engl. T ABUNDANT 5a 6 7 8 11 14 CAMPANULACEAE Wahlenbergia abyssinica (A. Rich). Thulin AH RARE 2 CAPPARACEAE Cadaba farinosa Forssk. S COMMON 4 Capparis tomentosa Lam. S RARE 4 15 Cleome hirta (Klotzsch) Oliv. AH COMMON 2 5 8 10 15 C. monophylla L. AH COMMON 2 5 8 13 RARE 3 Gynandropsis gynandra (L.) Briq. AH RARE 2 Maerua angolensis DC. T RARE 4 M. decumbens (Brongn.) De Wolf S COMMON 4 RARE 9 M. kirkii (Oliv.) F. White S COMMON 4 RARE 9 M. triphylla A. Rich. S COMMON 4 RARE 9 14 J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang'ui • Plants of Kiboko 13 CARICACEAE Carica papaya L. T COMMON 2 (cultivated) CELASTRACEAE Mayteruis heterophylla (Eckl. & Zeyh.) N. Robson S RARE 14 M. senega lens is (Lam.) Exell S RARE 17 CHENOPODIACEAE Chenopodium album L. AH RARE 2 15 COMBRETACEAE Combretum aculeaium Vent S COMMON 10 RARE 5b 8 C. apiculatum Sond. T ABUNDANT 11 13a RARE 1 2 C. nolle G. Don. T COMMON 11 13a Terminalia parvula Pampan T RARE 5a 9 11 14 T. kilimandscharica Engl. T RARE 5a 13a 14 COMMELINACEAE Aneilema johnstonii K. Schum. AH RARE 5a 10 Commelina africana L. var. boehmiana (K.Schum.) Bren an AH RARE 10 C. albescens Hassk. AH RARE 5a 6 10 11 13a C. benghalensis L. AH RARE 5b 9 10 13a 14 COMPOSITAE A canlhospermum hispidum DC. AH RARE 2 5b 6 10 11 13b 16 Aspilia mossambicensis (Oliv.) Wild PH COMMON 3 RARE 4 9 1 14 15 16 A. pluriseta Schweinf. PH COMMON 3 15 17 RARE 4 14 Athroisma psyllioides (Oliv.) Mattf. AH RARE 8 Bide ns pilosa L. AH RARE 2 5b B. schimperi Sch. Bip. AH ABUNDANT 15 Blepharispermum pubescens S. Moore S RARE 9 14 Conyza aegyptiaca (L.) Ait. AH RARE 2 5a Galinsoga parviflora Cav. AH RARE 2 Gynura miniata Welw. PH RARE 13b Launaea cornuta (Oliv. & Hiem) C. Jeffrey AH COMMON 2 5b 8 9 Schkuhria pinnata (Lam.) O. Ktze. AH COMMON 15 RARE 2 Sonchus asper (L.) Hill AH RARE 2 9 Sphaeranthus cyathuloides O. Hoffm. AH COMMON 12 16 RARE 2 Tagetes minuta L. AH RARE 2 Tridax procumbens L. AH COMMON 2 Vernonia sp. C of U.K.W.F. (Agnew) AH RARE 8 V. galamensis (Cass.) Less. ssp. petitiana (A.Rich.) M.G. Gilbert AH RARE 2 5b V. pleropoda Oliv. & Hiem S ABUNDANT 17 V. wakefieldii O. Hoffm. AH RARE 3 5b 6 7 8 9 11 16 J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang'ui • Plants of Kiboko 14 CONVOLVULACEAE Astripomoea hyoscyamoides AH COMMON 5b (Vatke) Verde. RARE 1 2 3 9 10 A. malvaceae (Klotzsch) Meeuse \ai.floccosa (Vatke) Verde. PH RARE 1 2 Ipomoea eriocarpa R. Br. AC RARE 2 I. kituiensis Vatke var. kituiensis PC RARE 5a 11 13a 1. mombassana Vatke PC ABUNDANT 9 I. obscura (L.) Ker-Gawl. AC RARE 2 13b CRASSULACEAE Kalanchoe lanceolata (Forssk.) Pers. AH RARE 2 CRUCIFERAE Farsetia stenoptera Hochst. AH RARE 2 15 CUCURBITACEAE Cucumis dipsaceus Spach AC RARE 2 5a 9 15 C. hirsutus Sond. PC RARE 9 C. prophetarum L. ssp. dissectus (Naud.) C. Jeffrey PC RARE 2 Kedrostis gijef (J.F. Gmel.) C. Jeffrey PC RARE 5a Lagenaria abyssinica (Hook.f.) C. Jeffrey AC RARE 8 Momordica boivinii Baill. PH RARE 2 CYPERACEAE Cyperus alternifolius L. PH ABUNDANT 15 17 C. articulatus L. PH ABUNDANT 15 C. bulbosus Vahl var. melanolepis Kukenth. AH RARE C. giolii Chiov. AH RARE 4 10 C. laevigatus L. PH ABUNDANT 15 C. longibracteatus Cherm. PH ABUNDANT 15 C. obtusiflorus V ahl AH RARE 1 2 3 7 8 C. rotundas L. AH RARE 1 2 3 7 10 Fimbristylis dichotoma (L.) Vahl. PH COMMON 15 F. hispidula (Vahl) Kunth PH ABUNDANT 15 Mariscus obsoletinervosus (Peter & Kuk.) Greenway AH RARE 5a EBENACEAE Euclea racemosa Murr. ssp. schimperii (A. DC.) F. White T RARE 17 EUPHORBIACEAE Acalypha fruticosa Forssk. S COMMON 14 RARE 9 10 15 A . indica L. AH RARE 2 Bridelia cathartica Bertol.f. S RARE 15 Croton dichogamus Pax S RARE 3 14 COMMON 5 C. scheffleri Pax T COMMON 4 Dalechampia ipomoeifolia Benth. PC COMMON 15 Erythrochocca bongensis Pax s RARE 15 Euphorbia cotinifolia Pax s RARE 15 J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang'ui • Plants of Kiboko 15 E. heterochroma Pax T COMMON 4 E. heterophylla L. AH RARE 2 E. inaequilatera Sond. AH COMMON 2 Flueggea virosa (Willd.) Voigt S RARE 2 15 Jatropha spicata Pax S RARE 10 Phyllanthus amarus Schumach. & Thonn. AH RARE 8 P. maderaspatensis L. PH RARE 2 Ricinus communis L. S RARE 15 Synadenium molle Pax S RARE 3 4 5b Tragia hildebrandtii Muell. Arg. PH RARE 12 T. subsessilis Pax PC RARE 2 12 GERANIACEAE Monsonia longipes R. Kunth. AH COMMON 15 12 16 RARE 2 GRAMINEAE Andropogon chinensis (Nees) Merr. P COMMON 2 RARE 5 6 11 Aristida adscensionis L. P ABUNDANT 2 COMMON 11 13 8 RARE 3 5 6 10 12 A. barbicollis Trin. & Rupr. A RARE 2 A. kenyensis Henr. A COMMON 2 1 13 RARE 7 16 A. mutabilis Trin. & Rupr. A RARE 2 Bothriochloa bladhii (Retz.) S.T. Blake P COMMON 5 B. insculpta (A. Rich.) A. Camus P ABUNDANT 16 COMMON 12 RARE 2 8 9 10 15 B. radicans (Lehm.) A. Camus P COMMON 12 15 16 RARE 5 6 7 9 11 13 16 17 Brachiaria deflexa (Schumach.) Robyns A COMMON 15 RARE 2 5 9 13 17 B. eruciformis (J. E. Smith) Grisenb. A COMMON 15 RARE 2 5 B. lachnantha (Hochst.) Stapf A COMMON 12 16 RARE 15 B. leersioides (Hochst.) Stapf A COMMON 2 5 (roadsides & fire breaks) RARE 6 7 10 15 16 B. leucacrantha (K.Schum.) Stapf A COMMON 8 (roadsides & fire breaks) RARE 2356789 11 15 B. reptans (L.) Gardner & Hubbard A RARE 2 9 15 Cenchrus ciliaris L. P ABUNDANT 3 (near Hunters Lodge) COMMON 10 15 17 RARE 1 2 5 6 7 8 9 11 12 14 17 Chloris gayana Kunth. P RARE 2 17 C. pycnothrix Trin. A RARE 2 15 12 C. roxburghiana Schult. P ABUNDANT 5 7 8 9 COMMON 13 14 RARE 2 3 6 10 11 12 C. virgata Sw. A COMMON 1 2 3 4 5 RARE 6 10 11 13 16 Chrysopogon plumulosus Hochst. P ABUNDANT 1 2 RARE 5 J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang'ui • Plants of Kiboko 16 Cymbopogon caesius (Hook. & Am Stapf) P C. pospichilii (K.Schum.) C.E. Hubbard P Cynodon dactylon (L.) Pers. P C. plectostachyus (K. Schum.) Pilg. P Cypholepis yemenica (Schweinf.) Chiov. A Dactyloctenium aegyptium (L.) Willd. A Dichanthium arum latum (Forssk.)Stapf var. paplillosum (A. Rich.) de Wet & Harlan P Digitaria abyssinica (A. Rich.) Stapf P D. gazensis Rendle P D. macroblephara (Hack) Stapf P D. milanjiana (Rendle) Stapf P D. velutina (Forssk.) P. Beauv. A Diheleropogon amplectens (Nees) W.D. Clayton P Dinebra polycarpha S.M. Phillips D. retroflexa (Vahl) Panzer A Echinochloa haploclada (Stapf) Stapf P Eleusine coracana (L.) Gaertn. A E. indica (L.) Gaertn. A Enneapogon cenchroides (Roem. & A Schult.) C. E. Hubbard Enleropogon macrostachyus P (A. Rich.) Benth. Eragrostiella bifaria (Vahl) Bor A Eragrostis aethiopica Chiov. A E. caespitosa Chiov. P E. cilianensis (All.) Lut. A E. cylindriflora Hochst. A E. rigidoir Pilg. P E. superba Peyr. P Eriochloa fatmensis (Hochst. & Steud.) W.D. Clayton A Eustachys paspaloides (Vahl) Lanza & Mattei P Harpachne schimperi A. Rich. A ABUNDANT 8 RARE 5 ABUNDANT 8 COMMON 12 RARE 5 16 ABUNDANT 1 9 COMMON 5 10 RARE 2 6 7 11 12 13 15 16 16 COMMON 3 (near Hunters Lodge) 10 RARE 6 9 RARE 12 3 4 COMMON 2 3 (roadsides & fire breaks) RARE 5 6 7 8 10 11 13 14 17 COMMON 12 15 16 RARE 2 RARE 5 6 14 ABUNDANT 5 6 7 13 14 COMMON 11 8 RARE 2 16 17 COMMON 6 11 RARE 7 14 COMMON 2 3 15 (roadsides & fire breaks) RARE 5 6 7 9 10 13 RARE 2 COMMON 2 12 15 COMMON 15 ABUNDANT 12 16 COMMON 17 RARE 2 (cultivated) RARE 2 COMMON 1 2 3 RARE 12 ABUNDANT 5b COMMON 5a 6 7 8 RARE 3 10 11 13 14 RARE 1 2 RARE 2 5 10 12 16 ABUNDANT 5a 14 10 COMMON 5a 14 10 RARE 6 7 8 11 16 COMMON 2 3 4 RARE 5 7 10 11 13 17 RARE 2 RARE 2 ABUNDANT 9 COMMON 5 7 8 11 13 14 RARE 2 3 6 10 COMMON 12 15 16 RARE 1 2 5a 6 11 13b RARE 2 J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang'ui • Plants of Kiboko 17 Heteropogon coniortus (L.) Rhoem. & Schult. P COMMON (alongside Mombasa road) RARE 2 5 6 8 11 17 Hyparrhenia hirta (L.) Stapf P RARE 2 Ischaemum afrum (J.F. Gmel.) Dandy P COMMON 16 (North Ranch Five-Comers) RARE 6 Leptocarydion vulpiastrum (De Not.) Stapf A COMMON 2 3 4 17 Leptochloa obtusiflora Hochst. P RARE 2 3 4 9 10 15 17 Leptothrium senegalense (Kunth) A COMMON 5b W.D. Clayton RARE 6 7 8 10 11 Lintonia nutans Stapf P RARE 15 Microchloa kunthii Desv. P RARE 2 Oropetium thomaemum (L.f.) Trin. A COMMON 4 RARE 2 5 8 10 11 15 Panicum atrosanguineum A. Rich. A RARE 5 12 16 17 P. coloratura L. P RARE 8 P. deustum Thunb. P COMMON 9 RARE 5 6 10 11 13 14 15 P. maximum Jacq. P COMMON 5 7 9 14 RARE 6 8 10 11 13 Paspalidium desertorum (A. Rich) Stapf A RARE 2 Pennisetum massaicum Stapf P ABUNDANT 12 15 16 COMMON 5 RARE 2 17 P. mezianum Leeke P ABUNDANT 12 15 16 RARE 3 8 P. setaceum (Forssk.) Chiov. P RARE 2 P. stramineum Peter P RARE 2 Rhynchelytrum repens (Willd.) C. E. Hubbard Rottboellia cochinchinensis (Lour.) P RARE 2 11 W.D. Clayton A RARE 2 3 9 15 Sehima nervosum (Rottler) Stapf P ABUNDANT 2 3 6 RARE 7 Setaria incrasata (Hochst) Hack. A COMMON 2 3 RARE 12 15 16 S. plicatilis (Hochst.) Engl. P COMMON 15 S.pumila (Pior.) Roem. & Schult A RARE 2 15 S. sphacelata (Schumach.) Moss P RARE 2 S. verticillata (L.) P. Beauv. A COMMON 4 RARE 2 9 10 12 15 Sorghum arundinaceum (Desv.) Stapf S . purpureo-sericeum (A. Rich.) A RARE 2 Ascher & Schweinf. A RARE 15 Sporobolus fimbrialus (Trin.) Nees P ABUNDANT 12 COMMON 5 8 14 RARE 2 6 7 10 11 17 S. helvolus (Trin.) Th. Dur. & Schinz P RARE 15 S. pellucid us HochsL P RARE 2 17 S. spicatus (Vahl) Kunth Tetrapogon cenchriformis (A. Rich.) P RARE 2 15 17 W.D. Clayton P RARE 6 12 16 T. tenellus (Roxb.) CV'"'. A COMMON 2 3 4 RARE 5 10 11 13 J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang'ui • Plants of Kiboko 18 Themeda triandra Forssk. P ABUNDANT 5a COMMON 7 11 RARE 5b 6 8 13 14 Trachypogon spicatus (L. f.) Kuntze P RARE 5a 6 11 Tragus berteronianus SchulL P COMMON 2 3 4 RARE 6 13 16 17 Tricholaena teneriffae (L. f.) Link. P ABUNDANT 1 2 RARE 3 Tripogon minimus (A. Rich.) Steud. P RARE 2 Urochloa panicoides P. Beauv. A RARE 2 15 ICACINACEAE Pyrenacantha kaurabassana Baill. PC RARE 5 LABI AT AE Hoslundia opposita Vahl S RARE 5 7 9 10 13 14 16 Leucas mollis Baker AH COMMON 17 RARE 12 15 16 Ocimum basilic um L. AH COMMON 2 8 RARE 3 5b 11 13 O. hadiense Forssk. S RARE 1 2 3 Orthosiphon parvifolius Vatke AH RARE 5 O. rubicundus Benth. AH COMMON 2 RARE 1 3 5 11 13 Plectranlhus barbatus An dr. S RARE 5a 11 14 P. caninus Roth AH RARE 2 P. sylvestris Guerke AH RARE 5a 14 LEGUMINOSAE Acacia brevispica Harms S ABUNDANT 3 COMMON 4 14 RARE 5 6 9 10 11 13 A. drepanolobium Sjostedt T ABUNDANT 12 16 RARE 15 17 A. hockii De Wild. S COMMON 2 RARE 1 A. mellifera (Vahl) Benth. S ABUNDANT 5b 10 RARE 3 5 6 7 8 9 11 16 17 A. nilotica (L.) Del. T COMMON 11 RARE 7 13b 16 A. Senegal (L.) Willd. T COMMON 5 6 7 14 RARE 2 3 8 10 11 12 A. tortilis (Forssk.) Hayne T ABUNDANT 5 6 14 COMMON 7 9 10 RARE 3 11 12 13 17 A. xanthophloea Benth. ABUNDANT 17 T RARE 15 Albizia amara (Roxb.) Boiv. ssp. sericocephala (Benth.) Bren an T RARE 3 5 6 7 8 10 11 A. anthelmirUica Brongn. S COMMON 5 6 13 J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang'ui • Plants of Kiboko 19 Atysicarpus glumaceus (Vahl) DC. A. rugosus (Willd.) DC. Cassia abbreviata Oliv. C. absus L. C. mimosoides L. C. occidentals L. Clitoria ternatea L. Crotalaria goodiiformis Vatke C. incana L. C. labumifolia L. C. polysperma Kotschy Dalbergia melanoxylon Guill. & Perr. D. vacciniifolia Vatke Dichrostachys cinerea (L.) Wight & Am. Dolichos sericeus E. Mey. Entada leptostachya Harms Indigofer a colutea (Burm.f.) Merrill I. lupatana Bak.f. I. mahndiensis GiUett Leucaena lalisiliqua (L.) Gillis Macrotyloma axillare (E. Mey.) Verde. Milletia leucantha Vatke Neonotonia wightii (Wight & Am.) Lackey Neorautanenia mitis (A.Rich.) Verde. Ormocarpum kirkii S. Moore 0. trachycarpum (Taub.) Harms Rhynchosia elegans A. Rich. R. minima (L.) DC. Sena bicapsularis (L.) Roxb. S. longiracemosa (V atke) Lock S. singueana (Del.) Lock Sesbania keniensis Gillett Tephrosia nana Schweinf. T. pumila (Lam.) Pers. T. subtiflora Bak. Vigna membranacea A.Rich. V. unguiculata (L.) Walp. Anlhericum subpetiolatum Bak. A. venulosum Bak. Asparagus africanus Lam. A. falcaius L. A. flagellar is (Kunth) Bak. Bulbine abyssinica A. Rich. Chlorophytum gallabatense Bak. Gloriosa superba L. Wurmbea tenius Bak. PH RARE 5 6 8 9 PH COMMON RARE 5 6 8 11 1 T RARE 7 14 AH RARE 5 15 AH RARE 9 16 S RARE 2 3 9 17 PC RARE 12 16 s RARE 11 AH RARE 2 9 AH RARE 2 6 AH RARE 2 5 9 10 T COMMON 2 RARE 3 4 5 11 S RARE 2 3 T COMMON 2 RARE 1 5 8 11 PC RARE 12 PC RARE 2 3 4 5 14 AH COMMON 9 RARE 5b 10 14 AH RARE 9 AH RARE 9 S RARE 2 8 PC RARE 5 11 14 T RARE 5 8 PC RARE 2 5b 13 14 PC RARE 4 5 S COMMON 2 3 14 S COMMON 2 3 RARE 11 PC RARE 17 PC RARE 2 5 10 s RARE 15 17 s RARE 5b T RARE 7 10 11 131 S RARE 17 HA/P COMMON 1 2 3 8 9 HA/P RARE 5 6 10 12 AH RARE 9 HA/P RARE 5 6 7 14 C A/P RARE 5 14 AH RARE 2 LILIACEAE HA/P RARE 5a 14 HA/P RARE 5a PC RARE 5a 14 PC RARE 3 4 PC RARE 2 5a 15 AH RARE 15 AH RARE 5 9 11 13b AH RARE 5a AH RARE 1 J. E. A. N. H. S. & Nat Mus. 80: (195) June 1990 Ndegwa wa Ndiang'ui • Plants of Kiboko 20 LOBELIACEAE Cyphia glandulifera A. Rich. PH RARE 1 LORANTHACEAE Plicosepalus sagittifolius (Sprague) Danser RARE (epiphyte on Acacia Senegal) MALPIGHIACEAE Caucanthus auriculatus (Radik.) Niedenzu PC RARE 9 11 13b 14 MALVACEAE Abutilon figarianum Webb. AH RARE 2 A. fruticosum Guill. & Perr. S RARE 17 A. mauritianum (Jacq.) Medic. S RARE 2 15 Hibiscus aponeurus Sprague & Hutch. PH COMMON 2 3 RARE 5 8 11 13 14 H. calyphyllus Cav. PH RARE 5 10 15 H. cannabinus L. AH RARE 12 15 16 H . palmatus Forssk. AH RARE 2 3 H. vitifolius L. AH RARE 2 3 13 Pavonia patens (Andr.) Chiov. S COMMON 2 3 5 RARE 4 11 14 Sida ovata Forssk. A/PH RARE 8 MELIACEAE Melia azedarach L. T RARE 5 9 (cultivated) M. volkensii Gurke T RARE 5 Trichilia emetica Vahl T RARE 15 17 MENISPERMACEAE Cissampelos mucronala A. Rich. PC COMMON 15 17 RARE 3 9 11 MORACEAE Ficus glumosa Del. T COMMON 15 RARE 2 NYCTAGINACEAE Commicarpus pedunculosus PH COMMON 9 17 (A. Rich.) Cuf. RARE 5b 11 14 C. plumbagineus (Cav.) Standi. PH COMMON 15 17 RARE 2 3 5b 6 7 10 OCHNACEAE Ochna ovata F. Hoffm. S RARE 5 OLACACEAE Ximenia americana L. T RARE 15 17 PALMAE Phoenix reclinata Jacq. S COMMON 15 PASSIFLORACEAE Adenia gummifera (Harvey) Harms PC COMMON 15 17 RARE 4 13 J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang'ui • Plants of Kiboko 21 Polygala sphenoptera Fresen. Oxygonum sinuatum (Meisn.) Dammer Helinus inlegrifolius (Lam.) Kuntze H. mystacinus (Ait.) Steud. Oldenlandia corymbosa L. var. caespitosa (Benth.) Verde. PerUanisia ouranogyne S. Moore Pentas parvifolia Hiem Vangueria madagascariensis Gmel. Zanthoxylum chalybeum Engl. Azima tetracantha Lam. Salvadora persica L. Allophyllus rubifolius (A. Rich.) Engl. Cardiospermum corindum L. Cycnium tubulosum (L. f.) Engl. Lycopersicon esculentum Mill. Solarium incanum L. S. nigrum L. S. renschii Vatke S. sessilistellatum Bitter Withania somnifera (L.) Dunal Dombeya kirkii Mast. D. rotundifolia (Hochst.) Planch. Hermannia uhligii Engl. Melhania velutina Forssk. Sterculia africana (Lour.) Fiori POLYGALACEAE A/PH RARE POLYGONACEAE AH RARE RHAMNACEAE PC RARE PC RARE RUBIACEAE AH RARE A/PH RARE S RARE S RARE RIJTACEAE T RARE SALVADORACEAE S COMON S COMMON RARE SAPINDACEAE S RARE PC COMMON SCROPHULARIACEAE AH RARE SOLANACEAE AH RARE S RARE AH RARE S RARE S RARE AH RARE STERCULIACEAE T RARE T RARE S COMMON RARE AH RARE T COMMON RARE 11 2 6 9 10 13b 2 3 5 6 7 9 10 2 9 2 5 9 4 16 I 2 3 8 11 (roadsides & firebreaks ) II 14 15 14 4 4 12 15 2 17 17 12 16 2 17 (cultivated) 2 3 5 7 8 9 10 11 13 16 2 17 4 15 9 17 4 2 3 5 8 2 6 9 10 11 13 16 17 2 5 3 4 2 11 14 PORTULACACEAE Talinum portulacifolium (Forssk.) Schweinf. PH RARE RESEDACEAE Caylusea abyssinica(FTesen.) Fisch. & Mey. AH RARE J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang'ui • Plants of Kiboko 22 TILIACEAE Corchorus trilocularis L. AH RARE 8 11 16 Grewia bicolox Juss. S COMMON 5 6 7 RARE 3 9 10 G. densa K. Schum. S COMMON 5 6 7 RARE 3 10 11 G.fallax K. Schum. T COMMON 5 14 RARE 3 6 7 G. similis K. Schum T COMMON 5 15 RARE 6 7 8 G. tembensis Fres. T COMMON 5 14 RARE 6 7 8 G. villosa Willd. S COMMON 5 6 8 RARE 2 7 9 Triumfetta flavescens A. Rich. S ABUNDANT 3 COMMON 2 RARE 1 8 10 TURNER ACE AE Streptopetalum hildebrandtii Urb. PH RARE 1 UMBELLIFERAE Sleganolaenia araliacea Hochst. T COMMON 3 4 RARE 2 5 16 VERBENACEAE Chascanum hildebrandtii (Vatke) Gillett PH RARE 2 Clerodendrum acerbianum (Vis.) Benth. & Hook.f. S RARE 2 C. myricoides (Hochst.) Vatke S RARE 2 5 11 Lanlana camara L. S RARE 5 15 L. rhodesiensis Moldenke PH RARE 5 L. viburnoides (Forssk.) Vahl PH RARE 5b Lippia javanica (Burm.f.) Spreng S RARE 2 3 4 L. kituiensis Vatke S RARE 5 17 Premna oligotricha Bak. S COMMON 11 13 RARE 5 Priva curtisiae Kobuski PH RARE 2 15 VITACEAE Cissus aphyllanlha Gilg. PC COMMON 3 4 5 C. cornifolia (Bak.) Planch. PH RARE 2 5 11 C. quadrangularis L. PC COMMON 4 C. rotundifolia (Forssk.) Vahl PC RARE 5 Cyphostemma adenocaule PC COMMON 5 14 (A. Rich.) Wild & Drum. RARE 1 C. serpens (A. Rich.) Desc. PC COMMON RARE 6 7 11 ZYGOPHYLLACEAE Tribulus terrestris L. AH COMMON 15 RARE 2 14 13 14 13 10 12 13 16 17 14 J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 Ndegwa wa Ndiang'ui • Plants of Kiboko 23 ACKNOWLEDGEMENTS I am grateful to the staff of the National Range Research Station, and especially to Mr. A. Tifow and Mr. Daniel Musau, for their invaluable help and cooperation during the field data collec- tion. Special regards go to Miss Margaret Muriuki, Mrs. Margaret Mwoloi and Miss Milcah Karanja for carefully and skillfully typing this paper with unequalled patience. Many thanks are due to Mr. S. Mathenge of the Botany Department, University of Nairobi, and staff of the East African Herbarium for their assistance in identifying some of the specimens. REFERENCES AGNEW, A. D. Q., (1974) Upland Kenya Wild Flowers. London, Oxford University Press. BLUNDELL, M., (1982) The Wild Flowers of Kenya. London, Collins.. BLUNDELL, M., (1987) Collins Guide to the Wild Flowers of East Africa. London, Collins.. COE, M. & COLLINS, N.M. (1986) KORA, An Ecological Inventory of Kora National Reserve, Kenya. London, Royal Geographic Society. DALE, I. R. & GREENWAY, P. J. (1961) Kenya Trees and Shrubs. London, Buchanan’s Kenya Estates Limited & Hatchards. London. GAKAHUl, C. G. Personal communication. LANGAT, R., WANJALA, B„ BRAUN, H. M. H, & RACHILO, J. R. (1975) Vegetation map of Kiboko Range Research Station. Ministry of Agriculture, Nairobi. LIND, E. M. & MORRISON, M. E. S. (1974) East African Vegetation London, Longmans Ltd. MICHIEKA, D. O. & VON DER POUN, B. J. A. (1977) Soils and Vegetation of Kiboko Range Research Station. A semi-detailed report 53 (Draft Ed.). Ministry of Agriculture, Nairobi. NDIANG'UI, N. (1984) Plant Communities of the Kiboko Area. In: Endangered Resources for Development ETMA/NES, Nairobi. PRATT, D. J., GREENWAY, P. J, & GWYNNE, M. D. (1966) A Classification of East African Rangelands: with an appendix on terminology. Journal of Applied Ecology 3: 369-382. PRATT, D. J., & GWYNNE, M. D. (1977) Rangeland Management and Ecology. New York, Robert E. Krieger Pub. Co., Inc. WHYTE, O. R. (1968) Grasslands of the Monsoon. New York, Fredrick A. Praeger. Received 19 June, 1989 Editor: C.F. Dewhurst Instructions for authors may be obtained from: The Hon. Secretary, E.A.N.H.S., P.O. Box 44486, Nairobi, Kenya Published by The EAST AFRICA NATURAL HISTORY SOCIETY •Typesetting by The Print Shop • Printed by AM REF J. E. A. N. H. S. & Nat. Mus. 80: (195) June 1990 QH / ://Xr JOURNAL OF THE EAST AFRICA NATURAL HISTORY SOCIETY AND NATIONAL MUSEUM December 1990 Volume 80, Number 196 A PROVISIONAL, ANNOTATED CHECKLIST OF THE BUTTERFLIES IN LAKE MANYARA NATIONAL PARK, ARUSHA REGION, TANZANIA. NORBERT J CORDEIRO P. O. Box 708, Moshi, Tanzania* f AUfi 1 3 WVl \ J ABSTRACT V :: : .'/'A- . X" Lake Manyara National Park is well known for its diverse habitats and large mammals. While little is understood about most of the smaller vertebrates and invertebrates, this investigation into the butterflies of the park has revealed over 180 species. Several of the following are of particular interest, either Decause of new extensions to their range or, their taxonomic differences as compared to other East African populations: Belenois margaritacea plutonica , Pieris brassicoides marghanita, Charaxes cithaeron kermethi, Ch. violetta melloni , Ch. hansalii baringana , a female aberration of Ch. achaemenes achaemenes (figured and described in this paper), Mimacraea marshalli , Aloeides conradsi talboti , Stugeta bowkeri nyanzana and Tuxentius stempfferi. INTRODUCTION Lake Manyara National Park (L.M.N.P.) is one of the areas of highest wildlife biomass in Africa (Coe et al., 1976; Loth & Prins, 1986) and is well known for its diverse fauna and flora. Although Watermeyer & Elliott (1943) initially described the general ecology, probably the most extensive earlier contribution to the ecology of L.M.N.P. is that of Douglas-Hamilton (1972). Greenway & Vesey-FitzGerald (1969), gave a detailed account of the vegetation, emphasizing the diverse habitat types and their corresponding species, whereas Loth & Prins (1986) described the physiography of the park, [see also Prins, 1988; Prins & Loth, 1988]. Most of the scientific research at L.M.N.P. has centred around vegctational changes caused by large herbivores (Vesey-FitzGerald, 1969 1973; Douglas-Hamilton, 1972; Mwalyosi, 1981 1983; Beckman & Prins, 1989; Prins & Beekman, 1989); although numerous ecological and behavioural studies have also been conducted on large mammals (Makacha & Schaller, 1969; Vesey- FitzGerald, 1969; Douglas-Hamilton, 1972; Weyerhauser, 1982; Kalemera, 1987; Beekman & Prins, 1989; Kalemera, 1989; Prins, 1989; Prins & Beekman, 1989; Prins & Iason, 1989). The changes in populations of large mammals have also been monitored over the years (Watson & Turner, 1965; Douglas-Hamilton, 1972; Mwalyosi, 1977; Boshe & Malima, 1986). Morgan-Davies (1964) recorded over 380 species of birds (TANAPA / AWF, 1986). The smaller vertebrates and invertebrate fauna, however, have received scant attention. * Present address: P. O. Box 0377, Hampshire College, Amherst MA01002 USA Norbert J. Cordeiro • Butterflies of Lake Manyara National Park, Tanzania 26 Only a few people have contributed towards butterfly records. Bams (1924) and Cooper collected in the region of L.M.N.P. as early as 1921 and 1935 respectively (according to specimen labels in the collections at the Natural History Museum, London (BMNH), (personal observations). Two-and-a-half decades later, Rydon & Morgan-Davies (1960a, b, and unpublished) collected butterflies in L.M.N.P., compiling two short checklists for the park’s authorities; they recorded 56 species. In the early 1970s, S. C. Collins (personal communication) also made a small collection of butterflies in the park. This paper is an account of the butterflies recorded in L.M.N.P. as a result of a five month study made between late July and early December 1987. The main content has been divided into two sections: (1) a checklist of all the species, and (2) brief notes on over 50 of them. The checklist is by no means an exhaustive study of the park’s butterfly fauna of the park. This is particulary evident in the families Lycaenidae, Acraeidae and Hesperiidae, which are poorly represented. There are also habitats that require exploring in greater depth than I had undertaken, e.g. the groundwater forest, the Marang Forest and the adjoining forest below the Rift Valley wall, as well as the vegetation of the escarpment. (Fig. 1). Description of the study area Lake Manyara National Park, is situated in northern Tanzania (centre of the park at 3°30'S, 35°60'E), it encompasses part of the eastern escarpment of the Rift Valley as well as the lake and a portion of the surrounding environment (Fig. 1). The lake lies at an altitude of approxi- mately 960m, while the Rift Valley wall generally rises to about 1000m above it and in some areas is even higher. L.M.N.P. is 330 km2 of which about two-thirds comprises the alkaline lake (Mwalyosi, 1981; TANAPA / AWF, 1986). The Marang Forest Reserve and an adjacent strip of land in the south of L.M.N.P. have been proposed for incorporation into the park boundaries in the near future (Loth & Prins, 1986). Annual rainfall is variable, ranging from 480mm [1969] to 1500mm [1968] (Mwalyosi, 1981) most of which falls between November and May (Prins & Loth, 1988). There is one long dry season from June to October (Prins & Loth, 1988). Within the Lake Manyara ecosystem there are five distinct vegetation zones, namely groundwater forest, marshland and reed beds, alkaline grassland, scrubland on the escarpment and the Acacia woodland (Greenway & Vesey-FitzGerald, 1969). METHODS As a result of the reorganization of the park museum, the new L.M.N.P. Rhopalocera collection which I re-established entailed the exploration of many vegetation zones of the park. [I replaced the moth-eaten collection made by A. M. Morgan-Davies which had been installed in the early 1960s (Rydon, in litt.)]. Collecting was undertaken on an irregular basis for between 1-3 days a week. To obtain additional records, I briefly studied the African collections in the Natural History Museum, London (BMNH) and the American Museum of Natural History (AMNH) in August 1988 and July 1989 respectively. I have followed Carcasson’s “Simplified Provisional Checklist of the Butterflies of the Afrotropical Region” in his Collins Handguide to the Butterflies of Africa (1981) (Hard-back edition) as the basis of this checklist. The list has been organized by families and genera except for the Lycaenidae which was placed at the end and the Hesperiidae at the beginning, following the arrangement of Kristensen (1976). For simplicity these groups have not been divided into subfamilies and tribes; however, the species have been arranged alphabethically so as to facilitate easy reference by the reader. Of special interest are Belenois margaritacea plutonica Joicey & Talbot, Pier is brassicoides J. EANHS & Nat. Mus. 80 (196) December 1990 27 Norbert J. Cordeiro • Butterflies of Lake Manyara National Park, Tanzania marghanita Hemming, Charaxes cithaeron kennelhi Poulton, Ch. violetta melloni Fox, Ch. hansalii baringana Rothschild, and a female aberration of Ch. achaemenes achaemenes Felder (which is figured and described in this papei), Mimacraea marshalli Trimen, Aloeides conradsi talboti Tite & Dickson, Stugeta bowkeri nyanzana Wichgraf, Tuxentius stempfferi Kielland. The significance of the above-named species to L.M.N.P. is discussed later. Other species marked with an asterisk (*) or a question mark (?) after the author(s’) name are also discussed further in the notes section and, where appropriate, I have added additional comments on the taxonomy, behaviour and special geographical ranges. Numbers after generic and specific names refer to species notes in the discussion. Important additional information was provided to me by J. Kielland, S. C. Collins and A. H. B. Rydon. RESULTS Lake Manyara checklist All specimens (except those recorded by other authors / collectors) are in the National Park museum or temporarily in my collection. Key to the abbreviations and symbols In the checklist, numbers preceding names refer to species mentioned in the discussion. Butterfly records other than my own * = specimens in the BMNH ** = various records from other authors or collectors *** = specimens in the AMNH Butterfly status at LMJ9.P. x = rare ox = uncommon o = common oo = very common Habitat abbreviations, ( after Carcasson, 1981 ) ag arid grasslands gf groundwater forest ah arid habitats h highlands aw arid woodlands hf highland forest c cultivated areas oh open habitats f forest rf riverine forest fm forest margin w woodland g grasslands Other abbreviations and symbols - A (?) after a habitat(s) abbreviation means that the habitat of the species in L.M.N.P. was not known to me as it was recorded by someone else. I have added the known habitat(s) in Africa according to Carcasson (1981). - A (?) representing the status of the taxon implies that it was recorded by another person, and information on its status in L.M.N.P. is unavailable. - A (?) after the author(s) name implies that the record is based on my observation and that it was not captured for verification. - ssp. denotes uncertainly on the sub-specific status of the taxon, or that it is currently being described by another author. J. EANHS & Nat. Mus. 80 (196) December 1990 Norbert J. Cordeiro • Butterflies of Lake Manyara National Park, Tanzania 28 Family, Genus & species Status Habitat HESPERUDAE Coeliades Hubner C. anchises anchises Gerstaecker (2) oo f Celaenorrhinus Hubner C. galenus Fabricius oo gf Tagiades Hubner T. flesus Fabricius oo w Eagris Guenee E. sabadius astoria Holland oo gf. w Sarangesa Moore S. motozi Wallengren 0 w S. phidyle Walker oo ah S. seineri seineri Strand o ah Netrobalene Mabille N. canopus Trimen (8) oo oh in f, rf Spialia Swinhoe S. spio Linnaeus**(9) oo oh Gomalia Moore G. elma Trimen o oh Metisella Hemming M. medea medea Evans (11) o gf Ampittia Moore A. capenas capenas Hewitson ox w Acleros Mabille A. ploetzi Mabille oo gf Zenonia Evans Z. zeno Trimen oo w Gegenes Hubner G. hottentota hottentota Latreille (15) o oh PAPILIONIDAE Papilio Linnaeus P. dardanus libullus Kirby 0 f P. demodocus demodocus Esper oo oh P. echerioides Trimen ssp. ox f P. nireus lyaeus Doubleday o oh in f, c Graphium Scopoli G. angolanus angolanus Goeze? (20) o oh G. antheus Cramer o f, fm G. leonidas leonidas Fabricius (22) oo w, f G. policenes Cramer**(23) ? f? PIERIDAE Catopsilia Hubner C.florella Fabricius oo oh Eurema Hubner E. brigitta brigitta Stoll oo oh E. desjardinsi marshalli Butler (26) oo w E. hapale Mabille o g E. hecabe solifera Butler oo oh, w Pinacopteryx Wallengren P . eriphia melanarge Butler oo oh, w Nepheronia Butler N. argia mhondana Suffert oo rf, f N. buqueli buqueli Boisduval oo ah Family, Genus & species Status Habitat N. thalassina sinalata Suffert OX rf, f Eronia Hubner E. cleodora dilatata Butler o w E. leda Boisduval oo w Colotis Hubner C. anlivippe zera Lucas oo oh C. aurigineus Butler oo oh C. aurora dissociatus Butler oo ah C. auxo increlus Butler (38) o w C. Calais Calais Cramer oo ah C. celimene celimene Lucas o oh C. chrysonome chrysonome ? a Klug* (41) C. danae pseudacaste Butler oo aw C. eris eris Klug (43) oo ah C. evagore antigone Boisduval oo ah C. evenina xantholeuca Sharpe oo w C. evippe complexivus Butler oo oh C. halimede australis T albot o aw C. hetaera ankolensis Stoneham oo oh C. hildebrandti Staudinger oo aw C. ione Godart oo w C. pallene pallene Hopffer** (51) ? oh? C. regina Trimen oo w C. vesta hanningtoni Butler oo oh C. vestalis castalis Staudinger (54) o ah Belenois Hubner B. aurota aurota Fabricius oo oh B. creona sever ina Stoll oo oh B. gidica Godart oo oh B. margaritacea plutonica X gf Joicey & Talbot (58) B thysa thysa Hopffer o w, fm B. zochalia agrippinides Holland o gf Pieris Schrank P. brassicoides marghanita X h Hemming (= meridionalis Joicey & Talbot) (61) Dixiea Talbot D. char ina liliana Grose-Smith o w, fm D. doxo cost at a Talbot o w D. or bona vidua Butler 0 oh D. pigea Boisduval o w, fm Appias Hubner A. epaphia contracta Butler oo w, f, rf A. sabina phoebe Butler o rf, f Mylothris Hubner M. chloris agathina Cramer oo w, fm M. ruppellii tirikensis Neave o w, f Leptosia Hubner L. alcesla inalcesta Bemardi oo f LIBYTHEIDAE Libythea Fabricius L. labdaca laius Trimen (71) o gf, fm J. EANHS & Nat. Mus. 80 (196) December 1990 29 Norbert J. Cordeiro • Butterflies of Lake Manyara National Park, Tanzania Family, Genus & species Status Habitat NYMPHAUDAE Charaxes Ochsenheimer Ch. achaemenes achaemenes Felder ((72) o w Ch. aubyni aubyni van Someren & Jackson (73) ox gf Ch. baumarmi tenuis van Someren o gf Ch. bohemani Felder (75) oo w Ch. brutus alcyone Stoneman oo f Ch. candiope candiope Godart (77) oo f, w Ch. cithaeron kennethi Poulton (78) o gf Ch. etesipe tavetensis Rothschild o f Ch. ethalion littoralis van Someren oo f, W Ch. hansalii baringana Rothschild (81)x aw Ch. jasius saturnus Butler (82) o w,c Ch. kirki kirki Butler (83) oo w Ch. pollux maua van Someren (84) ox gf Ch. varanes vologeses Mabille (85) oo w, fm Ch. violetta melloni Fox (86) 0 gf Ch. zoolina zoolina Westwood (87) oo w, fm Euxanthe Hubner E. wakefieldii Ward (88) o gf Hamanumida Hubner H. daedalus Fabricius oo oh Pseudacraea Westwood Ps. boisduvali trimem Butler**(90) ? w, rf, fm? Ps. lucretia expansa Butler o rf, fm Neptis Fabricius N. laeta Overlaet oo w N. saclava marpessa Hopffer oo w Cyrestis Boisduval C. camillus sublineata Lathy o gf, rf Byblia Hubner B. anvatara acheloia Wallengren oo oh B. ilithya Drury oo oh Neptidopsis Aurivillius N. ophione velleda Mabille oo f,w Eurytela Boisduval E. dryope angulata Aurivillius (98) oo w E hiarbas lita oo f Rothschild & Jordan (99) Hypolimnas Hubner H. misippus Linnaeus oo oh H. dubius wahlbergi Wallengren o f Salamis Boisduval S. anacardii anacardii Linnaeus (102) oo w,rf,oh in f S. parhassus Drury (103) oo f Junonia Hubner J. antilope Feisthamel o w J. archesia Cramer o w J. cuama Hewitson? ox w J. hierta cebrene Trimen oo oh Family, Genus & species Status Habitat J. limnoria taveta Rogenhofer o ah J. natalica natalica Felder oo w, fm J. octavia sesamus Trimen o w J. oenone oenone Linnaeus oo oh J. orithya madagascariensis Guenee 0 oh J. terea elgiva Hewitson oo f Catacroptera Karsch C. cloanthe cloanthe Stoll oo g Vanessa Fabricius V. cardui Linnaeus oo oh Antanartia Rothschild & Jordan A. abyssinica jacksoni Howarlh**(l 16)? h? Phalanta Horsfield Ph. phalantha aethiopica 00 w, oh Rothschild & Jordan ACRAEIDAE Bematistes Hemming B. aganice montana Butler ox f Acraea Fabricius A. acerata Hewitson o w A. anemosa Hewitson oo W, f A . braesia Godman o ah A. cabira Hopffer ? 0 fm, rf A. egina areca Mabille 0 f, rf A. encedon encedon Linnaeus oo oh A . eponina Cramer 00 oh A. esebria esebria Hewitson 00 f, rf, w A. johnstoni johnstoni Godman 00 f A. natalica natalica Boisduval 00 w, fm A. neobule neobule Doubleday**(129) ? oh? A. oreas oreas Sharpe o gf A. pudorella pudorella o w Aurivillius? (131) A. sotikensis Sharpe*(132) ? f Pardopsis Trimen P. punctatissima Boisduval? (133) o oh SATYRIDAE Gnophodes Westwood G. betsimena divers a Butler Melantis Fabricius M. leda helena West Bicyclus Kirby B. anynana anynana Butler B. safitza safitza Westwood Henotesia Butler H. perspicua Trimcn Ypthima Hubner Y. granulosa Butler (139) DANAIDAE Danaus Kluk D. chrysippus chrysippus Linnaeus oo oh Tirumala Moore T. petiverana Doubleday oo w, f 0 f, fm oo w oo w oo w oo w oo ag J. EANHS & Nat. Mus. 80 (196) December 1990 Norbert J. Cordeiro • Butterflies of Lake Manyara National Park, Tanzania 30 Family, Genus & species Status Habitat Amauris Hiibner A. albimaculaia interposita T albot o gf A. echeria meruensis Talbot** 7 hf? A. niavius dominicanus Trimen o gf LYCAEN1DAE Alaena Boisduval A. caissa caissa Rebel*** (145) 7 oh? A. ferrulineata Hawker-Smith** (146) o ah? Mimacraea Butler M. marshalli Trimen ssp.** (147) o gf Lachnocnema Trimen L. bibulus Fabricius (148) oo u Myrina Fabricius M. silenus ficedula Trimen o rf, w Aloeides Hiibner A . conradsi talboti 7 oh? Tite & Dickson** (150) Epamera Druce E. mimoseae rhodosense 7 w? Stempffer & Bennett** (151) E. tajoraca ertli Aurivillius** (152) ? aw? Stugeta Druce S. bow leer i nyanzana Wichgraf (153) o w Hypolycaena Felder H. pachalica Butler o aw H. philippus philippus Fabricius oo W, f Virachola Moore V. antalus Hopffer oo oh V. dinochares Grose-Smith oo w V. livia Klug o ah Anthene Doubleday A. amarah amarah Guerin de Meneville oo ah A. contrastata mashuna Stevenson oo ah A. definita definita Butler oo w A. indefinita Bethune-Baker oo f, rf A. kersteni Gerstaecker o w, f Family. Genus & species Status Habitat A. larydas Cramer OO aw A. otacilia otacilia Trimen o aw A. princeps princeps Butler oo w, aw Petrealea Toxopeus P. sichela sichela Wallengren o w Lampides Hiibner L. boeticus Linnaeus oo oh Cacyreus Butler C. lingeus Stoll oo w C. virilis Aurivillius oo w Leptotes Scudder L.jeanneli Stempffer* (171) ? w? L. pirithous pirithous Linnaeus oo w Tuxentius Larsen T. melaena melaena Trimen o aw 7 . stempfferi Kielland (174) X w Taracus Moore T. grammicus oo ag Grose-Smith & Kirby (175) Zizeeria Chapman Z. knysna Trimen o oh Acticera Chapman A. lucida Trimen o oh Zizula Chapman Z. hylax Fabricius oo oh Azanus Moore A.jesous Guerin de Meneville oo oh A. mirza Plotz oo aw A. moriqua Wallengren o oh A. nat alerts is Trimen ox oh Euchrysops Butler E. malathana Boisduval • oo oh E. osiris Hopffer oo oh E. subpallida Bethune-Baker oo oh Freyeria Courvoisier F. trochilus trochilus Freyer oo oh DISCUSSION Notes (2) Celaenorrhinus galenus. Understorey trees in the groundwater forest are favoured as territorial perches from which the male of this species swoops down on intruders. I observed that three or more males often occupied territories a few metres apart along road-sides, the edges of forest glades, and other open habitats within the forest. During such times, I saw two males “tumbling” over each other in the air, sometimes joined by a third or fourth party (n=17). Females crossing the paths of these males would be fought for by, usually, two males which would “tumble” over each other in the air and separate, one heading back towards its perch whilst the other pursued the female. Sometimes the female would avoid her pursuer, but on other occasions she would be forced to the ground in order to mate. J. EANHS & Nat. Mus. 80 (196) December 1990 31 Norbert J. Cordeiro • Butterflies of Lake Manyara National Park, Tanzania (8) Netrobalene canopus. Males of N. canopus are very territorial, occupying territories in open habitats throughout the groundwater forest. Each male perches on a favoured leaf of a branch that extends far out from the canopy of an understorey tree, often at approximately 2m from the ground. For five consecutive days I observed three males that shared territories on different branches of the same tree, also occupied by a male Graphium leonidas. They did not interfere with G. leonidas (or vice versa) but would fight amongst themselves for hours on end. First one male would fly into the territory of another and then an attack commenced whereupon the intruder was "tussled" with in the air. During such aerial “battles”, both would enter the territory of a third male, resulting in all three attacking each other. After several minutes of contest, one male would return to its perch. After further and short aerial combats, the other two would also retreat to their perches. (9) Spialia spio. This species was collected in L.M.N.P. (Rydon & Morgan-Davies, 1960a and unpublished). Three other species of this genus that might occur in L.M.N.P., especially along the escarpment, are S. colotes transvaaliae , S. mafa higginsi and S. zebra bifida. The two former species inhabit hilly and mountainous regions and have been recorded in the “Great Craters” (de Jong, 1978), referring to the Ngorongoro Highlands. The latter species has been recorded in Oldeani (de Jong, 1978), which is a part of these highlands. (11) Metisella medea is common in the groundwater forest and the forest at the southern end of L.M.N.P. However, I rarely found it flying in the woodlands at the Ndala Research Camp, suggesting that it might, occasionally, penetrate densely-wooded areas. (15) Gegenes hottentota hottentota. This species often flies together with G. nico brevicornis (Plotz), the latter insect being more common, according to Kielland (in litt.) who postulates that G. nico brevicornis should also occur in L.M.N.P. 18) Papilio echerioides ssp. A new subspecies of Papilio echerioides from the southern Kenya highlands and north and eastern Tanzanian mountains is currently being described by Kielland (in press). This species was rarely seen in the groundwater forest in the north of the Park. However, in the forest near the Yambi River area, the species flew in slightly greater numbers. Although I did not survey the Marang Forest on the escarpment, P. echerioides is much commoner in this higher altitude forest (Kielland, in litt.), and it is possible that a few individuals may migrate down into the lower forests. 20) Graphium angolanus angolanus. Not a positive identification because I did not manage to secure any specimens. The specimens that were flying at L.M.N.P. superficially resembled G. angolanus. As the species is well-spread across East Africa (D’Abrera, 1980; Carcasson, 1981) it is possible that it was G. angolanus and not any other Graphium of the pylades group. Kielland (in litt.) tells me that it is most definitely G. angolanus. (22) Graphium leonidas leonidas. The male of this species is territorial, perching on understorey trees in the forest and constantly driving off intruders. Its territorial behaviour is similar to the male of Euxanthe wakefieldii, which also occupies open habitats in the groundwater forest. Thus the two insects can often be found flying within the same patch of forest. I think that G. leonidas and E. wakefieldii confuse their predators by resembling each other in colour, pattern and behaviour (as well as their distasteful model Tirumala petiverana, which also flies in the same habitat). (23) Graphium policenes. Recorded by S. C. Collins (personal communication) as occurring in L.M.N.P. J. EANHS & Nat. Mus. 80 (196) December 1990 Norbert J. Cordeiro • Butterflies of Lake Manyara National Park, Tanzania 32 (26) Eurema desjardinsi marshalli. Typical E. desjardinsi occurs in Madagascar and the Comoros (D’Abrera, 1980; Carcasson, 1981). D’Abrera (1980) had recognised the African mainland subspecies as E. regularis, however, Berger (1980) subsequently revised the E. desjardinsi-regularis complex, raising both to species rank. Thus, the African mainland subspecies of E. desjardinsi has been given the next available name E. marshalli (Berger, 1981). Kielland (in lilt.) states that both E. desjardinsi and E. regularis sometimes fly in the same habitats, though E. desjardinsi is the commonest. He also mentioned that it is possible that E. regularis also occurs at L.M.N.P. (38) Colotis aurora dissociatus. D’Abrera (1980) had named this insect as C. eucharis dissociatus. Berger (1981) recently changed the name to C. aurora dissociatus. Several species of Colotis were preyed on by robber flies ( Alcimus spp., Diptera: Asilidae). Adult robber flies wait for their prey and then attack them in the air, relying primarily on sight (Shelly, 1987). Upon capturing prey species, robber flies then settle down and consume the body fluids of the victim (Shelly, 1987). At L.M.N.P., individual robber flies utilized open conspicuous sites in which to stake out their territories. For this reason I often observed robber flies along road-sides in the woodlands, and occasionally on forest paths. A robber fly would perch on grass at the road’s edge and, when it saw a flying insect it would fly out and attack it in the air. The robber fly would then pin the victim on the ground, secure its hold, and fly off into the grasses where it would begin to consume the body fluids. After feeding on the victim it would then drop it in the grass and fly off to its perch, ready to capture another insect. Slow-flying butterflies, like Appias spp., Belenois spp., Eurema spp., and some Lycaenids (see note 148), were among those fed upon after capture. (41) Colotis chrysonome chrysonome. There is a record of C. chrysonome in the BMNH collected by B. Cooper in July 1938 from “Ngaruka, north of L. Manyara” (personal observations) and below Mt. Oldeani in the Ngorongoro Highlands (Kielland, in litt.). (43) Colotis eris eris. A very common species, especially in the arid grasslands at the southern end of the lake. At times C. eris would frequent the Acacia woodlands but in smaller numbers. (51) Colotis pallene pallene has been recorded by T. A. Bams in the “District of Great Craters, [ii.-iii. 1921]” (Talbot, 1939), referring to the Ngorongoro Highlands region. (54). Colotis vestalis castalis is not a very common species at L.M.N.P. It occasionally flies among the Acacia woodlands and alkaline grasslands. (58) Belenois margaritacea plutonica is mainly a montane species, and for it to occur at a much lower altitude is an interesting record. I have also recorded this species at 900m in the Rau Forest Reserve in Moshi, about 800m below its normal altitudinal range on Mt. Kilimanjaro (Cordeiro, in preparation.). Both low altitude populations are most likely to be ecological variants. (61) Pieris brassicoides marghanita (= meridionalis). Only one specimen was observed and identified as belonging to this species, based on the underside pattern. I saw it in the lacustrine woodlands around the Ndala Research Camp (Eig. 1). This locality is much hotter and drier than its typical montane grassland (in association with forest) habitat in the Ethiopian highlands, and on Ngorongoro and Mt. Meru in northern Tanzania (Carcasson, 1964), thus it is unlikely that a pupulation exists at L.M.N.P. As P. brassicoides occurs higher up in the Ngorongoro Highlands (D’Abrera, 1980; Carcasson, 1981; Kielland, in litt., Rydon, in litt.) north-west of L.M.N.P.; it is likely that the specimen drifted downhill with the wind. Wind-assistance of lepidoptera on J. EANHS & Nat. Mus. 80 (196) December 1990 33 Norbert J. Cordeiro • Butterflies of Lake Manyara National Park, Tanzania mountains has been noted before by some authors (Salt, 1954; Robbins & Small, 1981). For example, in a study on wind-dispersal in Panamanian hairstreaks, Robbins & Small (1981) found that some high-altitude species were observed 5 km downwind from their normal habitats, this being due to the seasonal trade winds. Kielland (in litt.), however, thinks that its occurrence in L.M.N.P. “is very unlikely” and suggests that it should be collected for verification. Rydon and Collins are both of the same opinion. This taxon is interesting from a zoogeograph ical perspective as it is the only species of the Palaearctic genus Pieris occurring in Africa (Carcasson, 1964; D’ Abrera, 1980; Rydon, in litt.). It is unusual and a baffling phenomenon that this species is found below the equator and yet that it has strong affiliations with the Palaearctic butterfly fauna. (71) Libythea labdaca laius frequents river banks and the groundwater forest. It flies close to the ground, darting around in a typical Hesperiid fashion, often settling on damp sand to probe with its proboscis for moisture. (72) Charaxes achaemenes achaemenes. An aberrant female of Ch. achaemenes was baited in the lacustrine woodlands in the vicinity of the Ndala Research Camp. It was unusual in having an extensive blue-grey hind wing patch. Recently, Henning (1988: 228) figured a similar-looking aberration. Henning (in litt.) states that the extensive ‘grey’ hind wing patch in the figured specimen should actually be more bluish-green. The colouring of that plate is therefore somewhat misleading. Similar aberrations are known to occur in Ch.jasius saturnus (Henning, in litt.). As Henning has not given this aberration a name, I am taking the opportunity of doing so myself here: Charaxes a. achaemenes ab. glaucomaculata nov. (female) (Plate 1) (c.f. aberration figured in colour on p. 228 of S. F. Henning Charaxinae Butterflies of Africa, 1988. Description: Fore wing length: 36mm. Upperside : Pattern and colouration of f.w. typical Ch. a. achaemenes. Hind wing with a diffuse bluish-grey area extending inwards from between the anal angle and vein 6 (Ml). The bluish-grey obscuring to a certain extent the submarginal series of white spots and reaching almost as far as the middle of the wing; colouring and markings otherwise typical Ch. a. achaemenes. Underside: Typical Ch. a. achaemenes pattern. Holotype (female): Ndala Research Camp, Lake Manyara National Park, Arusha Region, Tanzania, 960m, 9. ix. 1987, N. J. Cordeiro. Specimen to be deposited in the Natural History Museum, London. Plate 1. Charaxes achaemenes achaemenes (Felder) (female) ab. glaucomaculata nov. I. EANHS & Nat. Mus. 80 (196) December 1990 Norbert J. Cordeiro • Butterflies of Lake Manyara National Park, Tanzania 34 (73) Char axes aubyni aubyni inhabits montane / sub-montane forests in East Africa. The popula- tion in the groundwater forest at L.M.N.P. is probably an ecological variant as it occurs in a habitat that is somewhat below its normal altitude range. (75) Charaxes bohemani. There was a battered specimen with no label in the park museum, which was probably collected by Morgan-Davies, as he had curated the insect collection during the early 1960s (Rydon, personal communication). This taxon, was not recorded in the Rydon & Morgan-Davies’s checklists (1960a, b and unpublished). It was very common in the Acacia woodlands at the southern end of the park but less common in the mixed woodlands in the north. (77) Charaxes candiope candiope is a very abundant species in the groundwater forest, but occurs in fewer numbers in the Acacia woodlands. It was observed feeding on exudations from Kigelia africana Lam. (Bignoniaceae) (n=35) throughout the park together with Euxanthe wakefieldii and Ch. varanes (see note 85). Charaxes candiope is a very aggressive butterfly, often molesting rivals of the same species or other members of the genus. Euxanthe wakefieldii was not disturbed by Ch. candiope, and both were frequently seen imbibing at the same wound (n=29). For interactions with Ch. varanes, see note 85. (78) Charaxes cithaeron kennethi. The L.M.N.P. population of Ch. cithaeron has been placed to kennethi by van Someren (1964: 232) who, however, qualified his action by saying that “the specimens from Arusha and Lake Manyara are less stable than typical coastal material.” L.M.N.P. is possibly the westem-most extent of this subspecies range. (81) Charaxes hansalii baringana. van Someren (1971, Map 2) records this species between Lakes Eyasi and Manyara, though he does not state the actual locality. I recorded Ch. hansalii in the lacustrine woodlands and on the escarpment where it apparently flies in very small numbers. Lequeux, (in litt .) tells me that this species is very local. In Rwanda he found it to be very common in a small area. (82) Charaxes jasius saturnus. This species flew in the woodlands and cultivated areas around the park, often probing the Castor Oil plants Ricinus communis Linn. (Euphorbiaceae) with its proboscis. (83) Charaxes kirki kirki. Rydon (1982), in his revision of the Charaxes viola group, has reinstated Ch. kirki to its original rank of full species removing its sub-specific status. Henning (1988) has accepted this revision. (84) Charaxes pollux maua. Typically a sub-montane / montane species, pollux was occasionally seen few in the lower altitude groundwater forest at the north end of the park. (85) Charaxes varanes vologeses was seen feeding on the sap of Kigelia africana with Ch. candiope and Euxanthe wakefieldii, along the forest margin at the north end of the lake (n=27 from September-December). Charaxes varanes showed an intolerant behaviour to the latter two species in that it always fed from a smaller sap source, away from Ch. candiope and E. wakefieldii. On most occasions when Ch. varanes came to feed, Ch. candiope would arrive to chase it away by a constant beating of wings (n=15). On three other occasions Ch. candiope did not exhibit rivalry towards varanes. I believe that the inferiority of Ch. varanes to Ch. candiope often led to the former species arriving at the food source to feed in the early mornings (0700-0900h) and evenings (1730-1830h) if weather permitted (n=21). Charaxes candiope was rarely present during these periods (n=2/21). Henning (1988) further provides evidence of the intolerance by competitors of Ch. varanes when he mentions that the butterfly establishes territories on the slopes rather than the J. EANHS & Nat. Mus. 80 (196) December 1990 35 Norbert J. Cordeiro • Butterflies of Lake Manyara National Park, Tanzania tops of hills, thus escaping molestation from hill-top territorial Charaxes. According to Rydon, (in litt.), however, Ch. lactetinctus Karsch (which maintains territories on hill-tops) chases Ch. varanes on the Tororo Hills (Uganda) and on the kopjes at Kabras (Kenya), even though Ch. varanes keeps lower down and tries to avoid contact with Ch. actetinctus. Nevertheless, on one occasion in November, I observed a battered specimen of Ch. varanes twice driving off a female of Ch. zoolina, suggesting that Ch. varanes is not inferior to all members of the genus, at least not to the smaller species, as I have observed elsewhere in eastern Africa. (86) Charaxes violetta melloni. van Someren (1966) mentions that the aggregate from “Tanzania: Newala, Iringa, Morogoro area, the higher zones of the Usambara Mts., and the area west of Kilimanjaro (Arusha-Meru)” represents the form/cline Ch. melloni Fox. More recently, this taxon was raised to subspecies rank (Henning, 1988). The population at L.M.N.P. probably represents the western and northern-most extent of range of this subspecies. (87) Charaxes zoolina zoolina , is a very common species of diverse habitats, and was observed feeding on exuding sap from two different tree species in September. In its lacustrine woodland habitat Ch. zoolina showed a preference for the juices of Maerua triphylla A. Rich. (Capparaceae), whereas in the groundwater forest it imbibed at wounds on Croton macrostachyus Del. (Euphorbiaceae) and a number of other unidentified forest trees. For further notes on inter-specific interactions see note 85. 88) Euxanthe wakefieldii. The behaviour and possible mimetic associations of this insect at L.M.N.P. has been recorded elsewhere (Cordeiro, 1988). Both sexes are partial to the exudations of Kigelia africana. The males are very territorial, driving away intruders from favoured perches at the edges of forests or in open areas in the forest. Females are uncommon, and are usually seen flying in the early mornings and late afternoons. For further information on the ecological relationships of this insect with some Charaxes, see notes 77 & 85. (90) Pseudacraea boisduvali trimeni. Although a common species throughout its range, I did not identify boisduvali at L.M.N.P., S. C. Collins (personal communication) mentioned that it does occur there. (98, 99) Eurytela dryope angulata and E. hiarbas lita fly in the understorey of forests and can often be seen feeding with the smaller species of Charaxes. When they do feed on the same tree as the larger Charaxes and beetles (Coleoptera), they are often to be found on the more distant wounds because the latter competitors tend to be very aggressive towards the smaller species of Nymphalidae. (102, 103) Salamis anacardii anacardii and S. parhassus. Open habitats within the forest are often favoured by these species in order to bask in the sun whilst perched on sun-lit branches. During periods of intense heat they take refuge in shaded spots or hang upside-down underneath large leaves. I came across both species feeding on elephant dung several times (n=7), together with some members of the Pieridae and/or Lycaenidae. (108) Junonia limnoria taveta. D’Abrera, (1980) mentions that this taxon occurs “in the extreme north of Tanzania”. I believe that the range of ssp. taveta in Tanzania extends from the north-east of the country westwards to the Great Rift Valley, as I have recorded this species from Kilimanjaro to Babati (Arusha region). (1 16) Antanartia abyssinica jacksoni. This species is common in forests of the Ngorongoro Highlands, north-west of the lake. I did not observe abyssinica in the groundwater forest of J. EANHS & Nat. Mus. 80 (196) December 1990 Norbert J. Cordeiro • Butterflies of Lake Manyara National Park, Tanzania 36 L.M.N.P. although Howarth (1966) mentions that it occurs on the western shore, possibly suggesting the Marang Forest where it does occur (Kielland, in lilt.). (127) Acraea johnstoni johnsloni was very common in the groundwater forest. It was also recorded from the lacustrine woodlands, where it was encountered infrequently together with A. esebria. It is typical that this sub-montane/montane butterfly was found in an arid woodland habitat at times, although Kielland (in litt.) notes that A. johnstoni often penetrates dense woodland when its forest habitat is not far away. (129) Acraea neobule neobule. This taxon was identified as A. terpsicore in Rydon & Morgan- Davies’s checklist (1960b, and unpublished). Kielland (in lilt.) notified me that Pierre (1978) recently separated A. neobule from A. terpsicore on the basis that the A. terpsicore has an Oriental distribution and different genitalia. (131) Acraea pudorella pudorella. I observed several specimens that resembled A. pudorella, flying in the mixed woodland throughout most areas in L.M.N.P. but did I not capture any specimens for verification. (132) Acraea sotikensis is probably a common species in the Marang Forest on the escarpment, and in the forest at the southern end of the lake. However, I did not manage to fully explore either habitats and thus record the species. Nevertheless, B. Cooper (coll. BMNH), in July 1938, obtained specimens of A. sotikensis “along [the] western shore in forest between lake and rift wall, Manyara”. (133) Pardopsis punctatissima. I might have seen this species flying in the arid grassland areas. Kielland, (in litt.) notes that P. punctatissima might inhabit the Maasai plains in northern Tanzania as it is a woodland and open-habitat species that occurs throughout Africa south of the Sahara. (139) Ypthima granulosa. Kielland, (1982) says that Y. granulosa inhabits “open, deciduous woodlands”. I found that it was more common in grassy-shrubby areas in the southern half of L.M.N.P., although it was also seen in smaller numbers in the Acacia woodlands. (143) Amauris echeria meruensis. Talbot, (1940) mentions that this taxon was collected by B. Cooper at “Lake Manyara, 3,000ft, vi-vii. 1937”. (145) Alaena caissa caissa occurs in “Manyara” [v. 1944, leg!, A. M. N. H.]. Kielland, (in litt.) notes that it is a very common species in this area. (146) Alaena ferrulineata. D’Abrera (1980) states that this insect has been recorded at “Ngorongoro Crater and Lake Manyara”. 147) Mimacraea marshalli. S. C. Collins (personal communication) was apparently the first person to collect this species in L.M.N.P. It is common in the understorey of the groundwater forest, often fluttering gently and settling on lichen-covered tree trunks. The Lake Manyara population is quite distinct in pattern from ssp. dohertyi Rothschild which occurs in the Kenya highlands (East of the Rift Valley) and at Mt. Meru in Tanzania. Rydon (in litt.) collected many specimens of the form dohertyi at Karamu and Lake Duluti in the foothills of Mt. Meru in April 1960. He only caught one specimen of f. marshalli which was larger than the specimens of f. dohertyi. Kielland (in litt.) thinks that the L.M.N.P. population is a cline between ssp. dohertyi and ssp. marshalli Trimen. J. EANHS & Nat. Mus. 80 (196) December 1990 37 Norbert J. Cordeiro • Butterflies of Lake Manyara National Park, Tanzania (148) Lachnocnema bibulus. This insect was often preyed upon by the Alcimus sp. For more information on the predatory behaviour of the robber fly, see note 35. (150) Aloeides conradsi talboti. Tile & Dickson, (1973) state that T. A. Bams and B. Cooper recorded B. talboti at the “north end of Lake Manyara” [ii. 1921 ] and on the “west shore of Lake Manyara” [ii. v. 1935] respectively. This taxon is interesting from a zoogeographical viewpoint, as it is one of three species of the genus Aloiedes located in Tanzania, the genus itself being mainly represented in southern Africa with over 30 species (D’Abrera, 1980). Aloeides conradsi has been recorded in the Ngorongoro Highlands (Tite & Dickson 1973; Kielland, in litt.), Singida and areas around Tabora (Kielland, in litt.) as ssp. talboti. West of Mt. Longido ssp. jacksoni occurs whereas [near] ssp. angoniensis occurs in western Tanzania at Kigoma and Mpanda (Kielland, in litt.). In southern Africa Aloeides ssp. generally inhabit grassland areas in a region of distinct seasons, quite unlike that of the equatorial regions. It is therefore of great interest that A. conradsi should be found some thousands of kilometres away to the north, in an environment that closely approaches the alpine climate. Aloeides conradsi has apparently adapted to moderately high-altitude grassland areas in eastern Africa, thereby living in weather conditions similar to those found in southern Africa. (151) Epamera mimoseae rhodosense. Recorded at L.M.N.P. by S. C. Collins (personal communication). He observed this species breeding on Loranthus spp. (Loranthaceae). (152) E. tajoraca ertli. S. C. Collins recorded this taxon feeding on Loranthus spp. at L.M.N.P. (153) Stugeta bowkeri nyanzana. D’Abrera (1980) states that the localities of this taxon in Tanzania are in the “southern and eastern shores of L. Victoria and Ukerewe Is.” Kielland (in litt.) states that this race also occurs in the region of the Northern Highlands. I found this insect to be very common in the lacustrine woodlands, often settling on bushes for long periods of time, apparently basking in the sun. (171) Leptotes jeanneli. B. Cooper collected this species at “Ngaruka” [2,800 ft, vi-viii, 1937], north of Lake Manyara. Kielland, (in litt.) mentions that the Leptotes species are very difficult to identify without dissection of the genitalia. However, he postulates that it is probable that L. pulchra Murray inhabits the marshy areas in L.M.N.P. (174) Tuxentius stempfferi. Larsen (1982) separated the African ‘ Castalius ’ from the Oriental Castalius on the basis that the genitalia of both groups were radically different. He proposed the new generic name Tuxentius for the African species. Kielland, (1976) described T. stempfferi which is closely related to T. melaena. Kielland, (in. litt.) identified a specimen that closely resembled typical T. stempfferi. This taxon has also been collected in nearby Qldeani in the Ngorongoro Highlands and as far away as Mikumi in Morogoro Region, (Kielland, 1976). That it has not been recorded in between the two distant localities is not surprising as Tanzania has been poorly collected and studied with regard to lepidoptera, and hence T. stempfferi may easily occur in many more places. (175) Tarucus grammicus. A very common species of the Acacia woodlands, has a preference for open, arid grassy habitats, as found in the Acacia woodlands which elephants have disturbed. The flight of this insect is weak and low, but due to its dark-brown upperside and zebra-patterned underside, I often lost sight of it in the long grass. When in flight, I think that the overall pattern and colours of T. grammicus enable it to escape from predators through this “disruptive colouration” - analagous to the effect created by the stripes of a zebra when seen from a distance. J. EANHS & Nat. Mus. 80 (196) December 1990 Norbert J. Cordeiro • Butterflies of Lake Manyara National Park, Tanzania 38 ACKNOWLEDGEMENTS I am especially grateful to Professor K. Hirji of the Serengeti Wildlife Research Institute (SWRI) and Mr. E. Kishe for giving me permission to work and conduct research in Lake Manyara National Park. Mr. C. Malima and Ms. P. Loesche I thank for their encouragement in my wildlife interests and for their support in this project whilst I was in Manyara. I am also thank- ful to the following who helped me in various ways: M. Boehnke, Y. Bouziane, M. Burengo, S. and L. Cordeiro, Dr. S. F. Henning, C. Gonzalez, M. Grossman, D. Harrison, Josephu and family, J. P. Lequeux, J. Timothy Mushi, E. Nyka, B. Orr-Wise, J. Osborne, and the Wardens and staff of L.M.N.P. I am grateful to my parents, Mr. and Mrs. H. J. Cordeiro, who encouraged my interests in Manyara. I am also grateful to Professor B. Wisner and the “Hampshire College Henry Luce Food, Resources and International Policy Program” for some financial assistance in the preparation and publication of this paper. My special thanks go to Mr. P. R. Ackery and Dr. F. H. Rindge for allowing me to study the African butterfly collections in the Natural History Museum, London and the American Museum of Natural History respectively. I express my deep gratitude to Mr J. Kielland who identified some of the difficult species and also contributed a lot of inform- ation on many of the species included in this list. Mr. S. C. Collins and Dr. A. H. B. Rydon also added a wealth of their knowledge to this paper for which I am very grateful. Initial drafts of this paper gready improved as a result of the constructive criticism and advice of Dr. C. D’Avanzo, Messrs. J. KieOand and P. Ackery, Dr. A. H. B. Rydon (who has also corrected and redrafted parts of this paper), and two anonymous referees, to all of whom I am greatly indebted. This is a Serengeti Wildlife Research Institute publication. REFERENCES Barns, T. A. (1924) Across the Great Craterland to the Congo. New York: Knopf. Beekman, J. H. & Prins H. H. T. (1989) Feeding strategies of sedentary large herbivores in East Africa, with emphasis on the African buffalo, Syncerus caffer. African Journal of Ecology 27: 129-147. Berger, L. A. (1980) Mission entomologique du Mus6e Royal de l’Afrique Centrale aux Monts Uluguru, Tanzanie. 25. Lepidoptera Papilionidae et Pieridae. Revue Zoologie Afrique 94 (4): 861-880. (1981) Les Papillions du Zaire. Weissenbruch: Bruxelles. Boshe, J. I. & Malima, C. (1986) Impact of Anthrax outbreak on the Impala population of Lake Manyara National Park, Tanzania. African Journal of Ecology. 24: 137-140. Carcasson, R. H. (1964) A preliminary survey of the zoogeography of African butterflies. East African Wildlife Journal 2: 122-157. (1981) Collins Handguide to the Butterflies of Africa. (Hard-back edition). London: Collins. Coe, M. J., Cumming, D. H. & Phillipson, J. (1976) Biomass and production of large African herbivores in relation to rainfall and primary production. Oecologia ( Berl .) 22: 341-354. Cordeiro, N. (1988) On the mimetic associations of Euxanthe wakefieldii (Ward) to its Danaid models (Lepidoptera: Nymphalidae, Danaidae). Bulletin of the East Africa Natural History Society 18(3): 30-33. D’Abrera, B. (1980) Butterflies of the Afrotropical Region. East Melbourne: Lansdowne Editions. Douglas-Hamilton, I. (1972) On the Ecology of the African Elephant. D. Phil, thesis, Oxford University. Greenway, P.J. & Vesey-FitzGerald, D. F. (1969) The vegetation of Lake Manyara National Park. Journal of Ecology 57: 127-145. J. EANHS & Nat. Mus. 80 (196) December 1990 39 Norbert J. Cordeiro • Butterflies of Lake Manyara National Park, Tanzania Henning, S. F. (1988) The Charaxinae Butterflies of Africa. Johannesburg: Aloe Books. Howarth, T. G. (1966) Revisional notes on the genus Antanartia (Lepidoptera:Nymphalidae). Bulletin of the British Museum (Natural History ) Entomology 18(2): 43pp. 5pls. de Jong, R . (1978) Monograph of the genus Spialia Swinhoe (Lepidoptera: Hesperiidae). Tijdschrift Voor Entomologie 121 (3): 23-146. Kalemera, M. C. (1987) Dry season diurnal activity of elephants in Lake Manyara National Park, Tanzania. African Journal of Ecology 25: 255-264. (1989) Observations on feeding preferences of elephants int the Acacia tortilis woodland of Lake Manyara National Park, Tanzania. African Journal of Ecology 27: 325-333. Kjelland, J. (1976) Some new and rare Rhopalocera from Tanzania (Lycaenidae: Satyridae). Entomologishe Berichteni 36: 105-112. (1982) Revision of the genus Ypthima in the Ethiopian Region excluding Madagascar (Lepidoptera, Satyridae). Tijdschrift voor Entomologie 125(5): 99-54. (in press.) The butterflies of Tanzania. Kristensen, N. P. (1976) Remarks on the family-level phytogeny of butterflies. Zeitschrift fur Zoologische Systemalik und Evolutionsforschung 14: 25-33. Larsen, T. B. (1982) The Butterflies of the Yemen Arab Republic. With a review of species in the Char axes viola- Group from Arabia and East Africa by A. H. B. Rydon. Biologiske Skrifter Danske Videnskabernes Selkskab 23(3): 1-61, Pis 1-2. Loth, P. E. & Prins, H. H. T. (1986) Spatial patterns of the landscape and vegetation of Lake Manyara National Park. ITC Journal 1986: 115-130. Makacha, S. & Schaller, G. B. (T969) Observations on lions in die Lake Manyara National Park, Tanzania. East African Wildlife Journal 7: 99-103. Morgan-Davies, A. M. (1964) Notes on some breeding birds of Lake Manyara N. P. Tanganyika Notes and Records No. 62: 96-104. Mwalyosi, R. B. B. (1977) A count of large mammals in Lake Manyara National Park. East African Wildlife Journal 15: 333,*335. (1981) Ecological changes in Lake Manyara National Park. African Journal of Ecology 19: 201-204. (1983) Utilization of Pastures in Lake Manyara National Park. African Journal of Ecology 21: 135-137. Pierre, J. (1978) Acraea from the Ivory Coast (Lepidoptera, Acraeidae). Bulletin de la Societie entomologique de France 83 (1/2): 3-22. Prins, H. H. T. (1988) Plant phenology patterns in Lake Manyara National Park, Tanzania. Journal of Bio geography 15: 465^480. 1989 Buffalo herd structure and its repercussions for condition of individual African buffalo cows. Ethology. 81: 47-71. & Bee km an J. H. (1989.) A balanced diet as a goal for grazing: the food of the Manyara buffalo. African Journal of Ecology 27: 241-259. & Iason, G. R. (1989) Dangerous lions and nonchalant buffalo. Behaviour 108: 262-295. & Loth P. E. (1988) Rainfall patterns as background to plant phenology in northern Tanzania. Journal of Biogeography 15: 45M63. Robbins, R. K. & Small G. B. Jr. (1981) Wind dispersal of Panamian Hairstreak butterflies (Lepidoptera: Lycaenidae) and its evolutionary significance. Biotropica 13 (4): 308-315. Rydon, A. H. B. (1982) Taxonomic notes on some members of the Charaxes viola group, with descriptions of three new species from Yemen Arab Republic and Ethiopia. In the Appendix to The Butterflies of the Yemen Arab Republic by T. B. Larsen. Biologiske Skrifter Danske Videnskabernes Selkskab 23 (3): 63-76. Pis. 3-5. J. EANHS & Nat. Mus. 80 (196) December 1990 Norbert J. Cordeiro • Butterflies of Lake Manyara National Park, Tanzania 40 & Morgan-Davies, A. M. (1960a) Checklist of butterflies received from A. M. Morgan-Davies Esq., F. Z. S. from Manyara in April 1960. unpublished. Ip. (1960b). Check-list of butterflies collected at Manyara in May 1960. unpublished. Ip. Salt, G. (1954) A contribution to the ecology of upper Kilimanjaro. Journal of Ecology 42: 375-423. Shelly, T. E. (1987) Natural history of three riparian species of robber flies in a Panamian forest (Diptera: Asilidae). Biotropica 19 (2) 180-184. Talbot, G. (1939) Revisional notes on the genus Colotis Hiibn. with a systematic list. Transactions of the Royal Entomological Society of London 88: 173-233. (1940) Revisional notes on the genus Amauris Hiibn. (Lep.) Transactions of the Royal Entomological Society of London 90 (10): 319-36. Snelson, D. (Ed.) (1986) Lake Manyara National Park. Tanzania National Parks and Kenya, Nairobi, African Wildlife Foundation. Tite, G. E. & Dickson, C. G. C. (1973) The genus Aloeides and allied genera (Lepidoptera: Lycaenidae). Bulletin of the British Museum ( Natural History ) Entomology 29 (5): 227-280, 5 pis. van Someren, V. G. L. (1964) Revisional notes on African Charaxes. (Lepidoptera: Nymphalidae). Part II. Bulletin of the British Museum (Natural History ) Entomology 15 (7): 181-235. 1966 Idem. Part III. Ibid. 18(3): 45-101. 1971 Idem. Part VII. Ibid. 26(4): 181-226. Vesey-FitzGerald, D. F. (1969) Utilization of the habitat by buffalo in Lake Manyara National Park. East African Wildlife Journal 7: 131-145. (1973) Animal impact on vegetation and plant succession in Lake Manyara National Park, Tanzania. Oikos 24: 314-324. Watermeyer, A. M. & Elliot, H. F. I. (1943) Lake Manyara. Tanganyika Notes and Records 15: 58-71. Watson, R. M. & Turner, M. I. M. (1965) A count of the large mammals of the Lake Manyara National Park: Results and Discussion. East African Wildlife Journal 3: 95-98. Weyerhauser, F. J. (1982) On the Ecology of the Lake Manyara Elephants. M. F. S. thesis, Yale University. The East Africa Natural History Society is most grateful to Barclays Bank of Kenya for their generous contribution towards the cost of publishing this Journal part. Manuscript received 11 April 1990 Editor: C. F. Dewhurst Published by: The East Africa Natural History Society • Typesetting by The Print Shop • Printed by AMREF J. EANHS & Nat. Mus. 80 (196) December 1990 Norbert J. Cordeiro • Butterflies of Lake Manyara National Park, Tanzania 41 Figure 1. Sketch Map of Lake Manyara National Park. (For more detail, see "Landscape ecological vegetation map of Lake Manyara National Park, Tanzania" [Loth & Prins 1986]). J. EANHS & Nat. Mus. 80 (196) December 1990 NOTICE TO CONTRIBUTORS Journal of the East Africa Natural History Society and National Museum Papers on Natural History subjects will be considered on the understanding that they are being offered exclusively to the Society and have not been submitted, at the same time, or previously, for publication elsewhere. Copyright is retained by the East Africa Natural History Society; references to contributions in the Journal may be made, but no part of the text, diagram, figure, table or plate may be reproduced without permission from the Editor. Such permission will only be granted with the approval of the authorfs). Two copies of papers for consideration should be sent to The Editor, Journal of the East Africa Natural History Society and National Museum, P.O. Box 44486, Nairobi, Kenya, and should be accompanied by a covering letter from the author, or in the case of multiple authors, from the author responsible for decisions regarding the text. Contributions may be sent as: 1 ) Preferably a floppy disk of either 5.25 or 3.5 inches in fully IBM compatible MS-DOS, carrying text processed in a commonly used wordprocessing language, or a floppy disk with the text prepared on an Apple Macintosh. If in any doubt, the author(s) should contact the Editor at the above address for further details. Three"near letter quality" printed copies on A4 size paper double-spaced should accompany the diskette. The author(s) should, of course, retain a back-up copy. 2) or as typescript, which should be on A4 size paper, typed on one side only, double spaced with margins of at least 5 cm on the left side of the paper to allow for editorial instuctions to the printer. Title: Should be brief and to the point. Authors name(s) and institution(s) where the research was carried out should be given, together with a suggested running page headline of not more than 50 characters including the names of author(s). A current address for correspondence, if applicable, should be included as a footnote. Abstract: To be on the first page and consist ofnot more than 150words. This should outline the major findings clearly and concisely without requiring reference to the text. Headings: Should be brief, and in the following format: "Abstract, Introduction, Methods, Results, Discussion, Acknowledgements and References". All headings to be typed on a separate line from subsequent text. Main headings to be typed in capital letters and centred on the page. Sub-headings to be at the left of the page, in lower case; further headings should be in italics or underlined (not bold) beneath them. The first letter of any headings should begin with a capital letter. Text: Numerals, dates & scientific units: Figures: Indent each new paragraph except those after a heading. Spelling should follow that of the Oxford English Dictionary. All pages must be numbered, at the top right hand edge. Prose numbers should be written in full up to and including ten, and as numerals thereafter, except at the end and beginning of a sentence. Dates should be written as "29 August", "10 — 13 January 1989" or "1969- 1989". Times should follow the 24 hour clock (e.g. 1500 h). All measurements must be in metric units. S.I. units and abbreviations should be used throughout with a space before and after the S.I. unit. Full stops should not be used after such units. The approximate position of Tables and Figures should be indicated in the left margin in pencil. The originals and two photocopies must be provided with each manuscript. Figures should be well drawn on pure white cartridge paper or good quality tracing paper. Lettering must be of very good quality (e.g. Letraset or good stencil). Typewritten or freehand lettering will not be accepted (without good reason). Each figure should have the authors' name, figure number and caption details written lightly on the reverse in pencil. Cite each figure in the text in consecutive order using arabic numerals. "Figure" should be abbreviated to "Fig.", except in the figure legends and at the beginning of sentences. Two separate caption sheets should accompany the typescript. Black and white photographs are acceptable if they have good definition and contrast and are produced on glossy paper. Consideration must be given by the author(s) to their final size when reduced for publication in the Journal which is B5. Tables: Each should be typed, double-spaced on a separate page and clearly numbered. Tables should be simple and self explanatory, and have a brief title above each one. They must be numbered consecutively with arabic numerals. Nomenclature: In cases where there are well known and internationally accepted English names, these should be used throughout the text. The scientific name should be given in full, underlined or preferably in italics, when quoted for the first time. The authority should also be quoted at this stage except in non-taxonomic ornithological papers. Only the generic name should begin with a capital letter. In subsequent reference to the name, use only the abbreviated generic name, with the full specific name but not the authority e.g .Spodoptera exempta (W alker); subsequently S. exempta. The authority should not be given if scientific names are used to describe an "association" or species complex, e.g. Acacia drepanolobium -Themeda triandra, wooded grassland. Type in capitals the first letter of the English names of species (e.g. Crowned Eagle, Grey-capped Warbler) but not of the higher taxa (e.g. eagles, warblers). References: Care should be taken to see that all references quoted in the text are also given in full, in alphabetical order, in this section. To avoid confusion, references to books should not be abbreviated and neither should they be underlined nor in italics. Journal names should be written in full, underlined or, if possible in italics. Personal communications should not be cited in the references, neither should they be abbreviated in the text. If there are two or more references cited by an author which were published in the same year, distinguish them by adding letters to the dates, e.g. Brown ( 1 972 a). For technical reasons author names in the reference list should NOT be typed in upper case. Authors are advised to refer to this issue of the Journal for the preferred format. Acknowledgments: As acknowledgements may imply endorsement of the data and conclusions in the authors' work, it is advisable to obtain permission before quoting persons who have made contributions to the study. Once the paper has been selected by the Editor for possible publication in the Journal, the manuscript will be refereed by two persons who are acknowledged specialists in that field before a decision is made on its acceptance. The author(s) will be sent one set of print-outs for the correction of typographical errors. Authors are urged to ensure that all changes to the text are incorporated into this copy, as no further changes can then be made. The author is entitled to 25 copies free of charge. Thirty free copies will be provided if there is more than one contributor. Further copies can be supplied at cost, provided the order is placed before printing has been undertaken. The East Africa Natural History Society is supported by membership subscriptions, and may be unable in some instances, due to financial constraints, to accept all submissions of value. While every effort will be made to minimise such restrictions, contributors are requested to seek financial support to assist the Society with the ever increasing costs of publication. Editor Journal East Africa Natural History Society and National Museum. January 1991 Referees: Proofs: Offprints: . * i ■ JOURNAL OF THE EAST AFRICA NATURAL HISTORY SOCIETY AND NATIONAL MUSEUM June 1991 Volume 81, Number 197 NATURAL INDICATORS OF SHALLOW GROUND WATER IN KIBWEZI DIVISION, KENYA MELVIN WOODHOUSE Environmental Health Unit African Medical and Research Foundation, ABSTRACT Human ability to locate sources of ground water using natural indicators is a skill developed from experience which improves with use. This ability, essential in many environments for survival, has to be nurtured. This paper describes the use of natural indicators to detect ground water in Kibwezi Division, Kenya. These skills have been developed and adapted during the period of recent settlement of the area in the early 1 970s. The local people have become self-reliant in siting wells with a success rate of 70% in part of the area. The purpose of this paper is to record these methods before they become redundant, replaced or forgotten. Already the reduction of natural tree cover in Kibwezi has reduced the variety of signs available. INTRODUCTION Kibwezi Division, situated some 200 km south east of Nairobi (Fig. 1) has erratic rainfall averaging 644 mm per year. The Division covers an area of 8000 km2 and experiences droughts during both rainy seasons in 60% of years (Fenner, 1982a). In the past, the area was covered with tangled scrub such as Acacia mellifera (Vahl) Benth. and was home for a wide variety of game. Maasai passed through the area and the Akamba people hunted there, although few made permanent settlements in the area in the last century. In 1891 the Lovedale Mission was established in Kibwezi near to the present site of the railway station. Due to an earth tremor in 1896 (Younge, 1977) the Kibwezi river, then the main source of water, increased its flow. The Mtito Andei river and the Masongaleni river were also important sources of water in the area and consequendy were stop-over points for Arab and Akamba traders. Mackinnon’s road reached Kibwezi in 1895, from where Sclater’s road began as the continuation towards Uganda. With the establishment of a road through the area, enterprising persons ventured into the inhospitable environment and attempted to grow rubber, to harvest Sansevieria and, more success- fully, to grow sisal. This all required permanent sources of water, both for agriculture and human consumption. The most reliable and readily available sources were the rivers. Undoubtedly wells were dug at prime sites not far from the visible surface water sources. Address for correspondence: AMREF, P.O. Box 30125, Nairobi, Kenya. Melvin Woodhouse • Natural indicators of ground water in Kibwezi 2 The railway line was built soon after the road and water sources for the steam engines, initially railed to the stations, were secured.'A number of stations to the east of Kibwezi were supplied via a pipeline from Umani spring in the nearby Chyulu Hills. The water from this pipeline still supports a considerable number of inhabitants. In 1918 a second major earth tremor reduced the flow of the Masongaleni river considerably, but it was partially restored in the 1930s. It became obvious that without adequate ground water any development in the area would have to depend totally on the rains or the rivers, of which only the Kibwezi appeared reliable, and of good quality. There was therefore considerable incentive to investigate ground water resources. The geological knowledge of the area was based on an early speculative, but perceptive, report (Brantwood Muff, 1908). It was not until 1954 that a more thorough investigation was carried out in a study which established the basis for the present-day understanding of the hydrogeology of the area (Temperley, 1955). The colonial administration was active during the 1930s drilling boreholes, as was the Kenya Sisal Company at Dwa farm. A more recent study has served to quantify the system suggested by Temperley (British Geological Survey et al., 1988). The Chyulu Hills, which form the southern boundary of the Division, rise 900 metres above the surrounding country. Benefiting from an average annual rainfall above 1200 mm, they form the focal point of hydrogeological studies. The Chyulu Hills are of recent volcanic origin and as a result of their high porosity there is no rainwater run off. The rainwater percolates down into the lava and meets an Archaean basement of the Kasigau and Kurase series. Water flows east on top of the basement to emerge at Mzima Springs, the source of Mombasa’s water supply. Temperley (1955) estimated that 80% of the rainfall on the Chyulus reappears at Mzima Springs and that the remaining 20% feeds the springs of the local rivers, showing that very little Chyulu water is recharging local aquifers. The basement system emerges from under the Chyulu lava and gives rise to the flat, red landscape characteristic of the area. Along the edge of the lava there exists a mosaic of gneisses which are poor but useable aquifers. It is these aquifers which are now being tapped by local people. Significant numbers of people began to settle in Kibwezi with the immigration of Akamba in the late 1960s, the previous restrictions on settlement now only applying to the Chyulu Hills. The availability of drinking water became a priority and the Government and non governmental organisations began to develop water resources in the area. With a present population of about 180,000 people a great demand for water exists. In a 1985 survey, the average distance people travelled to a water source was 4.4 km, the average amount of water obtained was eight litres per head per day, and a single journey took an average round-trip time of 90 minutes returning with 20 litres of water of dubious quality (Ferguson et al., 1988). Efforts are continuing in the Division to improve the water-supply situation via pipeline schemes, rainwater harvesting, boreholes and shallow wells. In the case of shallow wells, this is largely a community-based activity. The local people organize themselves into well groups, select a suitable site to dig and seek assistance with construction from agencies active in the area. Well-site selection using geophysical equipment has not been particularly successful in the past (Temperley, 1965) largely because no alternatives to resistivity techniques were available in Kenya. Well groups usually have only their own resources to call upon when siting a new well and consequently do not have the option of an elaborate geophysical survey. As a result, they have used natural indicators to help them. As the population of the Division increases and more land is cleared for farming, indicator trees, in particular, are becoming scarce. At the same time, advances have been made in geophysical techniques which are enabling cheaper and more accurate well siting to be done. The use of natural indicators may well become redundant. J. EANHS & Nat. Mus. 81 (197) June 1991 3 Melvin Woodhouse • Natural indicators of ground water in Kibwezi METHODS Over a six-year period the author became familiar with the current state of scientific knowledge of water resources in the Kibwezi area. Local knowledge was gathered informally through discussions with local Well group members and leaders. The African Medical and Research Foundation (AMREF) has held annual training workshops for groups constructing wells. There are now over 80 such groups. The workshops were a valuable forum for the exchange of local knowledge and served to consolidate the information gained through discussion. In particular, the close association between AMREF field staff and the Well groups they have assisted has led to a deeper insight into local methods for well siting. RESULTS There is increasing reliance upon machines to assist in locating ground water. In Zimbabwe in 1987, when using integrated geophysical techniques, a success rate of 90% was obtained for boreholes. When sites were allocated “logistically” the success rate was 50% (White, 1987). These interpretations still require human judgement to be made, whether computer assisted or not. In cases where people do not have access to modem technologies, they can only hope to use their own judgement to interpret their senses to find ground water. In Kibwezi a success rate of 70% has been obtained for siting wells in areas where natural indicators exist. The community is not reliant upon outside technical expertise and is in a position to continue unaided to find well sites. Where no alternatives exist there is little need to argue the merits of such an approach and the methods should only be judged by their success. Since the exploitation of ground water is essential to sustain human life, it would be expected that our senses do give indications of the location of ground water. Natural indicators of ground water in Kibwezi The indicators identified in Kibwezi may not be suitable for use in other areas. In fact in Kibwezi itself the indicators are not ubiquitous. The people of Kibwezi are currently making accurate identifications of the indicators discussed below, and this is reflected in the fact that many of the indicators have specific local names. Good knowledge of the environment has always been important for survival in Kibwezi, a fact still evident as well groups continue to develop their skills in locating ground water. For reasons of clarity the indicators are here put into six groups. Topography The flat topography of Kibwezi is punctuated only by resistant inliers and seasonal river channels. The more resistant basement geology has formed the watersheds of the area. The land drains north into the Athi river. As there is no run off from the Chyulu Hills, seasonal rivers rise where the lava ends, which having eroded lines of least resistance have preferentially cut into the softer gneisses, thus indicating their location. The gneisses to the east of the main road are less permeable and therefore poorer aquifers. The lack of definitive indicators of ground water on the Chyulu lava meant that Well groups applied their efforts away from the lava belt and initially concentrated on the larger seasonal river beds. The edges of lava flows are easily recognizable as they are several metres thick, end abruptly and have very little soil cover. The lava is black, often set against a red soil, and supports green vegetation for much of the year. The most productive wells are located in river basins generally not more than 2 km from the edge of the lava flow. Typical yieids from these wells of depths between 6.5 and 7.5 metres are above 33 m per day. J. EANHS & Nat. Mus. 81 (197) June 1991 Melvin Woodhouse • Natural indicators of ground water in Kibwezi 4 Without exception, the Well groups chose to dig in low-lying places, locating their wells in areas of softer gneiss rock and away from the basement watersheds. As more wells were dug, the people quickly realized that the closer they were to the watersheds the deeper they would have to dig, and there was the increased likelihood of a low yield. As experience grew, initial siting of wells began with a general survey to recognize the location of watersheds. Generally, the drainage channels are well defined and there was little difficulty in excluding potential basement sites and the watersheds as they bear different soils and vegetation. Some confusion occurs where seasonal water sources exist at the edge of basement outcrops. These sites have, without exception, failed to provide permanent water. Such sources are probably fed from local rainfall recharge on the outcrops, the water being stored in cracks. The storage volume is low and likely to be subject to evaporation losses. There are a number of old earth dams in the area which were explored as potential well sites. The dams generally only store water until September and are thus useful but inadequate during the dry season. As the dams dry out people make shallow excavations in the dam floors and along the walls thus gaining accurate information about the location of water. Frequently, groups went on to dig wells downstream of the dam walls. Presumably, the dams leak from the floor thus recharging a small area of local aquifer. Digging at the sides and upstream of the dams has so far proved un- successful. As most of the wells are dug along river channels, flood damage is a risk. Initially this risk was not a consideration in siting a well and certainly previous experience with shallow holes in river beds had resigned their users to the inevitable annual chore of re -excavation. With the location of good wells outside the channels, better practice quickly caught on. Sites could still be improved by a more thorough consideration of flood levels, easily done by checking for past flood debris caught up in riverine trees. Similarly, wells could also be better sited after consideration of the erosion patterns of the river — wells dug on the inside of meanders are less likely to be damaged in the future. Changing patterns of land use are also increasing the run off carried by seasonal rivers. I . and clearance for farming, resulting in less retention of rain water, is extending the drainage systems. It is now quite common to see drainage channels which have encroached upon sites of previously stable vegetation. Well groups should be particularly wary when they choose to dig a well near a river where there are mature trees being undercut in the middle of the channel; this is clear sign of rapidly increasing erosion. Geology The Kibwezi area can be divided into two distinct geological regions, the area underlain by basement system gneisses, and the volcanic region to the south of the railway (Saggerson, 1963). (Fig. 2) There are three main classes of rock of interest to potential well groups: lava, gneiss and undifferentiated basement rocks. These rocks have specific names in the Kamba language of Kibwezi; they are, respectively; kivuthi, ingee and nganza. Being agriculturalists, the Akamba who moved into Kibwezi were well aware of the potential of the various soil types encountered and had a good knowledge of the rock types in the area. Although this knowledge is more general than specific, it was quickly applied in the search for ground water. Experience gained from successful wells has enhanced common knowledge and soils and geology are now included amongst the natural indicators of ground water in common use. The siting of wells on lava is a very difficult task for which there is little local expertise. Suit- able geophysical techniques have yet to be found, although some very productive wells have been dug in lava flows. Being a haphazard process, few groups are willing to risk wasting their energies and resources digging in lava. It is important, however, to establish where a lava-flow ends when siting a new well. In following the courses of old rivers, which cut preferentially into the softer J. EANHS & Nat. Mus. 81 (197) June 1991 5 Melvin Woodhouse • Natural indicators of ground water in Kibwezi gneisses, lava flow boundaries point towards the location of suitable digging sites. The soils found near the edge of a lava flow are a hybrid between black cotton soil and red laterite soil. These soils are rarely deeper than five metres and the gneiss aquifers are found beneath them . At the edge of the lava flow, where water may seep into gneisses from local recharge, many of the tree species which are indicators of water may be seen. When river channels arise at the end of a lava flow, exposures of limestone may be seen. These were laid down in places where water collected over a long period of time, presumably showing historic drainage systems which cut into softer rock. They indicate suitable sites for well digging. Goats are quick to identify these outcrops and use them as salt licks. River channels which expose geology are surveyed by local Well groups. They look for the darker biotite-rich gneisses which are more productive aquifers. It is rare to find a homogeneous gneiss more than four metres thick. The gneiss hardens off and becomes less productive at depth until it is impossible to dig by hand. When banded with schists it can still be productive at depths of up to 26 metres. Due to the very mixed nature of these metamorphic rocks, local knowledge has not yet developed to the point at which accurate determination of possible well depth can be made from surface geological information alone. Once digging has begun, and the gneiss reached, its colour and hardness are interpreted to give a reasonably accurate prediction of the depth at which water may be found. In some cases, bars of gneiss can be seen cutting across river channels where they form good sites for sand dams. As these bars may have cut through productive gneiss, they help retain water in the buried gneiss. Wells have been successful downstream of these bars where presumably leakage is occurring and the bars are acting as sub-surface dams. Such areas are frequently explored as seasonal water holes by the local people. Undifferentiated basement outcrops forming inliers are not good sites for well construction. The soils in the vicinity of the outcrops are sandy, reddish and often contain large quartz fragments. Such soils are good negative indicators of water presence even when rock does not outcrop. On the eastern side of the main Mombasa road, the river channels are considerably larger; some have almost permanent baseflows which are held in the sandy river beds. The sands found in these locations are good for construction work and can be over 12 metres deep. Some productive wells have been sunk into these sands using concrete rings. Caution is needed in such places since not only are flash floods a major hazard but also the performance of the proposed well in the dry season must first be established. There are good murrains beneath the topsoil on the eastern side of the road. When people have dug pits for latrines here they have found that water has entered. These stony murram layers are usually not thicker than one metre, thus their storage capacity is small. So far a permanent dry- season source of water has not been found from these murrains. There are some clays and soils in the area which act as impermeable layers preventing rainwater from percolating downwards. Water is often taken from these clay pan areas, but the pans are heavily influenced by evapo transpiration and rarely remain wet throughout the dry season. As more experience is gained from wells that have been dug, the richer the understanding of the relationship between geology and ground water becomes. Obviously local knowledge of geology will not develop to a specialist level, nonetheless the present basic understanding is a significant contribution to well siting. Trees Table 1 gives the English, Kamba, Swahili and scientific names of trees found in the area. Longland (1952) refers to “ Mkuju ”, ( Ficus sur Forsk.) together with Acacia tortilis (Forsk.) as indicators of ground water in Tanzania. The use of trees as indicators of ground water is a very old practice, however, the usefulness of some species is probably only applicable to certain areas. For example, Longland’s (1952) observation that ground water is found at a depth of three times the height of the crown of A. tortilis has not been shown to be the case in Kibwezi. It is important to J. EANHS & Nat. Mus. 81 (197) June 1991 Melvin Woodhouse • Natural indicators of ground water in Kibwezi 6 realize that trees may be associated with a certain soil type and environment and not directly with the ground water table itself. A good general description of the woodland of Kibwezi is given in Fenner (1982a). The Kamba people were new to this area in the 1970s but knew most of the trees they came across. The names used locally for trees are not all of Kamba origin; many are similar to Taita names. Presumably these names were imported by the Akamba settlers. They had never needed to look at these trees in relation to ground water in this area before; rather they had to learn from experience. As the number of successful wells has increased, it has been possible to compare local tree cover with the degree of success. As a consequence, confidence and reliability have increased and now, once a general site has been selected, the tree species present are used as the principal indicator of the exact site to dig. The most successful tree indicator of good-quality ground water at shallow depth so far identified is Acacia robusta. (Taub.) Brenn (Figs. 3, 4) Where this tree is found and supported by topographical indicators, ground water of salinity below 1500 fiS1 can be obtained at depths of less than ten metres. The appearance of A. robusta can differ greatly depending upon its environment, and the anatomy of its non-reproductive parts is diverse. Acacia gerrardii (Benth.), Acacia xanthophloea (Benth.), and F. sur are all useful indicators of ground water at shallow depth and their presence in conjunction with A. robusta confirms a good site. Hyphaene compressa Gaeren is a very good ground water indicator but it is only found in Mangalete Sub-location of Kibwezi. Where present, water can be obtained at depths of less than seven metres. Grewia bicolor (Juss.) is frequently used as a forked stick by dowsers3. Even though found near river valleys, its presence has not been associated with ground water. Table 1 . Botanical names Generic/specific names English Kikamba Kiswahili Acacia gerrardii (Benth.) Gerrards acacia Muthithiu - Acacia mellifera (Vahl) Benth Hook-thorn Muthia Kikwata Acacia robusta (Taub.) Brenan Munina - Acacia tortilis (Forsk.) Hayne Umbrella thorn Mulaa, Kilaa Mgunga Acacia xanthophloea (Benth.) Fever tree Mulela Mukonge Cyperus papyrus L. Papyrus Ndoi Ndago mwitu Ficus sur (Forsk.) Cape fig Mukuyu Mkuyu Grewia bicolor (Juss.) Mulawa Mkone Hyphaene compressa (Gaeren) Newtonia hildebrandtii. (Vatke) Torre Doum palm Malala Mukami Mkoma Sansevieria. spp. Sphaeranthus cyathuloides. O. Hoffm. Sterculia rhynchocarpa (K. Sch.) Bowstring hemp Musonzouwa Kyusia Mkonge '(iS or microsiemen is a unit of the electrical conductivity of water and is proportional to its salinity. 3water diviner J. EANHS & Nat. Mus. 81 (197) June 1991 7 Melvin Woodhouse • Natural indicators of ground water in Kibwezi Some tree species have also been found to be useful negative indicators of ground water thus enabling an area to be excluded from a survey. Acacia mellifera (Vahl) Benth. and Sterculia rhynchocarpa (K. Sch.) generally exist where there is very little chance of finding shallow water. Thus, by elimination, the margins of aquifers can be inferred. Even so, S. rhynchocarpa can be seen growing close to seasonal river beds. The change in tree cover above different geological systems is readily observable. Newtonia hildebrandtii (Vatke) is a very useful indicator of water, but due to the high value of its timber for carving and charcoal it is now rarely found in the settled areas. Mature specimens of N. hildebrandtii can be seen near to the springs and dams protected on Dwa sisal estate. A rough estimate of the number of wells in Kibwezi with tree indicators in their immediate vicinity is 30%. As more land is cleared for agriculture it can be expected that in the not too distant future the natural tree cover of the area will have been removed. Herbaceous plants No herbaceous plants have been recognized in Kibwezi as reliable ground water indicators. Those which were considered were mainly wet-area plants and included marsh grasses, rushes and Cyperus papyrus L. and Sphaeranthus cyathuloides (O. Hoffm.) which can be found near most of the wells in the rainy season. The latter species appears to enjoy the moist freshwater environment resulting from retained run-off. Presumably it is for the same reasons that it is found growing near to granite inliers. Even so, S. cyathuloides does not thrive in areas of stagnant water and thus may indicate good quality local recharge, although this has not been confirmed by the author. Areas which retain some green cover during the dry season are likely well sites. The obvious exception is on lava flows where moisture can be retained in pockets of black soil and support some green cover for most of the year. Fauna In the past, the area supported abundant wildlife and many of the earlier settlers remember places where animals could be seen drinking water. Some people insist that Elephant ( Loxodonta africana ) (Blumenbach), Warthog ( Phacochoerus aethiopicus ) (Pallas) and Zebra ( Equus ( Hippotigris ) burchelli ) Gray, are all able to detect shallow ground water and dig for it. Even though some people say that their wells were sited where these animals used to drink, there is little consensus of opinion as to whether this is a true relationship or not. Bees (Hymenoptera: Apidae) are undoubtedly fond of wet areas and are very common around wells. Local bee-keepers say that bees will not occupy a hive if the nearest source of open water is more than 1.5 km away. However, bees are not regarded as a definitive indicator of ground water. It is also well known that termites do not build mounds in places which are wet. This can be seen particularly when trying to determine where spring lines exist when springs are non-emergent. The termite mound margin on the uphill side of a swampy area would normally mark the limit for attempted excavations. Perception by the local people There are numerous dowsers in the area and some have a good reputation having sited good wells. Many of them utilize natural indicators together with a forked stick (Fig. 5). Sticks of G. bicolor are commonly used. The dowsers use a simple technique. They have not been seen to make multiple passes over a specific site, but sometimes do give clients an indication of depth and yield but not of quality. There are also a number of very poor dowsers in the area. As considerable amounts of money can be made from dowsing, this is not surprising. At some particularly poor sites a dowser may specify a considerable depth to water, through a rock which cannot be dug by hand, thus maintain- ing his reputation, since the depth specified can never be reached. The dowsers encountered by the J. EANHS & Nat. Mus. 81 (197) June 1991 Melvin Woodhouse • Natural indicators of ground water in Kibwezi 8 author have all been men. It would be useful to study the abilities of the local dowsers in detail and to introduce them to the approaches and experiences of established dowsing societies. At a few wells people have said that a further reason for choosing the site was a sensation of warmth or cold when walking past the site at night. This may be linked to a microclimatic effect resulting from ground water or from the fact that the site was in a low-lying enclosed area. It is not disputed that such sensations exist. Table 2. Summary of natural indicators for detecting shallow ground water in Kibwezi Highly likely Near seasonal/permanent river beds. Within 2 km of the edge of the lava belt. Presence of A. robusta, A. gerrardii, A. xanthophloea Absence of termite mounds, S. rhynchocarpa and A. mellifera Likely Absence of basement outcrops. Near to a dam. ? Green cover in the dry season. In a low-lying area. Surface outcrops of mafic schist. Approved by a reputable dowser. Humid / cool sensation in still weather. Poor Presence of murram. On high ground. Near granite outcrops. Sandy/quartz soils. Water quality It is important, when promoting the methods mentioned, that as far as possible good-quality water is identified. This is due to the need for drinking water and because digging a well by hand is a major undertaking. The average salinity of the wells dug so far is 1900 pS. This is high because 17% of the wells had a salinity above 3000 pS. Water is often potable up to 2500 pS, whilst international standards may be set as low as 400 pS. At present an investigation is being carried out to relate salinity to soil type and vegetation and it is apparent that the high-salinity wells are restricted to a small geographical area. It is hoped that, should a relationship be identified, then the vegetation of the “salt aquifer” areas could be used as a negative indicator to prevent people from digging where such an aquifer exists. It is a major concern that the change in land use, which is resulting in the removal of natural bush cover, will increase evapotranspiration losses from the soil. This in turn may increase the salinity of the well water. The resultant increase in run off also represents the loss of a valuable resource as well as a threat to the stability of river morphology. There is a considerable wealth of local knowledge and awareness of the environment in Kibwezi. This may change as land pressure increases and new technologies are introduced. In such a fragile ecosystem the inhabitants represent the most powerful force to conserve or destroy the environment. Fortunately, the people of Kibwezi have an intimate awareness of their environment and it is encouraging to see the present trend of community-based conservation strategies emerge. The harnessing of the knowledge and attitudes of people in Kibwezi could tip the balance in favour of successful conservation of the area. J. EANHS & Nat. Mus. 81 (197) June 1991 9 Melvin Woodhouse • Natural indicators of ground water in Kibwezi ACKNOWLEDGEMENTS I thank the following people for their interest and keen observation which have contributed to this paper: Costas Emmanuel, John Kavali, Tom Konana, Samson Mutuku and George Wambua. Special thanks to Gerry Dekker for persevering with identifying the Acacia species and to Anne Bimie who provided very useful comments on the paper and produced the illustrations of A. robusta. REFERENCES Brantwood Muff, H. (1908) Report relating to the geology of the East Africa Protectorate. Colonial Reports misc. no. 45. HMSO. British Geological Survey, (1988) Ministry of Water Development, Institute of Hydrology (U.K), Ministry of Environment and Natural Resources. The Chyulu Hills Water Resource Study. WD/88/5C . Fenner, M. (1982a) Aspects of the ecology of Acacia-Commiphora woodland near Kibwezi. Journal East Africa Natural History Society and National Museum, no. 175, 1-12. (1982b) Features of the rainfall at Kibwezi, Kenya. Journal of the .East African Agriculture and Forestry Research Organisation. 45 (1) 83-91. Ferguson, A., Absolom, E., Kogi, W. & Omambia, D.(1985) Kibwezi Integrated Survey. African Medical and Research Foundation, Nairobi. Longland, F. (1952) Field Engineering. Dar es Salaam. HMSO. Saggerson, E.P. (1963) Geology of the Simba-Kibwezi area. Geological Survey of Kenya, Nairobi. Temperley, B.N. (1965) The location of ground water in areas of Archaean gneiss in which depth probes give resistivity curves of constant gradient. Proceedings of the International African Conference on Hydrology. CCIA. Nairobi. (1955) A study of the movement of ground water in lava covered country. Ministry of Environment & Natural Resources, Kenya, (unpublished). White, C. (1987) Borehole siting using geophysics. In: Developing World Water. (Ed.) J. Pickford. pp 106-113 Hong Kong: Grosvenor Press. Younge, B. (1977) The Uganda road. Kenya Past and Present, no. 8. Kenya Museum Society. Editor: C. F. Dewhurst Published by: The East Africa Natural History Society • Typesetting by The Print Shop • Printed by AMREF J. EANHS & Nat. Mus. 81 (197) June 1991 Melvin Woodhouse • Natural indicators of ground water in Kibwezi 10 J. EANHS & Nat. Mus. 81 (197) June 1991 11 Melvin Woodhouse • Natural indicators of ground water in Kibwezi Volcanic lava Metamorphic rock Laterite soils over basement granite Figure 2. Geology of Kibwezi Division N ♦ Figure 3. Acacia robusta mature tree J. EANHS & Nat. Mus. 81 (197) June 1991 m Melvin Woodhouse • Natural indicators of ground water in Kibwezi 12 Figure 4. Acacia robusta: flower cluster, leaf and pod J. EANHS & Nat. Mus. 81 (197) June 1991 13 Melvin Woodhouse • Natural indicators of ground water in Kibwezi Positive signs Likely signs Presence of A. robusta A. gerrardii A. xanthophloea Near to a river bed Within 2 km of lava belt Near to a dam Green cover in dry season Chyulu Hills Lava flow Seasonal river bed Latcritc soils Basement outcrop Negative signs Basement outcrops On high ground Sandy/Quartz rich soils Presence of A. mellifera Figure 5. Summary of natural indicators for detecting shallow ground water in Kibwezi J. EANHS & Nat. Mus. 81 (197) June 1991 NOTICE TO CONTRIBUTORS Journal of the East Africa Natural History Society and National Museum Papers on Natural History subjects will be considered on the understanding that they are being offered exclusively to the Society and have not been submitted, at the same time, or previously, for publication elsewhere. Copyright is retained by the East Africa Natural History Society; references to contributions in the Journal may be made, but no part of the text, diagram, figure, table or plate may be reproduced without permission from the Editor. Such permission will only be granted with the approval of the author(s). Two copies of papers for consideration should be sent to The Editor, Journal of the East Africa Natural History Society and National Museum, P.O. Box 44486, Nairobi, Kenya, and should be accompanied by a covering letter from the author, or in the case of multiple authors, from the author responsible for decisions regarding the text. Contributions may be sent as: 1) Preferably a floppy disk of either 5.25 or 3.5 inches in fully IBM compatible MS-DOS, carrying text processed in a commonly used wordprocessing language, or a floppy disk with the text prepared on an Apple Macintosh. If in any doubt, the author(s) should contact the Editor at the above address for further details. Three"near letter quality" printed copies on A4 size paper double-spaced should accompany the diskette. The author(s) should, of course, retain a back-up copy. 2) or as typescript, which should be on A4 size paper, typed on one side only, double spaced with margins of at least 5 cm on the left side of the paper to allow for editorial instuctions to the printer. Title: Should be brief and to the point. Authors name(s) and institution(s) where the research was carried out should be given, together with a suggested running page headline of not more than 50 characters including the names of author(s). A current address for correspondence, if applicable, should be included as a footnote. Abstract: Tobeonthefirstpageandconsistofnotmorethan 150words. This should outline the major findings clearly and concisely without requiring reference to the text. Headings: Should be brief, and in the following format: "Abstract, Introduction, Methods, Results, Discussion, Acknowledgements and References". All headings to be typed on a separate line from subsequent text. Main headings to be typed in capital letters and centered on the page. Sub-headings to be at the left of the page, in lower case; further headings should be in italics or underlined (not bold) beneath them. The first letter of any headings should begin with a capital letter. Text: Indent each new paragraph except those after a heading. Spelling should follow that of the Oxford English Dictionary. All pages must be numbered, at the top right hand edge. Numerals, Prose numbers should be written in full up to and including ten, and as numerals thereafter, except at the end dates & and beginning of a sentence. Dates should be written as "29 August", "10 - 13 January 1989" or scientific "1969 - 1989". units: Times should follow the 24 hour clock (e.g. 1500 h). All measurements must be in metric units. S.I. units and abbreviations should be used throughout with a space before and after the S.I. unit. Full stops should not be used after such units. The approximate position of Tables and Figures should be indicated in the left margin in pencil. Figures: The originals and two photocopies must be provided with each manuscript. Figures should be well drawn on pure white cartridge paper or good quality tracing paper. Lettering must be of very good quality (e.g. Letraset or good stencil). Typewritten or freehand lettering will not be accepted (without good reason). Each figure should have the authors' name, figure number and caption details written lightly on the reverse in pencil. Cite each figure in the text in consecutive order using arabic numerals. "Figure" should be abbreviated to "Fig.", except in the figure legends and at the beginning of sentences. Two separate caption sheets should accompany the typescript. Black and white photographs are acceptable if they have good definition and contrast and are produced on glossy paper. Consideration must be given by the author(s) to their final size when reduced for publication in the Journal which is B5. Tables: Each should be typed, double-spaced on a separate page and clearly numbered. Tables should be simple and self explanatory, and have a brief title above each one. They must be numbered consecutively with arabic numerals. Nomenclature: In cases where there are well known and internationally accepted English names, these should be used throughout the text. The scientific name should be given in full, underlined or preferably in italics, when quoted for the first time. The authority should also be quoted at this stage except in non-taxonomic ornithological papers. Only the generic name should begin with a capital letter. In subsequent reference to the name, use only the abbreviated generic name, with the full specific name but not the authority e.g .Spodoptera exempta (W alker) ; subsequently S. exempta. The authority should not be given if scientific names are used to describe an "association" or species complex, e.g. Acacia drepanolobium - Themeda triandra, wooded grassland. Type in capitals the first letter of the English names of species (e.g. Crowned Eagle, Grey -capped Warbler) but not of the higher taxa (e.g. eagles, warblers). References: Care should be taken to see that all references quoted in the text are also given in full, in alphabetical order, in this section. To avoid confusion, references to books should not be abbreviated and neither should they be underlined nor in italics. Journal names should be written in full, underlined or, if possible in italics. Personal communications should not be cited in the references, neither should they be abbreviated in the text. If there are two or more references cited by an author which were published in the same year, distinguish them by adding letters to the dates, e.g. Brown (1972 a). For technical reasons author names in the reference list should NOT be typed in upper case. Authors are advised to refer to this issue of the Journal for the preferred format. Acknowledgments: As acknowledgements may imply endorsement of the data and conclusions in the authors' work, it is advisable to obtain permission before quoting persons who have made contributions to the study. Once the paper has been selected by the Editor for possible publication in the Journal, the manuscript will be refereed by two persons who are acknowledged specialists in that field before a decision is made on its acceptance. The author(s) will be sent one set of print-outs for the correction of typographical errors. Authors are urged to ensure that all changes to the text are incorporated into this copy, as no further changes can then be made. The author is entitled to 25 copies free of charge. Thirty free copies will be provided if there is more than one contributor. Further copies can be supplied at cost, provided the order is placed before printing has been undertaken. The East Africa Natural History Society is supported by membership subscriptions, and may be unable in some instances, due to financial constraints, to accept all submissions of value. While every effort will be made to minimise such restrictions, contributors are requested to seek financial support to assist the Society with the ever increasing costs of publication. Editor Journal East Africa Natural History Society and National Museum. Referees: Proofs: Offprints: JOURNAL OF THE EAST AFRICA NATURAL HISTORY SOCIETY AND NATIONAL MUSEUM August 1991 Volume 81, No. 198 AN ETHNO-BOTANICAL STUDY OF GABRA PLANT USE IN MARSABIT DISTRICT, KENYA DANIEL STILES ‘AND ANEESA KASSAffrf#1 1 HS0%)/1 P.O. Box 23456 Nairobi, Kenya1 I Dtl* I 8 1991 Department of Literature University of Nairobi, Kenya ^Vi/B RARitb ABSTRACT This paper reports on the results of several research trips made to the Chalbi Desert area of Marsabit District, northern Kenya, between 1979 and 1983 to study various ecological and social aspects of Gabra life. We report here specifically on the preliminary results of an analysis of Gabra plant use. The research was conducted in collaboration with the UNESCO Integrated Project on Arid Lands and this study provides supplemental data for their Technical Report series of publications. INTRODUCTION The Eastern Cushitic Gabra are an Oromo-speaking people closely related to the Booran. One point of ideological differentiation between the two groups focuses on livestock: the Gabra depend mainly on camels and the Booran are cattle people. The approximately 30,000 Gabra occupy a large area about the size of Switzerland (40,000 km2) (Fig. 1) between Lake Turkana in the west and the Buie Dera plain in the east, the Mega escarpment in Ethiopia to the north and an ill-defined southern limit running from the Marsabit highlands northwest across the Chalbi Desert towards the Chari Ashe hills (Fig. 2). Territorial boundaries with neighbouring pastoral groups fluctuate (Stiles, 1981). The ecology of nomadic pastoralism is extremely complex. In a simplified way one could say that traditional pastoralists are in a never-ending search for pasture and water for their livestock, and that settlement distribution and movement are related to where these necessities can be found. The type, abundance and quality of plant species vary according to soils and rainfall, and also by the season. Different livestock animals have different forage needs in general, and these also vary according to the seasons (mainly defined as either wet or dry). People too have needs for plants, as medicine, for construction, food, for household utensils and tools, for ritual ceremonies and for firewood. The location of certain plant species is thus also of interest to man himself. Climate The climate of the Chalbi Desert region is the driest in East Africa. Rainfall is extremely variable from year to year, averaging roughly 150-200 mm annually at 500 m altitude and rising to around 1000 mm at 2000 m. Potential evaporation in the lowland areas is in the order of 2500 mm a year (Bake, 1983), 15 Stiles & Kassam • Gabra plant use Figure 1. Map of Kenya showing study area (in square) thus there is a very large water deficit, making rainfed cultivation impossible. Even in areas of over 1200 m altitude, agriculture is an uncertain occupation, as demonstrated by the Konso on the Hurri Hills where only one harvest out of three meets the needs of the people (personal observation). Northern Kenya experiences the eastern African system of northeast and southeast monsoons. The southeast monsoon, which originates over the Indian Ocean, brings the most moisture and the probability for rain is highest between the end of March and early June, with a peak in April. The northeast monsoon, originating over the Arabian region, brings less moisture with the highest probability of rain in November. July through September and December to March are normally very dry months (Ojany & Ogendo, 1982; Edwards et al. , 1979). In the Chalbi Desert the wind direction is almost always from the east or southeast and it blows very strongly, increasing potential evaporation and aeolian erosion. Temperatures are high, reaching 45°C in the shade during the driest months at 450 m altitude, though the mean daily maximum temperature at Kalacha (500 m) is about 38°C (Herbert Anderson, personal communication). J. EANHS & Net. Mus. 81 (198) August 1991 Stiles & Kassam • Gabra plant use 16 Figure 2. Map of area of northern Kenya occupied by the Gabra J. EANHS & Nat. Mus. 81 (198) August 1991 17 Stiles & Kassam • Gabra plant use Vegetation Some of the more detailed descriptions of vegetation in the Gabra area include Edwards (1940, 1945), Pratt et al. (1966), Pratt & Gwynne (1977) and FAO (1971). Herlocker (1979) presents the most detailed description and mapping of the plant communities and is the only study based on field work, except for a brief survey conducted by FAO. Following Herlocker (1979), the main primary and tertiary vegetation types sampled in our study were: Forest Shrubland Dwarf shrubland Lowland groundwater ( Hyphaene ) (2) 1 Evergreen (Suae da) (20) Evergreen with occasional perennial grasses (Salvadora with Sporobolus spicatus ) (22) Deciduous ( Acacia mellifera / A. seyal / Commiphora ) (26) Duosperma (34) Perennial grassland Upland ( Chrysopogon ) (46) and ( Panicum / Chrysopogon) (48) Wooded upland (Aristida / Chrysopogon / Pennisetum with Erythrina ) (51) Annual grassland - Bushed dwarf shrub: short (Aristida / Enneapogon with Acacia reficiens - Indigofera) (64) - Wooded dwarf shrub: short (Aristida with Acacia tortilis - Lagenantha ) (70) Barren land (73) The largest area of the study sampling zone was covered by type 64, then type 70, followed by type 34. The Chalbi Desert itself is type 73, with various halophytic plants occurring in some places (Fig. 3). Most of the study area falls within eco-climatic zone VI (very arid) of Pratt & Gwynne (1977), with a small area in the Hurri Hills in eco-climatic zone V (arid). Topography and soils The Chalbi Desert (Fig. 2) a former lake, is a depression of some 950 km2 in an area forming the sump of an interior drainage system covering 36,615 km2. The relatively flat surface of the Chalbi averages 450-480 m a.s.l. The ground rises to the northeast from lacustrine sediments onto volcanic plain which slopes up to the Hurri Hills (1685 m). This volcanic plain extends to the east and southeast and crosses to the north of the Marsabit highlands (Dida Galgallo plain). The Chari Ashe hills rise to 1 165 m on the western side of the Chalbi Desert, and to the southwest lies Mt. Kulal (2335 m), a former volcano. These highlands lie between the central and southern Chalbi Desert and Lake Turkana (410 m). The northern part of the Chalbi desert is separated from the lake by an intervening dissected lava plateau of approximately 700 m in altitude. The vegetation sampled in this study lies on one of four soil types (Sombroek et al., 1982): 1 . M7 - well drained rocky and stony Cambisols, with inclusions of rocky vents (lower part of Hurri Hills). 1 The numbers in parentheses are those used in the Herlocker (1979) tertiary vegetation types system. J. EANHS & Nat. Mus. 81 (198) August 1991 Stiles & Kassam • Gabra plant use 18 Figure 3. Distribution of tertiary vegetation types in the study area as defined by Herlocker (1979) J. EANHS & Nat. Mus. 81 (198) August 1991 19 Stiles & Kassam • Gabra plant use 2. P12 - light coloured Lithosols and calcic Xerosols developed on limestone lacustrine sediments (eastern edge of Chalbi Desert). 3. F8 - imperfectly drained calcic Xerosols developed on colluvium from various volcanic rocks, in many places with a boulder mantle (in broad fingering zones leading up to the Hurri Hills away from the Chalbi Desert and following drainage lines to the north of Marsabit away from the highlands). 4. R14 - well drained Lithosols and calcic Xerosols with a rocky and bouldery surface, developed on olivine basalts and pyroclastic rocks (completely surrounding the Hurri Hills, the Chari Ashe, Mt. Kulal and Marsabit, descending in places to the Chalbi Desert). The perennial grasslands are found on M7 souls, the annual grasslands (except for type 70) and the shrub and dwarf shrublands on F8 and R14 soils, and the Aristida / Acacia tortilis annual grassland (type 70) is found on the PI 2 soils. The Chalbi Desert itself is made up of poorly to very poorly drained, brown, saline Solonchak clays (P14 and 5 of Sombroek et al., 1982). Gabra environmental classification The Gabra, like other Oromo, have a complex system of classifying environmental and geographic features, which not only serve to describe the physical landscape, but also to express cultural concepts. For the purpose of land-use analysis, however the system can be greatly simplified and related to the vegetation and soil maps of the area (Herlocker, 1979; Sombroek et al., 1982) to derive four land-use zones:- I. Chalbi* - mostly barrenland (73) with Hyphaene (2) oases along the eastern margin at the base of the volcanic plain (R14) and around North Horr; saline Solonchak soils (pl4 and 5); 450-480 m average altitude. II. Basa - evergreen shrubland (20 and 22) and wooded dwarf shrub; short annual grassland (70); calcareous lacustrine Lithosols and calcic Xerosols (P12); 500-600 m. III. Buie - shrubland (26) on the higher part and lower down there is mainly bushed dwarf shrub annual grassland (64) and dwarf shrubland (34); either well or imperfectly drained Lithosols or calcic Xerosols derived from volcanic rocks, a lava cobble mantle common, sometimes a finer stony surface, punctuated by lava outcrops (F8 and R14); 600-1250 m. IV. Badda - perennial grasslands (46, 48 and 51) and woodlands and shrublands (woodlands were not sampled in this study); rocky and stony Cambisols (M7) and reddish brown, eutric Nitisols (M3 of Sombroek et al., 1982); 1250-1685 m. Each of these zones can be defined in relation to a primary opposition recognized by all Oromo, between lowlands ( gammoojji ) and highlands (badda), so that zones I, II, and the lower parts of III occur in the former category and the higher levels of III and IV occur in the latter category. * All Gabra terms used in this paper have been transcribed according to the system of notation used by Gragg (1982) except that for typographical convenience where he writes s, n, we use sh, ny to denote the same sound. When there is an existing convention of noting place names like chalbi, charri, etc., this has been retained. Q is transcribed k. The vernacular terms have been checked by Oromo speakers, but as there is no fixed system of writing, thus notadon varies and it should not be seen as definiuve. J. E^NHS & Nat. Mus. 81 (198) August 1991 Stiles & Kassam • Gabra plant use 20 Other geographic features distinguished by the Gabra are: dida, “plain”, c’albi ( chalbi ), being a type of plain generally barren and characterized by its salty soil ( muludde ); lafica , an open, flat savanna area; kurkura an area of dense pebbles opposed to the bale which is formed of volcanic cobbles and boulders; c'arri (c karri), a range of low-lying hills which occurs in the lowlands, tulluu ( tulu ), hills which occur in the highlands and k’ubi, hillocks which can occur anywhere. A mountain is gaara, and a crater gofa. Foothills are known as sarba, literally “calf’ (of leg). Golbo is a type of trough; large expanses of water, like lakes and rivers are termed galaana and river valleys and dry river beds are called laga. The Gabra are intimately aware of what each of these zones has to offer in the way of vegetation, water and soils and use this knowledge when planning their movement patterns. They measure the quality of pasture on the fluctuating scale of finna (fertility). This is a complex cultural notion which enters frequently into Gabra decisions on livestock herding strategies and some areas are known to be koshee (having finna) for certain stock types. METHODS Gabra pastoralists were requested to take plant samples in the area surrounding their settlement and return with them. Samples were collected in the vicinity of Kalacha, around four olla (nomadic camps) in the Arerite and Roba Gade areas, from various spots on the lava plateau ( bule ) and the lower parts of the Hurri Hills and from around Bubisa (Fig. 2). Each Gabra informant was asked a series of questions about the human uses, if any, of each plant, which livestock animals fed on the plant and how important it was in the diet according to season, and the areas where the plant could be found. The Gabra name for each plant was also recorded. The questioning was undertaken in English through a Gabra secondary-school student acting as interpreter. Plant-use information was also obtained when conducting studies of material culture and ritual and during periods of participant observation of Gabra daily life. The plant samples were put into a plant press and subsequently deposited at the East African Herbarium for identification. In analysing the results of four collection lists two major problems arose: (1) some plant samples given the same Gabra name by different informants received more than one species identification by the Herbarium, and (2) some plant samples given different names by the informants were identified as the same species by the Herbarium. There are several possible explanations for these apparent anomalies: (1) a Gabra informant misidentified the plant, (2) the Herbarium misidentified the sample, (3) the authors made a mistake in record-keeping before depositing the plants at the Herbarium, (4) some plants receive different names by the Gabra during different stages of growth or in different localities, and (5) some plants of similar appearance receive the same Gabra name. These possibilities are not mutually exclusive. The identifications posing problems, along with all the others, were checked against the vernacular names given in Dale and Greenway (1961) and Synott (1979), and in some cases Legesse’s Herbarium list (1981). Where either or all of these authors agreed with one of our matching pairs of Gabra - Herbarium identifications, we chose that one to include in our list (Table 1). Possible alternative identifications are noted in parentheses in the column containing the Gabra names. The Gabra do recognize higher orders of classification which group related plants into families, where most of the problematic identifications were found* The principal problem families were ( h ) ammeesaa and buuyyo; ( h ) ammeesaa corresponds to Burseraceae and other Herbarium identifications and those in Dale and Greenway (1961) indicate that * The reasons for these “confusions” from the botanical point of view may be explained by Gabra / Booran plant taxonomy. Field-work would suggest that when a plant “degenerates” (i.e. when it reproduces itself in a slightly modified form), it is given a different name. J. EANHS & Nat. Mus. 81 (198) August 1991 21 Stiles & Kassam • Gabra plant use there is some confusion in correctly identifying the various Commiphora species. The Gabra also include the trees Terminalia polycarpa / spinosa (Combretaceae) in (h) ammeesaa. Much more confusion was found in the d’umashoo family, which corresponds to the Capparaceae. Boscia, Cadaba and Maerua plant samples were identified interchangeably as deekuku, k’ad’u, k’alk’acca and ( [h ) afuursaa and could belong to almost any of the three genera. Buuyyo is the term for grass and thus corresponds to the Poaceae (Gramineae). Considering that samples were usually taken during the dry season when grass consists of little more than dry tufts, it was not always possible to get full identifications. The fact that multiple identifications (c.15%) occurred raises the question of the accuracy in single- collection studies. For this reason we feel that any ethno-botanical study needs to obtain two or more corroborating sample identifications to be considered valid. As not all of our samples meet that criterion, the list presented here should be considered provisional and subject to verification and possible revision. Plant identification and use There is always a question as to what is the best method to present a plant list. We have chosen alphabetical order by family because many of the users of this list will not be botanists, and for non- specialists it is the simplest method of reference. Hepper et al. (1981) chose to list the plants of Mt. Kulal, near our study area, by order of evolution and degree of relatedness of the families. The families were ordered and numbered. For those wishing information of this sort, the Hepper et al. numbers are reproduced here in the first column in Table 1. In general, the lower the number the more primitive the family, and the closer together the numbers the more closely related are the families. An alphabetical list of vernacular names is also appended. Table 1: A list of Gabra plants and their uses: (abbreviations used). Plant T — Tree Uses Co — Construction Sh — Shrub F — Food Dsh — Dwarf shrub Fw — Firewood L — Liane M — Medicine Tw — Twiner 0 — Cultural objects Tc — Tendril climber R — Ritual H — Herb V — Veterinary use G — Food: Grass Score 0 — Not eaten or used C — Camel food * No good data collected Ca — Cattle food + — Minor importance G — Goat food ++ — Moderate importance S — Sheep food +++ — Very important J. EANHS & Nat. Mus. 81 (198) August 1991 Stiles & Kassam • Gabra plant use 22 Animal Human No. Family, genus Gabra Plant and species name type C Ca G s V M F Co o R Fw 259 ACANTHACEAE Barleria acartihoides Vahl shiisha DSh -H-+ 0 + + 0 * 0 0 0 + 0 Barleria sp. Blepharis ciliaris (L.) maadeeka + 0 0 0 0 0 0 0 0 0 0 B.L. Burtt baarataa H ++ + ++ + + * 0 0 0 0 + B. linariifolia Pers. Duosperma ermophilum (Milne-Redh.) Napper saariima DSh ++ ++ ++ + 0 0 0 0 0 0 0 Ecbolium revolution (L.) k'atte DSh + 0 + + 0 0 0 0 0 0 0 C.B. Cl. {Lantana sp.) Ruellia patula Jacq. d'ad'ale + + + + 0 0 0 0 0 0 0 Indet. lakud'e Sh + + + + 0 0 0 0 0 0 0 313 AGAVACEAE Dracaena ellenbeckiana Engl. butte TSh 0 0 0 0 0 0 0 0 + 0 0 Sansevieria robusta (N.E. Br.) algge Sh 0 0 0 0 0 0 0 ++ + + 0 Jake 54 AIZOACEAE Trianthema salsoides Fenzl. k'ant'alaa H + 0 + 0 0 0 0 0 0 0 0 Zaleya pentandra (L.) Jeffrey laamisho H + -t- + + 0 0 0 0 0 0 0 63 AMARANTHACEAE Aerva javanica Schultes muk-illeensa H 0 0 0 0 0 + 0 0 0 0 0 (A. persica (Burm. f.)) sufki H 0 0 0 0 0 0 0 0 + 0 0 Merrill Digera muricata (L.) Mart. gelgedaana H ++ + ++ + 0 0 0 0 0 0 0 Pupalia lappacea (L.) Juss. matt’anne H + 0 + 0 0 0 0 0 + 0 0 var. velutina (Moq.) (Serico- Hook. f. comop sis sp.) Sericocomopsis hilde- jiibeete H ++ 0 ++ + 0 + 0 0 0 0 + brandtii Schinz ( Dasysphaera prostrata (Gilg & Schinz) 205 ANACARDIACEAE Rhus natalensia Krauss dabobbesa Sh + * + + ★ * + ★ * * + 230 APOCYNACEAE Acokanthera schimperi k’arraaru T 0 0 0 0 0 + 0 0 0 0 0 (A.E)C.) Schweinf. Adenium obesum (Forsk.) obbe Sh 0 0 0 0 0 + 0 0 0 0 0 Roem. & Schult. Carissa edulis (Forsk.) Vahl dagamsa Sh/L 0 0 0 0 0 + 0 0 0 0 0 ARISTOLOCHIACEAE Aristolochia bracteolata Lam. raafu ? 0 0 0 0 0 0 0 0 0 0 0 314 ARECACEAE (PALMAE) Hyphaene compressa meetti T 0 0 0 0 0 0 + ++ + 0 •f H. Wendle. J. EANHS & Nat. Mus. 81 (198) August 1991 23 Stiles & Kassam • Gabra plant use Animal No. Family, genus Gabra Plant and species name type C Ca G s V M F Co o R Fw 231 ASCLEPIADACEAE Calotropis procera (Ait.) k’obboo Sh 0 0 0 0 0 + 0 0 0 0 0 Ait. f. Caralluma speciosa (N.E. Br.) boorara H 0 0 0 0 0 + 0 0 0 0 0 N.E. Br. 195 BALANITACEAE Balanites aegyptiaca (L.) baddana T + 0 0 0 0 0 + 0 + + 0 Del. (B. orbicularis) 249 BORAGINACEAE Cordia sinensis Lam. Aschers mad’eera T/Sh + 0 + + 0 0 + ++ ++ +++ ++ (C. gharaf (Forsk.)) Heliotropium albohispidum Bak.. kokoomisha H ++ + + + + + 0 0 0 0 0 H. somalense Vatke H. subulatum (DC.) dubarraara H + 0 + + 0 0 0 0 0 0 0 Martelli 196 BURSERACEAE Boswellia hildebrandtii Engl. dakkara T + 0 + + 0 0 0 0 0 + + Commiphora cf. africana ammeesaa T/Sh + 0 + + + 0 0 + ++ 0 + (A. Rich). Engl. C. boiviniana Engl. dakd'aa T/Sh + ★ + * 0 0 + 0 + 0 + C. myrrha (Nees) k'umbi Engl. (C. ellenbeckii Engl.) T 0 0 0 0 + ++ 0 0 0 ++ + C. habessinica c'allankaa T + * + * * ★ + * + * * (O. Berg) Engl. C. erythrea (Ehrenb.) agarsu T + 0 + + 0 0 0 0 + 0 + Engl. C. incisa Chiov. waaraa T/Sh + 0 + + * + 0 0 + + + Commiphora sp. warab reeba T/Sh + 0 + + 0 0 0 + 0 0 + 146 CAES ALPINI ACE AE Delonix elata (L.) Gamble stikellaa T + 0 0 0 0 0 0 0 ++ 0 + 36 CAPPARACEAE Cadaba farinosa Forsk. deekuku Sh ++ 0 ++ ++ 0 0 0 0 0 0 0 (Maerua oblong ifolia (Forsk.) A. Rich.) C. mirabilis Gilg) C. gillettii) k'adu Sh ++ + ++ + 0 0 0 0 0 0 0 R.A. Graham (Boscial) Maerua angolensis DC.) k'alk'acca T + * * * * * * + * * * M. crassifolia Forsk. ( Boscia coriacea Pax) M. crassifolia Forsk.) d'umashoo Sh + 0 + + 0 0 0 0 0 0 ++ Human M. kaessneri Gilg. Bened.) J. EANHS & Nat. Mus. 81 (198) August 1991 Stiles & Kassam • Gabra plant use 24 Animal Human No. Family, genus Gabra Plant and species name type C Ca G s V M F Co o R Fw Maerua sp. (h) afuursaa T/Sh ++ + ++ + 0 0 0 0 0 0 + C Cadaba mirabilis ) Indet. k'ork'odda Sh + + + + 0 0 0 0 0 0 + 6i CHENOPODIACEAE Fadenia zygophylloides had'um H ++ 0 + + 0 0 0 0 0 0 0 Allen & Townsend (Gyroptera gilletti Bosch.) Suaeda monoica J.F. Gmel. d'uurtee T/Sh +++ 0 + + 0 0 0 0 0 0 + (Salsola den- droides Pallas) 121 COMBRETACEAE Combretum cf. denhardtiorum c'anc'ali Sh + 0 + + 0 0 0 0 + 0 + Engl. & Diels C. molle G. Don rukeesa T + 0 + + 0 0 0 0 0 0 + Terminalia spinosa Engl. T. polycarpa k’orobo T + 0 + + 0 0 0 0 0 0 + Engl. & Diels 280 COMMELINACEAE Commelina latifolia A. Rich. k'ayyu H + ++ + ++ 0 0 0 0 0 0 0 238 COMPOSITAE Aspilia mossambicensis (h) ada (Oliv.) Wild. (Vernonia wake- fieldii Oliv.) H + 0 + 0 0 + 0 0 0 0 0 251 CONVOLVULACEAE Seddera hirsuta Hall. f. gurbi + 0 + 0 0 0 0 0 0 0 0 0 103 CUCURBITACEAE Cucumis dipsaceus Spach buratte Tc 0 0 + + 0 0 0 0 0 0 0 C. prophetarum L. baram-barro Tc + + ++ + 0 0 0 0 0 0 0 Kedrostis gijef (J.F. Gmel) gaalle Tc ++ 0 ++ + 0 0 + 0 0 ++ 0 C. Jeffrey (i Cucumis sp.) CUPRESSACEAE Juniper us procera Endl. arru T 0 0 0 0 0 0 0 0 0 ++ 0 221 EBENACEAE Diospyros abyssinica (Hiem) lookko T * * * * * + + ★ + * * F. White 136 EUPHORBIACEAE Croton somalensis d'irri Sh + + + + + ★ * * * * * Vatke & Pax Euphorbia candelabrum addamma T 0 0 0 0 + 0 0 0 0 0 0 Kotschy E. cuneata Vahl (h) idaa Sh + 0 + + 0 0 0 0 0 0 0 E. tescorum Carter harkeena Sh 0 0 0 0 0 + 0 0 0 0 0 J. EANHS & Nat. Mus. 81 (198) August 1991 25 Stiles & Kassam • Gabra plant use No. Family, genus and species Gabra name Plant type Animal C Ca G S V Human M F Co O R Fw 264 LAMIACEAE (LABIATAE) Leucas pododiskos Bullock jilbeete kurooftu 293 LILIACEAE Asparagus africanus Lam. ergamssa Asparagus sp. okolle 132 MALVACEAE Pavonia zeylanica (L.) Cav. ilk’abata 23 MENIS PERM ACE AE Cocculus pendulus marmma H 0 0 0 0 0 Tc 0 0 0 0 0 Tc 0 0 0 0 0 H +0 + 00 Tw +0 + 00 + 0 0 0 0 0 0 0 0 ++ 0 0 0 0 0 + 0 0 0 0 0 0 0 0 0 0 0 0 0 0 (J.R. & G. Forst.) Diels 147 MIMOSACEAE A. etbaica Schweinf. (h) allak’abeessa T + 0 + + 0 * * ★ + * + A. goetzei Harms burraa T + 0 + + 0 + 0 0 0 0 + A. mellifera (V ahl) sap’ans T/Sh + 0 + + 0 0 0 + 0 0 + Benth. gurraaca A. nilotica var. subalata burk’uk’e T/Sh + 0 + + 0 0 0 + 0 0 + (L.) Del. A. nubica Benth. waanga Sh + 0 + + 0 + 0 0 0 0 + Acacia paolii Chiov. c’aac’anne S + 0 + + 0 0 0 0 ++ 0 + (A. horrida) Acacia reficiens Wawra sigirso T/Sh ++ 0 ++ + 0 0 0 + + + ++ ssp. misera (V atke) Brenan A. Senegal (L.) Willd. iddaad’o T/Sh + 0 ++ + 0 0 0 + 0 0 + A. seyal Del. var. fistula waac’c’u T 0 0 + 0 0 0 + + 0 0 + (Schweinf.) Oliv. A. tortilis (Forsk.) d’addaca T ++ 0 ++ + 0 0 + +++ ++ ++ +++ Hayne 167 MORACEAE Ficus glumosa Del. k’iltaa T/Sh 0 0 0 0 0 0 ★ 0 0 0 0 83 NYCTAGINACEAE Commicarpus helenae araddo. (J.A. Schultes) Meikle k’oraatti H + 0 + 0 + 0 0 0 0 0 0 gaala 182 OLEACEAE Ole a europaea L. var. africana ejersa T/Sh 0 0 0 0 0 0 0 0 + ++ + (Mill.) P.S. Green 253 OROBANCHACEAE Cistanche tubulosa (Schenk.) (h)armaac’a H 0 0 0 0 0 + 0 0 0 0 0 Hook.f. 148 PAPILIONACEAE Abrus schimperi Bak. wargidda H 0 0 0 0 0 0 0 0 + 0 0 ssp. africanus (Vatke) Verde. Crotalaria cf. dumosa (h)asura H + + + + 0 0 0 0 0 0 0 Franch. J. EANHS & Nat. Mus. 81 (198) August 1991 Stiles & Kassam • Gabra plant use 26 No. Family, genus and species Animal Gabra Plant name type C Ca G S V Human M F Co O R Fw Erythrina rotundata-obovata E. melanacantha) Harms) E. burttii Bak. f.) Indigofer a coerulta Roxb. var. occidentalis Gillett & Ali I. coerulea Roxb.) I. colutea (Burm. f.) Merrill I. clijfordiana Gillett I. insularis Chiov. I. spicata Forsk. waleena T 0 0 0 0 0 (h)asura harre H 0 0 0 0 0 (Cassia italica) 0 0 0 + 0 + 0 0 0 + 0 0 agaggaro H +++++++ 0 0 0 0 0 0 0 I. spinosa Forsk. k’ilt’ip’p’e Ds ++ + ++ ++ 0 0 + 0 0 0 0 Ormocarpum trichocarpum buutiyye Sh 0 0 + + 0 0 0 0 0 0 + (Taub). Engl. Rhynchosia minima (L.) DC. uube H/Tw + 0 ++ ++ 0 0 0 0 0 0 0 Vatovaea psuedolablab gaabbe Tw + 0 + + 0 0 + 0 + 0 0 (Harms) Gillett Vigna frutescens A. Rich. c’iimp’a H + + + + * 0 + 0 0 0 0 PASSIFLORACEAE Adenia venenata Forsk. POACEAE (GRAMINEAE) obbe Tc 0 0 + 0 0 0 0 0 0 0 0 Aristida adscensionis L. buuyyo biila G ++ +++ ++ ++ 0 0 0 ++ 0 0 + Aristida mutabilis Trin. & Rupr. Cenchrus ciliaris L. diilaleesa G + + + + * * He He He H= He C. pennisetiformis k'onc'orro G + + ++ ++ 0 0 0 0 0 0 0 Steud. C. setigerus Vahl buuyyo harre G + + + + 0 0 0 0 0 0 0 Chrysopogon plumulosus alala G + ++ + ++ 0 * He He Hi He He Hochst. Dactyloctenium bogdanii maa G + ++ + ++ 0 0 0 0 0 0 + S.M. Phillips Digitaria velutina (Forsk.) biila G + ++ + ++ 0 0 0 0 0 0 0 P. Beauv. Echinochloa haploclada geedi G + + + + * * He He He He He (Stapf) Stapf Leptothrium senegalense (Kunth) Clayton ilmmogora G + + + + 0 0 0 0 0 0 0 Panicum color atum L. Paspalidium desertorum laabbesa c'iraa G + ++ + + * * He He * He ★ (A. Rich.) Stapf (Cenchrus sp. Sporobolus helvolus). G + ++ + ++ 0 0 0 + 0 + 0 Pennisetum mezianum Leeke ogoona G + + + + * * ★ He * He * Sehima nervosum (Rottl.) sokhorsitu G * * * * * * * He * He * Stapf J. EANHS & Nat. Mus. 81 (198) August 1991 27 Stiles & Kassam • Gabra plant use Animal Human No. Family, genus Gabra Plant and species name type C Ca G s V M F Co o R Fw Setaria verticillata (L.) hank'arre G + + + + * * * + * * 4c P. Beauv. Sporobolus ioclados (Trin.) buuyyo fiinc'oo G + -t-+ + ++ 0 0 0 0 0 0 0 Nees S. spicatus (Vahl) harfuuk'a G + + + + 0 0 0 0 0 0 + Kunth Themeda triandra Forsk. buuyyo diimtu G + + + + * * ★ ♦ ♦ 4c 4C Indet. saatu G + ++ + ++ 0 0 + 0 0 0 0 190 RHAMNACEAE Ziziphus abyssinica A. Rich. k’urk’uura T/Sh + 0 + ♦ 0 0 + + + 0 + RUTACEAE Zanthoxylum chalybeum gaddaa T 4* ★ * * ★ + + 0 0 0 0 (Engl.) Kokw. 180 SALVADORACEAE Salvadora persica L. aadde T/Sh +++ 0 ++ + 0 + 0 + 0 0 + 252 SCROPHULARIACEAE Pseudosopubia hildebrandtii k’ors nyaata H 0 0 + 0 + 0 0 0 0 + 0 (V atke) Engl. 250 SOLANACEAE Lycium europaeum L. fursaa Sh + 0 + + 0 0 0 + + 0 + Solarium coagulans Forsk. (h)iddi, (h)iddi H 0 0 0 0 ★ 0 0 + 0 0 + (S. dubium Fres. ree, hiddi arado (5. coagulans Forsk.) (small version) H + 0 + 0 + 0 0 0 0 0 0 130 STERCULIACEAE Sterculia africana (Lour.) k’ararri T 0 0 0 0 0 0 0 0 ++ 0 + Fiori 128 7TLIACEAE Corchorus triocularis L. luuftoole (Farsetia steno- ptera Hochst.) H + + + + 0 0 0 0 0 0 0 Grewia tenax (Forsk.) Fiori d’eeka Sh + 0 + + 0 0 + 0 + + 0 G. trichocarpa A. Rich. G. bicolor Juss. (h)arorressa Sh + 0 + + 0 0 + ++ + 0 + G. villosa Willd. ogomdi S + 0 + + 0 0 + 0 + 0 + Triumfetta flavescens ic’iinni s + 0 + + 0 0 0 0 0 0 0 Hochst. VIOLACEAE Rinorea convallariiflora fit’o T 4c * * ★ * 4c * ★ ★ + 4c M. Brandt 193 VITACEAE Cyphostemma nierense rorroddo Tc 0 0 0 0 0 + 0 0 0 0 0 (Th. Fr. jr.) Desc. 66 ZYGOPHYLLACEAE Tribulus cistoides L. mogorree H + + + + 0 + 0 0 0 0 0 J. EANHS & Nat. Mus. 81 (198) August 1991 Stiles & Kassam • Gabra plant use 28 RESULTS Livestock The degree of importance of each plant as livestock forage is a subjective measure in this study based on the combined opinions of approximately a dozen informants. The importance of a plant in the diet of any particular animal is a function of its availability, i.e. its abundance, and of the dietary needs of the animals at any point in time. Due to the ever-changing nature of these variables, and the methodological problems involved in recording timed feeding observations, we feel that this method probably yields results as valid as timed feeding trials, with less chance of a bias due to local species availability in the area of the feeding trials. The data provided here are from samples collected within an area of approximately 2400 km2 covering a variety of ecological zones. The degree of importance of a plant to livestock diet as reported in Table 1 needs some qualification. It is not based entirely on what would be a measure of the weight dry matter ingested by the animal, as quantity is not always of the utmost importance to a pastoralist, particularly concerning the camel, for example, d’uurtee ( Suaeda monoica / Salsola dendroides ) is considered as a very important plant for camel health due to its high salt content, particularly during the dry season. Kilt’ip’p’e, ( Indigofera spinosa ), however, is rated as only moderately important. Measured in dry- matter weight, camels probably eat more k’ilt’ip’p’e, since it is more widespread and abundant than d’uurtee. This is certainly true during the rainy seasons. D’uurtee is more important, however, because of its food value, not quantity, and as such a pastoralist would be much more likely to plan his herding strategy at certain times to take into account the location of d’uurtee than he would of k’ilt'ip’p’e. This kind of analysis is another reason why we think that a qualitative evaluation of plants as forage has some usefulness. We also tried to make a distinction between “importance” and “liking” in our questioning of informants. For example, camels have a very strong liking for mad’ eera ( Cordia sinensis ), but because it is not very abundant in the study area it does not form a very important part of the camel diet (Table 1). Goats display the widest range in diet, eating 91 of the 140-150 plants listed (the number depends on how many alternative identifications might be valid). Camels come next with 88, followed by sheep with 78 and lastly by cattle with only 34. The seven grass species in Table 1, mentioned by informants but about which we had collected no first-hand information, can be added to the total number of plants eaten by each livestock type, as others have listed them (Pratt & Gwynne 1977; Field 1979 a & b; Sato 1980). Nothing can be said about their relative importance in the diet, however. This would bring the total number of species eaten to 98 for goats, 95 for camels, 85 for sheep and 4 1 for cattle. Table 2 presents a summary of the plants considered by our informants as being either “very important” or “moderately important” in the diet of each stock species. Table 2: The Very and Moderately important plants in the livestock diet in the eastern Chalbi Desert area. Camels Very important (+++): Moderately important (++); Barleria acanthoides Duosperma eremophilum Suaeda monoica (Salsola dendroides? ) Salvadora persica (Indigofera all species) Blepharis ciliairis / linariifolia Digera muricata Sericocomopsis hildebrantii Heliotropium albohispidum Cadaba farinosa C. gilletti Maerua sp. Fadenia zygophylloides Kedrostis gijef J. EANHS & Nat. Mus. 81 (198) August 1991 29 Stiles & Kassam • Gabra plant use Aristida adsecensionis / mutabilis Indigofera (many species) Acacia reficiens A. tortilis Cattle Very important (+++): Aristida adscensionis / mutabilis Moderately important (++): Commelina latifolia Cenchrus pennisetiformis Chrysopogon plumulosus Dactyloctenium bogdanii Digitaria velutina Paspalidium desertorum Sporobolus ioclados Goats Very important (+++): Indigofer a (all species) Moderately important (++): Blepharis ciliaris / linariifolia Duosperma eremophilum Digera muricata Sericocomopsis hildebrandtii Cadaba farinosa C. gillettii Maerua sp. Cucumis prophetarum Kedrostis gijef Cenchrus pennisetiformis Indigofera spinosa Indigofera (many species) Rhynchosia minima Salvadora persica Acacia reficiens A. paolii A. Senegal A. tortilis Sheep Very important (+++): Indigofera (all species) Moderately important (++): Cadaba farinosa Commelina sp. Aristida adscensionis / mutabilis Cenchrus pennisetiformis Chrysopogon plumulosus Dactyloctenium bogdanii Digitaria velutina Paspalidium desertorum Sporobolus sp. Indigofera (many species) Rhynchosia minima J. EANHS & Nat. Mus. 81 (198) August 1991 Stiles & Kassam • Gabra plant use 30 Camels have the highest number of “very important” species with four: shiisha ( Barleria acanthoides), called “food of the camel” by the Gabra, saariima ( Duosperma eremophilum ) d'uurtee (Suaeda monoica / Salsola dendroides ), and aadde ( Salvadora persica ). The genus Indigofera can be added to this class by combining agaggaro (several Indigofera species) and k’ilt’ip’p'e (I. spinosa ). Twenty species (14 Gabra taxa) were rated as “moderately important” in the camel diet. Two species of plants, Aristida adscensionis / mutabilis, both known as buuyyo biila in Gabra, are the only ones rated “very important” for cattle by the informants. This was undoubtedly due to their widespread abundance in the study area. Seven species were identified as being of “moderate importance”, only one of them ( Commelina sp.) not a grass. It is interesting to note that no plant was specifically identified as being “very important” in the diet of goats or sheep. However, as in the case with camels, the genus Indigofera could be included here by combining the two Gabra taxa of agaggaro and k’ilt’ip’p’e, each considered “moderately important.” Goats had 23 (18 Gabra) species names as “moderately important” and sheep had 17(12 Gabra) in that class. These results accord well with studies conducted by A.C. Field (1978), C.R. Field (1979a & b) and Sato (1980) on livestock feeding habits in northern Kenya, though these authors did not work in exactly the same study area. Little research was conducted in the area of traditional veterinary treatments, but eight plants were identified as being employed for this purpose, all of them of minor importance. Barataa ( Blepharis ) is burned and the ashes are spread over camel wounds: kokoomisha (Heliotr opium albohispidum ) leaves are chewed up and applied to a snake bite to reduce swelling; the resin ( aamp’e ) from (h) ammeesaa (C. africana ) is mixed with milk and applied to camels to remove ticks when they are in highland areas; k'umbi ( Commiphora myrha /ellenbeckii), the myrhh tree, yields a resin which has many uses. One use is as a ritual cure for anthrax and is practised by a few clans. The resin is chewed and then spat all around the animal enclosure; aaddaama ( Euphorbia candelabrum) is used to cure a camel disease called gaal malaa. A traditional doctor (c’iressa) must prepare and administer the medicine; araddo ( Commicarpus helenae) is chewed and spat into the nose of a calf as a decongestant; k’ors nyaata {Pseudo sopubia hildebrandtii) leaves are chewed up and the saliva put into the animal’s mouth to protect it from a curse by a budaa (person with the evil eye); and (h) iddi araddo (the small or young Solanum coagulans ) is used to treat a throat swelling disease called c’ilmale by burning it and passing the smoke under the animal’s throat. We cannot attest to the efficacy of any of these treatments. Human uses The Gabra use more species of plants for firewood than for any other purpose, with over 40 recorded Gabra taxa, though many more are certainly used. The most important, and preferred, wood is that of Acacia tortilis, followed by A. reficiens and Maerua crassifolia / kaessneri. Wood from other Acacia and Commiphora species is also commonly utilized as firewood, the frequency depending on local abundance. Salvadora persica (aadde) is not supposed to be used as firewood for reasons of aada (tradition), but as wood becomes more scarce younger women are sometimes using it. According to Gabra traditional beliefs, live wood should not be cut for use as firewood. In most cases this rule is adhered to, but in areas where dead wood is rare or absent, particularly on the lava plains ( bule ), branches from living trees and shrubs will be cut. The Gabra are very conservation minded in their use of wood in cooking fires. They use small amounts and usually pull unbumed faggots away from the centre of the hearth for re-use later on. The grasses listed under firewood are those most commonly used as tinder. In construction use, 18 Gabra taxa (21 species) have been recorded. The most important is A. tortilis ( d'addac’a ), the preferred thorn branch to make animal enclosures and the only one allowed for building the ritual naabo enclosure. Other Acacia species are also used in boma construction and A. reficiens ( sigirso ) shrub is most often used as the animal enclosure gate, replaced each year in a ceremony ( almado ). Since the Gabra build on average about 10 new animal enclosures a year for J. EANHS & Nat. Mus. 81 (198) August 1991 31 Stiles & Kassam • Gabra plant use camels and sheep / goats in the main family settlements {olla) and several more at the satellite camps (fora), the amount of woody vegetation consumed per year for this purpose is substantial. Live wood is almost always used; however, it serves a dual purpose as it is later used as firewood. The traditional Gabra house (mana) is also heavily dependent on plant matter. The house, or tent as it is sometimes called, is made by placing skins, cloth, mats and sometimes grass over a domed frame made of bent poles. The house poles (dediee, utubaa) are made from Cordia / sinensis {mad’ eera) or Grewia bicolor / trichocarpa {h) arorresa saplings. Wood for these poles is collected at certain ritually prescribed times of the year. The tops of houses are traditionally covered with thick grass-like mats made from Sanseveria {algge), the wild sisal, but because of the large amount of time needed to go to the Golbo / Gharri Ashe areas to collect it, the difficulty in processing it and its reduced abundance, other materials such as scrap aluminium and plastic sheeting are now being used. Sometimes grass thatch made from Aristida or, less commonly, Paspalidium / Cenchrus / Sporobolus {c'iira) is used to cover open patches on the roof or to plug air holes in the walls. Algge and c’iraa are also used to make rope and twine which are used to tie intersecting house poles and interior partition sticks together. Leaves of Hyphaene compressa {meetti), the doum palm, are important today for roofing of permanent wattle and daub housing, and the straight branches are sometimes employed in making the wattle frame and for bed poles. Commiphora wood is not very important in building, being mainly used in the construction of fences of ritual enclosures {naabo d’eeda and gosse). In the manufacture of objects of material culture ten Gabra taxa (32 species) were recorded. The most common type of object was a container or vessel, of which the Gabra have a wide variety. No attempt will be made here to present a full description of Gabra material culture as this is still subject to further research. The six most important plants used were: Cordia {mad’ eera), used to make a man’s walking stick {hororo), sometimes the {h) okkoo stick (used to shake Acacia pods {arbuu) from trees and to construct and repair the fence of the animal enclosure) and other ritual sticks; Commiphora {c’allankaa and (h) ammeesaa) used to make stools {kaara and barc’uma), water and fresh milk containers {soroora), and as a toothbrush stick {rigaaf, Delonix elata {sukellaa) has a soft wood used very often in carving to make camel bells {kokke), fat-storage containers {dibbe), soroora and a large wooden bead {qilinto) put around camel bulls’ necks during the mating season to protect them against the evil eye; Asparagus africanus {ergamsa) roots are used to weave milk containers {c’iic’oo), camel milking containers {gorfa), small containers used by children for carrying milk or water when herding, and the base and rim of soroora. The roots of another Asparagus species (or perhaps the same?) which the Gabra call okolle are used to weave the large (c.20 litres) water-transport and storage container {butte), which one commonly sees at well sites in Gabra country. Acacia paolii {c'aac’anne) is used to make various containers, and Sterculia africana {k’arrarri) is used to make cleaning “cloths” {sosso) by shredding and then soaking the bark in fat. The Gabra depend very little on wild plant foods. The 15 Gabra taxa (17 species) recorded were all rated by our informants as of “minimal importance”. The outside of the nut of the doum palm {k’one) is one of the most common wild plant foods, usually eaten by children as snacks, although in times of famine they may be eaten in great quantities by everyone. Wild berries produced by Cordia, Kedrostis, Ziziphus and Grewia, amongst others, are also enjoyed by people. The root of Vatovaea is eaten raw, mainly for its moisture, and the seedless pods of A. tortilis are eaten as a famine food. The other plants in the list are usually used to make infusions from the leaves or roots. Not marked in the food category were gums and resins, commonly chewed by the Gabra, belonging mainly to various Acacia and Commiphora species. Only 19 plants were recorded as being of medicinal use. Perhaps the Gabra do not depend as heavily as most East African peoples on herbal medicines (Kokwaro, 1976), but this list is incomplete. Investigations will have to be made with knowledgeable Gabra c’iressa and with waata (traditionally low caste hunter-gatherers), who are reputed to be very learned in the use of plant medicines, in order to discuss in detail Gabra medicinal plant use. Gabra medicine in any case seems to be based more on what one could call “faith healing”, often performed by people who are known as specialists for a certain part of the body because of their clan membership. Certain clans are associated with certain J. EANHS & Nat. Mus. 81 (198) August 1991 Stiles & Kassam • Gabra plant use 32 parts of the human body. Treatments often consist of ritualistic ceremonies which may or may not make use of plants; even when they do, the treatment may be no more than chewing k’umbi and spitting it on the afflicted person. Plants are very important in Gabra ritual life, which itself permeates all aspects of socio-economic behaviour. Due to the complexity of explaining plant use in this domain, and the large amount of data which we have, this subject will form the basis of a separate article. We also have the names for about 50 uncollected Gabra plants and their uses, which would add several species to each of the livestock- and human-use categories discussed above. CONCLUSIONS The Gabra have an intimate knowledge of plant types and distribution in their territory because their existence depends on it. Grass, herbs, shrubs and trees feed the livestock which supply the Gabra with milk, meal, blood, skins, a medium -of exchange, a repository of wealth and the basis of their social organisation. Plants are therefore of most importance in terms of livestock forage, but they are also essential for use as fuel, in construction and in the manufacture of material culture objects, though with the introduction of metal and plastic items this latter use is beginning to diminish. Plants are of less importance in medicine and as human food, but they are deeply involved in Gabra ritual life. Deforestation and land degradation from overpopulation and overstocking — desertification — is a very great threat to the future of the Gabra as nomadic pastoralists. Unless a lasting solution is found to the problem of desertification the Gabra will eventually not have the plants they need for survival. ACKNOWLEDGEMENTS We would like to express our appreciation to the Ford Foundation, the L.S.B. Leakey Foundation, the Kenya National Council for Science and Technology, the University of Nairobi and the Center for Field Research (Earthwatch) for financial support to our research. We are grateful to UNESCO-IPAL for logistical support and for research collaboration. However, the conclusions of this paper are the responsibility of the authors. We are grateful to the A.I.C. mission in Kalacha, and especially Mr and Mrs Simonson who are helpful in so many ways, and we would also like to thank the British Institute in Eastern Africa for the loan of equipment. We would also like to thank the many Gabra informants and interpreters who made the research possible. The staff of the East African Herbarium - National Museums of Kenya gave outstanding co-operation, particularly Ms. C. Kabuye who kindly checked and corrected the botanical names. J. EANHS & Nat. Mus. 81 (198) August 1991 33 Stiles & Kassam • Gabra plant use REFERENCES Bake, G. (1983) An analysis of climatological data from the Marsabit District of northern Kenya. IPAL Technical Report No. B-3. UNESCO, Nairobi. Dale, I.R. & Greenway, P.J. (1961) Kenya Trees and Shrubs. London and Nairobi: Hatchards. Buchanan’s Kenya Estates Ltd. Edwards, D.C. (1940) A vegetation map of Kenya with special reference to grassland types. Journal of Ecology 28: 377-385. - (1945) Horn of Africa (including Kenya ) Vegetation Map. Nairobi: Kenya Government Printer. Edwards, K.A., Field, C.R. & Hogg, I.G.G. (1979) A preliminary analysis of climatological data from the Marsabit District of Northern Kenya. IPAL Technical Report B-l, UNESCO, Nairobi. FAO (1971) Range development in Marsabit District, Kenya. AGP: SF/KEN 11. Working Paper 9. Field, A.C. (1978) ODM-IPAL sheep and goats project. Preliminary Report on the Impact of Sheep and Goats on the Vegetation in the Arid zone of Northern Kenya. IPAL Technical Report E-2, UNESCO, Nairobi. Field, C.R. (1979a) Preliminary report on ecology and management of Camels, Sheep and Goats in northern Kenya. IPAL Technical Report E-la, UNESCO, Nairobi. (1979b) The Food habits of Camels in northern Kenya. IPAL Technical Report E-lb, UNESCO, Nairobi. Gragg, G.S. (1983) Oromo Dictionary. East Lansing: African Studies Center. Hepper, F.N., Jaeger, P.M.L., Gillett, J.B. & Gilbert, M.G. (1981) Annotated check-list of the plants of Mount Kulal, Kenya. IPAL Technical Report D-3, UNESCO, Nairobi. Herlocker, D. (1979) Vegetation of southwestern Marsabit District, Kenya. IPAL Technical Report D-l. UNESCO, Nairobi. Kokwaro, J.O. (1976) Medicinal Plants of East Africa. East African Literature Bureau, Nairobi. Legesse, A. (1981) List of plants collected from Kenya (KI) and submitted by Prof. A. Legesse, Swarthmore College, Swarthmore, USA. Nairobi: East African Herbarium. Ojany, F.F.R. & Ogendo, R.B. (1982) Kenya, A Study in Physical and Human Geography, Nairobi: Kenya Literature Bureau. Pratt, D.J., Greenway, P.J. & Gwynne, M.D. (1966) A classification of East African rangeland with an appendix on terminology. Journal of Applied Ecology 3, 369-382. Pratt, D.J. & Gwynne, M.D. (1977) Rangeland Management and Ecology in East Africa. London: Hodder and Stoughton. Sato, S. (1980) Pastoral movements and the subsistence unit of the Rendille of northern Kenya: with special reference to camel ecology. Senri Ethnological Studies 6, 1-78. Sombroek, W.G., Braun, H.M.H. & van der Pouw, B.J.A. (1982) Exploratory soil map and Agro- Climatic zone map of Kenya, 1980. Expl. Soil Survey Report no. EL, Kenya Soil Survey, Nairobi. Stiles, D.N. (1981) The Gabra of Northern Kenya: past and future. Kenya Past and Present 13, 23-31. Synott, T.J. (1979) A report on the status, importance and protection of the montane forests. IPAL Technical Report D-2a, UNESCO, Nairobi. Tabuno, Fr. P. (1980) Practical use of plants by the Gabra people of Kenya. Unpublished manuscript, Marsabit Catholic Mission. J. EANHS & Nat. Mus. 81 (198) August 1991 Stiles & Kassam • Gabra plant use 34 APPENDIX Alphabetical list of Gabra plant names with their botanical equivalents. (Genus, species & family names) A aadde addaama agaggaro* agarsu alala (buuyyo alala) algge araddo arru B baddana baram-barro* barataa boorara buratte burk’uk’e burraa butte buutiyye buuyyo** — biila — diimtu — fiinc’oo — harre C’ c’aac’anne c’allankaa c’anc’ali c’imp’a c’iraa (buuyyo c’iraa) D dabobbesa dagamsa dak(a)d’aa dakkara deekuku diilaleesa dubarraara d’ad’ale d’addaca d’eekaa Salvadora persica L., SALVADORACEAE Euphorbia candelabrum Kotschy, EUPHORBIACEAE Indigofera coerulea Roxb.; I. colutea (Burm. f.) Merrill; I. cliffordiana Gillett; /. insularis Chiov.; I. spicata Forsk., PAPILIONACEAE Commiphora erythrea (Ehrenb.) Engl., BURSERACEAE Chrysopogon plumulosus Hochst., POACEAE (GRAMINEAE) Sansevieria robusta N.E. Br., AGAVACEAE Commicarpus helenae (J.A. Schultes) Meikle, NYCTAGINACEAE Juniperus procera Endl., CUPRESSACEAE Balanites aegyptiaca (L.) Del.; B. orbicularis Sprague, BALAN1TACEAE Cucumis prophetarum L., CUCURBIT ACEAE Blepharis ciliaris (L.) B.L. Burtt.; B. linariifolia Pers., ACANTHACEAE Caralluma speciosa (N.E. Br.) N.E. Br ASCLEPIADACEAE Cucumis dipsaceus Spach, CUCURBIT ACEAE Acacia nilotica ssp. subalata (L.) Del., MIMOS ACEAE Acacia goetzei Harms, MIMOSACEAE Dracaena ellenbeckiana Engl., AGAVACEAE Ormocarpum trichocarpum (Taub) Engl., PAPILIONACEAE Aristida adscensionis L.; A. mutabilis Trin. & Rupr.; Digitaria velutina, POACEAE (GRAMINEAE) Themeda triandra Forsk., POACEAE (GRAMINEAE) Sporobolus ioclados (Trin.) Nees, POACEAE (GRAMINEAE) Cenchrus setigerus Vahl., POACEAE (GRAMINEAE) Acacia paolii Chiov., A. horrida (L.) Willd., MIMOSACEAE Commiphora habessinica (O. Berg) Engl. BURSERACEAE Combretum cf. denhardtiorum Engl. & Diels, COMBRETACEAE Vignafrutescens A. Rich, PAPILIONACEAE Paspalidium desertorum (A. Rich.) Stapf; Cenchrus sp., Sporobolus helvolus (Trin.) Th. Dur. & Schinz, POACEAE (GRAMINEAE) Rhus natalensis Krauss, ANACARDIACEAE Carissa edulis (Forsk.) Vahl, APOCYNACEAE Commiphora boiviniana Engl., BURSERACEAE Boswellia hildebrandtii Engl., BURSERACEAE Cadabafarinosa Forsk., Maerua oblongifolia (Forsk.) A. Rich., CAPPAR ACEAE Cenchrus ciliaris L., POACEAE (GRAMINEAE) Heliotropium somalense Vatke, H. subulatum (DC.) Martelli, BORAGINACEAE Ruellia patula Jacq., ACANTHACEAE Acacia tortilis (Forsk.) Hayne, MIMOSACEAE Grewia tenax (Forsk.) Fiori, TILIACEAE J. EANHS & Nat. Mus. 81 (198) August 1991 35 Stiles & Kassam • Gabra plant use d’irri* d’umashoo d’uurtee E ejerssa ergamassa F fit’o fursaa G gaabbe gaalle gaddaa geeddi gelgedaana gurbi** H (h)addaa (h)ad’um (h)afuursaa (h)allak’abeesa (h)ammeesaa* (h)ank’arre (h)arkeena (h)arfuuk’a (h)armaac’a (h)arorressa (h)asura* — harre (h)idaa (h)iddi* — loonni — ree (aradab) I ic’iinni id’d’aad’o ilk’abate ilmmogora J jilbeete — kurroftu K kokoomisha Croton somalensis (Vatke) Pax EUPHORBIACEAE Maerua crassifolia Forsk., M. kaessneri Gilg & Bened., CAPPARACEAE Suaeda monoica J.F. Gmel., CHENOPODIACEAE Olea europea L. ssp. africana, (Mill.) P.S. Green, OLEACEAE Asparagus africanus Lam., LILIACEAE Rinorea convallariiflora M. Brandt, VIOLACEAE Lycium europaeum L., SOLANACEAE Vatovaea psuedolablab (Harms) Gillett, PAP1LIONACEAE Kedrostis gijef( J.F. Gmel.) C. Jeffrey, CUCURBITACEAE Zanthoxylum chalybeum (Engl.) Engl., RUTACEAE Echinochloa haploclada (Stapf) Stapf., POACEAE (GRAMINEAE) Digera muricata (L.) Mart, AMARANTHACEAE Seddera hirsuta Hall. f„ CONVOLVULACEAE Aspilia mossambicensis (Oliv.) Wild, Vernonia wakefieldii Oliv., COMPOSITAE Fadenia zygophylloides Allen & Townsend, Gyroptera gilletti Botsch, CHENOPODIACEAE Maerua sp., Cadaba mirabilis Gilg, CAPPARACEAE Acacia etbaica Schweinf., MIMOSACEAE Commiphora africana (A. Rich.) Engl., BURSERACEAE Setaria verticillata (L.) P. Beauv., POACEAE (GRAMINEAE) Euphorbia tescorum Carter, EUPHORBIACEAE Sporobolus spicatus (Vahl) Kunth, POACEAE (GRAMINEAE) Cistanche tubulosa (Schenk.) Hook, f., OROBANCHACEAE Grewia trichocarpa A. Rich., G. bicolor Juss., TILIACEAE Crotalaria cf. dumosa Franch, PAPELIONACEAE Indigofera coerulea Roxb. var. occidentalis Gillett & Ali, PAPILIONACEAE Cassia italica (Mill.) F.W. Andr., subsp. micrantha Brenan, CAESALPINIACEAE Euphorbia cuneata Vahl, EUPHORBIACEAE Solanum incanum L., SOLANACEAE S. coagulans Forsk., SOLANACEAE Triumfetta flavescens Hochst., TILIACEAE Acacia Senegal (L.) Willd., MIMOSACEAE Pavonia zeylanica (L.) Cav., MALVACEAE Leptrothrium senegalense (Kunth), Clayton, POACEAE (GRAMINEAE) Sericocomopsis hildebrandtii Schinz, Dasysphaera prostrata (Gilg & Schinz) Caraco, AMARANTHACEAE Leucas pododiskos Bullock, LAMLACEAE Heliotropium albohispAum Bak., BORAGINACEAE J. EANHS & Nat. Mus. 81 (198) August 1991 Stiles & Kassam • Gabra plant use 36 K’ K’ad’u k’alk’acca k’antallaa k’arrarri k’arrarru k’atte k’ayyu k’iltaa k’iltip’p’e k’obbo k’onc’orro k’oraatti gaala k’ork’odda k’orrobbo k’ors nyaata k’umbi k’urk’uura L laabbesa laamisho lakud’e lookko luuftoole M maa mad’eeka mad’eera marmma mat’t’anne meetti mogorre mookofa muk-illeensaa N nyaap’p’o O obbe ogomdi ogoono okolle R raafu*** rorroddo rukeesa Cadaba mirabilis Gilg, C. gillettii R.A. Graham; Boscia (?), CAPPARACEAE Maerua angolensis DC., M. crassifolia Forsk.; boscia coriacea Pax, CAPPARACEAE Trianthema salsoides Fenzl., AiZOACEAE Sterculia africana (Lour.) Fiori, STERCULIACEAE Acokanthera schimperi (A.DC.) Schweinf., APOCYNACEAE Ecbolium revolutum (L.) C.B. Cl., ACANTHACEAE Commelina latifolia A. Rich., COMMELENACEAE Ficus glumosa Del., MORACEAE Indigofera spinosa Forsk., PAPILIONACEAE Calotropis procera (Ait.) Ait. f., ASCLEPIADACEAE Cenchrus pennisetiformis Steud., POACEAE (GRAMINEAE) Commicarpus helenae (J.A. Schultes) Meikle, NYCTAGINACEAE Indet., CAPPARACEAE Terminalia spinosa. Engl., T. polycarpa Engl. & Diels, COMBRETACEAE Pseudosopubia hildebrandtii (Vatke) Engl., SCROPHULARIACEAE Commiphora coriacea Engl., C. ellenbeckii Engl., BURSERACEAE Ziziphus abyssinica A. Rich., RHAMNACEAE Panicum color atum L. POACEAE (GRAMINEAE) Zaleya pentandra (L.) Jeffrey, AIZOACEAE Indet., ACANTHACEAE Diospyros abyssinica (Hiem) F. White, EBENACEAE Cor chorus trilocularis L., TILIACEAE Farsetia stenoptera Hochst. CRUClFERAE Dactyloctenium bogdanii S.M. Phillips, POACEAE (GRAMINEAE) Barleria sp., ACANTHACEAE Cordia sinensis Lam., BORAGINACEAE Cocculus pendulus (J. R. & G. Forst,) Diels, MENISPERMACEAE Pupalia lappacea (L.) Juss, Sericocomopsis sp. (?). AMARANTHACEAE Hyphaene compressa Wendl., ARECACEAE Tribulus cistoides L., ZYGOPHYLLACEAE Croton dichogamus Pax, EUPHORBIACEAE Aerva javanica Schultes, AMARANTHACEAE Croton megalocarpus Hutch., EUPHORBIACEAE Adenium obesum (Forsk.) Roem. & Schult., APOCYNACEAE Adenia venenata Forsk., PASSIFLORACEAE Grewia villosa Willd., TILIACEAE Pennisetum mezianum Leeke, POACEAE (GRAMINEAE) Asparagus africanus Lam., LILIACEAE Aristolochia bracteolata Lam., ARISTOLOCHIACEAE Cyphostemma nieriense (Th. Fr. jr.) Desc., VITACEAE Combretum molle G. Don, COMBRETACEAE J. EANHS & Nat. Mus. 81 (198) August 1991 37 Stiles & Kassam • Gabra plant use S saariima Duosperma eremophilum (Milne-Redh.) Napper, ACANTHACEAE saatu Indet., POACEAE (GR AMINE AE) sap’ans gurraaca Acacia mellifera (Vahl) Benth., MIMOSACEAE sxgirso sokorsitu sukellaa suufki SH shiisha U uube W waac’c’u waaraa waanga waleena warab reeba waijidda Acacia reficiens, Wawra ssp.. misera (Vatke) Brenan, MIMOSACEAE Sehima nervosum (Rottl.) Stapf., POACEAE (GRAMINEAE) Delonix elata (L.) Gamble, Caesalpiniaceae Aerva javanica Schultes, AMARANTHACEAE Barleria acanthoides Vahl, ACANTHACEAE Rhynchosia minima (L.) DC., PAPILIONACEAE Acacia seyal Del. war. fistula (Schweinf.) Oliv., MIMOSACEAE Commiphora incisa Chiov., BURSERACEAE Acacia nubica Benth., MIMOSACEAE Erythrina melanacantha Harms, E. burtii Bak. f., E. rotundata-obovata Bak. f., PAPILIONACEAE Commiphora ellenbeckii Engl., BURSERACEAE Abrus schimperi Bak. ssp. africanus (Vatke) Verde., PAPILIONACEAE These are generics. The specific forms are labelled according to their use (liked by camel, goat etc.; incense; toothbrush, etc.), eco-zone ( badda , gammooji, bule, etc.), sometimes by colour (given by the soil) or reproductive characteristic (male, female). Buuyyo is the general word for grass and gurbi means “bush” or “shrub”. Both words sometimes form part of the name, followed by a qualifier. Raafu is the general work for plants whose leaves are cooked and eaten like spinach. It applies to more than one species. EDEN WILDLIFE TRUST Chm-ltublfl Trunt R^Rl»ler**d No. 278008 The East Africa Natural History Society is most grateful to the Eden Wildlife Trust for their very generous contribution towards the cost of publishing this journal part. Manuscript first received: 1984 Re-submitted: 10 December 1990 Editor: C F Dewhurst Published by: The East Africa Natural History Society » Typesetting by The Print Shop • Printed by AMREF J. EANHS & Nat. Mus. 81 (198) August 1991 NOTICE TO CONTRIBUTORS Journal of the East Africa Natural History Society and National Museum Papers on Natural History subjects will be considered on the understanding that they are being offered exclusively to the Society and have not been submitted, at the same time, or previously, for publication elsewhere. Copyright is retained by the East Africa Natural History Society; references to contributions in the Journal may be made, but no part of the text, diagram, figure, table or plate may be reproduced without permission from the Editor. Such permission will only be granted with the approval of the author(s). Two copies of papers for consideration should be sent to The Editor, Journal of the East Africa Natural History Society and National Museum, P.O. Box 44486, Nairobi, Kenya, and should be accompanied by a covering letter from the author, or in the case of multiple authors, from the author responsible for decisions regarding the text. Contributions may be sent as: 1 ) Preferably a floppy disk of either 5.25 or 3.5 inches in fully IBM compatible MS-DOS, carrying text processed in a commonly used wordprocessing language, or a floppy disk with the text prepared on an Apple Macintosh. If in any doubt, the authorfs) should contact the Editor at the above address for further details. Three"near letter quality" printed copies on A4 size paper double-spaced should accompany the diskette. The authors) should, of course, retain a back-up copy. 2) or as typescript, which should be on A4 size paper, typed on one side only, double spaced with margins of at least 5 cm on the left side of the paper to allow for editorial instuctions to the printer. Title: Should be brief and to the point. Authors name(s) and institution(s) where the research was carried out should be given, together with a suggested running page headline of not more than 50 characters including the names of author(s). A current address for correspondence, if applicable, should be included as a footnote. Abstract: Tobeonthefirstpageandconsistofnotmorethan 150words. This should outline the major findings clearly and concisely without requiring reference to the text. Headings: Should be brief, and in the following format: "Abstract, Introduction, Methods, Results, Discussion, Acknowledgements and References". All headings to be typed on a separate line from subsequent text. Main headings to be typed in capital letters and centered on the page. Sub-headings to be at the left of the page, in lower case; further headings should be in italics or underlined (not bold) beneath them. The first letter of any headings should begin with a capital letter. Text: Indent each new paragraph except those after a heading. Spelling should follow that of the Oxford English Dictionary. All pages must be numbered, at the top right hand edge. Numerals, dates & scientific units: Prose numbers should be written in full up to and including ten, and as numerals thereafter, except at the end and beginning of a sentence. Dates should be written as "29 August", "10 - 13 January 1989" or "1969 - 1989". Times should follow the 24 hour clock (e.g. 1500 h). All measurements must be in metric units. S.I. units and abbreviations should be used throughout with a space before and after the S.I. unit. Full stops should not be used after such units. The approximate position of T ables and Figures should be indicated in the left margin in pencil. Figures: The originals and two photocopies must be provided with each manuscript. Figures should be well drawn on pure white cartridge paper or good quality tracing paper. Lettering must be of very good quality (e.g. Letraset or good stencil). Typewritten or freehand lettering will not be accepted (without good reason). Each figure should have the authors' name, figure number and caption details written lightly on the reverse in pencil. Cite each figure in the text in consecutive order using arabic numerals. "Figure" should be abbreviated to "Fig.", except in the figure legends and at the beginning of sentences. Two separate caption sheets should accompany the typescript. Black and white photographs are acceptable if they have good J. EANHS & Nat. Mus. 81 (198) August 1991 definition and contrast and are produced on glossy paper. Consideration must be given by the authors) to their final size when reduced for publication in the Journal which is B5. Tables: Each should be typed, double-spaced on a separate page and clearly numbered. Tables should be simple and self explanatory, and have a brief title above each one. They must be numbered consecutively with arabic numerals. Nomenclature: In cases where there are well known and internationally accepted English names, these should be used throughout the text. The scientific name should be given in full, underlined or preferably in italics, when quoted for the first time. The authority should also be quoted at this stage except in non-taxonomic ornithological papers. Only the generic name should begin with a capital letter. In subsequent reference to the name, use only the abbreviated generic name, with the full specific name but not the authority e.g.Spodoptera exempt a (W alker); subsequently S. exempta. The authority should not be given if scientific names are used to describe an "association" or species complex, e.g. Acacia drepanolobium - Themeda triandra, wooded grassland. Type in capitals the first letter of the English names of species (e.g. Crowned Eagle, Grey-capped Warbler) but not of the higher taxa (e.g. eagles, warblers). References: Care should be taken to see that all references quoted in the text are also given in full, in alphabetical order, in this section. To avoid confusion, references to books should not be abbreviated and neither should they be underlined nor in italics. Journal names should be written in full, underlined or, if possible in italics. Personal communications should not be cited in the references, neither should they be abbreviated in the text. If there are two or more references cited by an author which were published in the same year, distinguish them by adding letters to the dates, e.g. Brown (1972 a). For technical reasons author names in the reference list should NOT be typed in upper case. Authors are advised to refer to this issue of the Journal for the preferred format. Acknowledgments: As acknowledgements may imply endorsement of the data and conclusions in the authors' work, it is advisable to obtain permission before quoting persons who have made contributions to the study. Once the paper has been selected by the Editor for possible publication in the Journal, the manuscript will be refereed by two persons who are acknowledged specialists in that field before a decision is made on its acceptance. The author(s) will be sent one set of print-outs for the correction of typographical errors. Authors are urged to ensure that all changes to the text are incorporated into this copy, as no further changes can then be made. The author is entitled to 25 copies free of charge. Thirty free copies will be provided if there is more than one contributor. Further copies can be supplied at cost, provided the order is placed before printing has been undertaken. The East Africa Natural History Society is supported by membership subscriptions, and may be unable in some instances, due to financial constraints, to accept all submissions of value. While every effort will be made to minimise such restrictions, contributors are requested to seek financial support to assist the Society with the ever increasing costs of publication. Editor Journal East Africa Natural History Society and National Museum. Referees: Proofs: Offprints: J. EANHS & Nat. Mus. 81 (198) August 1991 ' V 1h ‘ ^ V art £IIX fJH JOURNAL OF THE EAST AFRICA NATURAL HISTORY SOCIETY AND NATIONAL MUSEUM March 1992 Volume 82 No 199 STRUCTURE AND FUNCTION OF AFRICAN FLOODPLAINS JOHN J. GAII) FT* United States Agency for International Development, ABSTRACT In Africa, floodplains often cover enormous areas. They represent a formidable dry season refuge for the indigenous flora and fauna, but at the same time they have a large potential for the intensive, highly productive agriculture and hydropower production so desperately needed in Africa. The main topographic features of the larger floodplains are reviewed in this paper, along with a general insight into water relations, nutrient dynamics, productivity, species distribution and changes in vegetation induced by present management practice. The question is raised of whether floodplains will survive in the face of development, and a call is made for alternative management strategies. INTRODUCTION The inland water habitats of Africa make up about 450,000 km2 of the continent (Table 1). These habitats include seasonally inundated wetlands, such as swamp forest, peatland, mangrove swamp, inland herbaceous swamp and floodplain, as well as permanent water habitats. The habitat of most concern to us here is the floodplain, which is any region along the course of a river where large seasonal variation in rainfall results in overbank flooding into the surrounding plains. Some of these flooded plains are enormous and are equal in size to the world’s larges lakes (Tables 1 & 2). In African floodplains the typical complex vegetation mosaic is most obvious from the air, because they often cover enormous areas, stretching out as far as the eye can see. In this paper I can only touch briefly on the more important aspects of floodplains. In Africa they defy intensive study for a number of reasons, consequently this paper must be regarded only as a review. I hope it will provide some food for thought and a basis for future research. * Present address: AFR/ ARTS/FAR A, Room 2941, Dept of State, Washington DC 20523, U.S.A. The views and opinions expressed here are strictly those of the author, and in no way should they be inter- preted as the views and opinons of the U.S. Government or its official agencies. John J. Gaudct • Structure and function of African floodplains 2 Table 1 : Large open water habitats in Africa. Estimated area based on information in Beadle, (1974) and world atlases Water body Area (km2) Lake Victoria 75,000 Lake Turkana 32,500 Lake Malawi 24,000 Lake Kivu 16,000 Lake Chad 15,000 Lake Tanganyika 8,000 Lake Mobutu 4,700 Lake Edward 2,250 Lake Chilwa 700 Lake George 290 Lake Naivasha 240 Man-made lakes 27,000 Total open water 205,680 Floodplains 245,075 Total water habitats 450,755 METHODS General features The complexity of the vegetation mosaic seen in tropical floodplains is often the result of physical features relating to local substratum or to the effects of topography, which in turn affect the avail- ability of water and nutrients. Thus in order to understand the basic functions of the floodplain systems we must first look at their general features (see Welcomme, 1979). Topography The main topographic features within floodplains are either depressions, i.e. water bodies, or areas raised because of deposition, such as sand bars and mud flats. A typical feature is the levde, which is a raised berm or crest above the floodplain surface. It contains coarse material deposited as the flood flows over the top of the channel bank (Welcomme, 1979). Sutcliffe (1957) described the process whereby a silt-laden river, such as the White Nile, descending an appreciable gradient builds up its own bed and levdes until it actually flows at a level above the surrounding floodplain. As the bed slope flattens and flow rates decrease, suspended material is deposited more rapidly, levdes decrease in size, and the floodplain widens. Deposition is encouraged by vegetation grow- ing on the levdes. A good example is seen on the Cubango River after it leaves Angola and before it enters the Okavango Delta (Fig. 1), also the upper Zaire River with its associated vegetation (Fig. 2). Levdes are seen further downstream in the Okavango, and they are also evident where the Nile traverses the Sudan in the Sudd Swamp region. The levdes along the Nile River in the Sudd region trap the annual floods in depressions adjacent to the river. These are locally known as “toiches” (Fig. 3) and are fertile, seasonally-flooded habitats which support a wide variety of plant and animal species. J. EANHS & Nat. Mus. 82 (199) March 1992 3 John J. Gaudet • Structure and function of African floodplains Figure 1. Profile of levee and floodplain (molapo) along the Cubango River, Angola (after Smith, 1976) Table 2: Major African floodplains (after Welcomme, 1979 & Thompson, 1985). Location numbers refer to figure 6 Area of Area of Range of Location River high water low water conductance number system Region (W) (kmJ) (its. cm '') 1 White Nile Sudd Region, southern Sudan 92,000 10,000 20-500 2 Middle Congo Zaire and Congo 40,500 - - 3 Niger (see text) 25,980 5,980 31-70 4 Chari and Lake Chad swamps 13,800 _ Logone Yaeres floodplain 5.950 - 41-82 Sub total 19,750 5 Okavango Botswana internal delta 16,000 3,120 - 6 Kenamuke in S. Sudan Internal delta 13,950 - - 7 White Volta Volta Swamps 8,530 _ 41-124 and Oti in Ghana 4.810 - - Sub total 13,340 8 Senegal Main river swamps 4,560 500 72 Delta 7.970 - - Sub total 12,530 9 Lualaba SE Zaire, Kamulando depression in the Upemba Basin 11,800 7,040 141-255 10 Zambezi Barotse Plains in Zambia 9,000 700 57-126 11 Luapula Bangweula swamp, Zambia 8,800 - - 12 Rufiji Kilombero area in Tanzania 6,650 - - 13 Kafue Kafue flats in Zambia 6,000 - 130-320 14 Niger Niger River floodplain 4,800 1,800 - 15 Luena Liuwa plains in Zambia 3,500 - - 16 Benue Benue floodplain in Nigeria 3,100 1,290 - 17 Kafue Lukanga swamps N. of Kafue flats, Zambia 2,500 - - Sub total 56.150 Total area 290,200 30,430 J. EANHS & Nat. Mus. 82 (199) March 1992 John J. Gaudet • Structure and function of African floodplains 4 Figure 2. Sectional diagram of the upper Zaire River Basin (Shaba Province, Zaire) showing typical vegetation zonation associated with levees (after Thompson, 1985) As the level of the floodplain rises through deposition, the river eventually permanently breaches through some overspill notch in a lev6e, usually upstream. Flow is then diverted in a new direction across the floodplain. This leads to a complex system containing numerous water bodies. Welcomme (1979) distinguished the following: Lakes: large features which persist relatively unchanged over a number of years; Lagoons: water bodies which remain connected to the river throughout the year; Pools: smaller, more ephemeral bodies of water becoming isolated, drying out in some seasons; and Swamps: depressions where the soil remains saturated, more or less permanently covered with water HYDROLOGICAL CONDITIONS No floods Seasonal river spill floods Nearly permanent river floods LAND TYPE Hiqh land Toich Sudd VEGETATION Mixed Acacia or Palm forests with annual grasses pre-dominating Riverine swamp qrassland Papyrus swamp High river Low river ' f ^ 1 'W 1 IIMMI SOIL TYPE Clay and Loam soils Toich soils Sudd soils Figure 3. General profile of White Nile floodplain J. EANHS & Nat. Mus. 82 (199) March 1992 5 John J. Gaudet • Structure and function of African floodplains All these features can be seen in one portion of a floodplain on the Nile in southern Sudan (Fig. 4). Here two lakes. Lakes Nvong and Shambe, are seen along the course of the main river channel, the Bahrel Jebel, a branch of the White Nile. Perennial papyrus swamps line the river banks as well as lagoons and pools which are seen in profusion. Medium to deep flooded . Shallow flooded High ground (after Welcomme, 1979) Figure 4. One section of the White Nile Sudd region between Shambe and Kongor (68° N; 30-3 1°E) Two small lakes (Nuong and Shambe) are shown along with several towns The pools and lagoons in an African floodplain are dynamic ecosystems in themselves. Their physical features change dramatically with the wet and dry seasons when they often either dry up or fill up. This is clearly seen in a sequence shown for the Senegal River (Fig. 5), based on Reizer’s (1974) work, where he shows a flooding sequence which leaves behind only one large lagoon, Vindon Edi Lagoon, which persists for some time. J. EANHS & Nat Mus. 82 (199) March 1992 John J. Gaudet • Structure and function of African floodplains 6 Figure 5. Annual flooding cycle of Senegal River floodplain (after Reizer, 1974) J. EANHS & Nat. Mus. 82 (199) March 1992 7 John J. Gaudet • Structure and function of African floodplains Major African floodplains African floodplains of 1,000 kms1 2 or more are listed in Table 2 and located on Fig. 6. The largest are internal deltas which occur when a river system encounters some geological barrier causing lateral spread over very large alluvial plains. Good examples are: the Sudd region (Fig. 4), Yaeres Floodplain (Figs 7 & 8), Niger Central Delta (Fig. 9), and the Okavango Delta (Fig. 10). The total area of these major flooded wetlands, 290,000 km2 is reduced during low water periods by approxi- mately 85% to only about 43,500 km2, but as Thompson (1985) points out, the total areas flooded each year in continental Africa will be difficult to calculate. If the countless smaller regions are included, the estimated total would greatly exceed the below figure. Figure 6. Sketch map of Africa showing the location of major floodplain regions. Main African herbaceous wetlands and waterbodies located as follows: 1. Sudd Swamps of Upper Nile; 2. Middle Congo Swamp Region; 3. Niger Central Basin; 4. Lake Tchad Basin and Yaeres Floodplain; 5. Okavango Delta; 6. Kenamuke Swamp, Sudan; 7. Volta River; 8. Senegal Floodplain; 9. Upemba Basin, Lualaba River, 10. Barotse Plains; 11. Laupula Floodplain, Bangweula Swamp; 12. Rufiji Kilombcro Swamp; 13. Kafue Floodplain; 14. Niger River Floodplain; 15. Luiwa Floodplain, Zambia; 16. Benue Floodplain; 17. Lukanga Swamps. J. EANHS & Nat. Mus. 82 (199) March 1992 John J. Gaudet • Structure and function of African floodplains 8 Figure 7. Yaeres floodplain on the Chari and Logone rivers (after Carmouze et al„ 1983) Water and nutrients Water relations When dealing with such enormous regions in remote areas where only spot meteorological records are available, only rough estimates can be made of water relations. In large wet regions, the water balance would be complicated by local climatic conditions. Welcomme (1979) noted that precipi- tation on the floodplain itself saturates the soil, causing local flooding which often precedes the main overspill flood from the river. This is dramatically true in the Sudd region where the rain falls on impervious soil on either side of the river and then drains toward the channel. Extensive flooding may be evident even before the annual river flood occurs. The inundation of the upper Chari Basin and most other floodplains such as those of the Niger, Lualaba, Zambezi, Volta, Senegal and Kafue, follows similar patterns. However, there are excep- tions to the general rule. In the Okavango Delta, peak rainfall coincides with periods of low water and high evaporation in the floodplain; also, local rainfall is so erratic that no regular floodplain saturation may occur before the annual river flood. The water deficit here may be so large that on occasion no water comes out of the system. J. EANHS & Nat. Mus. 82 (199) March 1992 9 John J. Gaudct • Structure and function of African floodplains Seasonal Watercourses ► Mom directions of watftr -flow ► Minor direction of walwr -(low i Mom direction of migrating juveniles Secondary movements of juveniles Mo in dir©crt_ion o-f rh'»grcrt\r\g a ckiltS Secondary moviemcnt o-f odvi't^ 0 50 k m 1 1 I 1 1 1 1 Figure 8. Yaeres floodplain showing annual fish migration routes (after Welcomme, 1979) J. EANHS & Nat. Mus. 82 (199) March 1992 John J. Gaudet • Structure and function of African floodplains 10 Figure 9. Internal delta of the Niger River showing numerous lakes within the system (after Welcomme, 1979) Several workers concur that in Africa the evaporation rate in floodplains exceeds precipitation and in general a water loss can be expected during river passage through a floodplain (Table 3). This loss, 46-96% is due to evaporation from open water, evapotranspiration from emergent vegetation, and seepage loss. However, in most areas it is difficult to conceive of seepage loss as being significant because the fine clay deposits typical of floodplain areas would seem to act as a natural barrier to ground water loss. It is true that in some areas, such as the Okavango or Chari regions sand deposits are more prevalent and these would be conducive to ground water loss, and such ground water loss does occur in the Okavango region where the water tables slope away from the main flow channels. Water and mineral balance studies in the Lake Chad region also point to J. EANHS & Nat. Mus. 82 (199) March 1992 11 John J. Gaudct • Structure and function of African floodplains about 8% seepage loss in the sandy substrata in the southern part of the lake (Carmouze el al.. 1983). In many areas, including Lake Chad, the underlying substrate is a mud or compacted clay, or the surrounding soil may be so impervious, as in southern Sudan that water remains on the ground after rain, slowly moving downslope, a phenomenon referred to as “creeping flow” by local hydrologists. Evaporation and evapotranspiration (ET) in floodplains are even more controversial topics than seepage. Evaporation would occur from a variety of surfaces such as: open water, dry soil or wet mud and would be complicated by advected energy, which is high in open areas low inside the vegetation. Albedo also will vary with degree of wetness, burning of vegetation and other factors. The ET of papyrus has been shown to be less than that of open water (Rijks, 1969), possibly because it forms a closed canopy. On the other hand, open stands of an emergent macrophyte, e.g. Phragmites sp. will lose water at rates considerable in excess of those expected from open water (Imhof & Burian, 1972). 20°00° 20°15° Present downstream limit of continuous % Papyrus ^ Isolated or relict Papyrus ^ Limit of distribution of Phoenix palm Major rivers or river-beds . Approximate limit of (~\ areas currently liable \ to flooding 1 ' 20° 30° 22°15°E 22°30° 22%5 23°00° 23°15° 23°30° 23°45° 24°00° Figure 10. Distribution of papyrus and phoenix palm associations in the Okavango Delta (after Smith, 1976) J. EANHS & Nat. Mus. 82 (199) March 1992 John J. Gaudet • Structure and function of African floodplains 12 Table 3: Annual water balance in several floodplains (data from Wclcomme, 1979 and Thompson, 1976) Floodplain Station Precipitation (mm) Evaporation (mm) Input (mm) Output (xl010m3) Net Loss % Niger Bamako, Mali 1120 2150 7.11 3.82 46 Nile (Sudd) Wau, Sudan 1100 1900 2.70 1.40 48 Okavango Maun, Botswana 434 1800 1.18 0.06 95 Water chemistry and nutrient dynamics Generalizations about water chemistry in floodplains are difficult to make because of inter- and intrahabitat complexity. The most often measured floodplain parameter is conductivity (see Table 2), but measurements arc usually taken in open standing water and may not be a useful indicator of floodplain conditions. Conductivity in the main channel seems mostly influenced by: dilution with flood or rain water; solution input from land; swamps or floodplain drainage; evaporation; uptake by the biota; and absorption by sediments. Welcomme (1979) pointed out that all of these effects tend to produce higher conductivities in the dry season both in lagoons and river channels resulting in an inverse relationship between water depth and conductivity in African rivers. A secondary conductivity maximum due to solution input occurs when the river water invades the floodplain. The conductivity rise is usually accompanied by a pH drop due to acid swamp water input. The Niger central delta is an exception with a pH rise which remains unexplained. The conductivity rise and pH decrease is also accompanied by a decrease in dissolved oxygen which is typical during stagnant swamp or floodplain conditions. During dry seasons standing water in floodplains becomes stagnant, and this stagnation is strongly stratified due to a lack of strong wind effects and intense insolation. Sudden reductions in air temperature association with a cold front or a strong wind storm can break down this stratification. This results in an upwelling or “overturn” which can raise the oxygen deficit and cause chemically reduced bottom waters, and fish kills often result (Welcomme, 1979). Overturns and fish kills can also be caused during the entry of river floodwa- ter at the beginning of the flood season. This is an indication of how strongly reduced the bottom water can be in floodplains. One way of looking at changes in water chemistry in a floodplain is to compare the input to the output and assume all net changes can be ascribed to the overall system, that is, by using a “black box” approach. In floodplains or other large wetlands this approach is taken more out of futility than anything else. For example, longitudinal transects of water chemistry were done by Tailing (1957) and Bishai (1962) along a 1,000 km stretch of the Nile river where it passes through the Sudd region along the Bahr el Jabel. In this section of the river the results of both Tailing and Bishai show that conductivity, alkalinity, calcium and chloride levels were erratic and did not follow any distinct pattern. This agrees with a study by Gaudet (1978) who found that tropical swamps had little effect on the salinity of the through-flow, as long as the water passed freely through the system. Both Tailing and Bishai did find some effects on water chemistry along the 1 ,000km stretch of the Nile in that there were decreases in oxygen, sulphate and pH, and a rise in carbon dioxide, silicate, ammonia and iron in the river. Many of these changes could be explained on the basis of decomposition effects, the sort of general effects that are typical of mires and floodplains (Howard-Williams & Gaudet, 1985). A recent study was carried out in the Sudd Region to determine the ecological impacts of the Jonglei Canal (Anon, 1983). Part of the study was devoted to determining the effects of the Sudd on the chemistry of the water passing through. Because of the complexity of the system, earlier J. EANHS & Nat. Mus. 82 (199) March 1992 13 John J. Gaudct • Structure and function of African floodplains results as above and the enormous overriding effect of evaporation (a 50% loss of water occurs while it transits the Sudd) it was assumed that the effects of the whole system would be of limited magnitude. The study therefore concentrated on only a small portion of the whole Sudd. Two permanent sampling stations were set up, one at Bor, where the White Nile can be sampled prior to entering the Sudd and one downstream at Dhiam-Dhiam where a branch of the main White Nile river (the Atem) has passed through 75 km of floodplain and swamps. The effects noted (Table 4) were not much different from those noted in previous studies. The authors used the general model proposed by Gaudet (1979) for East African wetlands which proposes that three phases could be detected in the papyrus swamps on the northern edge of Lake Naivasha in Kenya, that is: stagna- tion, through-flow and drying. Apparently the Sudd region below Bor follows a remarkably similar pattern to the Lake Naivasha system. This even includes a rough correspondence of annual cycles. The results of the Sudd study show an overall decline in oxygen and pH, with increases in dis- solved carbon dioxide. A decrease also occurred in nitrate, phosphate and sulphate as expected. It is significant that high levels of dissolved silicon were found entering and leaving the system. All of these results are similar to those found by Gaudet working on papyrus swamps in Uganda and Kenya (Howard-Williams & Gaudet, 1985). Table 4: Summary of the processes at work in the Sudd wetlands, and the overall effects on water passing through the system (based on Anon, 1983) Phase I II III Status Stagnation Through -flow Drying Season Late dry Wet Late wet-Early dry Water balance Low river input, Minimal outflow Early pulses of river water inflow Minimal inflowwith continued drainage Processes in operation Decomposition and evapotranspiration Flushing due to influx of river water Stagnation begins Water chemistry increase, rise in phosphate and iron Turbidity decrease, colour Alkalinity increase Rise in conductance Effect on receiving waters Very little outflow Early flush of high alkalinity, later flush is more dilute General outflow of high conductance water The Sudd study concluded “the Sudd is unlikely to act as a long-term nutrient trap,” because most of the nutrient inputs into the Sudd will be taken up into the organic plant fraction and this will be recycled within the wetlands or flushed out into the river. However, the Sudd wetlands do have an ability to take up large amounts of nutrients on a short-term basis and this phenomenon is important in that the wetlands involved can act as a buffer to prevent large flushes of nutrients from entering receiving waters. The amounts that can be taken up can be estimated if we assume that much of the swamp and floodplain production is turned into detritus. This process would remove large amounts of soluble nutrients from the water passing through the system, and this in turn would affect the chemistry of water downstream. A rough estimate based on this assumption (Table 5) shows that the Sudd could have less effect on major inorganic ions such as Sodium (Na), Potassium (K), Calcium (Ca), J. EANHS & Nat. Mus. 82 (199) March 1992 John J. Gaudet • Structure and function of African floodplains 14 Magnesium (Mg) and Chloride (Cl), than on nutrient ions which are present at much lower concen- tration levels, such as Nitrogen (N), Phosphorous (P), Sulphur (S) and Iron (Fe). the nutrient ions would be even more strongly affected than this estimate because they would be in great demand by the river biota. In theory, the swamps and floodplain would completely remove nutrient ions from the river and flood water. In practice, however, the vegetation in these habitats probably takes up much of its Phosphorous (P), Sulphur (S) and Iron (Fe) from local floodplain soils and the large nitrogen demand is probably met by Nitrogen-fixation, a process that has been suggested by other authors to be a common feature of wetlands (Bristow, 1974). All of the nutrients so trapped in organic matter will be released later if and when the organic matter is degraded downstream. Table 5: Potential effect of one part of the Sudd on the chemistry of Nile river water Element Rate of removal by Vegetation2 Removal by from detritus1 (g/m2) (tonnes x 103) river3 (tonnes x 103) N 67.83 678 3 P 2.61 26 1 S 31.71 317 46 Fe 7.39 74 12 Na 7.48 75 230 K 5.823 58 81 Ca 1.13 11 173 Mg 3.65 37 55 Cl 0.70 7 152 1 Based on Gaudet (1978, 1979) 2 Assuming an immediate low water area of 10,000 km2 3 Concentration of Nile above Sudd region, based on Tailing (1957), Bishai (1962), and Kilham (1972). Also assumes a discharge of 23 x 109 m3 (Rzoska, 1976) Table 6: Effect of drought conditions on water chemistry and phytoplankton in a seasonally inundated part of the Okavango Delta (data from Reavell, 1979) based on 13-37 samples Total discharge from delta Total P Total N Temp Surface 02 Phytoplankton Date (m3s_1) (gl_1) (a) (gH) (b) (°C) (mgl-1) (cells H) iii/1972 — — — 28.2 — — vi/1972 — 60 425 17.7 6.9 — viii-ix/1972 9.32 95 604 21.4 6.3 — ii-iii/1973 0 171 994 29.2 4.5 54,038 v/1973 0 252 2,772 24.2 7.5 683,961 ix-x/1973 3.48 86 652 25.2 — 41,103 xii/1973 0.33 86 277 26.4 7.0 17,889 ii-iii/1974 4.04 119 176 26.4 7.6 8,545 vi-viii/1974 11.16 60 214 19.6 6.4 7,121 (a) Sol react. P plus org. P (b) Total dissolved inorganic only J. EANHS & Nat Mus. 82 (199) March 1992 15 John J. Gaudet • Structure and function of African floodplains i- Internal recycling may also be possible because of the large amounts of nutrient material trapped within floodplains. In many floodplains evaporation leads to an increase in concentration of salts. This is obvious during droughts, such as the dramatic drought in the Okavango delta in 1973. At this time evaporative loss was not balanced by local rain input, and Reavell (1979) noted that as the floodplain and swamp water contracted in to discontinuous pools there was an increase in total dissolved solids from 83 to 223 mg'1. Such water, when trapped during dry weather inside the floodplains, becomes a concentrated source of nutrients. This, along with the creation of still- water conditions, leads to an excellent growth environment for hypereutrophic algal species. Reavell (1979) documented this cycle in the Okavango delta (Table 6), and similar events are also known elsewhere in Africa. In the Chad basin when the Chari and Logone rivers overflow their banks and rapidly flood the Yaer6s floodplain, the same nutrient flush is seen (Beadle, 1974). At that time, the dried organic and mineral matter, along with ashes from bush fires and droppings from animal herds which previously inhabited the dry river beds, begins to decompose and to dissolve. Within a few days an explosive growth of phytoplankton occurs, followed by a zooplankton bloom and later by growth of macrophytes. In this floodplain, grasses and emergents (Polygonaceae and Alismataceae) occur in shallow water, while the genera Pistia, Nymphaea, Stratiotes and Utricularia occur in the deeper portions. The floods later subside and leave great numbers of fish stranded within the vegetation-choked water. Later still the area begins to dry down again and is then accessible for dry season animal grazing until the next flood, when this “instant eutrophication”, or nutrient build up, will again occur (Beadle, 1974). Productivity Production and biomass Floodplains are dominated by herbaceous monocotyledons, the most predominant genera being the tropical wetland grasses, Sporobolus, Echinochloa, Oryza, Vossia, Phragmiles, Paspalum, Miscanthidium, and Panicum. Sedges, especially Cyperus papyrus and Bulrush ( Typha sp.) are also common in the deeper floodplain water. Among these genera, the most well-studied are papyrus and bulrush. Net production rates determined for these two examples are very high. Papyrus net production on a sustained basis is 48-143 tonnes (dw) ha ’y1 (Thompson et al., 1979). For comparison, it should be noted that maize or sugar cane production can also be sustained at this rate, but only under heavy applications of artificial fertilizer. Papyrus maintains its high rate of production under natural, low nutrient conditions. Typha domingensis in the Lake Chilwa swamps averages 16 tonnes ha 'y1 (Howard- Williams & Lenton, 1975). Net production rates for Phragmites australis have been predicted at 58 tonnes ha 'y'1 if growing under tropical conditions (Thompson et al., 1979). Tropical wetland grasses ( Paspalum repens) are known to attain 12-16 tonnes ha 'y 1 (Junk, 1970), while the esti- mated net production of seasonally-flooded grasslands would be about 10-20 tonnes ha'y1 (Thompson, 1976). In addition to the dominant emergents, there is commonly a mixture of submerged and floating macrophytes at the river channel edge and in open areas of standing water. Indeed, pools and lagoons along the Nile and Zaire Rivers, and their attendant floodplains may be completely covered with Pistia stratiotes and/or Eichornia crassipes during stagnant water conditions in the dry season. E. crassipes alone is capable of 11-33 tonnes ha 'y1 (Westlake, 1975). Overall, most African floodplains should be capable of a sustained net production approaching 20 tonnes ha 'y '. On a continent-wide basis this would roughly amount to an annual net production of 0.8 x 109 tonnes using the area estimated at low water. For comparison, the total annual net production of tropical grasslands on a worldwide basis would be between 4 and 30 x 109 tonnes (Lieth, 1972) J. EANHS & Nat. Mus. 82 (199) March 1992 John J. Gaudet • Structure and function of African floodplains 16 The high rate of production in floodplains is certainly a factor in the overall productivity of these areas, especially in Africa where they support large numbers of domestic and wild animals, e.g. lech we ( Kobus ) in the Kafue River floodplains (Werger & Ellenbroek, 1980). Another factor here is the quality of floodplain grasses. Van Rensberg (1971) found the Kafue flats floodplain grasses were “grasses of high fodder quality, containing a relatively high content of crude protein. They are also highly palatable, even when dry.” In contrast, grass species from non-flooded areas lose their palatability on drying. Thus even in regions of similar grass production, the floodplain species will always be more attractive to herbivores. Vegetation Floodplain species The general vegetation of floodplains includes a large range of life forms. Within standing water bodies, pools and rivers inside the floodplain we can find typical hydrophyte associations (obligate water-plants). At the other extreme are the terrestrial plants that can tolerate a degree of flooding. The majority of the species that occur are emergent amphibious plants. Their habit ranges from herbaceous to woody, but they more often than not are emergent rhizomatous monocotyledons, predominantly grasses and sedges.They fall within the main classes of species for African wetlands established by Thompson (1985). Among the major wetland types listed, (Table 7) we are dealing here with the seasonal types (section 11(a) (2) A good starting point for the study of a typical floodplain flora is the partial list from Lebrun’s work in 1947 (Table 8). We see in this list a predominance of rhizomatous emergents (15 out of the 18 species). Many of the species mentioned are also common to perennial swamps. Thus, bulrush ( Typha sp.) and papyrus (C. papyrus) exist within the floodplains of large internal deltas. In addition to these typical species, it would also be appropriate to list several floating Sudd species which are common to floating meadows, or floating islands, such as Vossia cuspidata, Saccolepis interrupta , Echinochloa scabra (= E. stagnina ), Oryza longistaminata and Jardinea congoensis. These can often be seen inside floodplains during high water. Major floodplain associations The easiest way of describing floodplain associations has been the use of profiles drawn to scale, transecting the major channels or rivers concerned. The larger floodplains, however, defy descrip- tion using such methods. They often are simply traced from satellite imagery. The result is that we are left today with very general descriptions of African floodplains which often cover large geo- graphic areas. A case in point are the diagrams shown in Vesey-Fitzgerald (1973) (Fig. 11). He outlined several floodplains common to the valley grasslands along the Kafue River. In the upper part of the river, valley grasslands drain freely to allow a gradual drying and a distinct association develops (Fig. 1 1 a). However, there are many areas especially in the Kafue fiats where drainage is impeded because of silt deposition (Fig. 1 1 b). Here, in addition to the valley grasslands the floodplain supports associations dominated by E. pyramidalis, E. haploclada, and V. cuspidata and closer to the mainstream along the levies one finds E. scabra. J. EANHS & Nat. Mus. 82 (199) March 1992 17 John J. Gaudet • Structure and function of African floodplains Table 7: Main classes of plant species, and communities in African wetlands (Thompson 1985) I . MAJOR CLASSES OF SPECIES A. Monocotyledons 1. Graminae a. Large wetland grasses b. Small wetland grasses 2. Other herbaceous genera 3. Woody genera B. Dicotyledons 1 . Herbaceous 2. Woody C. Gymnosperms D. Pteridophytes E. Bryophytes II. MAJOR WETLAND TYPES A. Freshwater herbaceous wetlands 1. Permanent a. Reed swamps b. Bog fen and moorland 2. Seasonal a. Black clays, pans and dambos b. Floodplains and valley grasslands c. Alkaline grasslands B. Freshwater swamp forests 1. Permanent 2. Seasonal a. Floodplain forests b. Riverine and gallery forests III. MAJOR WETLAND PLANT COMMUNITIES A. Sudd communities B. Reed communities 1 . Vossia swamps 2. Cyperus swamps 3. Miscanihidium swamps 4. Cladium , Phragmites and Typha reed swamps C. Edaphic grasslands 1. Hydroseres 2. Secondary grasslands 3. Black clay, pan and dam bo communities 4. Alkaline grasslands and swamps D. Edaphic forests 1 . Permanent swamp forest 2. Seasonal swamp forest 3. Riverine and gallery forests secondary forests E. Montane wetlands J. EANHS & Nat. Mus. 82 (199) March 1992 John J. Gaudet • Structure and function of African floodplains 18 Table 8: Species common to flooded regions in Central Africa (after Lebrun 1947) sedges: Cyperus papyrus L. C. haspan L. C. articulatus L. C. dives Del Pycreus mundtii Nees. grasses: Leersia hexandra Swartz Phrag mites mauritianus Kunth. Echinochloa pyramidalis (Lam.) Hitch. & Chase E. scabra (Retz) Beauv. (= E. stagnina ) Vossia cuspidata (Roxb.) Griff. Oryza longistaminata A. Chev. & Roehr Paspalidium geminatum (Forssk.) Stapf Sporobolus robuslus Kunth. bulrush: Typha australis Schum. & Thonn. HERBS & SHRUBS: Aeschynomene elaphroxylon (Guill. & Perrot) Taub Mimosa asperata L. Pluchea ovalis (Pers.) DC. The typical occurrence of V. cuspidata and E. scabra close to the mainstream or in deep- flooding habitats will be discussed below in relation to the Sudd region. But first, it would be of interest to consider the most common successional types, hydroseres, within the floodplains. Thompson (1985) has pointed out that in the savanna regions of Africa, the most common hydrosere is Vossia-Oryza-Hyparrhenia. He noted that this sequence from an aquatic emergent grass to a typical savanna terrestrial type occurs in the Sudan, the Kafue Flats, the Bangweulu floodplain and generally throughout the West African river systems. The relative proportion of each component within the hydrosere varies from site to site, e.g., Vossia predominates in the fringing wetlands of the Zaire river, Oryza sp. dominates in the Kafue Flats, E. scabra is most important in the inland delta of the Niger River (Mali), and E. pyramidalis dominates in many of the Central African valley floodplains. Thompson (1985) also made the point that there is much variation within each locality. Many other grass species are usually found at any one site. A good example is shown in Fig. 12, where two Zambian floodplains are compared. We see here that in addition to the inevitable Vossia and E. scabra, there is a whole range of grass species within each floodplain. Unfortunately, in such a diagram there is a tendency to compress the species so that the mosaic pattern is not evident. In reality the mosaic or patchwork pattern is obvious and quite important. In the mosaic we can often see large differences in the distribution of non-dominant, minor species. Thompson pointed out that certain non-dominant floodplain grasses show great diversity, inhabiting a number of important niches. For example, Cynodon dactylon can exist as an understory grass within a com- munity dominated by E. pyramidalis, either as an associate of water-meadow grasses, as a sward- former in its own right, or as a member of the marginal community in a floodplain grass mosaic. Differences in species composition or dominance will also be related to local differences in soil J. EANHS & Nat. Mus. 82 (199) March 1992 19 John J. Gaudct - Structure and function of African floodplains type and flooding conditions. Regarding soil types, Reizer (1974) showed a clear difference in floodplain species along the Senegal River where there were large differences in soil types (Fig. 1, 3A, B, and C). Figure 11. A. Diagram of a section downslope into a valley with free drainage and distinct lev6es on either side of the main channel B. Diagram of the gradual change along a slope with impeded drainage. Both sections in central East Africa (after Vesey-Fitzgerald, 1973) Figure 12. Composite of two Zambian floodplains. Each site had different flooding regimes and different soil types but extensive grass associations occur in both cases (after Thompson, 1985, from Astle, 1965 and Van Rensberg, 1968) J. EANHS & Nat. Mus. 82 (199) March 1992 John J. Gaudct • Structure and function of African floodplains 20 Species distribution Certain large obvious species within floodplains have been studied because of their value as indicator plants. Smith (1976) in a study of the Okavango Delta used both papyrus and a phoenix palm as indicators. Because of the difference in rates of colonization between the two species their limits of distribution (see Fig. 10) at different periods of time indicated changes in permanent high water tables. Other indicator plants in floodplains which are often cited, are the salt-tolerant species that favour alkaline saline floodplains. In West African profiles, for example. Reizer (1974) described a Phragmites-Sporobolus-Tarnarix association which was clearly indicative of saline conditions. In fact, Sporobolus is often cited as the most adaptable genus in African alkaline saline floodplains (Thompson 1985). In Central and East African floodplains another useful indicator of saline habitats is Cyperus laevigatus. This is clearly seen in the profile taken by Lebrun (1947) across a small river emerging from a saline spring ( Fig. 13). found along levees in secondary channels in association with a small tree, Acacia nilolica. A comparison of lists of species from different African floodplains can be seen in Thompson (1985). His consideration of several lists from diverse regions within Africa shows a certain common trend in the continental floodplain flora. This is expected because of the wide distribution of many freshwater macrophytes. Species lists and/or general descriptions of floodplain vegetation can be found in the references cited in Table 9. Here some distinction is made between major floodplains associated with large rivers and those present on smaller rivers and lakes. The distribution of some tree species in floodplains is correlated with animal activity. Feely (1965) found that the hippopotamus ( Hippopotamus amphibius L.) influences the establishment of Acacia albida (Del.) along the Luanga and Zambezi river valleys in Zambia. These animals along with other wildlife eagerly seek out the foliage and ripe pods of A. albida. The seed is dispersed and germination enhanced by passage through their digestive system. As a result of hippopotamus travel inside floodplains, the young trees are regularly associated with newly deposited silt along the river meanders. Feely considers this species a pioneer on the levdes. Its roots tolerate submergence for several weeks or a few months, but the foliage must stay above the flood water for the young trees to survive. This is possible because of its fast growth rate during initial colonization (1.05 my1). Some work has been done on differentiation of species within the complex mosaic typical of floodplains. The work of Sutcliffe (1976) stands as a landmark in this respect. Within the Sudd region on the White Nile in the region south of Bor (Fig. 14) he differentiated the following communities: J. EANHS & NaL Mus. 82 (199) March 1992 21 John J. Gaudet • Structure and function of African floodplains a) grassland dominated by Phragmites mauritianus, containing Echinochloa pyramidalis and Oryza longistaminata', b) grassland composed of varying portions of E. pyramidalis and E. scabra c) semi-permanent swamp with E. scabra and Vossia cuspidata d) permanent papyrus swamp, and e) open water. Each of these communities graded into one another along the river, but a fairly pronounced boundary was evident between “shallow-flooded” species ( P . mauritianus, E. pyramidalis and O. longistaminata) and “deep-flooded” species (C. papyrus, V. cuspidata, and E. scabra). Off the main rivers in basins, the upstream region supported the shallow-flooded species, while the down- stream end of the basin contained the deep-flooded species. Sutcliffe (1976) also calculated the percent plant cover along 16 transects (Fig. 14) at different vertical heights, the level of the substrate above mean datum. He found a correlation between height and percent cover for shallow- flooded species. Here Phragmites occupied the highest level while E. pyramidalis and Oryza sp. were present at lower levels. Deep-flooded species, on the other hand, showed no evident correla- tion between level and cover. But in the deep-flooded communities the species composition depended on position rather than level, e.g. at the lower, wetter levels E. scabra was widespread in the floodplains, while Vossia sp. and Papyrus preferred sites near the rivers, with papyrus near the lower end of each basin. Table 9: References on floodplain vegetation associated with African water bodies (see also the annotated bibliography in Thompson, et al., 1985) Country Author 1 . MAJOR FLOODPLAINS ON MAJOR RIVERS White Nile Sudan El Hadidi (1976), Anon 0983) Niger Nigeria Cook (1968) Okavango Botswana Smith (1976) Zaire Zaire & C. Afr Republic Schmitz (1971), Germain (1952) Zambezi Zimbabwe Muller & Pope (1982) Volta Ghana Lawson (1963) Senegal Senegal Reizer (1974) Kafue Zambia Douthwaite & Van Lavieren (1977), Van Rensberg (1971) OTHER FLOODPLAINS A. Smaller rivers Ruzizi Zaire Germain (1952) Pongola. Tugela, South Africa Musil el at (1983), Berg & Tuskei Hamson (1964), Oliff (1959), AUanson (1961) B. Man-made lakes Lake Kainji Nigeria Chachu (1979) Lake Volta Ghana Hall etal. (1971) Lake Kariba Zimbabwe, Zambia Bowmaker (1973), Magadza (1970) Lake Nasser-Nubia Egypt, Sudan El Hadidi (1976) I^ake Cabora Bassa Mozambique Jackson & Davies (1976) I^ake Nyumba ya-Mungu Tanzania Welsh & Denny (1978) C. Natural lakes Lake Naivasha Kenya Gaudet (1977) Lake Sibaya South Africa Howard-Williams (1980) lake Chad Chad I^eonard (1969), litis & Lcmoalle (1979) lake Kivu Rwanda, Zaire Van der Ben (1959) lakes Mobutu & Edward Uganda, Zaire Van der Ben (1959) lake Chilwa Malawi Howard-Williams & Walker (1974) lake George Uganda Lock (1973) Lake Bangweulu Zambia Verboom (1975) lake Victoria Uganda, Kenya, Tanzania Lind & Morrison (1974) J. EANHS & Nat. Mus. 82 (199) March 1992 John J. Gaudet • Structure and function of African floodplains 22 Sutcliffe (1957) also examined the distinct boundary which exists between the shallow- and deep-flooded communities. He found that the distance of the boundary from the river edges varied considerably but followed the water-slope very closely. The water-slope depended on: (a) the maximum depth of flooding; (b) the period of flooding: and (c) the water-table, which in turn is dependent on the river level. He ruled out water-table as a water-slope factor affecting plant distribution, because the boundary between the communi- ties is quite distinct. If water-table were involved there would be little or no difference in the level at which each community occurs, and thus no clear boundary would exist. Also he found the period of flooding was not correlated with the boundary of these com- munities because the water surface during the period of his survey (10 April - 3 May 1952) did not coincide with the topography of the boundary. The maximum depth of flooding, on the other hand, was parallel to the bound- ary and appears to be the major factor determining this boundary between shallow- and deep-flooded communities. His levelling data indicated that sites where the annual depth of flooding does not exceed 130 cm for 10 days, or 1 18 cm for 30 days are ideal for shallow-flooded communities. Conversely, sites flooding in excess of 1 30 cm will only support deep-flooded communities. Within the shallow-flooded communities he noted that the proportion of Phragmites increased going away from the rivers. In most cases the distribution of Phragmites depended on the depth of available water especially during the dry season. This in turn depends on the period of flooding, thus Phragmites occupies sites which are flooded for short periods and dry out to a greater depth. In deep-flooded species, distribution is more complex as it is not related to substrate level. Vossia and E. scabra are found in quiet water. Because papyrus was found both above and below the minimum levels of flooding, Sutcliffe (1957) concluded that its distribution was not as dependent on depth or duration of flooding as the other two deep-flooded species. He found that papyrus was restricted to sites where the depth of flooding (the difference between minimum and maximum water level during flood) is less than 150 cm. Greater levels can be tolerated by the other deep- flooded species such as Vossia until a range of 3 m is reached at which point open water prevails. The ability of Vossia to dominate deep-flooding sites, i.e., hydrological ly unstable environments, was confirmed by Thompson (1985) who pointed out that it is reported from deep-flooding areas in Sudan, Chad, West African and East African rivers, Zaire and Zambia. Bor to Juba showing 16 transects, A-P. This region is south of that shown in Figure 4. J. EANHS & Nat. Mus. 82 (199) March 1992 23 John J. Gaudet • Structure and function of African floodplains Table 10: Comparison of two floodplains (based on Thompson’s UNDP review, 1974). Floodplain: White Nile Sudd region Okavango Delta, Local name: Toiche Molapo Flooding and local climate: 4-5 month lag peak rainfall in Uganda and maximum flooding in toiche. Maximum precipitation in region coincides with peak floods and low PE. 5 month lag between peak rainfall in Angola and maximum flooding in molapo. Maximum precipitation occurs during low water and peak PE. Hooding sequence: Flooding caused by water level in main channels overtopping natural levies. Flooding caused by overtopping of natural levees. Grasses in floodplain: Vetiveria nigritana Sporobolus pyramidalis Sorghuni lanceolatum P asp alum scrobiculatum Set aria anceps Oryza longistaminala Loudetia superba V. nigritana Sporobolus sp. Sorghum sp. P asp alum sp. Setaria sphacelata Oryza sp. Cymbogon plerinodes Digitaria eriantha Echinochloa scabra Wildlife: Many birds and smaller animals. Sitatunga, Lechwe, crocodile and hippopotamus still found. Many birds and smaller animals. Sitatunga, Lechwe, crocodile and hippopotamus evident. Fisheries (m tonnes yr'): «■* 400 Land use and human activity: 500,000 people and 700,000 cattle associated with region. Considerable use made of toiches for grazing promoted by burning. 12,000 people and 130,000 cattle in and around delta. Natural crown fires probably common, but otherwise burning is only practised within swamp areas for benefit of hunters and tourists . Sutcliffe (1957), by a series levelling and surveying measurements was able to explain the distribution of emergent vegetation on the basis of maximum depth and range of flooding. His explanation is much simpler in concept and operation and is more consistent with available data than earlier explanations based solely on depth and duration of flooding, and therefore is more appealing His general results were recently confirmed by the Mefit-Babtie study of the Sudd region (Anon., 1983). ** Presently being assessed (long term study underway) J. EANHS & Nat Mus. 82 (199) March 1992 John J. Gaudct • Structure and function of African floodplains 24 What is needed now is a more detailed and experimental approach which will distinguish between floodplain species using some system based on water relations. We also need to know more about the site preference of specific floodplain species, especially those floodplain grasses which were not included in this early work. CONCLUSIONS Floodplains under normal conditions The most impressive aspect of floodplains under normal conditions is the integration of physical and biological components within the system. Welcomme (1979) commented on this integration: “The traditional integrations of biological and human activities which make up the ecology of the floodplain river can only work if there is a community of micro-organisms, plants and animals which are adapted to the particular frequency of the environmental event which is fluctuating — in this case the flood.” Along large African floodplains the seasonal progression of events is similar. Thus, two examples north and south of The Equator, The White Nile Sudd (Figs. 4 & 14) and the Okavango delta (Fig. 10), although separated by 23-29 degrees of latitude follow similar cycles (Table 10). Fires occur in both and result in a fast recycle of ash nutrients but in a loss of volatile nutrients such as Nitrogen and Sulphur. The annual cycle shown diagram matically in Figure 15 reflects a very productive system that is under a series of complex control mechanisms. However, because of their high rate of natural production, floodplains are attractive to developers and in many places in Africa, floodplains are being brought under management. J. EANHS & Nat. Mus. 82 (199) March 1992 25 John J. Gaudct • Slructure and function of African floodplains Floodplains under management Welcomme ( 1979) gives many examples where water regulation, especially flood control, has allowed productive irrigation and fish farming to take place within floodplains. This is especially true in West Africa. In the closing section of this paper, 1 would like to look not at these well- known examples, but at some different cases, especially among major floodplains which have been put under stress as a result of river basin development. The subsequent results provide us with a valuable experience from which many African countries can benefit. The most publicized example of floodplain management in Africa is the enormous scheme which was started in the southern Sudan to conserve water by canalization of the main flow through the Sudd region on the White Nile. At present the scheme has been halted due to political disturbances in the south of the country. However, an ecological impact assessment was completed (Anon. 1983), and this outlined the major effects to be expected on completion of the canal. The basic options for management in this case will rest on future discharge levels in the White Nile itself. The canal can be operated either at high discharge levels such as have occurred since 1961 or at lower discharge levels such as oc- curred prior to 1961. Examples of both cases are shown in Fig. 16, and the resulting decrease in swamp and floodplain area is shown in Table 1 1 . Under the best possible conditions, that is with river discharge levels remaining as high as they arc now (in the 1980’s-90’s), the effects would be as follows: 1. The small lakes within the permanent wetlands will become shallower and susceptible to encroachment by vegetation. These lakes in the southern area of the Sudd will still remain perma- nent, however in the eastern area many will shrink in size and many will disappear; 2 Because of the decrease in turbidity of incoming water and the lower water depths that will occur, the growth of submerged vegetation will be encouraged along the banks and on the bed of the rivers flowing through the region. This will increase the retention of nutrients, but it will also encourage the growth of floating aquatic vegetation, such as water hyacinth ( E . crassipes ) and Vossia grass (V. cuspidata ); 3. The aquatic grass Vossia will increase considerably because of its tolerance to deep-flooding sites. It will overgrow open water in those areas subject to falling water levels. This will create blockages in the main channels of the Sudd. Such blockages do not occur at present because steamer traffic keeps the channels open. With the diversion of this traffic into the canal the block- ages by Vossia would be significant; 4. The area covered by Papyrus and Typha would be reduced because of contraction in their habitats, but as these permanent swamp types recede, their place will be taken by Oryza. The seasonally-flooded grasslands will also retreat and their place will be taken by grasses belonging to the genera Sporobolus and Hyparrhenia\ 5. There would be little effect on present day fishing grounds and fish stocks, because at present these are very much underutilized. With the increased growth of floating vegetation, channel blockage will occur and this will cause interference in the access of barges to the fish collecting centres. It will also cause interference to the access of canoes into some of the fishing grounds. These effects can be mitigated as the fishing resources will still be more than adequate. The canal may even improve fishing effort because of the increased communication and travel to markets outside the affected region. All of these effects may become more adverse if the effect of the canal is coupled with a fall in river discharge. If discharge levels decrease to those observed prior to 1961 , the floodplain area will be so reduced (Fig. 16D) that many of the floodplain lakes would be lost and thus fish populations would decline. In that case significant socio-economic problems would certainly arise. Another case in Africa is that of the Kafue flats (Fig. 17). Prior to the construction of the Kafue Gorge dam, the Kafue flats showed a classic floodplain profile (Fig. 17) and function. They had a gentle slope progressing from west to east with a loss of only 6 m elevation over 193 km. The soils in the flats are impervious clays. The Kafue river channel meanders over the fiats and divides into occasional branches. J. EANHS & Nat. Mus. 82 (199) March 1992 John J. Gaudet • Structure and funcuon of African floodplains 26 a) Swamp and floodplain as of 1980 b) Canal operation if high river discharge level were maintained c) Swamp and floodplain as it existed in 1952 d) Canal operation if river returned to pre 1961 discharges Swamp :HU| River flooded grassland □ All other vegetation types Figure 16. Comparison of swamp and floodplain during high water conditions on the Nile in 1980 and low water conditions (1953). The projected effects of the Jonglai Canal are shown (after Anon, 1983). J. EANHS & Nat. Mus. 82 (199) March 1992 27 John J. Gaudet • Structure and function of African floodplains Table 11: Estimated effects of the Jonglei Canal on areas of flooding in the Sudd region (after Anon. 1983). When in operation the canal will divert 20 million m3 per day. Data from Mefit-Babtie study (1983) swamp km2 area of decrease % floodplain km2 area of decrease % No canal, estimated areas in 1980 16,357 — 15,476 Operational canal, river remains at high discharge level 12,251 25.1 13,239 14.5 Operational canal, river returns to pre-1961 discharge level 3,688 77.5 7,624 50.7 The main river is bordered by natural levees so that when the river is running full, the water in the channel may be a metre higher than the surrounding land. Eventual overtopping of levees, flooding, subsequent natural drainage and drying completed the annual cycle (White, 1973). The aquatic vegetation at Kafue flats is found within five general communities: a) open water bodies up to 8m deep; b) lagoons and river edges flooding to 4-6 m supporting an association of V. cuspidaia. E. pyramidalis, E. scabra, Leersia hexandra and C. papyrus', c) flooded grassland, 1.5-6m depth dominated by O. longistaminata. d) shallow flooded grassland and levees with flooding depths of 0.25-25cm. Here tussock floodplain grasses occur such as Vetivaria nigritana and Seiaria avetiae', and finally e) floodplain marginal regions supporting a grass cover of Hyparrhenia rufa, Panicum coloratum, V. nigritana and Setaria sphacelala. Two dams were planned in order to achieve control over the Kafue river (Fig. 17). The first, Kafue Gorge dam would dam the river below the flats and would impound enough water to allow hydropower production during the subsequent four years until a second dam could be erected upstream above the flats (Fig. 17). This dam (the Itezhitezhi dam) would provide enough control over the whole river and floodplain system so that annual flooding-drying cycles could still be achieved within the floodplain. However, there would be a four year period between construction of the dams, during which time 2,000 km of floodplain would be inundated. This region to be inundated supported an Oryza floodplain grassland, where the principal species normally produced enough new growth each year to keep pace with the rising flood. The tips of the leaves and the inflorescence of Oryza must be maintained above the water level or it will die back. White (1973) predicted that four years of inundation would distrupt the normal cycle and that over such a long period death and decomposition of Oryza would result. He also cautioned that the open water expanse would be taken over by sudd-forming species with the result that the floodplain species would later require a very long period to be re-established. White (1973) recommended that as much of the annual flooding-drying cycle of the floodplains be maintained as possible over the 4-year period. Subsequently, the first dam was closed in 1972 and it was decided to maintain the floodplain cycles as much as possible. However,during the dry season in 1973 the Kafue River receded to one of its lowest recorded levels. The floodplain drained except for large lagoons, and the whole region was burned. Because of the expected increase in demand for water following this drought, and the need to have a larger storage for the subsequent year, it was decided to store more water. Thus, following the 1973-74 rains, the storage level was allowed to rise 1.2 m above normal. As a result 2,200 km of the floodplain was flooded (Magadza, 1977). The Oryza grass died off and a rapid growth occurred of sudd-forming species such as V. cuspidata in shallow water, and Aeschynomene elaphroxylon in open water. J. EANHS & NaL Mus. 82 (199) March 1992 John J. Gaudet • Structure and function of African floodplains 28 2. Itezhitezhi dam (after White, 1973) The second dam was finally closed in 1977, with the result that the flooded area became smaller during the peak wet season, but during the dry season the inundated region remained permanently flooded (Werger & Ellenbroek, 1980). These effects were not predicted by the early models of the dam operations (White, 1973). In addition, the second dam developed structural faults due to geological activity and subsequently had to be drained. At present over 40% of the original Kafue flats floodplain still remains inundated with little hope of any change in the future. Further downstream the Kafue enters the Zambezi which then flows east to the Indian Ocean. The Zambezi (at this point the mid-Zambezi) is dammed at Cabora Bassa in Mozambique and at Kariba between Zimbabwe and Zambia. Along a 100km stretch of the river between the two dams the Zambezi floodplain lies inside the Mana Pools Game Reserve (Fig. 18). In this floodplain (as in the Kafue flats) one typically finds Oryza grassland in the depressions and A. albida trees along the levees (Guy, 1977). This floodplain is of great importance to local wildlife during the dry season, it is the main grazing area of the mid-Zambezi Valley (Attwell, 1970). The whole floodplain region was earlier dramatically affected by the construction and operation of the Kariba dam in 1958. Since that time: a) lesser amounts of silt have been deposited (most being trapped by the dam); b) lower level of flooding prevails resulting in a lack of natural flooding; and c) out-of-season low volume floods are common. The terrestrial vegetation within the reserve is undergoing changes related to the changes in animal movement which in turn are caused by the changes in floodplain function. In the floodplain, even though there is still inadequate flooding, there is some evidence of colonization by A. albida on sandbanks and levees and regrowth of floodplain grasses, e.g. Oryza and Setaria species in the depressions. In other words, the floodplain vegetation is being re-established, but with much difficulty (Attwell, 1970). It has recently been proposed that a third dam be built mid-way between Kariba and Cabora Bassa, at Mpata Gorge (Fig. 18). This dam would inundate all of the riverine stands of A. albida and the Oryza grasslands mentioned above. In other words, it would eliminate the floodplain in this part of the mid-Zambezi (Muller & Pope 1982). J. EANHS & Nat. Mus. 82 (199) March 1992 29 John J. Gaudet • Structure and function of African floodplains Figure 18. Zambezi floodplain downstream from Lake Kariba showing the Mupata Gorge (after Atwell, 1970) The question now arises of whether or not any of the floodplains along the Kafue and Zambezi rivers will survive such development. Are there any development alternatives? In the case of the Kafue flats, even if the Itezhitezhi dam is put back into operation and even if it is operated in tandem with the Kafue Gorge dam, the floodplains may still not survive. The floodplain here needs an annual, ‘close-to normal’, flood-dry cycle. If this is to be achieved the two dams must be closely regulated in a careful, regular fashion which based on past experience, is obviously difficult to achieve. In the case of the mid-Zambezi floodplains, there is an alternative to the Mpata dam, that is to utilize another site located in a gorge on the Upper Zambezi, 35 km below Victoria Falls. The reservoir created by this dam (Batoka Gorge dam) would be confined within the deep gorge and would have minimal impact on any existing floodplain and on Victoria Falls. Based on past experience, it would seem that the choice of this site would be a much better alternative than those management schemes that could result in the complete elimination of a valuable natural resource, the Mana Pools floodplain. Manuscript first received: 1986 Re-submitted: 5 April 1991 Editor: C.F. Dewhurst J. EANHS & Nat Mus. 82 (199) March 1992 John J. Gaudet • Structure and function of African floodplains 30 REFERENCES Allanson, B.R. (1961) Investigations into the ecology of polluted inland waters in the Transvaal. Part 1. Hydrobiologia 18:1-76. Anonymous, (1983) Development studies in the Jonglei Canal area. Four vols. Mefit-Bebtie Glasgow, Khartoum and Rome. Attwell, R. (1970) Some effects of Lake Kariba on the ecology of a flood plain of the mid- Zamhezi valley of Zimbabwe. Biological Conservation 2:189-196. Beadle, L.C. (1974) The Inland Waters of Tropical Africa. Longman. London. Bishai, H. (1962) The water characteristics of the Nile in the Sudan with a note on the effect of Eichornia crassipes on the hydrobiology of the Nile. Hydrobiology. 19:357-382. Bowmaker, A.P. (1973) Hydrophyte dynamics in Mwenda Bay. Lake Kariba. Kariba Studies. 3:42-59. Bristow, J.H. (1974) Nitrogen Fixation in the rhizosphere of freshwater angiosperms. Canadian Journal of Botany. 52:217-221. Carmouze, J.P., Durand, J.R. & Leveque, C. (1983) Ecology and productivity of a shallow tropical ecosystem. The Hague. Dr. W. Junk Publ Chachu, R. (1979) The vascular flora of Lake Kainji. Proceedings of the International Conference Kainji Lake 2:479-487. Cook, C.D.K. (1968) The vegetation of the Kainji Reservoir site in northern Nigeria. Vegetation 15:225-243. Douthwaite, RJ. & Van Lavieren, L.P. (1977) A description of the vegetation of Lochinvar National Park, Zambia. Technical Report No. 34, National Council Scientific Research, Lusaka. El-Hadidi, M.N. (1976) The riverain flora in Nubia. In: Rzoska, J. (ed.) The Nile, Biography of an Ancient River. The Hague. Dr. W. Junk Publ. Feely, J.M. (1965) Observations on Acacia albida in the Luangwa Valley. Puku 3:67-70. Gaudet, J J. (1977) Natural drawdown on Lake Naivasha, Kenya and the formation of papyrus swamps. Aquatic Botany 3:1-47. (1978) Effect of a tropical swamp on water quality. Vehandlungen International Vereinigung fur Limnologie 20:2202-2206. (1979) Seasonal changes in nutrients in tropical swamp water. Journal of Ecology 67:953-981. Germain, R. (1952) Les associations vegetales de la plaine de la Ruzizi (Congo Beige) en relations avec le milieu. Inst. Nat. pour L’etude Agron. du Congo Beige (INEAC), ser. Sci. 52, Bruxelles. Guy, P.R. (1977) Notes on the vegetation types of the Zambezi Valley, Zimbabwe, between Kariba and Mpata Gorges. Kirkia 10:543-557. Hall, J.B., Pierce, M.P. & Lawson, G. (1971) Common plants of the Volta Lake. Publication of the Botanical Department, University of Ghana, Legon. Harrison, A. (1964) An ecological survey of the Great Berg River. In: D. Davis (ed.), Ecological Studies in Southern Africa. The Hague. Dr. W. Junk Publ. Howard- Williams, C. (1980) Distribution, biomass and role of aquatic macrophytes in Lake Sibaya. In: B. Allanson (ed.), Lake Sibaya. The Hague. Dr. W. Junk Publ. Howard- Williams, C. & Gaudet, JJ. (1985) The structure and functioning of African swamps. In: P. Denny (ed.). The Ecology and Management of African Wetland Vegetation. Dr. W. Junk Publ., Dordrecht. Howard- Williams, C. & Lenton, G. (1975) The role of the littoral zone in the functioning of a shallow tropical lake ecosystem. Freshwater Biology 5:445-459. Howard-Wiluams, C. & Walke, B.H. (1974) The vegetation of a tropical lake: classification and ordination of the vegetation of Lake Chilwa (Malawi). Journal of Ecology 62:831-854. Iltis, A. and Lemoalle, F. (1979) La vegetation aquatique du Lac Tchad. Special Publication, Reunion du Travail sur la Limnologie Africaine, Nairobi. ORSTOM, Paris. J. EANHS & Nat. Mus. 82 (199) March 1992 31 John J. Gaudet • Structure and function of African floodplains Imhof, G. and Burian, K. (1972) Energy flow studies in a wetland ecosystem -reed belt of the Lake Neusiedler Sea. IBP Special Publication, Austrian Acadamey of Science Jackson, P. & Davies, B. (1976) Cabora Bassa in its first year. Some ecological aspects and comparisons. Rhodesian Science News 10:128-133. Junk, W.J. (1970) Investigations on the ecology and production biology of the foating meadows (. Paspalum echinochloetum ) on the middle Amazon. 1. The floating vegetation and its ecology. Amazonia 2:449-495. Kjlham, P. (1972) Biogeochemistry of African lakes and rivers. Unpublished Ph.D. thesis, Duke University Durham, N.C. Koechlin, J. (1961) La vegetation des savannes dans le Sud de la R6publique du Congo. Institute de R6cherches Scientifique au Congo, Brazzaville. ORSTOM, Montpelier. Lawson, G. (1963) Preliminary report on the aquatic weeds in the Volta Basin. Volta Basin Re- search Project Technical Paper No. 6. University of Ghana. Legon. Lebrun, J. (1947) La vegetation de la Plaine Alluviale au Sud du Lac Edouard, 2 Vol. Institute des Parcs National du Congo. Beige, Bruxelles. Leonard, J. (1969) Aper^u sur la vegetation aquatique. In: Monographic Hydrologique du Lac Tchad. ORSTOM. N’Djamena. Lieth, H. (1972) Modeling the primary productivity of the world. INTECOL Bulletin 4:11-20. Lind, E. & Morrison, M.E.S. (1974) East African Vegetation. London. Longman. Lock, J.M. (1973) The aquatic vegetation of Lake George, Uganda. Phytocoenologia 1:250-263. Magadza, C. (1970) A preliminary survey of the vegetation of the shore of Lake Kariba. Kirkia 7: 253-267. Muller, T. & Pope, G. (1982) Vegetation report. Impact Assessment of the Proposed Mupata George Dam. Botany Department, University of Zimbabwe, Harare, Zimbabwe. Musil, C.F., Grunow, J.O. & Bornman, C.H. (1973) Classification and ordination of aquatic macrophytes in the Pongolo River pans, Natal. Bothalia 11: 181-190. Ouff, W. (1959) Hydrobiological studies on the Tugela River system. Part 1. The main Tugela River. Hydrobiologia 14: 281-385. Reavell, P. (1979) Limnological characteristics of the Okavango Delta and the effect of limiting factors on the phytoplankton biomass during 1972-1974 Abstracts of poster papers, S1L-UNEP Workshop on Limnology in Africa. Nairobi, Kenya. Reijks, D. (1969) Evaporation from a papyrus swamp. Quarterly Journal of the Royal Meteorologi- cal Society 95: 643-649. Reizer, C. (1974) Definition d'une politique d' amenagement des resources halieutiques d’un ecosysteme aquatique complexe par l' etude de son environment abiotique, biolique et anthropique. Le Fleuve Senegal Moyen et Inferieur. D.Sc. thesis, Fondation Universitaire Luxembougeoise, Arlon. Rzoskz, J. (1976) The Upper Nile swamps: a tropical wetland study. Freshwater Biology 4:1-30. Schmitz, A. (1971) La vegetation de la plaine de Lubumbashi (Haut-Katanga). Publ. INEAC Series Scientifique 1 13: 1-390. Smith, P.A. (1976) An outline of the vegetation of the Okavango drainage system. InrProceedings of the Okavango Delta Symposium. Botswana Society, Gaborone. Sutcliffe, J. (1957) The hydrology of the Sudd region of the upper Nile. D. Phil, thesis, Cambridge University, U.K. (1976) Hydrological study of the southern Sudan region of the upper Nile. Hydrological Sciences Bulletin reprinted in: Proceedings of the Okavango Delta Symposium. Botswana Society, Gaborone. Taixing, J.F. (1957) The longitudinal succession of the water characteristics in the White Nile. Hydrobiology 11:73-89. J. EANHS & Nat. Mus. 82 (199) March 1992 John J. Gaudet • Structure and function of African floodplains 32 Thompson, K. (1974) Environmental factors and aquatic plants in the Okavango Delta. Technical Report Project BOT/506. FAO, Rome. Thompson, K. (1976) The primary productivity of African wetlands with particular reference to the Okavango delta. In: Proceedings Okavango Delta Symposium. Botswana Society, Gaborone. Thompson, K. (1985) Emergent plants of permanent and seasonally-flooded wetlands. In: P. Denny (ed.). The Ecology and Management of African Wetland Vegetation. Dr. W. Junk Publ., Dordrecht. Thompson, K„ Shewry, P. & Woolhous, H. (1979) Papyrus swamp development in the Upemba Basin, Zaire: studies of population structure in Cyperus papyrus stands. Botanical Journal of the Linnaean Society 78: 299-316. Thompson, K., Howard-Williams, C. and Mitchell, D. (1985) A cross-indexed bibliography of African wetlands plants and vegetation. In: P. Denny (ed.). The Ecology and Management of African Wetland Vegetation. Dr. W. Junk Publ., Dordrecht. Van der Ben, D. (1959) La vdg<5tation des rives des Lacs Kivu, Edourd et Albert. In: Exploration Hydrobiologique des Lacs Kivu, Edouard et Albert (1952-1954). Institute Royale du Science et Nature de Belgique, Bruxelles Van Resnberg, J. 1971. The ecology of the Okavango Delta. FAO Technical Notes, No. 19, Rome. Verboom, W. (1975) List of plant species in the main vegetation types of the Bangweulu Basin. In: Grimsdall and Bell (eds.). Black Lechwe Research Project Final Report. Report AR1-NCSR/ TR31. Animal Production Research Lusaka, Zambia. Vesey-Fitzgerald, D.F. (1973) East African Grasslands. Nairobi, Kenya. East African Publishing House. Werger, MJ.A. & Ellenbroek, G.A. (1980) Water resource management and floodplain ecology: an example from Zambia In: J.l. Furtado (ed.), Tropical Ecology and Development Interna- tional Soc. Ecol., Kuala Lumpur. Welcomme, R.L. (1979) Fisheries Ecology of Floodplain Rivers. Longman, London. Welsh, R.P.H. & Denny, P. (1978) The vegetation of Nyumba ya Mungu reservoir, Tanzania. Biological Journal of the Linnaean Society. 10: 67-92. Westlake, D.F. (1975) Primary production of fresh water macrophytes.ln:Cooper (ed.) Photosyn- thesis and Productivity in Different Environments. Cambridge Univ. Press, Cambridge. White, E. (1973) Zambia’s Kafue hydroelectric scheme and its biological problems. In: Ackerman (ed.). Man-made Lakes. Their Problems and Environmental Effects. Geophysical Monographs Series 17, U.S. Geophysical Union, Washington, D.C. J. EANHS & Nat. Mus. 82 (199) March 1992 NOTICE TO CONTRIBUTORS Journal of the East Africa Natural History Society and National Museum Papers on Natural History subjects will be considered on the understanding that they are being offered exclusively to the Society and have not been submitted, at the same time, or previously, for publication elsewhere. Copyright is retained by the East Africa Natural History Society; references to contributions in the Journal may be made, but no part of the text, diagram, figure, table or plate may be reproduced without permission from the Editor. Such permission will only be granted with the approval of the author(s). Three copies of papers for consideration should be sent to The Editor, Journal of the East Africa Natural History Society and National Museum, P.O. Box 44486, Nairobi, Kenya, and should be accompanied by a covering letter from the author, or in the case of multiple authors, from the author responsible for decisions regarding the text. Contributions may be sent as: preferably a floppy disk of either 5.25 or 3.5 inches in fully IBM compatible MS-DOS, carrying text processed in a commonly used wordprocessing language, or a floppy disk with the text prepared on an Apple Macintosh. If in any doubt, the author(s) should contact the Editor at the above address for further details. Three “near letter quality” printed copies on A4 size paper double-spaced should accompany the diskette. The author(s) should, of course, retain a back-up copy. or a typescript, which should be on A4 size paper, typed on one side only, double spaced with margins of at least 5 cm on the left side of the paper to allow for editorial instructions to the printer. Title Should be brief and to the point. Authors name(s) and institution(s) where the work was carried out should be given, together with a suggested running page headline of not more than 50 characters including the names of author(s). A current address for correspondence, if applicable, should be included as a footnote. Abstract To be on the first page and consist of not more than 150 words. This should outline the major findings clearly and concisely without requiring reference to the text. Headings Should be brief, and in the following format: “Abstract, Introduction, Methods, Results, Discussion, Acknowledgements and References”. All headings to be typed on a separate line from subsequent text. Main headings to be typed in capital letters and centred on the page. Sub-headings to be at the left of the page, in lower case; further headings should be in italics or underlined (not bold) beneath them. The first letter of any headings should begin with a capital letter. Text Indent each new paragraph except those after a heading. Spelling should follow that of the Oxford English Dictionary. All pages must be numbered, at the top nght hand edge. Numerals Prose numbers should be written in full up to and including ten, and as numerals dates times and thereafter, except at the end dates & and beginning of a sentence. Dates should be scientific units written as “29 August”, “10-13 January 1989” or “1969-1989”. Times should follow the 24 hour clock (e.g. 1500 h). All measurements must be in metric units. SI units and abbreviations should be used throughout with a space before and after the SI unit. Full stops should not be used after such units. The approximate position of Tables and Figures should be indicated in the left margin in pencil Figures The originals and two photocopies must be provided with each manuscript. Figures should be well drawn on pure white cartridge paper or good quality tracing paper. Lettering must be of very good quality (e.g. Letraset or good stencil). Typewritten or freehand lettering will not be accepted (without good reason). Each figure should have the authors’ name, figure number and caption details written lightly on the reverse in pencil. Cite each figure in the text in consecutive order using arabic numerals. “Figure” should be abbreviated to “Fig.”, except in the figure legends and at the beginning of sentences. Two separate caption sheets should accompany the typescript. Black and white photographs are acceptable if they have good definition and contrast and are produced on glossy paper. Consideration must be given by the authorfs) to their final size when reduced for publication in the Journal, which is B5. Tables Each should be typed, double-spaced on a separate page and clearly numbered. Tables should be simple and self explanatory, and have a brief title above each one. They must be numbered consecutively with arabic numerals. Nomenclature In cases where there are well known and internationally accepted English names, these should be used throughout the text. The scientific name should be given in full, under- lined or preferably in italics, when quoted for the first time. The authority should also be quoted at this stage except in non-taxonomic ornithological papers. Only the generic name should begin with a capital letter. In subsequent reference to the name, use only the abbreviated generic name, with the full specific name but not the authority, e.g. Spodoptera exempta (Walker); subsequently S. exempta. The authority should not be given if scientific names are used to describe an “association” or species complex, e.g. Acacia drepanolobium - Theneda triandra , wooded grassland. Type in capitals the first letter of the English names of species (e.g. Crowned Eagle, Grey -capped Warbler) but not of the higher taxa (e.g. eagles, warblers) . References Care should be taken to see that all references quoted in the text are also given in full, alphabetical order, in this section. To avoid confusion, references to books should not abbreviated and neither should they be underlined nor in italics. Journal names should be written in full, underlined or, if possible in italics. Personal communications should not be cited in the references, neither should they be abbreviated in the text. If there are two or more references cited by an author which were published in the same year, distinguish them by adding letters to the dates, e.g. Brown (1972a). For technical reasons author names in the reference list should NOT be typed in upper case. Authors are advised to refer to this issue of the Journal for the preferred format. Acknowledgments As acknowledgements may imply endorsement of the data and conclusions in the authors’ work, it is advisable to obtain permission before quoting persons who have made contributions to the study. Once the paper has been selected by the Editor for possible publication in the Journal, the manuscript will be refereed by two persons who are acknowledged specialists in that field, before a decision is made on its acceptance. The author(s) will be sent one set of printouts for the correction of typographical errors. Authors are urged to ensure that all changes to the text are incorporated into this copy, as no further changes can then be made. The author is entitled to 25 copies free of charge. Thirty free copies will be provided if there is more than one contributor. Further copies can be supplied at cost, provided the order is placed before printing has been undertaken. Referees Proofs Offprints The East Africa Natural History Society is supported by membership subscriptions, and may be unable in some instances, due to financial constraints, to accept all submissions of value. While every effort will be made to minimise such restrictions, contributors are requested to seek Financial support to assist the Society with the ever increasing costs of publication. Editor Journal East Africa Natural History Society and National Museum. S' 5’ EDEN WILDLIFE TRUST Charitable Trust Registered No. 278008 The East Africa Natural History Society is most grateful to the Eden Wildlife Trust for their very generous contribution towards the cost of printing this journal part Published by: The East Africa Natural History Society • Typesetting by Development Communications Ltd • Printed by AMREF JOURNAL. OF THE EAST AFRICA NATURAL HISTORY SOCIETY AND NATIONAL MUSEUM June 1992 Volume 82 No 200 A FLIGHTLESS GRASSHOPPER OF THE GENUS UGANDA BOLIVAR, FROM THE ABERDARES NATIONAL PARK, KENYA (ORTHOPTERA, ACRIDIDAE) J. MARK RITCHIE International Institute of Entomology, 56 Queen’s Gate, London, SW7 5JR, UK. ABSTRACT A new species of the genus Uganda, U. darlingtonae, is described from moorland above 3000 m in the Aber dares National Park, Kenya. The habitat of the species is described and morphological characters distinguishing U. darlingtonae from the other montane species of the genus are described and figured with comparative measurements. INTRODUCTION The montane grasshopper fauna of Africa contains a number of short- winged species which are usually recognisably derived from lowland macropterous species, as for example in such genera as Coryphosima and Eyprepocnemis. Uvarov (1977) commented that Paracinema was replaced at higher altitudes by the closely related brachypterous genus Uganda. The genus Paracinema was last revised by Key (1936) who recognised two species, Paracinema luculenta Karsch and P. tricolor (Thunberg). P. luculenta occurs in West and Central Africa (Togo, Ghana, Sierra Leone, Zaire) while P. tricolor occurs throughout the wetter parts of Africa and into Madagascar and southern Europe. In East Africa, P. tricolor occurs as subspecies montana Key in Ethiopia (Wouramboulchi, nr. Djem Djem), at around 2800 metres (Key 1936). Specimens of the nominate subspecies have been found at around 2400 metres in the Impenetrable (Bwindi) Forest, SW Uganda (Ritchie, unpublished) and material intermediate between the two subspecies was recorded from the same altitude in Ethiopia (Key 1936). Both species of Paracinema are macropterous although in P.t. montana the tegmina and wings only just exceed the hind knees in length. The genus Uganda was described by Bolivar (1909) without any included species and was later placed by its author in the group Paracinemae (Bolivar 1914). He cited two included species of which one, the type species, U. kilimandjarica (SjOstedt), occurs at between 2300 and 4300 m on Mt. Kilimanjaro in Tanzania where the syntype series was collected at Kiboscho at 3000 m (Sjbstedt 1909). The other species, U. acutipennis Bolivar, was apparently collected at low altitude J. Mark Ritchie • A flightless grasshopper from the Aberdare Range, Kenya 34 on the Sesse Islands, Lake Victoria, Uganda (Bolivar 1914). It is not mentioned again in the literature after its original description by Bolivar. Since this species has not been recognised subsequently in collections from lowland southern Uganda, there must be some suspicion that it is not a member of the genus Uganda but a brachypterous representative of another acridine genus, such as Gymnobothrus or Gymnobothroides. The location of the type specimens is unknown to me, and no material referable to this species has been examined during this study. The new species of Uganda described here was first brought to my attention by Dr Johanna Darlington, and is named in her honour. Most of the specimens were collected from the vicinity of Muir’s Massif (0°17’S 36°37’E) at the northern end of the Aberdares Range, Kenya at altitudes ranging from 3250 to 3650 metres. Subsequently a single female was found to have been collected in 1934 at 3700 m on Ml Kinangop at the southern end of the Aberdare Range. RESULTS Key to montane species of Uganda Bolivar 1. Larger species (Table 1). Total length: male, 11.5 - 13.8mm (mean 12.99mm); female, 22.8 - 27.5mm (mean 24.32mm). Dorsum of pronotum with pale, raised, shiny lateral carinae, visible in metazona as well as prozona (Fig. 1). General colouration dark brown, with variable degree of green colour in some specimens; hind femur lower outer and inner areas and hind tibia distinctly orange-red. (Kilimanjaro) U. kilimandjarica (SjOstedt) 2. Smaller species (Table 2). Total length: male, 10.1 - 11.3mm (mean 10.6mm); female, 18.1 - 20.9mm (mean 19.4mm). Dorsum of pronotum with pale, raised, shiny lateral carinae visible in prozona only (Fig. 2). General colouration greenish or straw; hind femur inner area and hind tibia sometimes faintly tinged with orange-red but lower outer area of femur never orange-red. (Aberdares) U. darlingtonae sp. n. Table 1. Measurements (mm) of adult material of Uganda kilimandjarica (Sjdstedt) from Mt. Kilimanjaro, Tanzania Total length Head width* Pronotum length Tegmen length Femur length Femur length Male Mean 12.99 2.63 2.84 3.84 9.08 2.2 Range 11.55-13.8 2.4-2.9 2.7-3. 1 3. 4-4 .2 8.6-9.9 2.15-2.25 Number measured 7 7 7 7 5 5 Female Mean 24.32 4.62 4.69 5.89 14.28 3.29 Range 22.8-27.45 4.3-5.35 4.15-5.6 5.0-7.8 12.35-17.5 3.05-3.7 Number measured 9 7 9 9 8 8 * across genae J. EANHS & Nat. Mus. 82 (200) June 1992 35 J. Mark Ritchie • A flightless grasshopper from the Aberdare Range, Kenya 2mm Figure 1. Uganda kilimandjarica (Sjostedt) Figure 2. U. darlingtonae sp.n. head and pronotum, male, dorsal view J. EANHS & Nat Mus. 82 (200) June 1992 J. Marie Ritchie • A flightless grasshopper from the Aberdare Range, Kenya 36 Table 2. Measurements (mm) of all known adult material of Uganda darlingtonae sp. n. Total length Head width* Pronotum length Tegmen length Femur length Femur length Male Mean 10.57 2.19 2.39 3.39 8.37 1.84 Range 10.1-11.25 2.05-2.25 2.3-2.5 3.2-3.5 8.1-8.8 1.75-1.95 Number measured 4 4 4 4 4 4 Female Mean 19.4 3.7 3.79 5.15 12.73 2.84 Range 18.15-20.9 3.55-3.85 3.65-3.9 4.4-5.65 12.55-12.95 2.75-2.95 Number measured 5 5 5 5 5 5 * across genae Description of new species Uganda darlingtonae sp. n. Male. Small species (total length 10.1 - 11.3mm) (Table 2.). Antenna with 18 segments, distinctly shorter than length of head and pronotum combined (longer in Paracinema species). Interocular distance little more than one third as long as long axis of eye. Frons oblique, curved; frontal ridge constricted at apex to half its basal width, slightly sulcate at median ocellus, smoothly convex above, meeting vertex in acutely rounded curve; fastigium of vertex lanceolate, angular in front, rounded behind, forming a shallow depression with rounded margins; fastigial foveolae absent. Pronotum weakly tectiform with obtuse median carina, crossed by posterior transverse sulcus; lateral carinae straight in prozona, diverging forwards, converging and obsolescent before posterior sulcus; metazona with obtuse excurved outer lateral carinae bordered internally by dark fascia (Fig. 2); metazona slightly shorter than prozona, with rounded posterior margin. Mesostemal interspace rectangular, distinctly wider than long (longer than wide in Paracinema species). Elytra and wings reduced; elytra lateral, lanceolate, touching or slightly overlapping dorsally, with reduced venation and reticulation (Fig. 3), shiny, reaching to middle of third abdominal tergite; tympanum well developed, covered by elytron. Hind femur about 4.S times as long as maximum depth; lower lobes of hind knee rounded; hind tibia expanded in apical third, with 9 outer and 10 inner spines; tibial spurs unspecialised, arolium of normal size. Supra-anal plate spade-like, apically rounded, barely longer than wide; subgenital plate subconical with subacute apex. Cercus elongate, finger-like. Epiphallus variable (Figs. 4 & 5), of typical acridine form, subrectangular to rhomboidal, with widely-spaced ancorae and bilobate lophi; aedeagus slender and unspecialised, with ventral sub-apical lobe (absent in P. tricolor but present in U. kilimandjarica and P. luculenta). J. EANHS & Nat. Mus. 82 (200) June 1992 37 J. Mark Ritchie • A flightless grasshopper from the Aberdare Range, Kenya 2mm Figure 3. U. darlingtonae sp. n. thorax and left legmen, lateral view Figure 4. Figure 5. t . A 1 mm Figures 4 and 5: U. darlingtonae sp. n. epiphallus, dorsal view, showing the range of variability J. EANHS & Nat Mus. 82 (200) June 1992 J. Mark Ritchie • A flightless grasshopper from the Aberdare Range, Kenya 38 Colouration General colouration pale green fading to straw-coloured after preservation; antennae reddish- brown; head behind eyes and pronotum with longitudinal dark chocolate brown lateral fasciae (Fig. 2) continuing on elytra and abdominal tergites; lateral surfaces of pronotum, meso- and metanotum with indistinct dark markings; hind femur with dark longitudinal fascia along upper half of external medial area, fading to ground colour in lower half; internal ventral surface straw- coloured, sometimes faintly tinged with pale red; hind knee lunules dark brown; tibiae and tarsi straw-coloured with or without faint pale orange red tinge, spines black tipped. Female Similar to male but much larger and stockier (total length 18.1 - 20.9mm). Hind femur internal ventral surface and hind tibia distinctly flushed with pale orange-red. Ovipositor short with robust valves curving at apices. Measurements See Table 2. Material examined KENYA: Holotype male, 3 male, 4 female paratypes, 1 female nymph, Aberdares National Park, 0°14’S 36°35’30"E, AlchemillalEleusine/AndropogonlAgrostis moorland, SE foot of Chebuswa, alt. 3250 m, 19.iii.1987, (JM. Ritchie) (1 male paratype in British Museum (Natural History) (BMNH), London, remainder in National Museums Kenya (NMK), Nairobi); 1 female paratype, Aberdares Nat. He, north end, Muir’s Massif, 0°17’S 36°37’E, alpine moorland, alt 3650 m, 7.ix.l985 ( JF.E.C . Darlington) (NMK); 1 female paratype, Aberdare Range, Mt. Kinangop, 12000 ft, 30.ix.1934, (F.W. Edwards) (BMNH). DISCUSSION The pronotal morphology of the new species allies it with P. tricolor , especially in the presence of anterior dorso-lateral carinae which reach only as far as the first transverse sulcus (Fig. 2). In P. luculenta there arc no such carinae and in Uganda kHimandjarica the carinae continue posteriorly beyond the first sulcus (Fig. 1). The distinctive pair of dark longitudinal dorsal bands on the pronotum also allies U. darlingtonae to P. tricolor. However, the genitalia and general appearance of U. darlingtonae indicate a very close relationship with U. Idlimandjarica. The epiphallus and aedeagus of the two taxa are not readily distinguishable, suggesting a recent com- mon ancestry, perhaps during a period of climatic amelioration during the Pleistocene (<1 million years before present) which could have enlarged the available area of moorland habitat sufficiendy to permit exchange of populations between Mt Kilimanjaro and the Aberdare Range nearly 300 km apart The montane moorland habitat The Aberdares Range, together with Mt Kenya, 50km to the east, form the eastern highlands of Kenya separated from the Western highlands by the Gregory Rift Valley. The Aberdares are the highest mountain range in Kenya after Mt Kenya and Mt Elgon. The climate at above 3000m is extreme, with frequent night frosts. For much of the day the sun is concealed by cloud and rain falls almost daily. The temperature regime during March and April in a similar habitat on the west side of Mt Kenya was examined by Beck et al. (1981) who found that in an average period of J. EANHS & NaL Mus. 82 (200) June 1992 39 J. Mark Ritchie • A flightless grasshopper from the Aberdare Range, Kenya 24 hours, the air temperature at ground level was below 0°C for more than 6 hours and between 0° and 5° for a further 6 hours. Overall temperatures varied between -5°C and 25°C. At a depth of 20 mm below the soil surface the low temperatures were less severe, seldom falling below 0°C but averaging more than 12 hours between 0° and 5°C. The vegetation on Muir’s Massif is shown in Plate 1. This is typical of the habitat where U. darlingtonae was collected (3650 m) but with less bare ground and thicker vegetation than the slightly lower collecting locality at the foot of Chebuswa (3250 m). The afromontane moorland of the northern Aberdare Range is an open habitat with dwarf shrubs, herbs and grasses. In the more open areas of true moorland at the foot of Chebuswa hill the flora is dominated by grasses, of which 10 - 20% cover is accounted for by tussocks of Eleusine jaegeri Pilg., 20 - 30% by Andropogon amethystinus SteudL and 10% by Eragrostis schweinfurthii Chiov. Festuca abyssinica A. Rich, occurs occasionally and the dwarf shrub Alchemilla argyrophylla Oliv. is also present Up to 20% of the area is bare soil as a result of the activities of mole rats ( Jachyoryctes splendens (Riippell). Forbs include Anagallis cf. serpens D.C., Trifolium cryptopodium A. Rich, (common), Anthemis tigrensis A. Rich., Geranium arabicum Forssk. (frequent), Gnaphalium luteo- album L. and Trifolium rueppellianum Fiesen. In this habitat U. darlingtonae was found but was less common than another flightless grasshopper, Coryphosima sp„ which is perhaps the most noticeable invertebrate at this altitude. Within the open moorland habitat, some denser patches of vegetation occur, with up to 75% cover of A. argyrophylla and 20% cover of a small tussock grass, Agrostis cf. gracilifolia C.E. Hubbard. Herbs and bare ground make up the remaining 5%, with Euphorbia brevicornu Pax. (common), Satureja kilimands chari (Guerke Hedb.) (frequent), and Polygonum qfromontanum Greenway. T. ruepellianum and young Hypericum revolutum Vahl ssp. keniense (Schweinf.) N. Robson as occasional. Here also U darlingtonae was present, but only emerged from the dense Agrostis tussocks and Alchemilla and became active when the sun had been shining for half an hour and the air temperature had begun to feel pleasantly warm (c.l5°C). The onset of rain towards midday caused the insects to disappear once again. Near Chebuswa Hill the moorland grades into a Hagenia woodland JStoebe bushland mosaic with scattered trees of Hagenia abyssinica (Bruce) J.F. Gmel. Here the common shrubs are Stoebe kilimandscharica O. Hoffm. var densiflora O. Hoffm., H. keniense and A. argyrophylla , P. afromontanum , Erlangea fusca S. Moore, Hebenstretia dentata L. and Clutia kilimandscharica Engl.* The herbs include at least nine species and there are four scramblers and climbers. U. darlingtonae was not found within this formation though it may well occur. In such an extreme habitat it is likely that a graminivorous grasshopper like Uganda would tend to spend most of its periods of activity in more open areas with grasses offering greater opportunities to bask and feed during the infrequent periods of bright sunshine. ACKNOWLEDGEMENTS I am indebted to Dr Henk Beentje for his company on a collecting trip to the northern Aberdares and for identifying the principal plant species occurring in the moorland habitat of P. darlingtonae. Thanks are also due to Dr Johanna Darlington who provided a photograph of the moorland below Muir’s Massif. I am grateful to the Director, Wildlife Conservation and Management Department, Ministry of Tourism and Wildlife, Kenya, for permission to work in the Aberdares National Park. Mr. Quentin Luke kindly provided the plant authorities. The type specimen has been destroyed. This is probably Clutia robusta Pax, but would be called C. kilimandscharica Engl, if it could be proved that they are synonomous. J. EANHS & Nat. Mus. 82 (200) June 1992 J. Mark Ritchie • A flightless grasshopper from the Aberdare Range, Kenya 40 J. EANHS & NaL Mus. 82 (200) June 1992 Plate. 6. Muir’s Massif, northern Aberdare Range, Kenya, looking north west towards Lake 01 Bolossat. Afroalpine moorland, altitude 3650 m. Photo JP.E.C. Darlington. 41 J. Marie Ritchie • A flightless grasshopper from the Aberdare Range, Kenya REFERENCES Beck E., Rehder, H., Pongratz R., Schrhbe R. and Senser M. (1981) Ecological analysis of the boundary between the afroalpine vegetation types ‘Dendrosenecio Woodlands’ and 1 Sene do brassica-Lobelia keniensis community’ on Ml Kenya. Journal of the East Africa Natural History Society and National Museum 172: 1-11 Bolivar I. (1909) Observaciones sobre los Truxalinos. Boletin de la Sociedad Espanola de Historia Natural , Madrid 9: 285 - 296 Bolivar I. (1914) Estudios entomologicos, segunda parte. Trabajos Museo Nacional de Ciencias Naturales. Madrid. (Serie Zoologica) no 20. 1 10 pp. Key K.H.L. (1936) A revision of the genus Paracinema Fisch. (Orthoptera). Transactions of the Royal Entomological Society of London 85: (16) 379 - 399. SjOsted Y. (1909) Wissenschaftliche Ergebnisse der schwedischen Zool. Exped. nach dem Kilimandjaro, dem Meru und den umgebenden Masai steppen deutch-Ostafrikas, 1905 - 1906. 17. Orthoptera. 7. Acridiodea: pp. 149 - 99. Uvarov B.P. (1977) Grasshoppers and locusts. Vol. II. London, Centre for Overseas Pest Research. 613 pp. Manuscript received: 2 January 1992 Editor: C.F. Dewhurst J. EANHS & Nat Mus. 82 (200) June 1992 NOTICE TO CONTRIBUTORS Journal of the East Africa Natural History Society and National Museum Papers on Natural History subjects will be considered on the understanding that they are being offered exclusively to the Society and have not been submitted, at the same time, or previously, for publication elsewhere. Copyright is retained by the East Africa Natural History Society; references to contributions in the Journal may be made, but no part of the text, diagram, figure, table or plate may be reproduced without permission from the Editor. Such permission will only be granted with the approval of the author(s). Three copies of papers for consideration should be sent to The Editor, Journal of the East Africa Natural History Society and National Museum, P.O. Box 44486, Nairobi, Kenya, and should be accompanied by a covering letter from the author, or in the case of multiple authors, from the author responsible for decisions regarding the text Contributions may be sent as: preferably a floppy disk of either 5.25 or 3.5 inches in fully IBM compatible MS-DOS, carrying text processed in a commonly used wordprocessing language, or a floppy disk with the text prepared on an Apple Macintosh. If in any doubt, the author(s) should contact the Editor at the above address for further details. Three “near letter quality” printed copies on A4 size paper double-spaced should accompany the diskette. The author(s) should, of course, retain a back-up copy. or a typescript, which should be on A4 size paper, typed on one side only, double spaced with margins of at least 5 cm on the left side of the paper to allow for editorial instructions to the printer. Title Should be brief and to the point Authors name(s) and institution(s) where the work was carried out should be given, together with a suggested running page headline of not more than 50 characters including the names of author(s). A current address for correspondence, if applicable, should be included as a footnote. Abstract To be on the fust page and consist of not more than 150 words. This should outline the major findings clearly and concisely without requiring reference to the text. Headings Should be brief, and in the following format: “Abstract, Introduction, Methods, Results, Discussion, Acknowledgements and References”. All headings to be typed on a separate line from subsequent text. Main headings to be typed in capital letters and centred on the page. Sub-headings to be at the left of the page, in lower case; further headings should be in italics or underlined (not bold) beneath than. The first letter of any headings should begin with a capital letter. Text Indent each new paragraph except those after a heading. Spelling should follow that of the Oxford English Dictionary. All pages must be numbered, at the top right hand edge. Numerals, Prose numbers should be written in full up to and including ten, and as numerals dates, times and thereafter, except at the end dates & and beginning of a sentence. Dates should be scientific units written as “29 August”, “10-13 January 1989” or “1969-1989”. Times should follow the 24 hour clock (e.g. 1500 h). All measurements must be in metric units. SI units and abbreviations should be used throughout with a space before and after the SI unit. Full stops should not be used after such units. The approximate position of Tables and Figures should be indicated in the left margin in pencil Figures The originals and two photocopies must be provided with each manuscript. Figures should be well drawn on pure white cartridge paper or good quality tracing paper. Lettering must be of very good quality (e.g. Letraset or good stencil). Typewritten or freehand lettering wEl not be accepted (without good reason). Each figure should have the authors’ name, figure number and caption details written lightly on the reverse in pencil. Cite each figure in the text in consecutive order using arabic numerals. “Figure” should be abbreviated to “Fig.”, except in the figure legends and at the beginning of sentences. Two separate caption sheets should accompany the typescript. Black end white photographs are acceptable if they have good definition and contrast and are produced on glossy paper. Consideration must be given by the author(s) to their final size when reduced for publication in the Journal, which is BS. Tables Each should be typed, double-spaced on a separate page and clearly numbered. Tables should be simple and self explanatory, and have a brief title above each one. They must be numbered consecutively with arabic numerals. Nomenclature In cases where there are well known and internationally accepted English names, these should be used throughout the text. The scientific name should be given in full, under- lined or preferably in italics, when quoted for the first time. The authority should also be quoted at this stage except in non -taxonomic ornithological papers. Only the generic name should begin with a capital letter. In subsequent reference to the name, use only the abbreviated generic name, with the full specific name but not the authority, e.g. Spodoptera exempta (Walker); subsequently S. exempta. The authority should not be given if scientific names are used to describe an “association” or species complex, e.g. Acacia drepanolobium - Themeda triandra , wooded grassland. Type in capitals the first letter of the English names of species (e.g. Crowned Eagle, Grey -capped Warbler) but not of the higher taxa (e.g. eagles, warblers) . References Care should be taken to see that all references quoted in the text are also given in full, alphabetical order, in this section. To avoid confusion, references to books should not abbreviated and neither should they be underlined nor in italics. Journal names should be written in full, underlined or, if possible in italics. Personal communications should not be cited in the references, neither should they be abbreviated in the text. If there are two or more references cited by an author which were published in the same year, distinguish them by adding letters to the dates, e.g. Brown (1972a). For technical reasons author names in the reference list should NOT be typed in upper case. Authors are advised to refer to this issue of the Journal for the preferred format. Acknowledgments As acknowledgements may imply endorsement of the data and conclusions in the authors’ work, it is advisable to obtain permission before quoting persons who have made contributions to the study. Once the paper has been selected by the Editor for possible publication in the Journal, the manuscript will be refereed by two persons who are acknowledged specialists in that field, before a decision is made on its acceptance. The authors) will be sent one set of printouts for the correction of typographical errors. Authors are urged to ensure that all changes to the text are incorporated into this copy, as no further changes can then be made. The author is entitled to 25 copies free of charge. Thirty free copies will be provided if there is more than one contributor. Further copies can be supplied at cost, provided the order is placed before printing has been undertaken. Referees Proofs Offprints The East Africa Natural History Society is supported by membership subscriptions, and may be unable in some instances, due to financial constraints, to accept all submissions of value. While every effort will be made to minimise such restrictions, contributors are requested to seek financial support to assist the Society with the ever increasing costs of publication. Editor Journal East Africa Natural History Society and National Museum. STS’ EDEN WILDLIFE TRUST Charitable Trust Registered No. 278008 The East Africa Natural History Society is most grateful to the Eden Wildlife Trust for their very generous contribution towards the cost of printing this journal part | Publi«hed by: The East Africa N Mural Hiitory Society « Typeietting by Development Communication! Ltd » Printed by AMREF | JOURNAL OF THE EAST AFRICA NATURAL HISTORY SOCIETY AND NATIONAL MUSEUM June 1993 Volume 82 (201) A CHECK LIST AND IDENTIFICATION KEY FOR SUCCULENT PLANTS IN GENERAL CULTIVATION IN NAIROBI LEONARD E. NEWTON and PAUL K. MBUGUA Department of Botany, Kenyatta University, P.O. Box 43844, Nairobi, Kenya ABSTRACT Following a survey of succulent plants in the gardens of 50 organisations and institutions in Nairobi, a check-list of over one hundred commonly cultivated species and varieties is presented. An identification key is provided, with characters described in non-technical language. INTRODUCTION In its situation on the eastern side of the Kenya highlands, Nairobi has a climate that includes long periods of drought, during which unwatered lawns turn brown and many herbaceous plants die if neglected. For this reason, succulent plants have long been regarded as eminently suitable for cultivation in public and private gardens in the city. Succulent plants have water reserves in special tissues, developed in either the stems or the leaves, and are adapted by evolution for surviving long periods of drought. Leaf succulents are valued as evergreen foliage plants. In most stem-succulents the leaves are reduced or absent, and the stems are green as they have taken over the functions of the leaves. As the stems enlarge to accommodate the stored water they often grow into bizarre shapes. In addition to the decorative value of their stems and leaves, many succulent plants produce abundant colourful flowers. The succulent plants to be seen in cultivation in Nairobi include some indigenous species, which have been collected from the wild by residents. Most are exotic, the majority having been imported during early colonial days. The exotic species are mainly from the Republic of South Africa (R.S. A ), Madagascar, and the American continent. There are also some species from neighbouring countries in north-east Africa. With such diverse origins, there is no single guide to the identification of these plants. Indigenous species can be identified with the aid of the Flora of Tropica! East Africa (Turrill et al., 1952-). though accounts of some families that include succulent species, notably the Aloaceae, Asclepiadaceae and Compositae, have not yet been published. Commonly cultivated exotic species are mentioned in the Flora, but are not included in the identification keys, and are not described. In order to identify exotic species, therefore, it is necessary to turn to monographs, if they exist for the families or genera concerned, and to foreign floras, if the country of origin is known. Another problem is that it is not easy to prepare herbarium specimens of succulent plants. Consequently, most species are not well represented in herbaria, and many of the specimens that do exist give a very poor idea of what the living plant is like. This makes it difficult to identify succulent plants by comparison with herbarium specimens. It should also be borne in mind that in cultivation spontaneous hybridisation can occur between plants that originate from different places, and the resulting hybrids can cause confusion in attempts at identification unless their hybrid nature is recognised. L.E. Newton and P.K. Mbugua * Cultivated succulents in Nairobi 44 METHODS A survey of succulent plants in cultivation in Nairobi was carried out with the aim of determining the species grown, and preparing an identification key. Fifty sites were examined, including public gardens, private gardens, and gardens in the grounds of hotels and educational institutions. Samples of each species were collected for further study in cultivation at Kenyatta University. Voucher specimens have been deposited in the East African Herbarium (National Museums of Kenya, Nairobi) and their reference numbers are indicated after the names in the checklist. This survey concentrated on species in general cultivation. Specialist collections built up by gardeners interested in growing succulent plants as a hobby were excluded. Numerous literature sources were used for identifying the plants, but in the following check-list most names are those given by Backeberg (1977) for members of the Cactaceae, and Jacobsen (1977) for succulents in other families. The choice of these two works as standards for names was based on the fact that they represent comprehensive surveys of succulent plants, and are generally available in libraries. Later names are used for some species in the check-list, but earlier and better-known names are also given where appropriate. It should be remembered that plant names are often subject to change as a result of taxonomic research, and even some fascicles of the Flora of Tropical East Africa are already out of date. Exotic species were found to be poorly represented in the East African Herbarium, and as the type specimens for these species are scattered around the world’s herbaria final checking with type specimens to confirm the accuracy of the names was beyond the scope of this survey. The check-list includes a number of ‘borderline succulents’, i.e. plants that are only slightly succulent, or are not strictly succulent but have enlarged organs that resemble those of succulent plants. Examples are Chorisia speciosa and Plumeria acuminata, which are trees with thick trunks and branches. Such species are included if they are featured in succulent plant literature. The geographical origin is given for each species, if known. The origin of some species is unknown because they were already in cultivation, with no record of their origin, when coming to the attention of taxonomists. As far as possible, vegetative characters were used to construct the key, to facilitate identification of plants without flowers. Terminology is simplified as much as possible, to allow use of the key by readers without botanical training, and specialised technical terms have been used only where an alternative expression would be very lengthy. CHECK-LIST Agavaceae Aga\>e amaniensis Trel. & Nowell (origin unknown) {Mbugua 147) Originally described from material found in cultivation in Tanzania. All members of this genus are from the American continent, but the exact origin of this species is unknown. Aga\>e americana L. cv. Marginata (Mexico) {Mbugua 148) Agcn’e angustifolia Haw. cv. Marginata (origin unknown) {Mbugua 16) Agcn’e attenuata Salm-Dyck (Mexico) {Mbugua 81) Agcn’e bourgaei Trel. (Mexico) {Mbugua 88) Agcn’e expansa Jacobi (Mexico) {Mbugua 84) Agcn>e sisal ana Perr. (Mexico) {Mbugua 73) Grown commercially for the fibres in the leaves, but occasionally seen in gardens. Aizoaceae Aptenia cordifolia (L. f.) Schwant. (R.S.A.) {Mbugua 20) J. EANHS & Nat. Mus. 82 (200) May 1993 45 L.E. Newton & P.K. Mbugua « Cultivated succulents in Nairobi Carpobrotus edulis (L.) Bol. (R.S.A.) ( Mbugua 9) Lampranthus roseus (Willd.) Schwant. (R.S.A.) ( Mbugua 106) Aloaceae Formerly included in the Liliaceae, but to be treated as a separate family in the Flora of Tropical East Africa. Aloe bainesii Th. Dyer (R.S.A.) ( Mbugua 149) Aloe graminicola Reyn. (East Africa) ( Mbugua 85) Very close to A. lateritia , and possibly not a distinct species. Aloe lateritia Engl. (East Africa) ( Mbugua 70) Aloe nyeriensis Christian (East Africa) ( Mbugua 150) Aloe secundiflora Engler (East Africa) - Haworthia fasciata (Willd.) Haw. ( Mbugua 121) Apocynaceae Adenium obesum (Forsk.) Roem. & Schult. (East Africa) ( Mbugua 79) Plumeria acuminata Ait. (Mexico) ( Mbugua 141) Asclepiadaceae Caralluma dummeri (N.E.Br.) White & Sloane (East Africa) ( Mbugua 31) Stapelia leendertziae N.E.Br. (R.S.A.) ( Mbugua 54) Bomhacaceae Chorisia speciosa Saint Hil. (Brazil) ( Mbugua 144) Bromeliaceae Dyckia sulfurea C. Koch. (Brazil) ( Mbugua 104) Cactaceae Cereus peruvianus (L.) Mill. (South America) (Mbugua 98) Cereus peruvianus (L.) Mill. f. monstrosus DC. (South America) (Mbugua 56) Epiphyllum anguliger (Lem.) G. Don. (Mexico) (Mbugua 92) Epiphyllum hybrid (Mbugua 91) Heliocereus sp. (Guatemala/Mexico) (Mbugua 135) Mammillaria elongata DC. var. stella-aurata (Mart.) K. Sch. (Mexico) (Mbugua 117) Opuntia cylindrica DC. (Ecuador & Peru) (Mbugua 43) Opuntia durangensis Br. & R. (Mexico) (Mbugua 22) Opuntia microdasys (Lehm.) Pfeiff. (Mexico) (Mbugua 113) Opuntia prasina Speg. (Argentina) (Mbugua 97) Opuntia subulata Engelm. (Chile & Argentina) (Mbugua 58) Opuntia vulgaris Mill. (Central America) (Mbugua 39) Pachycereus orcuttii (K. Brand.) Br. & R. (USA) (Mbugua 44) Schlumbergera bridgesii (Lem.) Loefgr. (Brazil) (Mbugua 86) Formerly called Zygocactus truncatus (Haw.) K. Sch. Probably of hybrid origin. Commelinaceae Tradescantia sillamontana Matuda (Mexico) (Mbugua 82) J. EANHS & Nat. Mus. 82 (201) June 1993 L.E. Newton and P.K. Mbugua * Cultivated succulents in Nairobi 46 Compositae Senecio aizoides (DC.) Sch. Bip. (R.S.A.) ( Mbugua 11) Senecio crassissimus H. Humb. (Madagascar) ( Mbugua 61) Senecio hildebrandtii Bak. (Madagascar) ( Mbugua 64) Senecio jacobsenii Rowl. (East Africa) ( Mbugua 6) Senecio sempervivus (Forsk.) Sch. Bip. (East Africa) ( Mbugua 47) Crassulaceae Aeonium arboreum (L.) Webb & Berth, cv. Atropurpureum (Mediterranean region) ( Mbugua 87) Aeonium hcrworthii (SD.) Webb. & Berth. (Canary Islands) ( Mbugua 52) Cotyledon coruscans Haw. (R.S.A.) ( Mbugua 50) Coytledon orbiculata L. (R.S.A.) ( Mbugua 59) Crassula argentea Thunb. (R.S.A.) (Mbugua 65) Crassula multica\>a Lem. (R.S.A.) (Mbugua 7) Crassula perfoliata L. (R.S.A.) (Mbugua 15) Crassula portulacea Lam. (R.S.A.) (Mbugua 49) Crassula sarmentosa Harv. (R.S.A.) (Mbugua 17) Crassula schimperi Fisch. & Mey. (East Africa) (Mbugua 14) Echeveria Columbiana v. Poelln. (Columbia) (Mbugua 28) Echeveria pulvinata Rose (Mexico) (Mbugua 99) Echeveria tolimanensis Matuda (Mexico) (Mbugua 142) Graptopetalum macdougallii Alexander (Mexico) (Mbugua 138) Graptopetalum paraguayense (N.E.Br.) Walth. (Mexico) (Mbugua 76) Graptopetalum pusillum Rose (Mexico) (Mbugua 32) Kalanchoe beharensis Drake & Castello var. aureo-aeneus Jacobs. (Madagascar) (Mbugua 80) Kalanchoe beharensis Drake & Castello var. beharensis (Madagascar) (Mbugua 34) Kalanchoe delagoensis Eck. & Zeyh. (Madagascar) (Mbugua 1) Formerly called K. tubiflora (Harv.) Hamet Kalanchoe diagremontiana Hamet & Perr. (Madagascar) (Mbugua 143) Kalanchoe fedtschenkoi Hamet & Perr. (Madagascar) (Mbugua 3) Kalanchoe gastonis-bonnieri Hamet & Perr. (Madagascar) (Mbugua 2) Kalanchoe hametorum Hamet (Mozambique) (Mbugua 100) Kalanchoe longi flora Schltr. var. coccinea Mam. -Lap. (Tropical Africa) (Mbugua 23) Kalanchoe longiflora Schltr. var. longiflora (Tropical Africa) (Mbugua 5) Kalanchoe marmorata Bak. (East Africa) (Mbugua 42) Kalanchoe marnieriana Jacobs. (Madagascar) (Mbugua 89) Kalanchoe millotii Hamet & Perr. (Madgascar) (Mbugua 90) Kalanchoe nyikae Engl. ssp. nyikae (East Africa) As Kalanchoe hemsleyana Cuf. in Jacobsen. Kalanchoe pinnata (Lam.) Persoon var. calcicola Perr. (Tropical Africa) (Mbugua 145) Kalanchoe prolifera (Bowie) Hamet (Tropical Africa) (Mbugua 26) Kalanchoe pumila Bak. (Madagascar) (Mbugua 119) Kalanchoe rosei Hamet & Perr. (Madagascar) (Mbugua 48) Kalanchoe scapigera Welw. (Angola) (Mbugua 131) Kalanchoe thyrsiflora Harv. (R.S.A.) (Mbugua 94) Sedum dendroideum Moc. & Sesse (Guatemala, Mexico) (Mbugua 8) Sedum guatemalense Hemsl. (Guatemala) (Mbugua 45) Sedum pachyphyllum Rose (Mexico) (Mbugua 4) Sedum morganianum Walth. (Mexico) (Mbugua 122) Sedum nussbaumeranum Bitter (Mexico) (Mbugua 24) J. EANHS & Nat. Mus. 82 (200) May 1993 47 L.E. Newton & P.K. Mbugua • Cultivated succulents in Nairobi Sedum palmeri S. Wats. (Mexico) ( Mbugua 96) Dracaenaceae Sansevieria caulescens N.E.Br. (East Africa) ( Mbugua 63) Sansevieria suffruticosa N.E.Br. (East Africa) ( Mbugua 103) Sansevieria robusta N.E.Br. (East Africa) ( Mbugua 30) Often referred to as S. ehrenbergii, but in Kenya that species occurs only in Coast Province. Sansevieria trifasciata Prain cv. Hahnii ( Mbugua 102) Sansevieria trifasciata Prain var. laurentii (De Willd.) N.E.Br. (Congo Republic) ( Mbugua 72) Sansevieria trifasciata Prain var. trifasciata (Sri Lanka) ( Mbugua 12) Euphorbiaceae Euphorbia arbuscula Balf. f. (Socotra) ( Mbugua 105) Euphorbia bussei Pax var. kibwezensis (N.E.Br.) Carter (East Africa) ( Mbugua 41) Formerly known as E. kibwezensis N.E.Br. Euphorbia candelabrum Trem. (East Africa) ( Mbugua 101) Euphorbia milii Des Moulin (Madagascar) A very variable species. Many varieties have been named, but hybrids between these have appeared in gardens and certain identification is difficult. The following varieties appear to be cultivated in Nairobi. Euphoria milii Des Moulin var. bevilaniensis (Croiz.) Ursch & Leandri f. rubro-striata Drake & Castillo ( Mbugua 108) Euphorbia milii Des Moulin var. hislopii (N.E.Br.) Ursch & Leandri ( Mbugua 60) Euphorbia milii Des Moulin var. imperatae (Leandri) Ursch & Leandri (Mbugua 109) Euphorbia milii Des Moulin var. longifolia Rauh (Mbugua 107) Euphorbia milii Des Moulin var. splendens (Boj. ex Hook.) Ursch & Leandri (Mbugua 21) Euphorbia milii Des Moulin var. tulearensis Ursch & Leandri (Mbugua 112) Euphorbia obovalifolia A. Rich. (East Africa) (Mbugua 133) Euphorbia stenoclada H. Baill. (Madagascar) (Mbugua 95) Euphorbia tirucalli L. (East Africa) (Mbugua 37) Euphorbia 'heterochroma ’ (East Africa) Ten species are now recognised in this group, all formerly called E. heterochroma Pax or E. stapfii Berger. Specimens in cultivation collected in different areas represent different species. The one growing naturally in the southern Rift Valley of Kenya, near Nairobi, is E. scarlatina Carter. Jatropha podagrica Hook. (Guatemala) (Mbugua 78) Monadenium stapelioides Pax (East Africa) (Mbugua 136) Pedilanthus tithymaloides Poit. (Central America) (Mbugua 33) Synadenium grantii Hook. (East Africa) (Mbugua 47) Labiatae Coleus spicatus Benth. (India) (Mbugua 36) Liliaceae Bowiea volubilis Harv. & Hook. f. (East Africa to R.S.A.) (Mbugua 124) East African plants formerly distinguished as B. kilimandscharica Mildbr. Bulbine frutescens (L.) Willd. (R.S.A.) (Mbugua 38) Portulacaceae Portulaca cv. Grandiflora (origin unknown) (Mbugua 10) Probably of inter-specific hybrid origin, but parent species unknown J. EANHS & Nat. Mus. 82 (201) June 1993 L.E. Newton and P.K. Mbugua * Cultivated succulents in Nairobi 48 Vitaceae Cissus quadrangularis L. (East Africa) ( Mbugua 35) IDENTIFICATION KEY Water storage in fat, fleshy stems, leaves less fleshy and commonly deciduous, reduced or absent: stem succulents. Group I Water storage in fat, fleshy ± persistent leaves, which are at least more succulent than the stems: leaf succulents. Group II GROUP I - Stem succulents l Stem angled, flat or grooved 2 Stem not angled, flat or grooved, ± cylindrical 22 2. Stem with milky latex 3 Stem without latex 7 3. Stem spiny or thorny 4 Stem neither spiny nor thorny Euphorbia arbuscula 4. Stem 3 -angled 5 Stem with more than 3 angles 6 5. Stems and leaves variegated Euphorbia obovalifolia Stem and leaves not variegated Euphorbia bussei v. kibwezensis 6. Stem diameter over 7.0 cm, over 3.0 m high Euphorbia candelabrum Stem diameter about 1 .5—2.0 cm, under 3 m high Euphorbia ‘heterochroma ’ 7. Stem flattened or jointed 8 Stem angular, grooved or tuberculate 14 8. Stem spiny or thorny 9 Stem neither spiny nor thorny 12 9. Spines weak, up to 1 cm long, or absent 10 Spines strong, over 2 cm long Opuntia vulgaris 10. Spines present, accompanied by a few minute barbed bristles 11 Spines absent, stem with numerous minute yellow barbed bristles Opuntia microdasys 11. Stem joints over 35 cm long, spines more than 1 cm long Opuntia prasina Stem joints less than 22 cm long, spines less than 1 cm long Opuntia durangensis 12. Stem with hairs 13 Stem without hairs Epiphvllum hybrid 13. Stem divided into short joints, 4-7 cm long Schlumbergera bridgesii Stem not divided into short joints Epiphyllum anguliger 14. Stem spiny or thorny 1 5 Stem neither spiny nor thorny 19 15. Stem with more than 4 vertical grooves, no aerial roots 16 Stem with only 3 vertical grooves, and aerial roots Heliocereus sp. 16. Young spines covered in woolly outgrowths Cereus peruvianus Young spines not covered in woolly outgrowths 17 17. Spines 1-4 together 18 Spines more than 4 together 21 18. Stem tips with needle-shaped succulent leaves, 2. 5-4.0 cm long Opuntia subulata Stem tips with leaves 1-2 cm long, that soon fall Opuntia cylindrica 19. Stem with tendrils, climbing Cissus quadrangularis Stem without tendrils, erect or creeping 20 J. EANHS & Nat. Mus. 82 (200) May 1993 49 L.E. Newton & P.K. Mbugua • Cultivated succulents in Nairobi 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. Stem hairy, with teeth 0.2 cm long Stapelia leendertziae Stem not hairy, with conical teeth up to 1.4 cm long Caralluma dummeri Stem with parallel, vertical and continuous grooves, weak spines Pachycereus orcuttii Stem with discontinuous, zig-zag grooves, very firm spines ....Cereus peruvianus f. monstrosus Stem thorny or spiny 23 Stem neither thorny nor spiny 32 Stem with milky latex 24 Stem without latex 31 Stem with numerous tubercles, and 2-3 small backward curved spines per tubercle Monadenium stapelioides Stem without tubercles 25 Inflorescence bright to dull red 26 Inflorescence bright to greenish-yellow 30 Stem 2 cm or more in diameter, leaf more than 10 cm long Euphorbia milii v. hislopii Stem less than 2 cm diameter, leaf less than 5 cm long 27 Floral bracts yellowish-red striate Euphorbia milii v. bevilaniensis f. rubro-striata Floral bracts entirely red or yellow 28 Leaves more than 3 times longer than wide Euphorbia milii v. longifolia Leaves less than 3 times longer than wide 29 Stems 5 mm diameter Euphorbia milii v. imperatae Stems at least 10 mm diameter 30 Thorns soft, 5-10 mm long Euphorbia milii v. tulearensis Thorns firm, more than 10 mm long Euphorbia milii v. splendens Tree with conical thorns on trunk; leaves palmate Chorisia speciosa Dwarf plant with clusters of needle-like spines on stem tubercles; leaves absent Mammillaria elongata v. stella-aurata Stem with milky latex 33 Stem without milky latex 36 Leaves variegated 34 Leaves (when present) not variegated 35 Stem with whitish-cream vertical bands Pedilanthus tithymaloides Stem without bands, pinkish to greenish bark Svnadenium grantii Leaves persistent, to 20 cm long Plumeria acuminata Leaves short-lived, when present 1 .0-1.5 cm long Euphorbia tirucalli Green shoots non-succulent, growing from bulb Bowiea volubilis Green shoots succulent and not distinct from storage organs 37 Leaves 20 cm or more long, lobed, stem surface rough Jatropha podagrica Leaves less than 2 cm long, simple spatula-shaped, stem surface smooth Adenium obesum GROUP II - Leaf succulents 1. Leaves simple 2 Leaves compound 75 2. Leaves thorny or spiny, sometimes with apical spine 3 Leaves neither thorny nor spiny (but may have a hardened tip, e.g. Sansevieria spp.) 14 3. Margin thorny or spiny 4 Margin neither thorny nor spiny, but apical spine present Agave sisalana 4. Plant with obvious stem above ground 13 Plant without obvious stem above ground 5 5. Leaves tough, fibrous, not snapping cleanly when folded 6 Leaves scarcely fibrous, snapping cleanly when folded 9 J. EANHS & Nat. Mus. 82 (201) June 1993 L.E. Newton and P.K. Mbugua * Cultivated succulents in Nairobi 50 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. Leaves with longitudinal coloured bands Leaves without coloured bands Leaves with longitudinal ridges, slightly rough Leaves without ridges, smooth Leaves 2. 4-2. 6 m long Leaves 0.5-0. 6 m long Leaves with white spots Leaves without spots Lower leaf surface with white oblong spots except near tip Lower leaf surface with white spots for whole length Leaves with numerous, small greenish-yellow vertical lines Leaves without lines Leaf surface rough, white-gray Leaf surface smooth, grayish-green Stem covered with old leaves Stem not covered with old leaves Leaves ± cylindrical, may be strap-shaped at base Leaves with at least 1 flat surface Leaves produced in opposite pairs Leaves produced singly Stem with hairs at nodes Stem without hairs Leaf with grayish-black spots Leaf without spots Leaf tip rounded Leaf tip pointed Stem hanging Stem upright Leaves with thick waxy bloom, greenish-white Leaves without waxy bloom, bright green Leaves sheathed Leaves not sheathed Leaves containing fibres Leaves not containing fibres Leaf with groove near base Leaf with groove extending from base to over half-way up Leaves 9-15 cm long Leaves 60-90 cm long Stem with spines Stem without spines Leaves with thick waxy bloom Leaves without waxy bloom Leaves with 1-2 flat surfaces (underside may be rounded) . Leaves with 3 flat surfaces (trigonous, especially at tips) ... Leaf margin regular Leaf margin irregular Leaf under surface ± rounded (esp. in young leaves) Leaf upper and under surfaces ± flat Leaves hairy or scaly •. Leaves smooth, without hairs or other ornamentation 7 11 Agave amaniensis 8 . Agave americana cv. Marginata Agave angustifolia cv. Marginata 10 Aloe secundiflora Aloe lateritia Aloe graminicola Dyckia sulfurea 12 Agave expansa Agave bourgaei Aloe nyeriensis Aloe bainesii 15 27 16 18 Portulaca cv. Grandiflora 17 Kalanchoe delagoensis Crassula schimperi 19 21 Sedum morganianum 20 Sedum pachyphyllum Sedum guatemalense 22 25 23 Bulbine frutescens Sansevieria robust a 24 Sansevieria suffruticosa Sansevieria caulescens Opuntia subulata 26 Senecio aizoides Senecio hildebrandtii 28 74 29 59 30 37 31 32 J. EANHS & Nat. Mus. 82 (200) May 1993 51 L.E. Newton & P.K. Mbugua * Cultivated succulents in Nairobi 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. Leaves hairy, giving grayish colour Leaves not hairy Upper leaf surface ± concave (esp. at base) Upper leaf surface completely flat Leaf with keel, edge sharp Leaf without keel, edge rounded Leaf 8-15 cm long Leaf 5.0-6. 5 cm long Leaf tip extended into a bristle Leaf tip not extended into a bristle Leaves with thick waxy bloom Leaves without waxy bloom Leaves with fibres Leaves without fibres Leaves with gray-green transverse bands Leaves without bands Leaves with yellowish-white stripes along the margins . Leaves without stripes Leaves with parallel sides, to 90 cm long Leaves with rounded sides, to 10 cm long Stem lying on ground Stem upright Plant hairy Plant not hairy Leaves produced in opposite pairs, tips pointed Leaves produced singly, tips rounded Plant with milky latex Plant without latex Leaves variegated gray-green Leaves not variegated Leaves produced in opposite pairs Leaves produced singly Leaf tips rounded Leaf tips pointed (at least in young stage) Leaf surface marked with minute spots Leaf surface without minute spots Leaves with pinkish-green tips Leaves dark green throughout Leaf base extended as two rounded lobes below attachment to stalk Leaf base entirely above attachment to stalk Leaves clasping the stem, 10-15 cm long, 7-9 cm wide Leaves not clasping the stem, 7-9 cm long, 4-6 cm wide Leaves with velvety hairs Leaves not hairy Leaves with stalks Leaves without stalks Plant with thick waxy bloom Plant without waxy bloom Echeveria pulvinata Haworthia fasciata 33 36 34 Cotyledon coruscans 35 Graptopetalum pusillum Echeveria tolimanensis Graptopetalum paraguayense Echeveria Columbiana Sedum nussbaumeranum 38 41 39 Agave attenuata ..Sansevieria trifasciata v. laurentii 40 Sansevieria trifasciata v. trifasciata .... Sansevieria trifasciata cv. Hahnii 42 44 Tradescantia sillamontana 43 Aptenia cordifolia Senecio jacobsenii 45 46 Pedilanthus tithymaloides Monadenium stapelioides 47 55 48 50 49 52 Crassula portulacea Crassula multica\>a Kalanchoe marnieriana 51 Kalanchoe thyrsiflora Cotyledon orbiculata Kalanchoe scapigera 53 54 Crassula perfoliata Kalanchoe hametorum Crassula argentea I. EANHS & Nat. Mus. 82 (201) June 1993 L.E. Newton and P.K. Mbugua • Cultivated succulents in Nairobi 52 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. Leaves without stalks 56 Leaves with stalks 57 Leaf tip pointed Graptopetalum macdougallii Leave tip rounded Sedum palmeri Leaf surface homy Senecio crassissimus Leaf surface not homy 58 Leaf veins visible on both surfaces Senecio sempervivus Leaf veins not visible on upper surface, only major vein visible on underside Sedum dendroideum Stem with 4 angles 60 Stem ± cylindrical 61 Plant hairy Coleus spicatus Plant not hairy Kalanchoe longiflora v. longiflora Leaves stalked 62 Leaves not stalked 71 Leaves smooth 63 Leaves with fine hairs 70 Leave base extended as two rounded lobes below attachment to stalk 64 Leaf base entirely above attachment to stalk 65 Leaves ovate Kalanchoe fedtschenkoi Leaves lanceolate Kalanchoe rosei Leaf stalk attached to lower surface of blade 66 Leaf stalk attached to edge of blade 67 Leaf margins wavy Kalanchoe beharensis v. beharensis Leaf margins toothed Kalanchoe diagremontiana Leaves with thick waxy bloom 68 Leaves without waxy bloom 69 Leaf margin teeth with bluish-pink base (especially on old leaves) Kalanchoe hemsleyana Leaf margin teeth same colour as blade at the base Kalanchoe pumila Leaf surface marked with minute spots Crassula sarmentosa Leaf surface without minute spots Kalanchoe longiflora v. coccinea Leaves and stem (esp. young stems) velvety Kalanchoe beharensis v. aureo-aeneus Leaves and stems with tufts of soft hairs Kalanchoe millotii Leaves produced in opposite pairs 72 Leaves produced singly 73 Leaf tip with pointed elongation Kalanchoe gastonis-bonnieri Leaf tip rounded Kalanchoe marmorata Leaves light gray-green Aeonium haworthii Leaves dark -brownish purple Aeonium arboreum cv. Atropurpureum Leaves 2^4 cm long Lampranthus roseus Leaves 8-10 cm long Carpobrotus edu’.is Stem 4-angled (esp. in youngest parts), diameter 4-6 cm Kalanchoe prolifera Stem ± cylindrical, diameter 2-3 cm Kalanchoe pinnata v. calcicola J. EANHS & Nat. Mus. 82 (200) May 1993 53 L.E. Newton & P.K. Mbugua » Cultivated succulents in Nairobi ACKNOWLEDGEMENTS We are grateful to the proprietors of the 50 organisations and institutions who allowed free access to their gardens, and allowed removal of living material for further study. This project was made possible by a postgraduate scholarship awarded to PKM by Kenyatta University. REFERENCES Backeberg, C. (1977). Cactus Lexikon. Poole: Blandford Press. Jacobsen, H. (1977). Lexikon of Succulent Plants. 2nd. edn. Poole: Blandford Press. Turrill, W.B. etal. (ed.) (1952-continuing). Flora of Tropical East Africa. London: Crown Agents. (Published as separate family fascicles, with various editors. Later fascicles edited by R.M. Polhill and published by Balkema, Rotterdam.) EDEN WILDLIFE TRUST Charitable Trust Registered No. 278008 The East Africa Natural History Society is most grateful to the Eden Wildlife Trust for their very generous contribution towards the cost of printing this Journal part Published by: The East Africa Natural History Society Printed by AMREF J. EANHS & Nat. Mus. 82 (201) June 1993 NOTICE TO CONTRIBUTORS Journal of the East Africa Natural History Society and National Museum Papers on Natural History subjects will be considered on the understanding that they are being offered exclusively to the Society and have not been submitted, at the same time, or previously, for publication elsewhere. Copyright is retained by the East Africa Natural History Society; reference to contributions in the journal may be made, but no part of the text, diagram, figure, table or plate may be reproduced without permission from the Editor. Such permission will only be granted with the approval of the author. Three copies of papers for consideration should be sent to The Editor, Journal of East Africa Natural History Society and National Museum, P.O. Box 44486, Nairobi, Kenya, and should be accompanied by a covering letter from the author, or in case of multiple authors, from the author responsible for decisions regarding the text. Contributions may be sent as: • preferably a floppy disk of either 5.25 or 3.5 inch in fully IBM compatible MS-DOS, carrying text in a commonly used wordprocessing language, or a floppy disk with the text prepared on an Apple Macintosh. If in any doubt, the author(s) should contact the Editor at the above address for further details. Three “near letter quality” printed copies on A4 size paper double-spaced should accompany the diskette. The author(s) should, of course, retain a back-up copy. • or a typescript, which should be on A4 size paper, typed on one side only, double spaced with margins of at least 5 cm on the left side of the paper to allow for editorial instructions to the printer. Title Should be brief and to the point. Author name(s) and institution(s) where the work was carried out should be given, together with a suggested running page headline of not more than 50 characters including the name(s) of the author(s). A current address for correspondence, if applicable, should be included as a footnote. Abstract To be on the first page and consist of no more than 150 words. This should outline the major findings clearly and concisely without requiring reference to the text. Headings Should be brief, and in the following format: “Abstract, Introduction, Methods, Results, Discussion, Acknowledgments and References”. All headings to be typed on a separate line from subsequent text. Main headings to be typed in capital letters and centred on the page. Sub-headings to be at the left of the page, in lower case; further headings should be in italics or underlined (not bold) beneath them. The first letter of any headings should be a capital letter. Text Indent each new paragraph except those after a heading. Spelling should follow that of the Oxford English Dictionary. All pages must be numbered, at the top right hand edge. Numerals, dates, times and scientific units. Prose numbers should be written in full up to and including ten, and as numerals thereafter, except at the end dates and beginning of a sentence. Dates should be written as “29 August”, “10-13 January 1989” or “1969-1989”. Times should follow the 24 hour clock ( e.g . 1500 h). All measurements must be in metric units. SI units and abbreviations should be used throughout with a space before and after the SI unit. Full stops should not be used after such units. The approximate position of tables and figures should be indicated in the left margin with pencil. Figures The original and two photocopies must be provided with each manuscript. Figures should be well drawn on pure white cartridge paper or good quality tracing paper. Lettering must be of very good quality (e.g. Lettraset or good stencil). Typewritten or freehand lettering will not be accepted (without good reason). Each figure should have the author’s name, figure number and caption details written lightly on the reverse in pencil. Cite each figure in the text in consecutive order using Arabic numerals. “Figure” should be abbreviated to “Fig.”, except in figure legends and at the beginning of sentences. Two separate caption sheets should accompany the typescript. Black and white photographs are acceptable if they have good definition and contrast and are reproduced on glossy paper. Consideration must be given by the author(s) to their final size when reduced for publication in the Journal, which is B5. Tables Each table should be typed, double-spaced on a separate page and clearly numbered. Tables should be simple and self-explanatory, and have a brief title above each one. They must be numbered consecutively with Arabic numerals. Nomenclature. In cases where there are well known and internationally accepted English names, these should be used throughout the text. The scientific name should be given in full, underlined or preferably in italics, when quoted the first time. The authority should also be quoted at this stage except in non- taxonomic ornithological papers. Only the generic name should begin with a capital letter. In subsequent references to the name, use only the abbreviated generic name, with the full specific name but not the authority, e.g. Spodoptera exempta (Walker); subsequently S. exempta. The authority should not be given if scientific names are used to describe an “association” or species complex, e.g. Acacia drepanolobium - Theneda triandra, wooded grassland. Type in capitals the first letter of the English names of species (e.g. Crowned Eagle, Grey-capped Warbler), but not of higher taxa (e.g. eagles, warblers). References. Care should be taken to see that all references quoted in the text are also given in full, in alphabetical order, in this section. To avoid confusion, references to books should not be abbreviated and neither should they be underlined nor in italics. Journal names should be written in full, underlined or, if possible, in italics. Personal communications should not be cited in the references, neither should they be abbreviated in the text. If there are two or more references cited by an author which were published in the same year, distinguish them by adding letters to the dates, e.g. Brown (1972a). For technical reasons author names in the reference list should not be typed in upper case. Authors are advised to refer to this issue of the Journal for the preferred format. Acknowledgements. As acknowledgements may imply endorsement of the data and conclusions in the authors’ work, it is advisable to obtain permission before quoting persons who have made contributions to the study. Once the paper has been selected by the Editor for possible publication in the Journal, the manuscript will be refereed by two persons who are acknowledged specialists in that field, before a decision is made of acceptance. The author will be sent one set of printouts for the correction of typographical errors. Authors are urged to ensure that all changes to the text are incorporated into this copy, as no further changes can then be made. The author is entitled to 25 copies free of charge. Thirty free copies will be provided if there is more than one contributor. Further copies can be supplied at cost, provided the order is placed before printing has been undertaken. The East African Natural History Society is supported by membership subscriptions, and may be unable in some instances, due to financial constraints, to accept all submissions of value. While effort will be made to minimise such restrictions, contributors are requested to seek financial support to assist the Society with the ever increasing cost of publication. Editor, Journal East Africa Natural History Society and National Museum. Referees Proofs Offprints H 11 K JOURNAL Iti OF THE EAST AFRICA NATURAL HISTORY SOCIETY AND NATIONAL MUSEUM December 1993 Volume 82 (202) ALPINE VERTEBRATES OF MOUNT KENYA, WITH PARTICULAR NOTES ON THE ROCK HYRAX Truman P. Young Calder Conservation and Ecology Center, Fordham University Box K, Armonk, NY, 10504, U.S.A. Matthew R. Evans* Department of Zoology, University of Cambridge Downing Street, Cambridge, CB2 3EJ, U.K. INTRODUCTION Various observers over the years have contributed to our knowledge of Mount Kenya’s alpine vertebrates. There is only limited evidence that traditional cultures visited alpine Mount Kenya (Coe 1967), and no records of what they found there. Sharpe (1900) and Thomas (1900) made the first lists of animals on Mount Kenya, followed shortly by Loring and Heller (in Roosevelt 1910). Hollister (1919) provided a description of mammals collected from East Africa, including those of Mount Kenya. However, it was not until Moreau (1944) that a systematic description of the vertebrate fauna was attempted. Since then, a great deal has been added to our understanding of the vertebrates of alpine Mount Kenya. Coe and Foster (1972) made a considerable contribution to our knowledge of the mammals of the northern slopes, in particular the smaller mammals (rodents and shrews). Williams (1978) provided an extensive list of birds and mammals of Mount Kenya National Park. Coe (1967), Coe and Sale (1971), Fayad (1981), and Young (1991) gave brief summaries of the alpine faunas of Mount Kenya and Kilimanjaro. Numerous other authors have provided information on individual species and groups. Mount Kenya has been the subject of considerable botanical, zoological, and ecological research over the last thirty years (see Rehker 1989, Young 1990). Nonetheless, it has been half a century since the last comprehensive description of Mount Kenya’s alpine vertebrate fauna (Moreau 1944). It is therefore appropriate to update our knowledge of the alpine vertebrates of Mount Kenya. In this review, we have drawn on published literature and on personal communications with both visitors and those familiar with the mountain. However, much of the information below is drawn from personal experience. Between 1977 and 1990, TPY spent over 500 days and nights above treeline on Mount Kenya. Most of this time was spent in the Teleki Valley, but also included one to several visits to * Present address: Edward Grey Institute of Field Ornithology, Department of Zoology, Oxford University South Parks Road, Oxford, OX1 3PS, U.K. T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya 56 each of the Hobley, Hinde, Gorged, Hohnel, Mackinder, Liki, Sirimon, Kazita, and Marania Valleys. In 1989-90, MRE spent over 200 days on Mount Kenya, mostly in the Teleki Valley. In the descriptions below, the ‘research camp’ (of TPY) was located at 4180 m in the Teleki Valley, near the Ranger Station. ‘Northern slopes’ refers the the area above treeline between the Hinde and the Sirimon Valleys (mainly the drainages of the Marania and Kazita rivers). The ‘Treeline’ occurs at 3000-3300 m. Elevations are probably accurate to ± 100 m. Except where noted, all of the observations below are ours. At least 1 5 of the 112 species listed below have not been previously recorded from alpine Mount Kenya. In addition, species previously thought to be only alpine visitors have been identified as at least partly resident in the alpine zone ( e.g ., Slender-billed Starling, Lion and Zorilla). There are also several new altitudinal limits for Africa (Spotted Hyaena, Zorilla, Bongo, Sykes’ and Colobus Monkeys, Yellow-crowned Canary, and Rana wittei). We have accepted the nomenclature of specialised literature (reptiles and amphibians: Loveridge 1957; birds: Britton 1980, Lewis and Pomeroy 1989; mammals: Dorst and Dandelot 1972 and Kingdon 1974). ALPINE VEGETATION OF MOUNT KENYA The Mount Kenya forest varies with aspect. It is wettest on the south-eastern slopes and driest on the northern slopes, with bamboo favoring the wettest sites (Hedberg 1951, Young 1990). However, at higher altitudes (above treeline), the western slope (Naro Moru track) and southern slope (Kamweti track) are wetter than either the northern slope (Sirimon and Timau tracks) or eastern slope (Chogoria track). In general, rainfall increases with altitude up to 2500-3200 m, and then decreases with altitude (Winiger 1986). Numerous attempts have been made over the years to describe the vegetation of Mount Kenya (Hedberg 1951, 1964; Coe 1967), but only recently has an alpine vegetation map been produced (Rehder et al. 1988, 1989) and a quantitative analysis of plant community composition done (Young and Peacock 1992). Common to all descriptions is the recognition of the importance of altitude and topography on the alpine vegetation. A well defined timberline occurs at 3000-3300 m on all but the northern slopes. (On the northern slopes we have arbitrarily set 3000 m as the lower limit for alpine records). Many plant species have their lower or upper altitudinal limits at timberline. Immediately above the forest boundary there is a band of ericaceous scrub whose composition and breadth vary with aspect. This vegetation band is narrowest and least diverse on the wetter and steeper slopes to the west and south, and is widest and contains the greatest diversity of shrubs on the drier and gentler northern and eastern slopes. Vast areas between the Marania and Hinde Valleys are covered by this little-explored vegetation. Fires are a regular occurrence in this habitat (Bill Woodley, Phil Snyder, personal communication), perhaps occuring at intervals of several years (Bongo Woodley, personal communication). At higher altitudes (3400-3700 m) this ericaceous scrub merges gradually into the classic afroalpine vegetation, with its characteristic giant rosette plants ( Lobelia and Senecio spp). The more level valley bottoms and (except at great altitude) ridges are wetter and contain a more lush vegetation than the slopes, which are characterised by sparser vegetation cover and drought resistant species (Young and Peacock 1992). At nearly all alpine sites above the heather, the tussock grass Festuca pilgeri is the dominant plant. In some sites it forms virtual monocultures of large tussocks, especially on wetter aspects between 3600 and 4100 m. The greatest diversity of plant species appear to be at middle alpine altitudes (3900^1200 m), at least on the western slope. Above 4100 m, the vegetation gradually becomes sparser, and plants are rare above 4500 m. J. EANHS & Nat. Mus. 82 (202) December 1993 57 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya In addition to altitude and topography, the age of Senecio keniodendron stands is a key determinant of plant community composition. This species occurs as single-sized (and presumably single-aged) stands on mid-altitude slopes. In the Teleki Valley, the understoreys of younger stands are characterised by Festuca pilgeri and Helichrysum spp, and the understoreys of older stands are characterised by nearly monospecific stands of Alchemilla spp (Young and Peacock 1992). The latter appear to be favorite resting spots for Grimm’s Duiker. In addition, virtually all major drainages contain running water throughout the year, and there are numerous permanent tarns, occurring at altitudes of 3000 to 4500 m. MOUNT KENYA’S ALPINE VERTEBRATES Fish (3 spp) Salvelinus alpinus (Arctic Char, American Brook Trout) Not native; stocked in Lake Hohnel in 1949, but apparently failed to breed there (Copley 1953, Watson 1988). They are extinct there now (John Omira Miluw'i, personal communication). Salmo trutta (Brown Trout) Not native; stocked in the Ontilili, Hinde, and Sirimon rivers (Coe and Sale 1971, Watson 1988). Salmo gairdnerii (Rainbow Trout) In addition to the two species above, Mills (1971) reports that Rainbow Trout were introduced to Kenya. This species is present in the Teleki Valley (Naro Moru River), and perhaps also in the Marania and Kazita rivers. The upper Naro Moru River in Teleki Valley has been heavily fished in the past (J. Omira, personal communication), but was supporting fair numbers of fish in 1989-90, when MRE was given a gravid female weighing 1 kg. The trout in the Marania and Kazita rivers are numerous and under-sized (TPY, personal observation). Bongo Woodley reports that this species has been recently released into Rutundu, Alice, Ellis and Carr Lakes, where they do not breed due to a lack of flowing water. At these lakes, 1-2 kg fish are commonly caught, and a 3.5 kg trout was caught in Lake Alice (Bongo Woodley, personal communication). Amphibians (2 spp) Phrynobatrachus kinangopensis Angel Loveridge (1957) reports this species as high as 3350 m. In addition to this and the following species, Tom Madsen (personal communication) suggests there are others just above timberline on the Sirimon Track. Rana witte i Angel This species is the common larger (body length 5 cm) frog in the lower alpine on the Timau Track, where it breeds in the Kazita River. It also occurs on the Sirimon Track (T. Madsen, personal communication). A sighting of probably this species by Alan Smith and TPY at 3800 m in the Hinde Valley is apparently an altitude record for an amphibian in East Africa. Bongo Woodley reported frogs from Lake Alice and the Nithi River in August 1993 (personal communication). J. EANHS & Nat. Mus. 82 (202) December 1993 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya 58 Reptiles (6 spp) Algyroides alleni Barbour (Alpine Meadow Lizard) Algyroides alleni is common throughout the drier parts of the mountain. TPY has encountered them in the Sirimon, Liki, Marania, Kazita, Hinde, Gorges, and Hobley Valleys at altitudes of 3400^1600 m, but not in the wetter Teleki or Hohnel Valleys. However, MRE has seen one just down the ridge from Two Tarn Hut. Their upper limit is represented by a population just south of Kami Hut at the head of the Mackinder Valley, and may be an altitude record for an East African reptile. At this site, the lizards live amongst the rocks, whereas at lower altitudes they are usually found in and on grass tussocks of Festuca pilgeri. Coe (1969) found eggs of this species at 3800 m. Mabuya irregularis Lonberg (Skink) TPY found this skink at 3600 m in dry scrubby grassland 4 km north-west of Ithanguni Peak in 1982. Its alpine distribution is not known. The genus is currently under taxonomic review, and it is possible that the alpine form is a distinct species (T. Madsen, personal communication). Chameleo schubotzi Stemfeld (Kenya Side-striped Chameleon) The type of this species (then C. bitaeniata schubotzi) in 1912 is recorded as being collected from 14,000 feet (4200 m) on Mount Kenya (Loveridge 1957), but such an extreme height seems unlikely. In June 1978 Alan Smith and TPY found an individual at 3850 m on the ridge west of the Hobley Valley. It was on the leaf of a Senecio keniensis (brassica) rosette about 0.5 m tall. James Hebrard (1981) reports that they are not uncommon in the ericaceous scrub just above treeline on Mount Kenya, and prefer the scrub layer less than two meters above the ground. Chameleo hohnelii Steindachner (Hohnel’s Chameleon) This species appears to be restricted to the ericaceous scrub just above treeline, and prefers the vegetation more than two meters above the ground (Hebrard 1981). Vipera hindei Boulenger (Hinde’s Viper) This snake, which is reportedly not uncommon in the alpine grassland of the Aberdare Mountains (A. McKay, personal communication), also occurs in alpine Mount Kenya. Individuals have been reported from the northern slopes by Raymond Hook (Moreau 1944), Phil Snyder, and Nigel Trent (personal communications). With Tim Tear, TPY photographed a small individual on the Timau Track at 3400 m in September 1982. Bongo Woodley found two Hinde’s Vipers (13 and 20 cm long) on the shore of Lake Alice in March 1993 (personal communication). Psammophylax variabilis multisquamis Loveridge (Striped Grass Snake) Reported from Mt Kenya’s ‘high grass moorlands’ (Sprawls 1978). Bongo Woodley saw a long thin snake along the Nithi River above the roadhead on the Chogoria Track in March 1993 (personal communication), perhaps of this species. Birds (58 spp) Tachybaptus (Poliocephalus) ruficollis capensis (Pallas) Salvadori (Little Grebe or Dabchick) Reported by Moreau (1944) from Lake Ellis (3450 m). J. EANHS & Nat. Mus. 82 (202) December 1993 59 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya Scopus umbretta Gmelin (Hamerkop) Reported from the Chogoria Track above treeline in March 1984 by Tim Tear (personal communication). Anastomus lamelligerus Temminck (Open-billed Stork) On at least two occasions. Open-billed Storks have visited the upper Teleki Valley. Coe and Sale (1971) reported a dead individual at the head of the valley (4180 m). In September 1976, a group of several birds landed in the valley in the vicinity of the Teleki Hut (4000 m) and, apparently unable to regain flight the next day, died there (John Omira Miluwi, personal communication). TPY saw several of these dead birds in 1977. Open-billed storks are regional migrants (Williams and Amott 1980), and may occasionally mistake the boggy alpine valley bottoms for hospitable habitat, with fatal results. Ciconia ciconia L. (White Stork) Reported as an occasional visitor on migration by Williams (1978). Bostrychia (Lampribis) olivacea akeleyorum (Dubois) Chapman (Green Ibis) This high forest bird, recorded from 2000-3700 m (Britton 1980) is suspected of being at least partly nocturnal (Williams and Amott 1980). Small groups are frequently heard and seen at dusk flying above the forest near the Met Station (3000 m) on the Naro Moru Track. Jackson (1938) quotes Ackeley as seeing them on “Mount Kenya from 6,000-12,000 feet (timber line)”. In November 1979, John Imhof reported several pre-dawn encounters with honking, long-necked birds flying down the Teleki Valley at 4000 m. These birds may have been Green Ibises. Anas sparsa leucostigma (Eyton) Rtippell (African Black Duck) There are several resident pairs of Black Ducks on the tarns at 4000^1400 m. TPY found an abandoned nest with five eggs at Teleki Tarn (4300 m). It was located at the water’s edge and made out of grass (Festuca pilgeri ), lined with down, and covered by a layer of grass. TPY also saw a pair of adults with ducklings at Thompson’s Tam (4400 m). Coe (1967) reports other records of Black Ducks nesting on alpine tarns. They have also been seen on the Nithi River near Lake Ellis (Bongo Woodley, personal communication) Oxyura maccoa Eyton (Maccoa Duck) TPY found a male and three females on Lake Rutundu (3000 m) in September 1982. Sagittarius serpentarius Miller (Secretary Bird) TPY has seen only one individual, on the Timau Track at 3400 m in 1980, although other sightings by Bongo Woodley, Helen Young and Vince Fayad (personal communications.) between 1978 and 1993 imply they may not be uncommon along the Timau and Sirimon Tracks. Raymond Hook also reported one at 3960 m (Moreau 1944). Gyps rueppellii Brehm (Riippell’s Vulture) Occasional wanderer to the moorlands (D.A. Turner, personal communication) Gypaetus barbatus meridionalis (L.) Keys. & Bias. (Lammergeyer) Lammergeyers have been repeatedly sighted along the Sirimon Track near Sendeyo and Tereri Peaks. We have each seen one flying high above the Teleki Valley, in different years. These birds may be resident. J. EANHS & Nat. Mus. 82 (202) December 1993 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya 60 Circus aeruginosus L. (Eurasian Marsh Harrier) Reported as a migrant visitor by Williams (1978). Circus macrourus Gmelin (Pallid Harrier) Reported as a migrant visitor by Williams (1978). Circus pygargus L. (Montagu’s Harrier) Reported as a migrant visitor by Williams (1978). Reported as high as 3700 m by Meinertzhagen (1937). Aquila rapax Temminck (Tawny Eagle) Occasional visitor to the moorlands (D.A. Turner, personal communication) Aquila nipalensis Hodgson (Steppe Eagle) Occasional visitor to the moorlands (D.A. Turner, personal communication) Aquila verreauxii Lesson (Verreaux’s Eagle) The distribution of Verreaux’s Eagle in the alpine zone of Mount Kenya is patchy, and apparently related to the distribution of the Augur Buzzard. To our knowledge, the two species do not co-occur in any locality on Mount Kenya. Verreaux’s Eagles are residents in the Gorges Valley, and can be seen in the Teleki Valley above 4200 m. We have both seen aggressive encounters between these two species. Whenever a Verreaux’s Eagle entered the main Teleki Valley, it was almost immediately set upon by one or two resident Augur Buzzards and was driven out. On one occasion TPY saw an Augur Buzzard drive a Verreaux’s Eagle away from a hyrax the latter had cornered on the ground. Verreaux’s Eagles reportedly feed mainly on hyrax, and this appears to be the case on Mount Kenya. TPY has seen several unsuccessful attacks on hyrax on the Teleki Valley, and flushed a Verreaux’s Eagle off a recently killed hyrax in the Gorges Valley. The hyrax had been disemboweled. Buteo (rufofuscus) augur Riippell (Augur Buzzard) Augur Buzzards are common above treeline on Mt Kenya. We have seen them in all valleys except Gorges. Alan Smith and TPY found a nest built between the rosettes of a Senecio keniodendron Fries & Fries plant five m tall at 4200 m in the Hinde Valley. The nest was composed mostly of twigs and was about 50 cm in diameter. It contained three eggs. The nest and eggs were first visited in June 1978, and had been abandoned by July. Chris Laine (personal communication) and MRE have also seen Augur nests in S. keniodendron trees. In August 1980 TPY observed a nesting pair at 4000 m in the Liki North Valley. Their nest was located high on a vertical rock face. They appeared to be feeding young. Our observations in the Teleki Valley and the examination of feeding perches (often dead S. keniodendron trees) and pellets indicate Augur Buzzards prey mainly on the rat, Otomys orestes. Coe ( 1 967) and Peter Hetz (personal communication) report cases of Augur buzzards taking (young) hyrax, but we have seen no aggressive or apprehensive encounters between these two species. Two pairs of Augur Buzzards lived in the upper Teleki Valley (above 4000 m) in the late 1970s, each pair with a territory estimated at three to four km2. Four pairs were present from 3900 m in the Teleki Valey in 1989. They defended these against other Augur Buzzards and against Verreaux’s Eagles (see above). There are several melanistic birds above treeline on Mt Kenya, but they do not appear to be more common there than lower on the mountain. Buteo buteo L. (Common Buzzard) Occasional visitor to the moorlands (D.A. Turner, personal communication) A. Forbes-Watson saw flocks above Sirimon Track at 3000-3200m in the 1950s (Bongo Woodley, personal communication). J. EANHS & Nat. Mus. 82 (202) December 1993 61 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya Buteo (tachardus) oreophilus Hartert & Neumann (Mountain Buzzard) Moreau (1944) reported them out over the moorlands as high as 3500 m. Melierax (Micronisus) gabar Daudin (Gabar Goshawk) MRE saw an individual in the Teleki Valley in February 1990. Milvus migrans Boddaert (Black Kite) Reported as an accidental visitor by Williams (1978), probably as a passage migrant. Falco naumanni Fleischer (Lesser Kestrel) Reported as a migrant visitor by Williams (1978). Falco tinnunculus L. (Kestrel) Regular visitor to the moorlands on migration (D.A. Turner, personal communication). Francolimis jacksoni Ogilvie-Grant (Jackson’s Francolin) Jackson’s Francolin has been described as mainly a forest bird, with the Montane Francolin replacing it in the moorlands (Williams and Amott 1980, Lewis and Pomeroy 1989). Flowever, we have found Jackson’s Francolin to be common along the Naro Moru Track up to 4000 m, and have not seen any other francolin in the alpine zone of Mount Kenya (see also Nievergelt et al. 1987). Jackson’s Francolin is also found on the Timau Track. At both localities, their calls can sometimes be heard late in the day. Francolinus psilolaemus Gray (Moorland [Montane] Francolin) This species is reported to be the Moorland Francolin on Mount Kenya by Williams and Arnott (1980), but we have seen only Jackson’s Francolin above treeline (see also Neivergelt et al. 1987). D.A. Turner reports that this species is “known only from the northern slopes of the mountain where the wheat fields approach the lower level of the moorlands, with no forest belt in between. The birds seen on Mt Kenya probably belong to the race theresae, reported as high as 3900m by Meinertzhagen (1937)”. This is the Shelley’s Francolin ( Francolinus shelleyi ) of earlier literature. Sarothrura affinis antonii Smith (Chestnut-tailed [White-spotted] Pygmy Crake) Reported by Meinertzhagen (1937) from 3700 m and by Moreau (1944), though there are no recent records from Mt Kenya (D.A. Turner, personal communication) Fulica cristata Gmelin (Red-knobbed Coot) TPY saw two pairs of Red-billed Coots on Lake Rutundu (3100 m) in September 1982. One pair had five very young chicks. Also reported from Lake Ellis by Moreau (1944). Vanellus melanopterus Cretzschmar (Black-winged Plover) Reported as an uncommon visitor to the moorlands by Williams (1978). Gallinago nigripennis Bonaparte (African Snipe) The African Snipe is often flushed from wet moorlands at altitudes up to 3700 m (according to D.A. Turner, to 4000m). We have seen them on the Timau, Sirimon, and Naro Moru Tracks. Tringa nebularia Gunnerus (Greenshank) Reported as a visitor by Williams (1978), probably as an annual visitor. J. EANHS & Nat. Mus. 82 (202) December 1993 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya 62 Tringa ochropus L. (Green Sandpiper) TPY saw a Green Sandpiper in September-November of each of 1979, 1980 and 1981, in the wetlands at the head of the Teleki Valley, probably on migration. Each visit lasted several days, and all may have been by the same individual. Actitis hypoleucos L. (Common Sandpiper) Tim Tear saw Common Sandpipers at Lake Rutundu in September 1982. Columba guinea L. (Speckled Pigeon) Pairs and small flocks of Speckled Pigeons appear to be residents in the Kazita and Hobley Valleys as high as 4000 m. Tyto capensis Smith (Cape Grass Owl) Reported from the marshy hollows on moorlands by Williams (1978). No recent records. Asio capensis Smith (African Marsh Owl) Reported from moorlands by Williams (1978). No recent records. Asio otus graueri L. (Long-eared Owl) Reported from thickets on moorlands by Williams (1978). D.A. Turner reports a “female collected at 3350 m in Hagenia woodand high on the Naro Moru Track 10 September 1961, while a large owl seen by the Park Warden (Bongo Woodley) flying over giant heath just above tree line in July 1992 may have been this species.” Bubo capensis mackinderi (Smith) Sharpe (Mackinder’s Eagle-owl) Common although rarely seen, Mackinder’s Eagle-owls were often heard at night as high as 4200 m. Sessions (1972) reports them from 2440^1270 m. When encountered by day, they are relatively unafraid of man, and can sometimes be approached to within a few feet (see also Sessions 1972). TPY has seen hyrax mobbing a Mackinder’s Eagle-owl that had landed in the vicinity of their burrows in the Teleki Valley. MRE’s analysis of ~50 pellets revealed a diet consisting of almost exclusively Otomys, with remains of shrews found in three pellets, and a duiker jaw in another. Sessions (1972) reports that “at the head of the Teleki Valley on Mt. Kenya I found nearly every pellet to contain the bone of the rock hyrax”. Caprimulgus poliocephalus Riippell (Montane Nightjar) TPY heard nightjars flying around a camp at 4200 m Hobley Valley at dusk in 1978. Tim Tear saw one at dusk while camping at 3500 m in the Kazita Valley in 1982. Other possibilities include C. abyssinica (reported by Moreau (1944) from 3500 m) and C. europaeus. D.A. Turner reports C. poliocephalus as being common in timberline and ericaceous habitats (personal communication), and both he and A. Forbe-Watson (personal communication) believe it to be the present species. Apus aequatorialus von Muller (Mottled Swift) Resident alongside Alpine Swifts on alpine cliffs (D.A. Turner, personal communication) Apus melba africana L. (Alpine Swift) Alpine Swifts can be found along streams and tarns to 4300 m or higher, particularly around Two Tam Hut. MRE has seen them in the Gorges Valley. J. EANHS & Nat. Mus. 82 (202) December 1993 63 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya Schoutedenapus (Apus) myoptilus Salvadori (Scarce Swift) Williams (1978) reports that Scarce Swifts “probably nests in crags on alpine moorlands”. D.A. Turner reports this species as commonly recorded over the moorlands (personal communication). Merops sp. (unidenitfied bee-eater) MRE has seen a flock of unidentified bee-eaters flying over the Teleki Valley in March 1989. Possibilities include the migrants M. apiaster L. (Eurasian Bee-eater) and M albicollis Vieillot (White- fronted Bee-eater). Hirundo rustica L. (Eurasian Swallow) Occurs regularly over the moorlands on migration (D.A. Turner, personal communication). TPY sighted a single individual in the upper Teleki Valley in early 1978. MRE caught one in the Teleki Valley in March 1989. This species is also listed by Williams (1978). Riparia paludicola duels (Vieillot) Reichenow (African Sand Martin) African Sand Martins are apparently not uncommon along river courses on the northern slopes. TPY found individuals at nesting burrows in the banks of the southern Kazita River at 3800 m in 1982. We have also seen this species along the Naro Moru river around 4100 m. There are abandoned nesting holes, apparently of this species, in the bare vertical stream banks. Psalidoprocne pristoptera Riippell (Black Rough-wing Swallow) Reported by Moreau (1944) as breeding only in the forest, but foraging as high as 4100 m. MRE has seen them foraging above the river in the Teleki Valley at 3900 m. Oenanthe oenanthe L. (Northern Wheatear) TPY sighted a single individual (on migration?) in the upper Teleki Valley in 1978. This species is also listed by Nievergelt et al. (1987) at 3800 m and by Williams (1978). Cercomela (Pinachroa) sordida ernesti (Riippell) Sharpe (Alpine or Mountain Chat) A common resident of alpine Mount Kenya, Mountain Chats are aggressive camp scavengers, and have been reported as high as 4570 m (Meinertzhagen 1937). Coe (1969) reported a nest in a tussock of Festuca pilgeri, and MRE found three nests in Festuca tussocks. TPY found three nests excavated from the dense layers of leaves retained around Senecio keniodendron trunks, and one between the living leaves of a S. keniodendron rosette. All of the nests we have found contained three eggs. The nests were 1-2 m above the ground. We have seen these birds eating both seeds and insects, and visiting the flowers of Lobelia deckenii keniensis (Young 1982). White pollen slashes can often be seen on their foreheads where L. d. keniensis flowering is commbn. Although reported to be an alpine endemic (Williams and Amott 1980), TPY has seen them as low as the town of Naro Moru (2000 m). Bradypterus cinnamoneus Riippell (Cinnamon Bracken Warbler) Reported by Moreau (1966) as high as 3850 m on Mount Kenya. Phylloscopus trochilus L. (Willow Warbler) MRE caught an individual in the Teleki Valley in February 1989. Cisticola hunteri Shelley (Hunter’s Cisticola) Hunter’s Cisticola, easily identified by its distinctive duet, is common throughout the lower alpine zone, wherever taller ericaceous scrub is available (up to 4000 m). J. EANHS & Nat. Mus. 82 (202) December 1993 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya 64 Anthus cervinus Pallas (Red-throated Pipit) An occasional visitor to the moorlands on migration (D.A. Turner, personal communication). An individual Red-throated Pipit visited the research camp in the Teleki Valley in September-October of 1979 and 1980. Macronyx sharpei Jackson (Sharpe’s Longclaw) Vagrant to the lower moorlands (D.A. Turner, personal communication). Reported by Raymond Hook as high as 3950 m (Moreau 1944). Onychognathus tenuirostris raymondi (Riippell) Meinertzhagen (Slender-billed Chestnut-winged Starling) Coe (1967) reports this species as a daily migrant from the forest to the alpine zone, usually in flocks of less than a dozen. This agrees with TPY’s observations in the years 1978-80. However, during the May-August 1977 field season, an entirely different situation prevailed. Flocks of up to 100 starlings were regularly seen in the upper Teleki Valley. These large flocks were seen mobbing the resident Augur Buzzards on several occasions. Three starling nests were discovered in the rocky cliffs of the Naro Moru river around 4100 m. One of these nests contained three eggs. When visited six weeks later (in late July), all these nests were empty. The starlings fed most frequently on Lobelia deckenii keniensis inflorescences. White pollen slashes were often seen on their foreheads (see also Meinertzhagen 1937). These inflorescences produce copious nectar and harbor large numbers of insects. The density of reproductive L. d. keniensis plants in 1977 was more than double the density in any of the next four years. In 1978-1983, L. d. keniensis inflorescences continued to be visiting starlings’ main food source in the upper Teleki Valley. The situation of 1977 was repeated in 1983-84, another mass flowering year for L. d. keniensis. In June 1983, TPY discovered an active starling nest at 4300 m in the Teleki Valley. It was located in a deep crevice at the top of a 15 m waterfall. Adults constantly flew into and out of this crevice, and the cries of chicks could be heard. These starlings normally nest around waterfalls in the forest (Williams and Amott 1980). It appears the Slender-billed Chestnut-winged Starlings, which normally nest in the forest and only make occasional forays into the alpine, are attracted by the mass flowering of L. d. keniensis in 1977 and 1983 to become short-term residents of the alpine zone, and even nest there. Jackson (1938) cites Mackinder as reporting breeding birds high in the Hohnel Valley (in 1899). A. Forbes- Watson found starling nests at 4200 m in the Hinde Valley in early 1955, under rock overhangs (Bongo Woodley, personal communication). MRE has seen small flocks flying up the valley, and larger flocks (up to 50 birds) flying down. He has seen them feeding around the cliffs below Lewis Glacier, and on the eastern side of Point Lenana. Nectarinia johnstoni johnstoni Shelley (Scarlet-tufted Malachite Sunbird) This sunbird is a common resident on Mount Kenya, visiting both Lobelia telekii and L. deckenii keniensis (Young 1982) throughout the alpine zone up to 4300 m, and Protea kilimandscharica in lower alpine areas on the northern slopes. They also hawk for small insects. Evans (1991) considers their breeding season to be December to April. Williams (1951) reports their nesting periods as January-February and July-August. Mackinder found nestlings in August (Moreau 1944). TPY has seen nesting birds in December, January, June, and July. MRE believes their breeding season to be highly variable, e.g. the end of November in 1989-90 an the beginning of January in 1990-91. In agreement with Williams (1951) and Coe (1967), these nests usually contained only a single egg or nestling. MRE found only four nests (out of 100) with two eggs. Above 4100 m, the most common nesting sites TPY found were Festuca pilgeri grass tussocks, although there was one in an old S. keniodendron inflorescence, and one in an Erica arborea shrub. For other examples of the first two, J. EANHS & Nat. Mus. 82 (202) December 1993 65 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya see Photos 1 and 2 in Williams (1951). MRE, working at lower elevations, found Erica shrubs to be the most common nesting sites, followed by Lobelia telekii inflorecences, grass tussocks, and Senecio keniodendron inflorescences. In his description of their nests, Williams reports that they were lined with ‘vegetable down’. TPY identified similar material as the felty pubescence from the undersides of Senecio keniensis (brassica) leaves., and deposited a sample nest at the National Museums of Kenya. Males are strongly territorial, without obvious seasonality. TPY estimated territory sizes around his research camp at 1000-2000 m2, larger than the 60-250 m2 reported by Williams (1951) and Coe (1961). Territory size may vary with environmental quality: territories were smaller in a year with high densities of L. telekii inflorescences than in a year with lower inflorescence density (Evans 1991). TPY found non-teiritorial males congregated in noisy groups of up to 10 on several occasions, accompanied by a few females and often some Mountain Chats (see also Williams 1951). Recently, an experimental study has been carried out in the Teleki Valley on the costs and adaptive values of the elongated tail feathers and bright red pectoral tufts (Evans 1991, Evans and Thomas 1992, Evans and Hatchwell 1992a, 1992b). Nectar inia tacazze jacksoni Stanley (Tacazze Sunbird) We have found Tacazze Sunbirds on Mount Kenya visiting both Lobelia spp at altitudes up to 4000 m. Serinus canicollis Swainson (Yellow-crowned Canary) The absence of these canaries from Alpine Mount Kenya puzzled Moreau (1944), who found it “astonishing to me because on the great mountains of northern Tanganyika the species ascends to the limits of vegetation and beyond”. In response, Raymond Hook stated: “In my opinion on both Kenya and Aberdares, they go as high as the Compositae go. I would have said that they ate practically every kind of compositae, including the Giant Groundsel [Senecio keniodendron ]” (Moreau 1944). TPY had never seen canaries above treeline before 1979. In 1978-79, the Giant Groundsel of Mount Kenya gregariously flowered, a rare occurrence (Smith and Young 1982). When these plants began to set seed, canaries began to appear in the lower alpine. By November 1979, there were hundreds of Yellow- crowned Canaries visiting fruiting stands of S. keniodendron up to 4200 m, feeding on the ground. A few were even seen feeding (on windswept seeds?) on the Lewis Glacier at 4700 m. The canaries disappeared from the alpine zone shortly thereafter. MRE saw occasional single birds in 1989-90 at 3800^1000 m. Serinus striolatus striolatus Riippell (Streaky Seedeater) The Streaky Seedeater is a common resident throughout the alpine zone. TPY found an abandoned nest composed mainly of grass atop a Festuca pilgeri grass tussock at 4200 m in the Teleki Valley. It contained one dead chick and one unhatched egg. MRE has also seen many nests started, but none successful. Corvus albicollis Latham (White-naped Raven) Alan Smith reported a White-naped Raven at the research camp (4180 m) in 1984. Moreau (1944) also reports them as high as 3660 m. This species is a common camp visitor on the Shira Plateau on Kilimanjaro. Mammals (43 spp) Graphiurus murinus raptor Dollman (African Dormouse) The Dormouse is commonly found in alpine huts, where its loud chattering vocalisations keep visitors awake at night. TPY saw an individual on the rocks below Lake Michaelson (3800 m) during the day in J. EANHS & Nat. Mus. 82 (202) December 1993 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya 66 September 1978. Dormice have been trapped up to 4180 m both on the northern slopes by Coe and Foster (1972) and in the Teleki Valley by TPY. Dendromus insignis percivali Heller (Striped Tree Mouse) Common in the alpine zone on all aspects at altitudes up to 4330 m (Moreau 1944, Coe and Foster 1972). This was the most common small rodent trapped at night at TPY’s research camp (4180 m). Bongo Woodley reports rodents from ‘Black Hole Bivvy’ (4550 m) above the Teleki Valley in September 1991 (personal communication), perhaps of this species. Lophuromys flavopunctatus Thomas (Harsh-furred Mouse) Coe and Foster (1972) report this species is common on the northern slopes as high as 4100 m. MRE had a family living near in his lavatory at 3900 m. Lemniscomys striolatus massaicus Pagenstecher (Striped Grass Mouse) Not previously reported from the alpine zone (but see Moreau 1944), TPY saw one on the rocks near Lake Michaelson, and MRE has seen them feeding on Lobelia telekii seeds in the Teleki Valley. Rhabdymus pumilo diminutus Thomas (Four-striped Grass Mouse) This species was trapped as high as 3800 m on the northern slopes by Coe and Foster (1972). Praomys (Hylomyscus) denniae (Climbing Wood Mouse) Reported by Fayad (1981) from alpine Mount Kenya, and perhaps seen by MRE in a Senecio keniodendron stand in 1989. Tachyoryctes splendens (rex) Riippell (Mount Kenya Mole Rat) The Mount Kenya Mole Rat is common throughout the northern slopes and the Hinde Valley at altitudes up to 4050 m on the northern slopes, and 3750 elsewhere (Jarvis and Sale 1971; personal observation) Their mounds are up to 6 m in diameter, and can be recognised at a distance by the strikingly different vegetation growing on them, dominated by Alchemilla spp (Jarvis and Sale 1971, Coe 1969, Coe and Foster 1972; personal observation). In some localities, the mounds are virtually overlapping. The animals themselves are rarely seen. There has been debate recently about whether the origin of ‘mima- type’ mounds in Kenya is due to the activities of termites or mole rats (Gakahu and Cox 1984, Martin 1988). There are no termites above treeline on Mount Kenya or the Aberdares, but this does not mean that all alpine mounds are caused by Mole Rats. Frost-thaw processes in the alpine zone can also produce vast areas of mounded terrain (Baker 1967). They feed mainly on roots and subterranean stems. They appear to be the only mole rat in Kenya (out of three genera) that stores food. Roots and leaves of the alpine plant Haplosciadium abyssinicum were found in a food store at 3600 m on Mount Kenya (Jarvis and Sale 1971). Otomys entrances into Tachyoryctes burrows have been found on Mount Kenya (Jarvis and Sale 1971). Otomys orestes orestes Thomas (Groove-toothed Rat) The Groove-toothed Rat is probably the most common alpine rodent on Mount Kenya, occurring at all altitudes and aspects. It is the most common food item identifiable in the scat and pellets of Leopard, Augur Buzzard, and Mackinder’s Eagle-owl. These rodents are often seen during the day along their conspicuous runs. TPY has seen several chases between adults, and they may be territorial. Reported as high as 4750 m (Moreau 1944, Coe 1967). J. EANHS & Nat. Mus. 82 (202) December 1993 67 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya Otomys tropicalis tropicalis Thomas (Groove-toothed Rat) This second alpine Otomys species is reported from 2280^4175 m (Hollister 1919). Coe and Foster (1972) found only Otomys orestes, and question the presence of O. tropicalis. Hystrix sp. (Porcupine) Fayad (1981) reports alpine porcupine, and Moreau (1944) found a quill at 3500 m above Nanyuki. Lepus sp. (Hare) The existence of a hare in the lower moorlands of Mount Kenya was suggested by Raymond Hook (Moreau 1944) who twice found their dung at 3350 m. Phil Snyder has reported seeing a hare just above timberline on several occasions. He described them as relatively large, and yellowish. (Unidentified bat) An unidentified species of bat occurs at Lake Rutundu (3000 m), perhaps the African Pipistrelle (Pipistrellus nanus) reported by Williams (1978) from high forest. Crocidura allex alpina Heller (Alpine Pygmy Shrew) Apparently common in the alpine zone. Coe and Foster (1972) report it as high as 4100 m on the northern slopes. TPY trapped an individual at night at 4180 m in the Teleki Valley in early 1978, and saw another in daylight at 3800 m in early 1980. Crocidura fumosa Thomas (Dusky Shrew) Coe and Foster (1972) trapped a single individual at 3800 m on the northern slopes, and report it to be common at lower altitudes. They also list C. turba. Crocidura turba zaodon Osgood (Dusky Shrew) Coe and Foster (1972) suggest that some of their Crocidura specimens may have been this subspecies. Reported by Moreau (1944) as high as 3260 m on Mount Kenya in forest. Myosorex (Surdisorex) polulus Hollister (Mole Shrew) Trapped by Coe and Foster (1972) at 3920 m on the northern slopes. Moreau (1944) reported it at altitudes of 2750-3680 m. Kingdon (1974), reporting it at altitudes of 2800-3600 m, states that it is the only endemic mammal on Mount Kenya. Cercopithecus mitis Wolff (Sykes’ Monkey) In November 1979, TPY was shown a Sykes’ Monkey found in the rocks below the Darwin Glacier (4600 m) by Vince Fayad. Its head and limbs were missing, and it was in a mummified state. Local climbers state that it had been at that location for at least a year. In December 1990, another Sykes’ Monkey was reported from Top Hut (4800 m) by Clive Ward. It was complete, and only recently dead. We cannot explain the presence of these individuals (or the following) so high on the mountain. Colobus polykomos Oken (Abyssinian Black-and-white Colobus Monkey) A ‘mummified’ colobus on Point Peter (4700 m) has been reported by several mountaineering parties. It can only be this species. Lycaon pictus Temminck (African Wild Dog or Hunting Dog) Coe (1967) reports several packs as high as 4250 m, preying upon zebra and eland. As recently as the 1 960s packs of Wild Dogs were taking sheep off the moorlands above Embori Farm, but not in the J. EANHS & Nat. Mus. 82 (202) December 1993 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya 68 1970s or early 1980s (Bill Halstead, personal communication). In January 1992, John Temple reported Wild Dogs on the Timau (3000 m) and Chogoria (3500 m) tracks (personal communication). Canis mesomelas Schreber (Black-backed Jackal) An unspecified species of jackel was reported by Raymond Hook at altitudes of up to 3660 m (Moreau 1944), and Fayad (1981) listed that Black-backed Jackals from the alpine zone. lctonyx striatus Perry (Zorilla) Zorilla are apparently residents of the alpine zone of Mount Kenya. Raymond Hook reported them as high as 3660 m (Moreau 1944). Coe (1969) collected one at 4270 m on the northern slopes that had been feeding on Otomys rats. On the night of 7 October 1979, a visitor to TPY’s research camp (4180 m) saw a live Zorilla. N. Barrah saw a Zorilla at the nearby Teleki Valley Ranger Station in September 1993, also at night (Bongo Woodley, personal communication). TPY found a Zorilla skull in the rocky moraine below Mackinder’s Camp (4150 m) in 1978. This is now on deposit at the Department of Osteology, National Museums of Kenya (ref. # OM 6432). MRE saw one at his camp near this same site in February 1989. Judging by the repeated presence of a distinctive odor, especially early in the morning, TPY believes that a Zorilla lived in the rocky outcrop below the 12,000 foot rain gauge (at -3600 m) on the Naro Morn Track, in the late 1970s. Genetta tigrina Schreber (Large-spotted Genet) Reported by Fayad (1981) from alpine Mount Kenya. Herpestes paludinosus G. Cuvier (Marsh Mongoose) Coe (1967, 1969) reported a pair of Marsh Mongooses from 3500 m in the Gorges Valley. Herpestes sanguineus Ruppell (Black-tipped [Slender] Mongoose) Fayad (1981) reports this species from the alpine zone. Crocuta crocuta Erxleben (Spotted Hyaena) Spotted Hyaena are apparently regular, if infrequent, visitors to the alpine zone. TPY has seen seen tracks as high as 4000 m in the Teleki Valley, and they have been reported much higher. Tracks that TPY saw on the Lewis Glacier (4800 m) in 1979 were probably also of hyaena. In 1986, Bongo Woodley found tracks coming up from the Teleki Valley, crossing near Austrian Hut (4800 m) and descending into Hobley Valley (personal communication). Bill Halstead reports that they visit the moorlands particularly during the rains. MRE had one around his camp (4100 m) daily in February-April 1989. Acinonyx jubatus Schreber (Cheetah) Nigel Trent reported seeing a cheetah at around 3500 m along the Timau Track in 1982. There have also been sightings of Cheetah from the Sirimon Track at and above 4000 m (Campbell 1983). Felis serval Schreber (Serval Cat) Serval Cats are common in the alpine zone of the Aberdare Mountains, and may also occur on Mount Kenya (see Moreau 1944). Felis lybica Forster (Wild Cat) The Wild Cat is known from the northern slopes up to 3800 m (Coe and Foster 1972), and may be common there (Coe 1969). J. EANHS & Nat. Mus. 82 (202) December 1993 69 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya Panthera leo L. (Lion) Lions are apparently occasional visitors to the alpine, and may stay for several weeks, such as the individuals that were seen repeatedly around the Sirimon roadhead in 1977. It is also possible that this is part of a small resident population. TPY has seen pug marks at 3600 m on the Naro Moru Track. Panthera pardus L. (Leopard) Leopard are resident in the alpine zone of Mount Kenya. We have found their tracks, droppings and cave shelters throughout the alpine. TPY has encountered a leopard only once in the alpine zone, but there have been numerous sightings of leopards by others (Martin Otieno, John Omira Miluwi, Phil Snyder, Lew Awodey), including a melanistic individual in the Teleki Valley (John Omira Miluwi). MRE has seen the pug marks of a female and her cub in the Teleki Valley. They apparently cross passes as high as 4800 m (Coe 1967). TPY has found two duikers (one in the Teleki Valley at 4300 m, one in the Hinde Valley at 4100 m) apparently killed by leopards. Examination of several leopard dens (overhanging rocks) revealed large numbers of hyrax bones. Scats collected on Mount Kenya contained occasional hyrax bones and often the remains of Groove-toothed Rats. Procavia johnstoni mackinderi Thomas (Mount Kenya Rock Hyrax) Hyrax are the most conspicuous mammals of alpine Mount Kenya. There are several hundred in the upper Teleki Valley alone, where TPY estimates the average density to be in the neighborhood of 20-100 animals per km2. A description of their habits is found in Coe (1962) and Sale (1965). Kingdon (1971) reports them as occurring at 3200-4650 m. They have been reported as high as 4700 m (Moreau 1944) and even higher (Dorst and Dandelot 1972), but we have seen no evidence of them living higher than 4300 m. Although most individuals are dark brown, there was one adult male on the south side of the Teleki Valley at 4200 m during the 1980s that had very pale fur. ‘White’ hyrax ( Heterohyrax brucei ) also occur in the Serengeti (Hoeck 1982). The Mount Kenya subspecies has evolved darker and longer fur than its lowland relatives, apparently as adaptations to the extreme cold (Coe 1962, 1967; Sale 1967). In addition, they can often be seen basking in the sun and huddling together, traits they share with other Rock Hyrax. On Mount Kenya, the distribution of rock hyrax is delineated by the presence of appropriate shelter — the rocky moraines in which hyrax make their homes. We have never seen hyrax more than 150 m from these rocky shelters, and believe that their overall numbers are limited by the availability of these moraines. TPY did once find tracks of a hyrax high on a ridge between the Teleki and Hoehnel Valleys far from any established colony. This may have been a dispersing male, if comparisons to ecologically similar marmots are appropriate, and these animals are likely to be easy prey in the open (Van Vuren 1990). TPY has found one small isolated moraine that was once inhabited by hyrax, judging by the skulls present, but that no longer supported them. Local extinction of small isolated populations may not be a rare event in hyrax (Hoeck 1982) or other species (Smith 1980, Berger 1990), and more isolated areas may be only rarely colonised (Smith 1980). In the vicinity of inhabited moraines the vegetation is closely cropped, and hyrax colonies can be recognised from a distance by the bright green vegetation surrounding the rocky outcrops, which contrasts with the pale, uncropped alpine grassland. Mahaney and Boyers (1983) suggest that this close cropping by hyrax and groove-toothed rats can produce bare areas of solifluction desert by allowing greater cooling of the soil. Hyrax eat a wide variety of plant species, but seem to avoid certain plants ( Anthoxanthum nivale, Sedum ruwenzoriense, Carduus keniensis ) that are therefore common near active burrows (see also Coe 1962). Mahaney and Boyer (1983) examined hyrax dung microscopically, and found that 92% of the identifiable fragments were monocotyledons, and only 8% were dicotyledons. J. EANHS & Nat. Mus. 82 (202) December 1993 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya 70 Experimental removal of spines from C. keniensis leaves renders them readily palatable to hyrax (Young and Smith, in press). Descriptive and experimental evidence indicates that hyrax may limit the distribution of the shrub Alchemilla argyrophylla around colonies (Young and Smith, in press). They occasionally feed on the toxic Lobelia spp (Coe 1962, Young 1985), and even Giant Senecio (Young and Smith, in press), especially during dry periods. Rock hyrax are mainly diurnal and feed mostly at mid-morning and mid-afternoon (Sale 1965, Coe 1969). They seem to avoid rainy weather, but TPY has occasionally seen hyrax feeding both on moonlit nights (see also Coe 1962 and Dorst and Dandelot 1972) and during rain, particularly if there was rain during the normal afternoon feeding period (4-5 PM). The most important predators of alpine rock hyrax on Mount Kenya are undoubtedly leopards, which apparently stalk colonies (John Omira Miluwi, personal communication). Verreaux’s Eagles also take hyrax in certain localities. Mackinder thought that Mackinder’s Eagle-owls were important predators (Moreau 1944), but we have examined numerous Eagle-owl pellets and found no identifiable hyrax remains. Coe ( 1 967) reports seeing an Augur Buzzard take a hyrax, but we have seen no aggressive behavior between these species. TPY has observed several females that had lost all of the hair on their rumps, and the skin itself was raw to the point of bleeding. We do not know what causes this, but it matches to symptoms of sarcoptic mange reported by Hoeck (1982) in Serengeti hyrax. As elsewhere (Sale 1965, Hoeck 1982) the hyrax colonies of Mount Kenya consist of a single adult male and differing numbers of adult females and immatures. TPY has have seen several fights between males, often resulting in the loss of blood and substantial clumps of hair. Older males are often heavily scarred around the face and ears. Both sexes contest over food. In these encounters, animals will often turn their backs to each other and try to push the other away. Female Mount Kenya Rock Hyrax give birth synchronously within a colony, with up to six litters appearing within a period of less than two weeks. Most commonly, these births were in June or December. Each colony has only one birth season per year, and it is usually the same month each year. Similarly, in the Serengeti, individual colonies had one birth season per year (Hoeck 1982). For P. johnstoni, this was in March to May. For Heterohyrax brucei, individual colonies had birth peak in either May-June or December-January, as in the Mount Kenya Procavia. An interesting exception occurred in the colony near TPY’s camp in 1979. In the previous year, the young appeared in June, but the resident male was displaced by another in late 1978, and the subsequent brood was not bom until September 1979. Hyrax have very long gestation periods (7.5-8 months), qnd it is possible that the change in dominant males disrupted the pregnancies of the colony’s females. On at least two occasions, hyrax have been seen giving birth above ground during the day (John Omira Miluwi, personal communication, and an anonymous observer, personal communication). This is in conflict with earlier information about hyrax (Sale 1965), and needs confirmation. However, given the likely low sanitation of hyrax burrows and the low risk of diurnal predation, such birthing behavior may be reasonable. In both cases, the observers were attracted by the many adult and yearling hyrax that gathered around the scene. Young hyrax are very playful, and seem to spend the majority of their time in play. They suckle for at least several months. TPY has seen one pair suckling at the age of nine months. Unless mothers tolerate other females’ young. Mount Kenya hyrax do have at least occasional litters of more than one (but see Coe 1967). Coe (1962) reported two kinds of vocalizations, but we distinguish at least seven: • A long, loud call, usually beginning with coarse mewing sound and ending with a series of coughing noises, is given by multiple individuals when people (personal observation) or leopards (J. Omira, personal communication) are seen at a distance. Upon hearing this call, mountain J. EANHS & Nat. Mus. 82 (202) December 1993 71 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya rangers have initiated successful scans of nearby slopes for leopard. When foraging hyrax hear this call, they stop and look and then often run back to the safety of their burrows. This call and the next are the ones most often heard by people in the vicinity of colonies. • A very similar ‘long call’ is given by territorial males, particularly in the evening, and may serve to advertise occupancy of colonies. These two calls are distinguished by their context and participants. • A loud, sharp call is given when an eagle attacks. This produces an instantaneous reaction among foraging hyrax, which dive for the nearest cover. • When foraging in close proximity, adult hyrax make a quiet, grunting ‘contact call’. • Juveniles make high-pitched squeaks in a variety of situations. • Adult females also make a high-pitched squeaking noise when involved in antagonistic interactions with other adults females or the dominant male. This may be an appeasement vocalisation. • During serious fights and chases, adult males (and sometimes females) make loud grunting noises. Fourie (1977) has documented over twenty noises made by captive P. capensis, of which three were limited to females giving birth, and two were not vocalisations (sneezing and teeth gnashing). Loxodonta africana Blumenbach (African Elephant) Elephants are apparently regular visitors to alpine Mount Kenya. Moreau (1944) reports two well-worn alpine elephant trails on the alpine northern slopes. TPY has seen them eating Giant Groundsel ( Senecio keniodendron) at 4000 m in the Teleki Valley. Numerous visits to the alpine zone of northwest Mount Kenya occurred in the late 1970s and early 1980s (Mulkey et al. 1984). It is not known whether this is strictly a recent phenomenon. A dead elephant was found at 4600 m in the Hinde Valley in 1944 (Taffe 1944). This animal was identified as a female by Michael Rainey in 1979 (personal communication). Diceros bicornis L. (Black Rhinoceros) Rhinos have been sighted on several occasions above the treeline. Lew Awodey found a recently dead individual (horn intact) near the Sirimon river at 3700 m in 1978 (personal communication). In 1983, Nigel Trent reported three resident animals above Timau (personal communication). They may be extinct there now, although they still occur in.the forest. Equus burchelli Gray (Burchell’s Zebra) There is at least one resident herd of zebras on the northern slopes, and probably several. They can be commonly seen along the Timau Track. TPY has seen zebra tracks throughout the northern slopes, as far south as the northern ridge of the Hinde Valley and as high as 3700 m. Coe (1969) reports their droppings as high as 4300 m. An aerial survey in June 1993 counted 299 zebras between Mbara Crater and the Kazita River (Bongo Woodley, personal communication). J. EANHS & Nat. Mus. 82 (202) December 1993 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya 72 Alcephalus buselaphus cokii Pallas (Kongoni or Coke’s Hartebeest) Seen along the Timau Track repeatedly in 1980 as high as 3700 m (Helen Young, personal communication). Raymond Hook found a skeleton at 4000 m on the northern slopes (Moreau 1944). Cephalophus nigrifrons Gray (Black-fronted Duiker) Williams (1978) reports this species in bamboo zone and on moorlands. Silvicapra grimmia altivallis L. (Grimm’s or Bush Duiker) Grimm’s Duikers are resident throughout the alpine, and are most common on slopes were Alchemilla spp are found (see Young and Peacock 1992), perhaps because these shrubs provide cover (Coe 1969) or food (King 1975). On a single day in 1978, TPY saw seven individuals in the Hinde Valley, and MRE from one position saw twelve in the lower Teleki Valley. They usually occur singly or in pairs. On the Aberdares and Kilimanjaro, they occur at densities of about two per km2 (King 1975). MRE has seen prolonged chases between individuals, running round and round in a particular area. TPY found two Grimm’s Duikers that were apparently victims of leopard predation, one each in the Teleki (4300 m) and Hinde (4100 m) Valleys. In January 1992, John Temple reported a dead duiker (probably this species) high (-4900 m) on the Lewis Glacier (personal communication). Oreotragus oreotragus Zimmermann (Klipspringer) Reported by Williams (1978) and Fayad (1981), but we tend to agree with Moreau (1944) that they do not occur in the alpine zone of Mount Kenya. Dorst and Dandelot (1972) report that they occur up to 13,000 ft in Ethiopia, and Haltenorth and Diller (1980) up to 4000 m. Raphicerus campestris Thunberg (Steinbuck) Although Moreau (1944) believed that Steinbuck were merely visitors to the northern slopes, we concur with Coe (1969) and Coe and Foster (1972) that they are probably resident. Whether they “replace Grimm’s duiker” there, as the latter authors suggest, is open to question. TPY has seen both steinbucks and duikers at 3000-3500 m along the Timau Track. Boocercus eurycerus Ogilby (Bongo) In 1979, Vince Fayad and TPY found the skeleton of an adult male bongo at ‘American Camp’ at 4300 m in the Teleki Valley. We did not find the skull, and identification was made from a femur and lower jaw by the Department of Osteology of the National Museums of Kenya (Nina Mudida, personal communication). This location is far from the normal forest/bamboo habitat of bongo, but Haltenorth and Diller (1980) say, “seasonal wandering around to particular feeding places is possible; also from forest to forest or mountain to mountain”. Taurotragus oryx Pallas (Eland) There are up to several herds of eland on the northern slopes. A large herd can often be seen at the Sirimon track roadhead. TPY has seen young in these herds, and believes they are breeding residents (but see Coe 1969). A herd of 70 elands, including many subadults, was seen from the air at ‘Brigands Retreat’ in late 1990 (Bongo Woodley, personal communication). Syncerus caffer Sparrman (African Buffalo) African Buffalos are regular visitors to alpine Mount Kenya. The carcass of a bull was found at 4700 m (Ross 1911). They are common visitors to Lake Hoehnel (4200 m), and Martin Otieno has seen them there on several mornings. We have seen droppings as far up the Teleki Valley as 4150 m. Buffalo tracks and droppings can be seen on all game trails on the mountain, both in forest and alpine. Buffalo J. EANHS & Nat. Mus. 82 (202) December 1993 73 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya maintain a trail between the Teleki and Hausberg Valleys at an altitude of about 4100 m. Their visits to the Teleki Valley are mainly nocturnal, but M. Peacock and TPY saw a lone bull at 3900 m on the north slope of the valley at midday in January 1980. DISCUSSION The alpine environment on Mount Kenya is severe, with nightly freezing likely throughout the year, common dry periods, and low plant productivity (Hedberg 1964, Coe 1967, Smith and Young 1987). Nonetheless, a wide variety of vertebrate species occur there, and many seem to thrive. Some do so by using the alpine ecosystem only occasionally; others are permanent residents. Residents exhibit a variety of evolutionary adaptations and behaviours that help them to cope with the unique challenges of this environment. These include longer hair, darker colouration, basking in available sunlight, diurnal foraging patterns, and perhaps lowered reproductive rates. Alpine Mount Kenya is a semi-isolated island ecosystem. The alpine habitats are very different from the forest surrounding them. This has led to a considerable amount of endemism and vicariance among the alpine plants (Hedberg 1957) and invertebrates (Salt 1987), but not among the alpine vertebrates. Of the species in Table 1, only the Mole-shrew (Myosore x polulus ) is endemic to Mount Kenya, with vicariants on other East African mountains (Kingdon 1974). The Mount Kenya Rock Hyrax ( Procavia johnstoni mackinderi) is an endemic subspecies. Although the alpine habitats of Mount Kenya are island-like, there are three ways in which the alpine zone is accessible to animals from other ecosystems. First, birds can fly over the forest to reach the alpine zone. These may be altitudinal migrants (Yellow-crowned Canary), regional migrants (Open- billed Stork) or palearctic migrants ( e.g ., Harriers, Wheatear, European Swallow, Willow Warbler), as well as occasional visitors (e.g.. Black Kite, Secretary Bird). The list of avian migrants and visitors is likely to increase in the future. Second, resident forest species may venture out into the alpine moorlands, and even become partial residents there. Some of these visitors are difficult to explain, such as the Bongo, the two Sykes’ Monkeys, and the Colobus Monkey. For others, the lower alpine zone may provide a suitable foraging habitat (e.g., Mountain Buzzards, Black-fronted Duikers, Green Ibises). Elephants, Buffalos and Slender-billed Starlings appear to be regular visitors from the forest to high in the alpine zone, with the latter occasionally nesting there. Third, the treeless gap on the northern slopes has provided savanna species the opportunity to visit and colonise alpine Mount Kenya. This may be why Mount Kenya is the only alpine area in East Africa with populations of rock hyrax. It is possible that some of the larger mammals on the nothem slopes were local migrants trapped by increasing development at lower altitudes, but their continuing success in the alpine suggests that this need not be the case. These species include Lion, Zebra, Eland, and Kongoni. Two highly endangered species. Black Rhino and Wild Dog, apparently still occur on Mount Kenya, the former perhaps restricted to the forest. There are several other species whose status is unclear. These include those that have only rarely been reported (either because they are rarely seen or because they are nocturnal) or may have been misidentified: Striped Grass Snake, Green Ibis, Maccoa Duck, Pygmy Crake, Montane and Shelley’s Francolins, Long-eared Owl, Cape Grass Owl, Marsh Owl, an unidentified nightjar, Sharpe’s Longclaw, Climbing Mouse, one of the Groove-toothed Rats (Otomys tropicalis), an unidentified hare, Porcupine, an unidentified bat. Jackal, Marsh Mongoose, Slender Mongoose, Genet Cat, and Serval. As a conservative estimate, the confirmed resident vertebrate fauna of alpine Mount Kenya is as follows: two introduced fish, two frogs, two lizards, two chameleons, one snake, nineteen birds, seven rodents, three shrews, and ten larger mammals, for a total of 48. Two areas of the mountain in particular J. EANHS & Nat. Mus. 82 (202) December 1993 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya 74 are in need of more detailed information. The northeastern (and southwestern) slopes have only rarely been visited by biologists, who have concentrated their attention on the areas along the main tracks (Naro Moru, Sirimon, and to a lesser degree Chogoria). Also, the areas just above the forest, the ericaceous and lower alpine habitats, have been less studied relative to higher alpine areas. It is hoped that future studies of these areas will both increase and refine our knowledge of Mount Kenya’s alpine vertebrates. ACKNOWLEDGEMENTS Over the years we have benefitted from the knowledge, the assistance, and the hospitality of numerous people, among them John Omira Miluwi, Martin Otieno, Lynne Isbell, Bill Woodley, Bongo Woodley, Phil Snyder, Tim Tear, Lee Fonvielle, David Kariuki, Vince and Chris Fayad, Helen Young, Mary Peacock, Ron and Rae Nelson, Simon and Libba Marion, Paul and Margie Robinson, Michael Rainey, Iain Allen, Alan Smith, Nigel Trent, John Temple, Alec Forbes-Watson, Tom Madsen, John Imhof, Alex McKay, Nina Mudida, Steve Mulkey, and Steven Wahome. We also thank Lynne Isbell, W. Mahaney, and two anonymous reviewers for assistance with the manuscript. D.A. Turner provided numerous detailed notes on the birds. This work was supported in part by the Arthur K. Gilkey Fund of the American Alpine Club to TPY and a SERC studentship to MRE. REFERENCES Baker, M.J. 1967. Geology of the Mount Kenya area. Geological Survey of Kenya, Report #79. Nairobi. Berger, J. 1990. Persistence of different-sized populations — an empirical assessment of rapid extinctions in Bighorn Sheep. Cons. Biol. 41:91-98. BRITTON, P.L. 1980. Birds of East Africa. East Africa Natural History Society, Nairobi. Campbell, L. 1983. Moorland cheetah. Bull E. Afr. Nat. Hist. Soc. 12:34-35. COE, M.J. 1961. Notes on Nectarina johnstoni on Mount Kenya. The Ostrich 32:101-103. COE, M.J. 1962. Notes on the habits of the Mount Kenya hyrax ( Procavia johnstoni mackinderi Thomas). Proc. Zoo/. Soc. Lond. 138:639-644. COE, M.J. 1967. The ecology of the alpine zone of Mount Kenya. W. Junk, The Hague. COE, M.J. 1969. Microclimate and animal life in the equatorial mountains. Zoologica Africana 4:101-128. COE, M.J. & J.B. Foster. 1972. The mammals of the northern slopes of Mount Kenya. J. E. Afr. Nat. Hist. Soc. 131:1-18. Coe, M.J. & J.B. Sale. 1971. The flora and fauna of Mount Kenya and Kilimanjaro. In: Mitchell, J. (ed.). Guide to Mount Kenya and Kilimanjaro (2nd ed.) Mountain Club of Kenya, Nairobi. COPLEY, H. 1953. The introduction of the American Brook Trout ( Salvinus fontinalis) to Kenya. J. E. Afr. Nat. Hist. Soc. 93:35-36. DORST, J. & P. DANDELOT. 1972. A field guide to the larger mammals of Africa. Collins, London. Evans, M.R. 1991. The size of adornments of male scarlet-tufted malachite sunbirds varies with environmental conditions, as predicted by handicap theories. Anim. Behav. 42:797-803. Evans, M.R. & B.J. Hatchwell. 1992a. An experimental study of male adornment in the scarlet- tufted malachite sunbird: I. The role of pectoral tufts in territorial defence. Behav. Ecol. Sociobiol. 29:413-421. J. EANHS & Nat. Mus. 82 (202) December 1993 75 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya Evans, M.R. & B.J. Hatchwell. 1992b. An experimental study of male adornment in the scarlet- tufted malachite sunbird: II. The role of the elongated tail in mate choice and evidence for a handicap. Behav. Ecol. Sociobiol. 29:421^127. Evans, M.R. & A.L.R. Thomas. 1992. The aerodynamics and mechanical effects of elongated tail feathers in the scarlet-tufted malachite sunbird: measuring the cost of a handicap. Anim. Behav. 43:337-347. FAYAD, C.C. 1981. The flora and fauna of Mount Kenya and Kilimanjaro. Pp. 39-65 in Allen, I. (ed.), Guide to Mount Kenya and Kilimanjaro (3rd ed.) Mountain Club of Kenya, Nairobi. FOURIE, P.B. 1977. Acoustic communication in the rock hyrax, Procavia capensis. Z. Tierpsychol. 44:194-219. Gakahu, C.G. & G.W. Cox. 1984. The origin of mima mound terrain in Kenya. Afr. J. Ecol. 22:3 1-42. HAETENORTH, T. & H. DlLLER. 1980. A field guide to the mammals of Africa including Madagascar. Collins, London. Hebrard, J.J. 1981. Sympatry among Kenyan chameleons. Abstract. Am. Zool. 21:920. HEDBERG, O. 1951. Vegetation belts of the East African mountains. Svensk. Botanisk Tidskrift 45:140-202. HEDBERG, O. 1957. Afroalpine vascular plants: a taxonomic revision. Symbolae Botanica Uppsalensis 15:1-249. HEDBERG, O. 1964. Features of afroalpine plant ecology. Acta Phytogeographica Suecica 49:1-144. HOECK, H.N. 1982. Population dynamics, dipsersal and genetic isolation in two species of hyrax. Z. Tierpsychol. 59:177-210. Hollister, N. 1919. East African mammals in the United States National Museum. Pt. II, Rodentia, Lagomorpha, and Tubulidentata. Bull. U.S. Nat. Mus. 99. JACKSON, F.J. 1938. Birds of Kenya Colony and the Uganda Protectorate. Gurney & Jackson, London. JARVIS, J.U.M. & J.B. SALE. 1971. Burrowing and burrow pattern of East African mole rats. J. Zool. 163:451^179. King, D.G. 1975. The afro-alpine grey duiker of Kilimanjaro. J. E. Afr. Nat. Hist. Soc. 152:1-9. KlNGDON, J. 1971. East African Mammals. Vol. I. Academic Press, London. Kingdon, J. 1974. East African Mammals. Vol. II. Academic Press, London. Lewis, A. & D. POMEROY. 1989. A bird atlas of Kenya. A. A. Balkema, Rotterdam. LOVERIDGE, A. 1957. Check list of the reptiles and amphibians of East Africa. Bull. Mus. Comp. Zool. 117:151-362 (XXXVI). Mahaney, J. & M.G. Boyer. 1983. Herbivores and landform origin in the Mount Kenya Afroalpine area. E. Afr. agric. For. J. 49:41-52. Martin, G.H.G. 1988. Mima mounds in Kenya — a case of mistaken identity. Afr. J. Ecol. 26:127-133. MEINERTZHAGEN, R. 1937. Some notes on the birds of Kenya Colony, with especial reference to Mount Kenya. Ibis 14:731-760. Mills, D. 1971. Salmon and Trout. Oliver and Boyd. Edinborough. MOREAU, R.E. 1944. Mount Kenya: a contribution to the biology and bibliography. J. E. Afr. Nat. Hist. Soc. 18:61-92. MOREAU, R.E. 1966. The bird faunas of Africa and its islands. Academic Press, NY. MULKEY, S.S., A.P. SMITH, & T.P. YOUNG. 1984. Predation by elephants of Senecio keniodendron in the alpine zone of Mount Kenya. Biotropica 16:246-248. Nievergelt, B., G. Anzenberger, B. Stucki, & R. Zingg. 1987. Ein zoologisch/botanisches Hohenprofil in der Westflanke des Mount Kenya. Vein. Natur. Gesell. Zurich. 132:53-64. REHDER, H., E. Beck & J.O. Kokwaro. 1988. The afroalpine plant communities of Mount Kenya (Kenya). Phytocoenologia 16:243-263. J. EANHS & Nat. Mus. 82 (202) December 1993 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya 76 REHDER, H., E. Beck & J.O. KOKWARO. 1989. The afroalpine plant communities of Mount Kenya (Kenya). Comment of the vegetation map of Mount Kenya. Phytocoenologia 17:287-288. REHKER, J. 1989. Bibliography of East African mountains. Laikipia Reports 13. Institute of Geography, University of Berne, Switzerland. ROOSEVELT, T. 1910. African Game Trails. Charles Scribner’s Sons, NY, NY. ROSS, W.M. 1911. Two finds on Mount Kenya. J. E. Afr. Nat. Hist. Soc. 1 1 :60. SALE, J.B. 1965. The feeding behaviour and ecology of rock hyraces (Genera Procavia and Heterohyrax ) in Kenya. E. Afr. Wildl. J. 3:1-18. SESSIONS, P.H.B. 1972. Observations on the Mackinder’s Eagle-Owl Bubo capensis mackinderi Sharpe on a Kenya Farm. J. E. Afr. Nat. Hist. Soc. 138:1-20. SHARPE, R.B. 1900. On the birds collected during the Mackinder expedition to Mount Kenya. Proc. Zool. Soc. Lond. 1900:596-609. SMITH, A.T. 1980. Temporal changes in insular populations of the pika ( Ochotona princeps). Ecology 61:8-13. SMITH, A.P. & T.P. Young. 1982. The cost of reproduction in Senecio keniodendron , a giant rosette plant on Mount Kenya. Oecologia 55:243-247. SMITH, A.P. & T.P. YOUNG. 1 987. Tropical alpine plant ecology. Ann. Rev. Ecol. Syst. 18:1 37-1 58. Sprawls, S. 1978. A checklist of the snakes of Kenya. J. E. Afr. Nat. Hist. Soc. 167:1-18. Taffe, M.P. 1944. Elephant on Mount Kenya. J. E. Afr. Nat. Hist. Soc. 18:93. Thomas, O. 1900. List of mammals acquired by Mr. Mackinder during his recent expedition to Mount Kenya, East Africa. Proc. Zool. Soc. Lond. 76:173-180. Van Vuren, D. 1990. Dispersal of yellow-bellied marmots. PhD Diss. University of Kansas, Lawrence. WATSON, R.W.M. 1988. Trout in Kenya. Bull. E. Afr. Nat. Hist. Soc. 17:45^19. WILLIAMS, J.B. 1951. Nectarinia johnstoni: a revision of the species, together with data on plumages, moults and habits. Ibis 93:589-595. WILLIAMS, J.B. 1978. A field guide to the national parks of East Africa. Collins, London. WILLIAMS, J.B. & N. ARNOTT. 1980. A field guide to the birds of East Africa. Collins, London. WlNlGER, M. 1986. Causes and effects of the thermo-hygric differentiation of Mount Kenya. Geographica Bernensia, African Studies Series A1 :3 1 — 45. Young, T.P. 1982. Bird visitation, seed set, and germination rates in two Lobelia species on Mount Kenya. Ecology 68: 1983-1986. Young, T.P. 1985. Lobelia telekii herbivory, mortality, and size at reproduction: variation with growth rate. Ecology 66:1879-1883. YOUNG, T.P. 1990. Mount Kenya forests: an ecological frontier. Geographica Bernensia, Afripan Studies Series A 8 : 1 97-20 1 . YOUNG, T.P. 1991 . Flora, fauna, and ecology of Mount Kenya and Kilimanjaro. Pp. 37-49 in Allen, I. (ed.), Guide to Mount Kenya and Kilimanjaro (4th ed.) Mountain Club of Kenya, Nairobi. YOUNG, T.P. & M.M. PEACOCK. 1992. Giant senecios and the alpine vegetation of Mount Kenya. J. Ecology 80:141-148. YOUNG, T.P. & A.P. Smith, (in press). Herbivory on alpine plants of Mount Kenya. To appear in: P. Rundel & A.P. Smith (eds), Tropical Alpine Environments: Plant Form and Function. Cambridge University Press, Cambridge. J. EANHS & Nat. Mus. 82 (202) December 1993 77 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya Table 1. Alpine vertebrates of Mount Kenya. Upper altitudinal limits are based on the literature (questionable records not included) and personal observations. Lower altitudinal limits for Mount Kenya are included when they are known, but not if the records merge into a forest or lowland population. Status: First letter(s) indicate nature of occurrence: r= resident, v= visitor, m= migrant, i= introduced, x=accidental. An asterisk indicates evidence of breeding (courtship, nesting, young). The second letter(s) indicate abundance; a= abundant, c= common, o = occasional, r= rare, e= locally extinct. SPECIES (1 12 spp) Latin Name Altitude (m) Status Fish (3 spp) Char Salvelinus alpinus up to 4300 i, e? Brown Trout Salmo trutta up to 4000? i*, c Rainbow Trout Salmo gairdnerii up to 4200 i*,c Amphibians (2 spp) Frog Phrynobatrachus kinangopensis up to 3350 r, c? Frog Rana wittei up to 3800 r, c? Reptiles (6 spp) Alpine Meadow Lizard A 1 gyro ides alleni 3400-4600 r*, a Skink Mabuya irregularis up to 3600 r?, or Kenya Side-striped Chameleon Chameleo schubotzi up to 4200 r, c Hohnel’s Chameleon Chameleo hohnelii up to 3400 r, c? Hinde’s Viper Vipera hindei 3000-3400 r, o Striped Grass Snake Psammophylax variabilis ? ?, ? Birds (58 spp) Lesser Grebe (Dabchick) Tachybaptus ruficollis up to 3200 9 9 ' ? * Hamerkop Scopus umbretta up to 3200 V?, ? Open-billed Stork Anastomus lamelligerus up to 4150 x, r White Stork Ciconia ciconia ? x, ro Green Ibis Bostrychia olivacea up to 4100? v?, ro African Black Duck Anas sparsa up to 4400 r*, c Maccoa Duck Oxyura maccoa up to 3000 ?, ? Secretary Bird Sagittarius serpentarius up to 3500 v, o Ruppell’s Vulture Gyps rueppellii ? V, 0 Lammergeyer Gypaetus barbatus up to 4300 r, o Eurasian Marsh Harrier Circus aeruginosus ? m, ro Pallid Harrier Circus macrourus ? m, ro Montagu’s Harrier Circus pygargus ? m, ro Tawny Eagle Aquila rapax ? V, 0 Steppe Eagle Aquila nipalensis 7 v, o Verreaux’s Eagle Aquila verreauxii up to 4500 r, o Augur Buzzard Buteo augur up to 4400 r*, a Common Buzzard Buteo buteo 9 v, 0 Mountain Buzzard Buteo oreophilus up to 3500 ?, ? J. EANHS & Nat. Mus. 82 (202) December 1993 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya 78 Table 1. continued Alpine vertebrates of Mount Kenya. SPECIES Latin Name Altitude (m) Status Gabar Goshawk Melierax gabar 4000? m, r Black Kite Milvus migrans ? v, r? Lesser Kestrel Falco naumanni ? m, ro Kestrel Falco tinnunculus ? m, o Jackson’s Francolin Francolinus jacksoni 3400-4000 r, a Moorland Francolin Francolinus psilolaemus up to 3900 r?, ? Chestnut-tailed Pygmy Crake Saruthrura affinis up to 3700 r?,? Red-knobbed Coot Fulica crist at a up to 3000 r*, ? Black-winged Plover Vanellus melanopterus ? v|? African Snipe Gallinago nigripennis 3450-3800 r, a Greenshank Tringa nebularia ? m, r Green Sandpiper Tringa ochropus up to 4 1 80 m, r Common Sandpiper Actitis hypoleucos up to 3000 9 9 * > * Speckled* Pigeon Columba guinea up to 4000 rv, o Cape Grass Owl Tyto cape ns is ? r?, ? African Marsh Owl Asio capensis ? r?,? Long-eared Owl Asio otus ? r?,? Mackinder’s Eagle Owl Bubo capensis up to 4200 r, c Montane Nightjar Caprimulgus poliocephalus up to 4200 9 9 * 5 * Mottled Swift Apus aequatorialis ? r, o Alpine Swift Apus melba up to 4500 r*, c? Scarce Swift Schoutedenapus myoptilus ? r*?, o Unidentified bee-eater Merops sp. 4000? m, r Eurasian Swallow Hirunda rustica up to 4100 m, ro African Sand Martin Riparia paludicola up to 4150 r*, c Black Roughwing Swallow Psalidoprocne prist opt era up to 4 1 00 v, om, Northern Wheatear Oenanthe oenanthe up to 41 50 m, ro Alpine (Hill) Chat Cercomela sordida 2000-4570 r*, a Cinnamon Bracken Warbler Bradypterus cinnamoneus up to 3850 v?,o Willow Warbler Phylloscopus trochilus 4000? m, r Hunter’s Cisticola Cisticola hunteri 2600^1200 r*, a Red-throated Pipit Ant bus cervinus up to 4 1 80 m, r Sharpe's Longclaw Macronyx sharpei up to 3950 9 9 • > * Slender-billed Chestnut-winged Starling Onychognathus tenuirostris up to 4250 v & r*,o-i Scarlet-tufted Malachite Sunbird Nectarinia johnstoni 3450^1300 r*, a Tacazze Sunbird Nectarinia tacazze up to 4000 r?, co Yellow-crowned Canary Serinus canicollis up to 4700 v, r-c Streaky Seed-eater Serinus striolatus 2750^1200 r*, a White-naped Raven Corvus albicollis up to 4180 v, r Mammals (43 spp) African Dormouse Graphiurus murinus up to 4180 r, c Striped Tree Mouse Dendromus ins ignis up to 4330 r, a J. EANHS & Nat. Mus. 82 (202) December 1993 79 T.P. Young and M.R. Evans • Alpine vertebrates of Mt Kenya Table 1. continued Alpine vertebrates of Mount Kenya. SPECIES Latin Name Altitude (m) Status Harsh- furred Mouse Lophuromys flavopunctatus up to 4100 r,? Striped Grass Mouse Lemniscomys striolatus up to 4300 r?,? Four-striped Grass Mouse Rhabdymus pumilo up to 3800 r,? Climbing Mouse Praomys (Hylomyscus) denniae ? r, a Mount Kenya Mole Rat Tachyoryctes splendens (rex) up to 4050 r, c Groove-toothed Rat Otomys orestes up to 4750 r, a Groove-toothed Rat Otomys tropical is up to 4175 9 9 * 5 * Porcupine Hystrix sp. up to 3500 ?, ? Hare Lepus sp. up to 3350+ 9 9 • 5 • Bat (Pipistrellus nanus)? up to 3000 9 9 • ? * Pygmy Shrew Crocidura alpina up to 4 1 80 r, c Dusky Shrew Crocidura fumosa up to 3920 r,? Dusky Shrew Crocidura turba up to 4 1 00 9 9 • 5 * Mole Shrew Myosorex (Surdisorex) polulus up to 3920 r, ? Shrew Soncus infinitesimus ? r?,? Sykes’ Monkey Cercopithecus mitis up to 4800 x, r Black and White Colobus Colobus polykomos up to 4700 x, r Hunting Dog Lycaon pictus up to 4250 v(r), r Black-backed Jackel Canis mesomelas up to 3660 9 9 • 5 * Zorilla Ictonyx striatus up to 4270 r, o Large-spotted Genet Genetta tigrina ? 9 9 * 5 * Marsh Mongoose Herpestes paludinosus up to 3500 ?, ? Slender Mongoose Herpestes sanguineus ? 9 9 * 5 * Spotted Hyaena Crocuta crocuta up to 4800 V, 0 Cheetah Acinonyx jubatus up to 4000+ v, r Serval Felis serval ? 9 9 * > * Wild Cat Felis lybica up to 3800 r, c Lion Panthera leo up to 3600 v (r?), o Leopard Panthera pardus up to 4600 r*, c Rock Hyrax Procavia johnstoni 3450-4700 r*, a African Elephant Loxodonta africana up to 4600 v, o Black Rhinoceros Diceros bicornis up to 3700 r, e? Burchell’s Zebra Equus burchelli up to 4300 r*, c Kongoni Alcephalus buselaphus up to 4000 r?,o Black-fronted Duiker Cephalophus nigrifrons ? 9 9 • 5 * Bush Duiker Silvicapra grimmia 3250-4300 (-4900) r, c Klipspringer Oreotragus oreotragus ? ?, ? Steinbuck Raphicerus campestris up to 3500 r, o Bongo Boocercus eurycerus up to 4300 x, r Eland Taurotragus oryx up to 3800 r*, c African Buffalo Syncerus caffer up to 4700 v, c J. EANHS-& Nat. Mus. 82 (202) December 199-3 NOTICE TO CONTRIBUTORS Journal of the East Africa Natural History Society and National Museum Papers on Natural History subjects will be considered on the understanding that they are being offered exclusively to the Society and have not been submitted, at the same time, or previously, for publication elsewhere. Copyright is retained by the East Africa Natural History Society; reference to contributions in the journal may be made, but no part of the text, diagram, figure, table or plate may be reproduced without permission from the Editor. Such permission will only be granted with the approval of the author. Three copies of papers for consideration should be senf to The Editor, Journal of East Africa Natural History Society and National Museum, P.O. Box 44486, Nairobi, Kenya, and should be accompanied by a covering letter from the author, or in case of multiple authors, from the author responsible for decisions regarding the text. Contributions may be sent as: • preferably a floppy disk of either 5.25 or 3.5 inch in fully IBM compatible MS-DOS, carrying text in a commonly used wordprocessing language, or a floppy disk with the text prepared on an Apple Macintosh. If in any doubt, the author(s) should contact the Editor at the above address for further details. Three “near letter quality” printed copies on A4 size paper double-spaced should accompany the diskette. The author(s) should, of course, retain a back-up copy. • or a typescript, which should be on A4 size paper, typed on one side only, double spaced with margins of at least 5 cm on the left side of the paper to allow for editorial instructions to the printer. Title Should be brief and to the point Author name(s) and institution(s) where the work was carried out should be given, together with a suggested running page headline of not more than 50 characters including the name(s) of the author(s). A current address for correspondence, if applicable, should be included as a footnote. Abstract To be on the first page and consist of no more than 1 50 words. This should outline the major findings clearly and concisely without requiring reference to the text. Headings Should be brief, and in the following format: “Abstract, Introduction, Methods, Results, Discussion, Acknowledgments and References”. All headings to be typed on a separate line from subsequent text. Main headings to be typed in capital letters and centred on the page. Sub-headings to be at the left of the page, in lower case; further headings should be in italics or underlined (not bold) beneath them. The first letter of any headings should be a capital letter. Text Indent each new paragraph except those after a heading. Spelling should follow that of the Oxford English Dictionary. All pages must be numbered, at the top right hand edge. Numerals, dates, times and scientific units. Prose numbers should be written in full up to and including twelve, and as numerals thereafter, except at the end dates and beginning of a sentence. Dates should be written as “29 August”, “10-13 January 1989” or “1969-1989”. Times should follow the 24 hour clock (eg. 1500 h). All measurements must be in metric units. SI units and abbreviations should be used throughout with a space before and after the SI unit. Full stops should not be used after such units. The approximate position of tables and figures should be indicated in the left margin with pencil. Figures The original and two photocopies must be provided with each manuscript. Figures should be well drawn on pure white cartridge paper or good quality tracing paper. Lettering must be of very good quality (e g. Lettraset or good stencil). Typewritten or freehand lettering will not be accepted (without good reason). Each figure should have the author’s name, figure number and caption details written lightly on the reverse in pencil. Cite each figure in the text in consecutive order using Arabic numerals. “Figure” should be abbreviated to “Fig.”, except in figure legends and at the beginning of sentences. Two separate caption sheets should accompany the typescript. Black and white photographs are acceptable if they have good definition and contrast and are reproduced on glossy paper. Consideration must be given by the author(s) to their final size when reduced for publication in the Journal, which is B5. Tables Each table should be typed, double-spaced on a separate page and clearly numbered. Tables should be simple and self-explanatory, and have a brief title above each one. They must be numbered consecutively with Arabic numerals. Nomenclature. In cases where there are well known and internationally accepted English names, these should be used throughout the text. The scientific name should be given in full, underlined or preferably in italics, when quoted the first time. The authority should also be quoted at this stage except in non-taxonomic ornithological papers. Only the generic name should begin with a capital letter. In subsequent references to the name, use only the abbreviated generic name, with the full specific name but not the authority, e g. Spodoptera exempta (Walker); subsequently S. exempta. The authority should not be given if scientific names are used to describe an “association” or species complex, e g. Acacia drepanolobium - Theneda triandra, wooded grassland. Type in capitals the first letter of the English names of species (e.g. Crowned Eagle, Grey-capped Warbler), but not of higher taxa (e g eagles, warblers). References. Care should be taken to see that all references quoted in the text are also given in full, in alphabetical order, in this section. To avoid confusion, references to books should not be abbreviated and neither should they be underlined nor in italics. Journal names should be written in full, underlined or, if possible, in italics. Personal communications should not be cited in the references, neither should they be abbreviated in the text. If there are two or more references cited by an author which were published in the same year, distinguish them by adding letters to the dates, e g. Brown (1972a). For technical reasons author names in the reference list should not be typed in upper case. Authors are advised to refer to this issue of the Journal for the preferred format. Acknowledgements. As acknowledgements may imply endorsement of the data and conclusions in the authors’ work, it is advisable to obtain permission before quoting persons who have made contributions to the study. Once the paper has been selected by the Editor for possible publication in the Journal, the manuscript will be refereed by two persons who are acknowledged specialists in that field, before a decision is made of acceptance. The author will be sent one set of printouts for the correction of typographical errors. Authors are urged to ensure that all changes to the text are incorporated into this copy, as no further changes can then be made. The author is entitled to 25 copies free of charge. Thirty free copies will be provided if there is more than one contributor. Further copies can be supplied at cost, provided the order is placed before printing has been undertaken. The East African Natural History Society is supported by membership subscriptions, and may be unable in some instances, due to financial constraints, to accept all submissions of value. While effort will be made to minimise such restrictions, contributors are requested to seek financial support to assist the Society with the ever increasing cost of publication. Editor, Journal East Africa Natural History Society and National Museum. Referees Proofs Offprints EDEN WILDLIFE TRUST Charitable Trust Registered No. 278008 The East Africa Natural History Society is most grateful to the Eden Wildlife Trust for their very generous contribution towards the cost of printing this Journal part Published by: The East Africa Natural History Society Printed by ABE Graphics, P.O. Box 62008, Nairobi Heckman BINDERY, INC. Bound-Tb-Please* JULY 03 ! N. MANCHESTER, INDIANA 46962